Anesthesia Analgesia September 2009

Editorial
Spend Time on Patients and Families or on Documentation?

Armand R. J. Girbes, MD, PhD* Jan G. Zijlstra, MD, PhD
From the *Department of Intensive Care, University Hospital VU Medical Centre, Amsterdam; and Department of Intensive Care, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands. Accepted for publication May 19, 2009. Address correspondence and reprint requests to Armand R. J. Girbes, MD, PhD, Department of Intensive Care, VU Medical Centre, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands. Address e-mail to arj. girbes@vumc.nl or a.girbes@planet.nl. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af802c
lthough it seems quite obvious that every aspect of medical care and treatment needs good documentation, we all probably believe that our documentation is insufficient from time to time. In daily practice, we observe much more than we could ever document. Medical care of a patient is the result of interaction between the patient, the doctor, and the medical team, including nurses. This is clearly the most important and decisive factor for the patient, yet it is a process that cannot be documented easily in round figures. A patient can receive perfect medical care, whereas the documentation may have flaws.1 In this issue, Spronk et al.2 report on their review of charts of 347 patients who died during intensive care unit (ICU) admission or shortly after discharge from the ICU. The authors focus on documentation around withdrawal and withholding of treatment in two ICUs. A minority of 77 patients died while receiving full life-support, and for 4 patients, a restriction in life-support order was already present on admission. For 206 of the 266 patients who died after a change in life-support order, the patients wishes regarding life support were not documented. In one third of the cases, it was not documented which staff physician was involved in the end-of-life decision. However, although the involvement of the senior staff is not documented, it does not mean that, in reality, they were not involved (personal communication). The structure in both hospitals, with training of medical specialists, is such that in practice the staff is involved in such decisions, provided that supervision of junior doctors is adequate. There is no evidence in the Spronk et al. report that data were insufficient to ensure patient care, although this was not investigated. The authors deserve much credit for their willingness to report a practice in their hospital, which at first glance is not favorable. Documentation is probably poor in many other (teaching) hospitals. However, in view of the importance of the subject, end-of-life decisions, one would not dispute that good quality documentation is required. The challenge is, of course, to find a solution. Because documentation is time consuming and in general doctors are under time pressure when performing their clinical work, there is a constant competition between write and document and treat the patient and talk to the patient and family. Therefore, we should not document for the sake of documentation but first define the purpose and then create the documentation system. Documentation has several purposes. First, for immediate patient care, it helps in physician-to-physician communication, with transfer of medically relevant information. This generally takes place during rounds in the ICU, when every patient is discussed.3 Because of the fact that a range of professionals are present at rounds simultaneously, different kinds of contributions are received. In addition to written information, oral information is supplied, which by nature can be more extensive. A system that ensures transfer of information is of paramount importance in intensive care medicine. The goal of documentation is of course to improve continuity of care and to prevent loss of information. When a doctor hands over the care of a patient to another doctor, it is clear that to some extent information will be lost, no matter how good the documentation is. In intensive care medicine, continuity of care is crucial, and an important
691
Vol. 109, No. 3, September 2009
aspect of this is transfer and thereby documentation of patient data, including considerations to institute or change certain treatments, performance of diagnostic tests, differential diagnostic considerations, provision of information to the patient and family, and social and ethical issues. A second reason for documentation is to provide hard data, mainly for surveillance and diagnostic purposes. Nowadays, in the ICU, automated patient data management systems make storage of many patient data easy, and, for example, circulatory and respiratory functions data are generated and documented automatically. This is also true for treatments given, ventilator settings, perfusors for vasoactive drugs, laboratory values, etc. However, even with automated data collection, the physician must still write down diagnostic and therapeutic considerations, discussions with the patient, and findings of physical examinations. One might expect important data regarding the care of the patient to be recorded in the patient record (electronic). The paper by Spronk et al.2 in this issue of Anesthesia & Analgesia demonstrates that this is not the case. Whether the patient was incapacitated at the moment of the end-of-life decision, and which staff member was involved, was poorly documented. Clearly, patient data management systems do not improve documentation of more complicated processes. The third reason we record data is for quality assurance. We want to evaluate whether processes are
performed in accordance with our standards. And fourth, in an environment where legal issues and notably legal blame plays a major role, written documentation may also be collected for potential legal defense against charges. To determine whether documentation is insufficient, we need to define which data, considerations, and (motivated) decisions are important enough to be documented. It would also be helpful then to define a minimum dataset. Once defined, action can be taken to improve a possibly failing documentation system. Lack of data does not mean that the process of care is of poor quality or is not performed according to the standard but, to find that out, further research is required. To prevent lawsuits through documentation is an almost unreachable target. Documentation is important but should be performed without losing sight of the purpose. REFERENCES
1. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA 2007; 297:831 41 2. Spronk PE, Kuiper AV, Rommes JH, Korevaar JC, Schultz MJ. Practice of and documentation on withholding and withdrawing life support: a retrospective study in two Dutch intensive care units. Anesth Analg 2009;109:841 6 3. Pronovost P, Berenholtz S, Dorman T, Lipsett PA, Simmonds T, Haraden C. Improving communication in the ICU using daily goals. J Crit Care 2003;18:715
692
Editorial
ANESTHESIA & ANALGESIA
Editorial
Not All Strikes Are Easy to Call

Michael G. Fitzsimons, MD, FCCP Keith H. Baker, MD, PhD
From the Department of Anesthesia and Critical Care, Massachusetts General Hospital, Boston, Massachusetts. Accepted for publication April 23, 2009. Address correspondence and reprint requests to Michael G. Fitzsimons, MD, FCCP, Department of Anesthesia and Critical Care, Massachusetts General Hospital, 55 Fruit St., Boston, MA 02114. Address e-mail to mfitzsimons@partners.org. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc826
he decision to enter the field of anesthesia after recovery from substance abuse has been a controversy in the anesthesiology community. The incidence of abuse has been estimated at 1% of faculty members and 1.6% of residents in training programs.1 This incidence is similar to that reported by Ward et al.2 for the decade between 1970 and 1980. Unfortunately, death or near-death is often the event which identifies the physician as a substance abuser.1 The traditional view has been that recovery from substance abuse by anesthesiologists, particularly those who abuse opioids, is poor. Menk et al.3 demonstrated that only one third of residents who abuse opioids successfully reenter the specialty of anesthesia. Collins et al.4 reported that only 46% of residents who returned to training ultimately completed the program. Fry5 reported that only 15% of residents were successful in their effort to reenter the specialty of anesthesia. All of these studies are accompanied by frequent death during relapse that range from 9% to 31%. These dismal statistics have led many in the anesthesia community to call for either redirection of rehabilitated residents into lower-risk specialties or, more drastically, for the anesthesia leadership to adopt a one strike, youre out policy.6 Drs. Skipper, Campbell, and DuPont7 are to be congratulated on providing a major ray of hope with their article in this issue of the journal. They offer encouraging evidence in an area where the literature is highly pessimistic for the impaired anesthesiologist. Their 70% success rate for physicians treated at state Physicians Health Programs (PHPs) demonstrates that recovery is possible and even likely. Despite this finding, several aspects of their paper must be addressed. The article reveals that only 16 of 42 PHPs participated in the study. The primary reasons for nonparticipation were lack of resources and/or regulatory impediments. The outcome rate for programs not participating is unknown, but it is unlikely that programs which declined to participate based on lack of resources would have outcomes as good as those that have more resources. In 2007, the California State Auditor evaluated the Medical Board of Californias Physician Diversion Program and determined that standards were not being met. On July 26, 2007, the Medical Board voted to close the program. There are no universal or federal standards for all programs and thus outcomes may vary by state. In all likelihood, well funded and organized programs have the resources to produce better results. The American Medical Association Council on Mental Health recognized that treatment rather than discipline should be the goal for the impaired physician. By 1980, nearly all state medical societies had authorized or implemented a physician treatment program. Well funded and organized programs have developed better treatment models over the years. The Oregon experience revealed that 75% of those who were treated were either stable or improved during an 8-yr follow-up.8 The Massachusetts Medical Societys Physician Health Service reported a 75% success rate as defined as continuous or complete abstinence from any use of alcohol or drugs for a minimum of 3 years.9 The prior surveys by Menk et al. and Collins et al. that document poor outcomes do not address what would have happened if the individuals had been referred to a physicians health service. They did not report whether a contract had been signed or whether the physician was compliant with a monitoring program, all keys
693
to successful recovery. The major flaw in comparing the data from Menk et al. and Collins et al. to the current data is that residents were the primary subjects in the earlier papers, and yet this very population was appropriately excluded from the current report because of small numbers. Extreme caution must be taken when generalizing the current results using board-certified anesthesiologists and applying those results to residents. A resident must still negotiate the stressful period of training. After residency, the nascent graduate then has many more years of actual anesthesia practice and exposure to the same substances to which he or she became addicted. Finally, the trainee may not have developed the professional support system that an older anesthesiologist has. Anesthesiologists are frequently described as being overrepresented in PHPs.9,10 The assumption is often made that this is due to a higher rate of substance abuse disorders among anesthesiologists. There is clearly a frequent incidence of death among substance-abusing anesthesiologists likely because of the high potency and low therapeutic windows of such drugs, including opioids, propofol, and volatile anesthetics.11,12 We may classify any use of occupational drugs as substance abuse or misuse because of our fear of death to the individual. Nearly 20% of training programs have reported a pretreatment fatality.4 The authors compare multiple variables between anesthesiologists and other physicians in an attempt to find differences. They found that anesthesiologists in treatment programs are more likely to abuse opioids and have a higher incidence of IV drug abuse, yet they have a lower incidence of a positive drug screen while undergoing treatment. When a large number of variables is compared among groups, it becomes increasingly likely that some difference will occur because of chance alone. In particular, with a significance value set to P 0.05, we expect about 1 in 20 test items to be abnormal even when there is no true difference between groups. Fortunately, the main findings just mentioned were significant with P values 0.01 and thus likely represent trustworthy findings. Importantly, some of the measured variables may covary. Take, for example, the finding that anesthesiologists are more likely to use narcotics and that they are also more likely to use IV drugs. These findings are unlikely to be independent of each other and thus may represent a single difference between anesthesiologists and nonanesthesia physicians. Questions remain as to the overall likelihood of recovery for an anesthesiologist who has engaged in
substance abuse. It is clear that trained anesthesiologists who enter well-funded and supported PHPs and sign a contract including intensive monitoring have a high likelihood of recovery and return to successful practice. We are indebted to Skipper et al.7 for showing that a positive outcome can be obtained. More research is needed regarding the success of treatment of residents and fellows who develop such disorders. A simple one strike, youre out policy when applied to all providers does not take into account these new data by Skipper et al. Their data focused only on anesthesiologists in practice. If addicted anesthesia residents are referred to the same types of physician health service programs as studied by Skipper et al. and complete a full course of treatment, they may enjoy the same positive outcomes. We believe that residents in anesthesia without coexisting psychiatric disorders, polysubstance abuse, and a family history of substance abuse and who complete a PHP should be considered for reentry into the specialty in consultation with the residents addictionologist and psychiatrist. REFERENCES
1. Booth JV, Grossman D, Moore J, Lineberger C, Reynolds JD, Reves JG, Sheffield D. Substance abuse among physicians: a survey of academic anesthesiology programs. Anesth Analg 2002;95:1024 30 2. Ward CF, Ward CG, Saidman LJ. Drug abuse in anesthesia training programs: a survey1970 1980. JAMA 1983;250:9225 3. Menk EJ, Baumgarten K, Kingsley CP, Culling RD, Middaugh R. Success of reentry into anesthesiology training programs by residents with a history of substance abuse. JAMA 1990;263:3060 2 4. Collins GB, McAlister MS, Jensen M, Gooden TA. Chemical dependency treatment outcomes of residents in anesthesiology: results of a survey. Anesth Analg 2005;101:1457 62 5. Fry RA. Chemical dependency treatment outcomes of residents. Anesth Analg 2006;103:1588 6. Berge KH, Seppala MD, Lanier WL. The anesthesiology communitys approach to opioid- and anesthetic-abusing personnel. Anesthesiology 2008;109:762 4 7. Skipper GE, Campbell MD, DuPont RL. Anesthesiologists with substance use disorders: a 5-year outcome study from 16 state Physician Health Programs. Anesth Analg 2009;109:891 6 8. Shore JH. The Oregon experience with impaired physicians on probation. JAMA 1987;257:2931 4 9. Knight JR, Sanchez LT, Sherritt L, Bresnahan LR, Fromson JA. Outcomes of a monitoring program for physicians with mental and behavioral health problems. J Psychiatr Pract 2007;13:2532 10. Talbot GD, Gallegos KV, Wilson PO, Porter TL. The Medical Association of Georgiss Impaired Physicians Program Review of the first 1000 physician: analysis of specialty. JAMA 1987;257: 292730 11. Wischmeyer PE, Johnson BR, Wilson JE, Dingmann C, Bachman HM, Roller E, Tran ZV, Henthorn TK. A survey of propofol abuse in academic anesthesia programs. Anesth Analg 2007;105:1066 71 12. Wilson JE, Kiselanova N, Stevens Q, Lutz R, Mandler T, Tran ZV, Wischmeyer PE. A survey of inhalational anaesthetic abuse in anaesthesia training programmes. Anaesthesia 2008;63:616 20
694
Editorial
Editorial
Barbarians at the Gate

Warren S. Sandberg, MD, PhD
From the Department of Anaesthesia, Harvard Medical School; and Department of Anesthesia and Critical Care, Massachusetts General Hospital, Boston, Massachusetts. Accepted for publication May 18, 2009. Supported by department funds of the Department of Anesthesia and Critical Care, Massachusetts General Hospital. Address correspondence and reprint requests to Warren S. Sandberg, MD, PhD, Department of Anesthesia and Critical Care, Massachusetts General Hospital, 55 Fruit St., Jackson 4, Boston, MA 02114. Address e-mail to wsandberg@partners.org. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af803e
nesthesiology is at a crossroads. In some settings compensation outstrips revenue, supported by stipends to anesthesia groups from hospitals. At the same time, the number of anesthesiologists relative to demand seems to be declining. According to a draft report by the RAND corporation, there is an imminent shortage of anesthesiologists,1 and this will certainly exert further upward pressure on compensation. Meanwhile, government payers are signaling that reimbursements must decline. What will be the outcome of this developing conflict? In this issue of Anesthesia & Analgesia, Kheterpal et al.2 report the results of their most recent survey of workforce and finances in academic anesthesiology programs. Their notable findings are a continuing desire to hire more anesthesiologists into academic practice and another increase in the subsidy from hospitals to anesthesia groups (now $109,000/year per faculty full-time equivalent [FTE]) simply to maintain the status quo in terms of filled faculty positions. The figure represents the difference between revenue per faculty FTE generated by the department and the cost per faculty FTE borne by these same academic departments. One might worry that this subsidy represents the salary support required to keep the departments staffed, although the data of Kheterpal et al. do not prove causation. However, this dismal state of affairs is distressing for obvious reasons; academic departments train the residents who are our future. Academic anesthesia departments do virtually all of the research in anesthesiology, and they provide much of the care for complex cases. In the continuing gap between compensation and revenue, one can discern a concealed but fundamental challenge to anesthesiology, namely a disruptive change that could upend the foundational expectations about how anesthesia care is provided in the United States and perhaps elsewhere. To understand this potential threat, one must understand the notion of the disruptive innovation. Disruptive innovation was described in terms of products and the companies that make and buy them by Clayton Christensen in The Innovators Dilemma.3 However, the construct also applies to medical specialties and the services they provide. Generally stated, any product has a range of users who need differing degrees of performance from the product. This is certainly true in anesthesiology, where cardiac surgeons, for example, need more capabilities from their anesthesiologists than surgeons who specialize in minor outpatient procedures. Companies that make the most capable, reliable (i.e., high performance) products meet the needs of the high-end users who demand the most from such products. Such high-end customers are willing to pay high margins for performance, and thus are the companys best customers. Successful companies improve their products by responding to the needs of their high-end customers and seeking the most profitable opportunities. Figure 1 demonstrates, in graphic terms, the general relationship between the performance demanded from products by high-end users and the capabilities achieved by the makers of such products. Required performance of a product, technology (or medical specialty) is shown as two parallel lines on Figure 1. The upper line indicates the capability and performance required to meet the needs of the most demanding users. The lower line represents the capabilities required to meet the most basic needs of the users of a given product. The shaded zone between the lines
695
Figure 1. Relationship between the

performance demanded from products by high-end users and the capabilities achieved by the makers of such products. Adapted from Ref. 3, with permission from Harper Collins Harvard Business School Press.
represents the continuum between the minimum necessary and the highest required performance. Through diligence in attending to customer needs and investment in research, products almost always evolve as fast as or a little faster than the needs of the most demanding customers. Or, from a different perspective, there will always emerge a customer who can use all of the available capability in the best example of a product and will be willing to pay high margins. Companies aggressively pursue improved capabilities in their products to meet the needs of their best customers, illustrated by the top pointed line in Figure 1. They do this by pursuing a course of sustaining innovation, wherein the products capabilities steadily improve, but still the core product itself continues to resemble the original on the dimensions of performance that the vast majority of users care about. Reflection on the history of technological, educational, and scientific development in anesthesiology indicates that the specialty has been on a steady course of sustaining innovation since its inception. At the other end of the continuum, low-end users are often problematic for makers of high-end products. If there is more than one potential supplier of products who meet low-end customers needs, then all of the suppliers must compete on price, and profit margins must be low. In other words, for these low-end users, the product is simply a commodity for which the key distinction is price, rather than performance. Certainly, commodity users are not willing to pay a premium for increased capabilities that they do not need. Protecting their own financial interests, commodity customers are always on the lookout for a way to get their needs met for a lower cost. Occasionally, a new, different technology appears in the market that just meets the needs of the commodity users of the original technology. If the price is right, the low-end users will switch to the new, less capable technology. Initially, everyone is happy, including the purveyors of the original product. The
696
Editorial
new technology cannot compete at the high or even the middle end of the market, so it is not an obvious threat. Getting rid of the low-end commodity customers frees the industry leaders to focus on best meeting the needs of its high margin, high-end user customers. Over time, performance of new technology improves as it follows its own path of sustaining innovation, illustrated by the lower pointed solid line in Figure 1. If the price continues to be right, the new technology or product begins to take customers from the industry leader as its own capabilities inevitably improve. Eventually, the new technology captures the entire market. To see how this analogy might apply to anesthesiology, consider the two curves in Figure 1 to be the skills and knowledge required to give anesthesia for liver transplantation (top curve) and laparoscopic cholecystectomy (bottom curve). I chose these examples because they happen to be the two procedures that I do frequently, and so considering these may lessen the offense that could be given by presuming to establish more general hierarchies of difficulty in anesthesia practice. Through improved technology, training, and education, including subspecialty fellowships, the field of anesthesiology readily produces graduates who, as a group, are more than capable of handling the toughest liver transplant cases and for whom the laparoscopic cholecystectomy is mere childs play. What do we have to worry about? In advising would-be disruptive innovators where to look for opportunities, Christenson suggests looking for jobs to be done that are not currently well covered by current solutions. The Kodak FunSaver Camera is given as an example: at a time when film photography was being replaced by increasingly capable digital cameras, no one expected these buy-at-the-drugstore disposable cameras to be very successful. However, they provided a cheap way to take snapshots at a moments notice, after which the user just dropped the
whole thing off to be developed. The job to be done was on-demand photography. The FunSaver was a step on the pathway to the true disruptive innovation that allows on-demand, anytime snapshots: the digital camera that is ubiquitously present in the equally ubiquitous mobile phone. Analogously, there are numerous jobs to be done wherein patients have procedures entailing some form of discomfort or physiologic perturbation, for which anesthesiologists are not a perfect fit. This is not to say that anesthesiologists cannot do these cases, but rather that the multispecialty, fellowship-trained anesthesiologist is not analogous to the FunSaver or the phone camera. We are completely capable of performing these cases, but we bring too large a tail of complexity, equipment, expectations, and, above all, cost. Low-end consumers of anesthesia services regard the ability to safely produce a deeply sedated or anesthetized patient who is happy at the end of the procedure as a commodity, where the key differentiators compensation and fit with the procedure area workflow boil down to cost. The current commodity users of anesthesia services are ready for a solution that better meets their needs, and the products, technologies, and services that could provide this solution are already visible, albeit faintly. Consider the case of fospropofol. The initial marketing plan for this drug was as an alternative to propofol that could be administered by nonanesthesiologists for use during uncomfortable diagnostic and therapeutic procedures. Reasoning that the active agent would have the same risk profile as IV propofol, the American Society of Anesthesiologists encouraged the Food and Drug Administration to apply the same warning requiring administration by clinicians with expertise in airway management (i.e., anesthesiologists) as is found on the label for propofol.* This was probably the right position to take in the current environment, given the capabilities of alternative practitioners, devices, etc., for managing the airway in accidentally oversedated patients. But the important lesson is that proceduralists want to have a drug such as fospropofol that they could use without an anesthesiologist. Some aspect of current anesthesiology practice and management of the procedural suite is not meeting the needs of the proceduralists. There are unmet needs for anesthesia care in the segments of the market that are not commonly regarded as central customers of anesthesia services. Other potentially disruptive technologies have been demonstrated. For example, Bispectral Indexmediated closed loop control of isoflurane4 or propofol57 provides better performance on some metrics than human operators. Similar demonstrations of closed loop control for muscle relaxants have been published.8 Web searching the term McSleepy
*Available at: http://www.asahq.org/Washington/ASAfospro pofolcomments4-23-08.pdf for full text of the written comments to the FDA. Accessed May 13, 2009. Vol. 109, No. 3, September 2009
reveals some eye-opening text, images, and video. Others are developing automated methods to alert clinicians of developing hemodynamic perturbations.9 None of the developers of these technologies is proposing that they replace the anesthesiologist in the operating room (OR). That suggestion will come from the financial stakeholders in anesthesia care. Anesthesiologists are expensive. The total annual cost of an academic anesthesiologist reported by Kheterpal et al. is $605,000. Academic programs are unable to generate enough revenue to cover this compensation, in part because academic surgical cases are long and ORs are not optimally utilized.10,11 An increase in the number of anesthetizing locations over the past decade means that the relative number of available providers decreases, and compensation must increase to ensure service. Of course, this scenario also motivates a search for less expensive alternatives. Current trends in anesthesiology employment and education do not bode well for efforts to control costs. Because of the down economy with less hiring by practices, recent graduates are flocking to fellowships, furthering the process of subspecialization. This increases the trainees capabilities, with an attendant increase in the compensation they can command as a benefit of subspecialty certification. If the RAND forecasts prove accurate, these highly qualified individuals will be negotiating compensation amid a scarcity of providers. The field of anesthesiology is aggressively moving upmarket. Although this seems financially exciting for anesthesiologists, it amplifies the pressure for the users of more basic (or commodity) anesthesia services to find an alternative source of anesthesia care for their patients. If a less expensive, albeit less capable but still sufficient, alternative becomes available, will commodity users of anesthesia care pay extra for the super-capable product? Experiences from other industries say no. Ongoing rising cost of the available high-end product will motivate the search for lower cost alternatives and whet the payers appetites for implementation. Less expensive does not mean certified registered nurse anesthetists, who are comparable with anesthesiologists in many ways, including cost, but rather a nonanesthesiologist alternative way to provide what is essentially anesthesia care. One does not need to look very far to see potential replacements. Critical care nurses are essentially providing anesthesia when they transport an intubated, ventilated, sedated, and pressor-dependent patient from the intensive care unit for a diagnostic procedure, for example. There are also the technological developments mentioned earlier. The very number of such emerging new technologies, pharmaceuticals, and proposed alternatives to anesthesiologists as providers should sound the alarm that there is something wrong with the way the specialty currently serves its commodity users. Instead, all parties appear happy to find ways for
2009 International Anesthesia Research Society
697
anesthesiologists to get out of the commodity end of the market. This is a big mistake for anesthesiologists. Business history has countless examples of companies that abandoned a low-yield market sector to a disruptive technology only to be knocked off completely some years later by the very same low-end innovation. The problem is that the seemingly limited performance technology, product or service will inevitably become more capable. It will only be a matter of time before McSleepy can handle maintenance for basic general anesthetics and then more complex ones. The continuing move upmarket is painting anesthesiology into a corner. If an alternative care model or technology that allows nonanesthesiologists to produce anesthetized patients gains acceptance, the specialty of anesthesiology will face a choice: contract the boundaries of our practice, leaving behind the simpler cases, or find ways to compete more effectively for the lower end of the anesthesia market. This is an opportunity to leverage technology to become more efficient in our use of personnel and their intellect. Seizing this opportunity implies that we should regard technologies, such as anesthesia information management systems, decision support tools, closed-loop control for drug administration, as well as new care team models as the topics of enthusiastic research and development. Choosing this course moves OR and procedure room anesthesiology closer to the model used in intensive care units, where anesthesiologists assure that care is provided, but only personally provide care when their unique skills cannot be replaced. It is tempting to raise the objection that the practice of medicine is not subject to the same kinds of market forces and disruption that sent silver-halide film to join the dinosaurs, and this is to a limited degree true. State regulations, the label warnings on anesthetic drugs, liability risk, and hospital policies all provide inertia that slows the rate of change in who and what can provide anesthesia to the point that change is almost imperceptible. However, with sufficient incentive to alter the status quo plus a little luck, anesthesiologists protected status can be altered by the stroke of a pen, as evidenced by the gubernatorial opt-out provision from physician supervision of nurse anesthetists, published in 2001. Cost, protectionism, and unmet needs provide the incentive for change. The considered position taken by the specialty over innovations such as fospropofol could be seen as protectionist or obstructive. Embracing new technologies as something that enables anesthesiologists to do more with fewer people sends a completely different message than does opposing technological development. Kheterpal et al.2 publish a simple observation, but one that should trouble us. Compensation and hospital support to anesthesiology departments cannot continue to increase forever in the face of determination to
For reference, see http://www.nsf.gov/news/news_summ.jsp? cntn_id104259. Accessed May 13, 2009.
reduce health care costs. But there is not much history of high-end producers voluntarily decreasing the price of their flagship product. Instead, the product is replaced by an upstart. How to avoid this fate is not completely clear. Anesthesiologists should strive to be facilitators of patient care and patient flow in all environments. When departments receive support from hospitals, the reasons for giving the support should be explicitly agreed upon, and the work to be done specified so that the value in exchange to cost is widely understood.12 Anesthesiologists can perform many nonclinical value-added activities that benefit the hospital and its patients. Just as it is appropriate for these to be compensated, so too should departments that receive subsidies be visible throughout the facility performing these important nonclinical organizational and leadership functions. Figure 1 should always be in our minds. New technologies, some of them disruptive, will always continue to appear and challenge our status quo. Occasionally, these can apparently be beaten back. In 2005, a robot called Penelope worked as the scrub technician during a live surgery on a patient at New York Presbyterian Hospital. The robot functioned without an operator, recognizing the instruments optically and responding to voice commands. This event provoked one trade publication to run the headline Controversial robotic arm tested on its front page.13 The new technology was received rather coolly by OR personnel. Fast forward to 2009 and one finds the device now positioned for use in the central sterilization area, where its optical instrument recognition, counting accuracy, speed, and unbeatable duty cycle are to be used in instrument picking, counting, and kit packing. This clever strategy gives the robot a place in the workflow to grow and develop along its sustaining pathway. Eventually it will make it into the OR. REFERENCES
1. Byrd J, Peterson MD. The ghost of study past: new RAND study shows shortage of anesthesiologists. ASA Newsl 2009;73:39 40 2. Kheterpal S, Tremper KK, Shanks A, Morris M. Seventh and eighth year follow-up on work force and finances of the United States anesthesiology training programs: 2007 and 2008. Anesth Analg. In press 3. Christensen CM. The innovators dilemma. New York: Harper Collins, 1997 4. Locher S, Stadler KS, Boehlen T, Bouillon T, Leibundgut D, Schumacher PM, Wymann R, Zbinden AM. A new closed-loop control system for isoflurane using bispectral index outperforms manual control. Anesthesiology 2004;101:591 602 5. Absalom AR, Kenny GN. Closed-loop control of propofol anaesthesia using bispectral index: performance assessment in patients receiving computer-controlled propofol and manually controlled remifentanil infusions for minor surgery. Br J Anaesth 2003;90:737 41 6. De Smet T, Struys MM, Neckebroek MM, Van den Hauwe K, Bonte S, Mortier EP. The accuracy and clinical feasibility of a new bayesian-based closed-loop control system for propofol administration using the bispectral index as a controlled variable. Anesth Analg 2008;107:1200 10
698
Editorial
7. Struys MM, De Smet T, Versichelen LF, Van De Velde S, Van den Broecke R, Mortier EP. Comparison of closed-loop controlled administration of propofol using Bispectral Index as the controlled variable versus standard practice controlled administration. Anesthesiology 2001;95:6 17 8. Edwards ND, Mason DG, Ross JJ. A portable self-learning fuzzy logic control system for muscle relaxation. Anaesthesia 1998; 53:136 9 9. Ansermino JM, Daniels JP, Hewgill RT, Lim J, Yang P, Brouse CJ, Dumont GA, Bowering JB. An evaluation of a novel software tool for detecting changes in physiological monitoring. Anesth Analg 2009;108:873 80 10. Abouleish AE, Dexter F, Epstein RH, Lubarsky DA, Whitten CW, Prough DS. Labor costs incurred by anesthesiology groups because of operating rooms not being allocated and cases not being scheduled to maximize operating room efficiency. Anesth Analg 2003;96:1109 13; table of contents
11. McIntosh C, Dexter F, Epstein RH. The impact of servicespecific staffing, case scheduling, turnovers, and first-case starts on anesthesia group and operating room productivity: a tutorial using data from an Australian hospital. Anesth Analg 2006;103: 1499 516 12. Dexter F, Epstein RH. Calculating institutional support that benefits both the anesthesia group and hospital. Anesth Analg 2008;106:544 53; table of contents 13. Dunbar C. Controversial robotic arm tested: device responds to surgeons voice commands for instruments. AORN Connections 2005;3:1,4
699
Cardiovascular Anesthesiology
Section Editor: Charles W. Hogue, Jr.
Perioperative Echocardiography and Cardiovascular Education

Section Editor: Martin J. London
Hemostasis and Transfusion Medicine

Section Editor: Jerrold H. Levy
Lipopolysaccharide Evokes Resistance to Erythropoiesis Induced by the Long-Acting Erythropoietin Analogue Darbepoetin Alfa in Rats
Peter Brendt, MD* Ariane Horwat* Simon T. Scha fer, MD* Sven C. Dreyer, MD* Joachim Go thert, MD Ju rgen Peters, MD*
BACKGROUND: Anemia is common in patients with sepsis but its mechanism is unknown. We tested the hypothesis that effects on erythropoiesis evoked by darbepoetin alfa (DA), a long-acting erythropoietin analog, are diminished by lipopolysaccharide (LPS). METHODS: We performed a prospective, controlled, randomized animal study (male Lewis rats n 44). The interventions we used were intraperitoneal injection of Escherichia coli LPS (10 mg/kg) or vehicle followed by either DA (25 g/kg) or vehicle (four experimental groups). Blood and reticulocyte counts and variables of iron metabolism were measured at baseline and 3 and 14 days after interventions. RESULTS: Animals treated with DA alone showed an eightfold increase in reticulocyte count from baseline on Day 3, whereas no increase was seen in animals administered LPS or LPS/DA. On Day 14, the red blood cell count and hemoglobin concentration had increased by approximately 10% from baseline (P 0.001) in the DA group but had decreased after LPS on Days 3 and 14 (P 0.05) and in animals administered LPS/DA. Consumption of iron was seen on Day 3 in the DA group but not after LPS or LPS/DA combined. Values of ferritin and transferrin did not change between groups. CONCLUSION: LPS abolishes erythropoiesis and iron use evoked by DA and this is accompanied by a decrease in hemoglobin concentration and red blood cell concentration. Accordingly, endotoxin suppresses DAs ability to increase erythropoiesis.
(Anesth Analg 2009;109:70511)
inety-five percent of intensive care patients suffer from anemia within 3 days of their intensive care unit (ICU) admission and 44% receive packed red blood cells.1 Although mechanisms of anemia vary and include the primary disease (e.g., bleeding) and blood loss due to phlebotomy,2 anemia can be a risk factor for morbidity and mortality, especially in patients with cardiac disease.3 Of interest, patients with sepsis usually have a hemoglobin concentration lower than other patients admitted to the ICU and its concentration continues to decrease,4 so that septic patients are more likely to require packed red blood cell transfusion.4 Production of red blood cells in the bone marrow is generated by pluripotent stem cells and is tightly
From the *Klinik fu r Ana sthesiologie und Intensivmedizin; Universita t Duisburg-Essen; and Klinik fu r Ha matologie, Universita tsklinikum Essen, Universita t Duisburg-Essen, Essen, Germany. Accepted for publication March 29, 2009. This work is attributed to the Klinik fu r Ana sthesiologie und Intensivmedizin, Universita t Duisburg-Essen, Universita tsklinikum Essen, Essen, Germany. Address correspondence to Dr. Peter Brendt, Klinik fu r Ana sthesiologie und Intensivmedizin, Universita tsklinikum Essen, Hufelandstr. 55, Essen D-45122, Germany. Address e-mail to Peter.Brendt@uniduisburg-essen.de. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc80f
regulated by erythropoietin (EPO), the main hematopoietic growth factor. EPOs interaction with red blood cell precursors is mediated by the cellular EPO receptor resulting by up-regulation of the antiapoptotic protein Bcl-xL in progenitor cells in an increase in the number of peripheral red blood cells.5,6 The signal transduction cascade distal to the EPO receptor includes signal transducer and activator of transcription 5, Ras/mitogen-activated protein kinase, and phosphoinositide-3 kinase/Akt pathways, which has been studied in depth.7,8 Since its introduction into clinical practice nearly two decades ago, EPO has rapidly become a principle drug to treat anemia. Darbepoetin alfa (DA), an EPO analog with modified glycosylation and a threefold longer circulating half-life than pharmacologic EPO, stimulates erythropoiesis in the same manner as endogenous EPO.9,10 Lipopolysaccharide (LPS) stimulation decreases erythropoiesis in rats because of reduced EPO synthesis.11 However, in patients with septic shock, the endogenous serum EPO concentration is increased,12,13 apparently independent of the hemoglobin concentration.13 No relationship or a negative relationship was observed between serum EPO levels and blood hemoglobin concentrations in septic patients.14 This raises the question whether
705
sepsis/endotoxinemia per se can interfere with erythropoiesis and alter the response to erythropoietic grows factors. Because relevant information is not available, we tested the hypothesis that LPS alters the erythropoietic effects of DA in an animal model.
METHODS
This study was conducted in accordance with German governmental regulations complying with the European Community guidelines for the use of experimental animals. The experimental protocol was approved by the local animal care committee, and animals were treated in accordance with the guidelines of the American Physiological Society and the Guide for the Care and Use of Laboratory Animals (National Institute of Health publication 8523, revised 1996). Male Lewis rats weighing 300 480 g (Harlan-Winkelmann, Borken, Germany) were kept in accordance with animal welfare guidelines and supplied ad libitum with commercial rat chow and tap water.
count was expressed in relation to 1000 counted red blood cells. Free iron concentration was measured by spectrophotometry in heparinized plasma after reduction to the ferrous form and complexation with ferrozine (ADVIA 1650 colorimetric test apparatus, Bayer AG, Leverkusen, Germany). In addition, ferritin and transferrin plasma concentrations were measured in heparinized plasma using a nephelometer (BN2, Dade Behring, Marburg, Germany) and a commercially available standard kits (Dade Behring, Marburg, Germany).
Statistical Design and Analysis

Data are expressed as means sd and statistical analyses were performed with SPSS13.0 (SPSS, Chicago, IL). An a priori power analysis was performed with the software G*Power 3.0.10. We determined the group size required to detect differences in reticulocyte counts in response to interventions after 3 days (effect size: 1; -error: 0.05; power: 0.95). The mean of the control group for reticulocyte concentration was set as 11.6, based on values in the literature,15 with a standard deviation of 3.6. We looked for an increase in reticulocyte concentration with DA to at least 15. Using these data, the calculated sample size was 24 (4 6). To improve our margin of safety, we increased the number of animals. We expected that 30% 40% of the rats treated with LPS or LPS/DA would die but none of the animals receiving DA or placebo. To account for these losses, we therefore a priori increased the number of animals in the two former groups. Accordingly, we designed and started the experiments with 44 animals (vehicle n 8, DA n 8, LPS n 14, LPS/DA n 14) per group in a randomized fashion. Specifically, rats were selected at random from cages by an animal care technician not otherwise involved in the experiments and later randomly assigned to prefilled syringes containing drugs for treatment. As expected, animals in the groups treated with LPS or LPS/DA died (n 5 in both groups). This left nine surviving rats in these groups (mortality 36%), one more than the presumed required group size. In the vehicle group, one rat showed baseline values of reticulocyte count and hemoglobin above the twofold SD of the other animals and hence was not used in the data analysis. Changes of values of variables from baseline over time (within group) were tested with the general linear model for repeated measurements and pairwise comparisons. Differences in mean values of changes from baseline among groups were determined using the general linear model (SPSS) using univariate analysis, analysis of covariance, and pairwise comparisons, with factors for treatment and baseline values as covariate (analysis of covariance). The following a priori null hypotheses were tested: there are no differences in values of variables (1) among groups after 3 days of treatments, (2) within treatment groups between baseline and after 14 days
Experimental Preparation
In this randomized study including four experimental groups, we examined the effect on erythropoiesis of DA or vehicle in a sepsis-like condition as induced by Escherichia coli LPS (serotype O127:B8, Sigma-Aldrich, Taufkirchen, Germany). At baseline, orbital sinus blood (approximately 0.8 mL) was sampled from every rat during a brief isoflurane anesthetic. After this procedure, the animals were randomized to the four treatment groups. Animals received either LPS (10 mg/kg in 500 L of saline) or 500 L pyrogen-free saline (IP). Immediately after LPS or saline injection, either DA 25 g/kg or 500 L pyrogen-free saline was injected IP. Seventy-six hours and 14 days later blood samples were drawn again, as previously described.
Measurements
Blood Count and Reticulocyte Count Immediately after blood withdrawal values of hematological variables (hemoglobin concentration, hematocrit, erythrocyte, leukocyte, and platelet counts, mean corpuscular hemoglobin, mean corpuscular volume [MCV], mean corpuscular hemoglobin concentration) were measured using a dedicated animal hematology autoanalyzer (Scil Vet abc Ha matologieTM, Scil Animal Care Company GmbH, Viernheim, Germany). For measurements of reticulocyte count, blood was stained with brilliant cresyl blue (Merck, Darmstadt, Germany) mixing 25 L of blood and 25 L of brilliant cresyl blue solution. After 20 min of incubation, a thin smear was prepared on a microscope slide and allowed to dry in air for 30 min. Reticulocytes were then counted by two examiners (blind to allocation of the animals to experimental groups) using oil immersion and a 1000-fold magnification (H600 Hund GmbH, Wetzlar, Germany), results averaged, and reticulocyte
706
Endotoxin Suppresses Erythropoiesis
Figure 1. Effect of darbepoetin alfa (DA), lipopolysaccharide (LPS), LPS/DA combined, and vehicle on reticulocyte count at
baseline and on days 3 and 14. DA markedly increased the reticulocyte count on Day 3 and this effect had vanished on Day 14. This effect on erythropoiesis by DA was abolished by LPS. Data represent individual values and means. *P 0.05 compared with vehicle and to baseline.
(recovery) of treatments, and (3) over time between the control (vehicle) group and treatment groups. Statistical significance was assumed with an -error P of 0.05.
(on Days 2, 2, 3), one rat during blood withdrawal on Day 3, and another rat died on Day 9. These rats were particularly lethargic and showed diffuse internal bleeding on gross postmortem examination.
RESULTS
General Effects of LPS and DA
Rats receiving LPS, combined with DA or vehicle administration, developed signs of sepsis, such as diarrhea, piloerection, and lethargy, whereas animals administered vehicle or DA did not. Furthermore, 3 days after the application of LPS, the platelet count had decreased by approximately 70% (P 0.0001) in both the LPS/vehicle and the LPS/DA groups. This sepsis-like condition persisted for 5 to 7 days, and surviving animals thereafter returned to normal conditions. None of the animal died as a consequence of DA or vehicle administration during the observation period. In contrast, in the LPS/DA and the LPS/vehicle groups five rats died in each group leaving nine rats for data analysis in the LPS/DA and LPS/vehicle groups. All animals in the LPS/DA group died before blood withdrawal on Day 3 (on Days 3, 2, 3, 2, 2). In the LPS/vehicle group, three rats died before Day 3
Effects on Erythropoiesis of LPS, LPS/DA, DA, and Vehicle

Rats that received DA alone showed a marked increase in reticulocyte count (Fig. 1), both when compared with the other groups on Day 3 and when compared with baseline (P 0.001). This profound eightfold increase in reticulocyte count after DA was completely abolished by the administration of LPS in the LPS and the LPS/DA groups. Consequently, after 14 days, the red blood cell concentration increased by 11.9% in the DA group (P 0.05, Fig. 2). In contrast, in the LPS and the LPS/DA groups, red blood cell concentration slightly decreased on Day 3 and after 14 days compared with the vehicle group (P 0.05). The hemoglobin concentration in the LPS and the LPS/DA groups decreased by 13.8% 9.8% and 8.2% % 7.3% within 3 days, respectively and remained lower than the vehicle and DA groups (P 0.05). The hemoglobin concentration remained unchanged over
707
Figure 2. Effect on red blood cell concentration of darbepoetin alfa (DA), lipopolysaccharide (LPS), LPS/DA combined, and vehicle at baseline, and on Days 3 and 14. Although DA had increased red blood cell concentration on Day 14, both LPS and LPS/DA groups showed a decreased red blood cell concentration on Days 3 and 14. Data represent individual values and means. *P 0.05 compared with vehicle. #P 0.05 compared with baseline. the observation period in vehicle-treated rats. In contrast, in the DA group the hemoglobin concentration had increased by 8.8% 5% on Day 14 (P 0.005, Table 1). Red blood cells in LPS and LPS/DA-treated animals had an increased (P 0.006) volume (mean difference: 3.5 fL and 3.1 fL for LPS and LPS/DA, respectively) compared with vehicle on Day 14. Animals treated with DA alone showed an increase in MCV versus baseline and the other groups (mean difference: 5 fL) on Day 3 (P 0.05). For all groups during the observation period, MCV and mean corpuscular hemoglobin are shown in Table 1. LPS/DA, and DA-treated rats increased in each group over time (P 0.05).
DISCUSSION
Our findings demonstrate that LPS abolishes erythropoiesis and iron utilization evoked by DA and this is accompanied by a decrease in hemoglobin concentration and red blood cell concentration after LPS administration. Accordingly, endotoxin suppresses DAs action with an inability to evoke erythropoiesis. These data suggest that endotoxin mediates suppression of the erythropoietic activity of endogenous EPO. An advantage of using IP LPS injection as an experimental model is that many effects of LPS are well known, the model is standardized, and for approximately 6 days LPS injection results in a disease state similar to human sepsis. Conversely, sepsis is likely not confined to LPS alone and findings might be different in humans with prolonged sepsis or forms of sepsis not involving endotoxin. Furthermore, the timing and characteristics of cytokine secretion after LPS administration may differ among different species as well as among patients with sepsis.16,17 The mortality rate (36%) seen in our experimental setting closely
Effect on Iron Metabolism of LPS, LPS/DA, Vehicle, and DA

The iron concentration on Day 3 was lower in the DA group than in the other groups (P 0.001) and 86% 15% lower when compared with its baseline (P 0.001). In contrast, iron did not decrease after administration of LPS alone or LPS in combination with DA when compared with the baseline values. There were no differences in values of ferritin and transferrin among vehicle, LPS, and LPS/DA groups on Days 3 and 14. However, the ferritin values in LPS,
708
Table 1. Effect of DA, LPS, LPS/DA Combined, and Vehicle on Hematological and Iron Metabolism Variables at Baseline and on Days 3 and 14 Vehicle (n 7)
Hemoglobin (g/dL) Baseline Day 3 Day 14 Mean corpuscular hemoglobin (pg) Baseline Day 3 Day 14 Mean corpuscular volume (fL) Baseline Day 3 Day 14 Free iron plasma concentration (ug/dL) Baseline Day 3 Day 14 Transferrin plasma concentration (g/L) Baseline Day 3 Day 14 Ferritin plasma concentration (ug/L) Baseline Day 3 Day 14
DA darbepoetin alfa; LPS lipopolysaccharide. * P 0.05 compared with vehicle on the same experimental day. P 0.05 compared with baseline value within group.
LPS (n 9)
15.9 0.8 13.6 1.4* 14.8 0.5* 17.3 0.5 17.5 0.9 17.3 0.7* 50.3 0.9 49.8 1.1 50.7 2.1* 260 62 238 45 250 33 0.71 0.29 0.60 0.21 0.78 0.38 54 41 53 29 125 72
DA (n 8)
15.9 0.5 16.0 0.6 17.3 1.0* 17.1 0.4 18.1 1.1 16.6 0.5 50 1.2 53.1 1* 49.1 0.6 234 33 61 42* 269 62 0.79 0.266 0.83 0.32 0.77 0.32 37 11 59 20 97 7
LPS/DA (n 9)
15.7 1.0 14.4 0.7* 14.7 0.8* 17.1 1.1 17.2 0.8 16.7 0.6 49.3 1.8 48.9 1.6 50.2 2.1* 261 56 169 48 266 55 0,88 0.30 0.97 0.28 0.72 0.37 30 12 68 33 110 58
15.3 1.5 15.3 0.4 15.2 0.9 16.2 1.7 17.2 1.1 16. 0.8 48.6 2.3 48.1 2.4 47.1 1.1 193 26 223 44 221 33 0.92 0.32 0.96 0.35 0.91 0.44 96 93 48 14 67 24
mirrors the outcome of septic shock in humans.18 In any case, LPS affects and disrupts numerous physiologic pathways and cytokine production. Therefore, the effect seen in our study could be a result of these manifold changes and potential mechanisms need to be explored on a cellular and molecular level. Nevertheless, our animal model proved useful for demonstrating LPS-mediated interference with activity of DA and could be used for further investigations into the mechanisms of anemia accompanying sepsis. After LPS administration, we found abolition of DAs erythropoietic effect despite treatment with a DA dose of 25 g/kg, i.e., a dose resulting in an eightfold increase in reticulocyte count in the DA alone group. However, we did not establish a dose-effect relationship and this very large dose of DA used could still have been too small to alter LPS-evoked suppression of erythropoiesis. There are several general explanations for the decrease in hemoglobin concentration seen after LPS administration in our experiments. On one hand, LPS could decrease red blood cell and hemoglobin concentration by shortening red blood cell survival. On the other hand, LPS could directly interfere with erythropoiesis. Although we are not aware of data showing decreased red blood cell survival in LPS-evoked septic states, LPS is able to bind to human erythrocytes and reduce their deformability,19 which could shorten red blood cell survival. Our data, however, strongly suggest decreased erythropoiesis as the major cause, because DA, sparking erythropoiesis in vehicle animals, was unable to
initiate erythropoiesis and increase reticulocyte concentration when given together with LPS. This argues for LPS to inhibit erythropoiesis at some stage. This is further supported by the observation that the iron concentration decreased after DA administration associated with an increased reticulocyte count indicative of erythropoiesis but did not change after administration of either LPS or LPS/DA. Thus, after LPS administration, the bone marrow was unable to pick up iron and use iron when stimulated by this large dose of DA. Several mechanisms, alone or in combination, could be responsible for this, such as impaired proliferation of erythroid precursor cells due to changes in the complex balance among apoptosis, transcriptional, and epigenetic controls over the differentiation state, decreased expression of EPO receptors, perturbed DA/EPO signal transduction, decreased release of cells into the circulation, or cytotoxic effects on hematopoietic cells of LPS or cytokines. Anemia, often seen in sepsis, may result from a disturbed iron homeostasis. Decreased serum iron and transferrin concentration are the hallmark of anemia of chronic disease and the ferritin concentration is normal or increased.20 In volunteers, however, injection of a tiny dose of LPS (2 ng/kg) evoked a decrease in serum iron concentration for 22 h, whereas the ferritin concentration remained unchanged.21 In our experiments, we saw significant changes of ironrelated metabolism. First, we found increased iron consumption after administration of DA alone but not
709
in the LPS and the LPS/DA groups. Therefore, disturbed iron metabolism as a mechanism for the impaired ability of DA to increase the reticulocyte count after LPS administration is possible but seems to be different from the typical anemia of chronic disease with decreased iron and transferrin concentrations. Furthermore, we saw an increase over time in the ferritin concentration after administration of LPS, LPS/DA, and DA alone. This is in line with data from septic patients who showed an increase in ferritin levels.22 The increase in ferritin in the DA group was interpreted due to increased intestinal iron absorption, which would be in line with data after EPO treatment in rats23 and humans.24 Of interest, the observation of suppression of DAs effects after administration of LPS in our experiments contrasts with the effects of DA or EPO in inflammatory diseases other than sepsis, which respond to exogenously administered EPO.25 Among intensive care patients, trauma patients with multiple organ failure respond to EPO (600 U/kg 3 times weekly).26 However, ICU patients in general appear to show a blunted response to EPO but they still respond.27 In ICU patients, EPO increased the hemoglobin concentration and this may decrease the need for transfusions.28,29 Anemia related to inflammatory disorders like rheumatoid arthritis has been explained by increased apoptosis of the erythroid lineage and has been linked to tumor necrosis factor .30 In a rodent model, this could be mitigated by DA.31 Indeed, tumor necrosis factor is overexpressed in the acute phase of severe sepsis and this could be one mechanism leading to suppression of DAs effects. Furthermore, interferon-, an important mediator in Gram-negative sepsis,32 downregulates mRNA expression of the EPO receptor, indicating that the number of EPO receptors can influence apoptosis of erythroid precursor cells.33 In hematopoietic progenitor cells of patients with myelodysplastic syndromes, LPS induces apoptosis via the toll-like receptor-434 and this could be another link to the effects of LPS on bone marrow in sepsis. Regardless of these speculations, our data clearly show that LPS suppresses DAs ability to evoke erythropoiesis. Our findings demonstrate that LPS abolishes erythropoiesis and iron utilization evoked by DA and this is accompanied by a decrease in hemoglobin and red blood cell concentration. Accordingly, endotoxin induces resistance to DA with an inability to increase erythropoiesis and this could be a link to general EPO resistance in sepsis. REFERENCES
1. Corwin HL, Gettinger A, Pearl RG, Fink MP, Levy MM, Abraham E, MacIntyre NR, Shabot MM, Duh MS, Shapiro MJ. The CRIT Study: anemia and blood transfusion in the critically ill current clinical practice in the United States. Crit Care Med 2004;32:39 52 2. Smoller BR, Kruskall MS. Phlebotomy for diagnostic laboratory tests in adults. Pattern of use and effect on transfusion requirements. N Engl J Med 1986;314:12335
3. Hebert PC, Wells G, Tweeddale M, Martin C, Marshall J, Pham B, Blajchman M, Schweitzer I, Pagliarello G. Does transfusion practice affect mortality in critically ill patients? Transfusion Requirements in Critical Care (TRICC) Investigators and the Can Critical Care Trials Group. Am J Respir Crit Care Med 1997;155:1618 23 4. Nguyen BV, Bota DP, Melot C, Vincent JL. Time course of hemoglobin concentrations in nonbleeding intensive care unit patients. Crit Care Med 2003;31:406 10 5. Orkin SH, Weiss MJ. Apoptosis. Cutting red-cell production. Nature 1999;401:433, 435 6 6. Silva M, Benito A, Sanz C, Prosper F, Ekhterae D, Nunez G, Fernandez-Luna JL. Erythropoietin can induce the expression of Bcl-x(L) through Stat5 in erythropoietin-dependent progenitor cell lines. J Biol Chem 1999;274:221659 7. Socolovsky M, Constantinescu SN, Bergelson S, Sirotkin A, Lodish HF. Cytokines in hematopoiesis: specificity and redundancy in receptor function. Adv Protein Chem 1998; 52:14198 8. Wojchowski DM, Gregory RC, Miller CP, Pandit AK, Pircher TJ. Signal transduction in the erythropoietin receptor system. Exp Cell Res 1999;253:14356 9. Joy MS. Darbepoetin alfa: a novel erythropoiesis-stimulating protein. Ann Pharmacother 2002;36:118392 10. Elliott S, Lorenzini T, Asher S, Aoki K, Brankow D, Buck L, Busse L, Chang D, Fuller J, Grant J, Hernday N, Hokum M, Hu S, Knudten A, Levin N, Komorowski R, Martin F, Navarro R, Osslund T, Rogers G, Rogers N, Trail G, Egrie J. Enhancement of therapeutic protein in vivo activities through glycoengineering. Nat Biotechnol 2003;21:414 21 11. Frede S, Fandrey J, Pagel H, Hellwig T, Jelkmann W. Erythropoietin gene expression is suppressed after lipopolysaccharide or interleukin-1 beta injections in rats. Am J Physiol 1997;273: R106771 12. Krafte-Jacobs B, Bock GH. Circulating erythropoietin and interleukin-6 concentrations increase in critically ill children with sepsis and septic shock. Crit Care Med 1996;24:14559 13. Abel J, Spannbrucker N, Fandrey J, Jelkmann W. Serum erythropoietin levels in patients with sepsis and septic shock. Eur J Haematol 1996;57:359 63 14. Tamion F, Le Cam-Duchez V, Menard JF, Girault C, Coquerel A, Bonmarchand G. Erythropoietin and renin as biological markers in critically ill patients. Crit Care 2004;8:R328 R335 15. RZik S, Beguin Y. Serum soluble transferrin receptor concentration is an accurate estimate of the mass of tissue receptors. Exp Hematol 2001;29:677 85 16. Tiao G, Rafferty J, Ogle C, Fischer JE, Hasselgren PO. Detrimental effect of nitric oxide synthase inhibition during endotoxemia may be caused by high levels of tumor necrosis factor and interleukin-6. Surgery 1994;116:3327; discussion 337 8 17. Deitch EA. Animal models of sepsis and shock: a review and lessons learned. Shock 1998;9:111 18. Laterre PF, Levy H, Clermont G, Ball DE, Garg R, Nelson DR, Dhainaut JF, Angus DC. Hospital mortality and resource use in subgroups of the Recombinant Human Activated Protein C Worldwide Evaluation in Severe Sepsis (PROWESS) trial. Crit Care Med 2004;32:220718 19. Poschl JM, Leray C, Ruef P, Cazenave JP, Linderkamp O. Endotoxin binding to erythrocyte membrane and erythrocyte deformability in human sepsis and in vitro. Crit Care Med 2003;31:924 8 20. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med 2005;352:101123 21. Kemna E, Pickkers P, Nemeth E, van der Hoeven H, Swinkels D. Time-course analysis of hepcidin, serum iron, and plasma cytokine levels in humans injected with LPS. Blood 2005; 106:1864 6 22. Garcia PC, Longhi F, Branco RG, Piva JP, Lacks D, Tasker RC. Ferritin levels in children with severe sepsis and septic shock. Acta Paediatr 2007;96:1829 31 23. Kong WN, Chang YZ, Wang SM, Zhai XL, Shang JX, Li LX, Duan XL. Effect of erythropoietin on hepcidin, DMT1 with IRE, and hephaestin gene expression in duodenum of rats. J Gastroenterol 2008;43:136 43 24. Skikne BS, Cook JD. Effect of enhanced erythropoiesis on iron absorption. J Lab Clin Med 1992;120:746 51
710
25. Means RT Jr, Krantz SB. Progress in understanding the pathogenesis of the anemia of chronic disease. Blood 1992;80:1639 47 26. Gabriel A, Kozek S, Chiari A, Fitzgerald R, Grabner C, Geissler K, Zimpfer M, Stockenhuber F, Bircher NG. High-dose recombinant human erythropoietin stimulates reticulocyte production in patients with multiple organ dysfunction syndrome. J Trauma 1998;44:3617 27. van Iperen CE, Gaillard CA, Kraaijenhagen RJ, Braam BG, Marx JJ, van de Wiel A. Response of erythropoiesis and iron metabolism to recombinant human erythropoietin in intensive care unit patients. Crit Care Med 2000;28:2773 8 28. Silver M, Corwin MJ, Bazan A, Gettinger A, Enny C, Corwin HL. Efficacy of recombinant human erythropoietin in critically ill patients admitted to a long-term acute care facility: a randomized, double-blind, placebo-controlled trial. Crit Care Med 2006;34:2310 6 29. Corwin HL, Gettinger A, Pearl RG, Fink MP, Levy MM, Shapiro MJ, Corwin MJ, Colton T. Efficacy of recombinant human erythropoietin in critically ill patients: a randomized controlled trial. JAMA 2002;288:282735
30. Papadaki HA, Kritikos HD, Valatas V, Boumpas DT, Eliopoulos GD. Anemia of chronic disease in rheumatoid arthritis is associated with increased apoptosis of bone marrow erythroid cells: improvement following anti-tumor necrosis factor-alpha antibody therapy. Blood 2002;100:474 82 31. Coccia MA, Cooke K, Stoney G, Pistillo J, Del Castillo J, Duryea D, Tarpley JE, Molineux G. Novel erythropoiesis stimulating protein (darbepoetin alfa) alleviates anemia associated with chronic inflammatory disease in a rodent model. Exp Hematol 2001;29:12019 32. Silva AT, Cohen J. Role of interferon-gamma in experimental gram-negative sepsis. J Infect Dis 1992;166:3315 33. Taniguchi S, Dai CH, Price JO, Krantz SB. Interferon gamma downregulates stem cell factor and erythropoietin receptors but not insulin-like growth factor-I receptors in human erythroid colony-forming cells. Blood 1997;90:2244 52 34. Maratheftis CI, Andreakos E, Moutsopoulos HM, Voulgarelis M. Toll-like receptor-4 is up-regulated in hematopoietic progenitor cells and contributes to increased apoptosis in myelodysplastic syndromes. Clin Cancer Res 2007;13:1154 60
711
The Reduced Anticoagulant Effect of Fondaparinux at Low Antithrombin Levels

Carl-Erik Dempfle, Prof Dr med* Julia Eichner* Nenad Suvajac* Parviz Ahmad-Nejad, Dr med Michael Neumaier, Prof Dr med Martin Borggrefe, Prof Dr med*
BACKGROUND: Low antithrombin levels may compromise the anticoagulant effect of heparin and heparin-related compounds, such as fondaparinux. METHODS: We compared the anticoagulant effect of 10 concentrations of fondaparinux added to plasma samples with normal range (n 25, antithrombin 95.4% 9.2%) and low antithrombin (n 22, antithrombin 45.5% 13.2%) levels, using the Heptest coagulation assay. RESULTS: Heptest clotting time was shorter at any given fondaparinux concentration in the antithrombin-deficient samples, indicating less anticoagulant effect than in the group with normal antithrombin levels. At a high fondaparinux concentration, a saturation effect is observed with no further increase in Heptest clotting time. Addition of antithrombin concentrates results in a shift of the dose-response curve. When antithrombin concentrate was added, Heptest clotting time increased up to a fondaparinux concentration of 10 g/mL. CONCLUSIONS: In the conventional prophylactic and therapeutic dose range, not only treatment with antithrombin concentrates but also an increase in fondaparinux dose normalizes the anticoagulant effect. A saturation effect is observed at high fondaparinux concentrations. Higher levels of antithrombin lead to an exaggerated effect of fondaparinux on Heptest.
(Anesth Analg 2009;109:7126)
eparins cause a conformational change in antithrombin, increasing the inhibitory potency of this serine protease inhibitor approximately 1000-fold.1 Intensive care patients often display a diminished anticoagulant response if standard doses of heparin are used.2,3 Unfractionated heparin binds avidly to a large number of plasma proteins, some of which behave as acute phase proteins,4,5 resulting in a decreased anticoagulant response in many intensive care patients. Low antithrombin levels are a frequent finding in critically ill patients and also may result in a reduced anticoagulant effect of heparin and heparin-like compounds.6,7 This reduced heparin sensitivity has mainly been discussed concerning the use of unfractionated heparin in the context of extracorporeal circulation and renal replacement therapy, but it may also be relevant concerning treatment with fractionated heparins or heparin pentasaccharides, and for prevention of venous thromboembolism in critically ill patients. Fondaparinux is a synthetic heparin pentasaccharide with a plasma half-life of 1321 h, and minimal
From the *I. Department of Medicine, and Institute for Clinical Chemistry, University Medical Centre Mannheim, Mannheim, Germany. Accepted for publication April 17, 2009. Address correspondence and reprint requests to Carl-Erik Dempfle, I. Department of Medicine, University Medical Centre Mannheim, Theodor Kutzer Ufer 1-3, D-68167 Mannheim, Germany. Address e-mail to carl-erik.dempfle@umm.de. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae94b0
intra- and intersubject variability in dose-responses.8 Fondaparinux binds nearly exclusively to antithrombin8 and targets the inhibitory activity of antithrombin to factor Xa.9 Fondaparinux is eliminated by the kidneys.10 Conventional therapeutic plasma levels are in the range of 0.52 g/mL.8 Fondaparinux has little or no effect on activated partial thromboplastin time (aPTT) and activated clotting time (ACT), but the inhibitory effect of the fondaparinux-antithrombin complex can easily be determined by antifactor Xa assays. Fondaparinux has been approved for prophylaxis and treatment of venous thrombosis and pulmonary embolism, and various other indications. In patient populations with high risk for thrombosis, fondaparinux significantly reduces the incidence of venous thrombosis, and this reduction is not associated with an increase in bleeding complications.11,12 To our knowledge, there are no published data concerning the effect of low antithrombin levels on the anticoagulant effect of fondaparinux. Therefore, we designed in vitro experiments to determine how low antithrombin levels affect the anticoagulant activity of fondaparinux measured by Heptest. We also evaluated the effects of antithrombin supplementation on the effect of fondaparinux. Supraphysiological levels of antithrombin have been administered to septic patients for its potential antiinflammatory effect.13 We chose the Heptest coagulation assay14 as a target parameter, because this assay is sensitive to low levels of fondaparinux. In addition, Heptest can also
712
be performed easily on small laboratory coagulation analyzers, which are used in intensive care units. A special version of the Heptest, the Heptest Hi, is used for measurement of high levels of fondaparinux up to 10 g/mL.
METHODS
Plasma Samples
Residual citrated plasma from clinical routine diagnostics was used anonymously and samples were identified according to the results table from the coagulation analyzer. No patient-related data were recorded. According to the local ethics committee, no informed consent is required for this approach. Criteria for selection of samples were absence of unfractionated or low molecular weight heparin therapy indicated on the electronic laboratory request form, normal range antifactor Xa assay, indicating absence of unfractionated heparin, low molecular weight heparin and fondaparinux, normal range prothrombin time and aPTT, and normal fibrinogen level. Samples were grouped according to the antithrombin level determined by the routine chromogenic assay from DadeBehring Diagnostics, Marburg, Germany. The low antithrombin group consisted of plasma samples with an antithrombin level 60%. The normal range group consisted of 25 plasma samples and the low antithrombin group consisted of 22 plasma samples.
Fondaparinux dilution 5 L was mixed with 45 L of plasma Samples A, B, or C in coagulometer tubes. Concentrations of fondaparinux given in the Results section, Table, and Figures are final concentrations in this mixture. After 2 min of incubation at 37C, 50 L of Heptest factor Xa reagent was added. After an additional 2 min (Heptest) or 30 s (Heptest HI), 50 L of Recalmix regent from the Heptest assay kit, containing phospholipids, calcium, and coagulation factor V, was added and the coagulometer started. The coagulometer device detected clotting automatically when the motion of the steel ball immersed in the test solution was impaired by clot formation.
Statistical Analysis
Mean, standard deviation, median values, and interquartile ranges were calculated for all groups. Wilcoxons signed rank test was used for comparison of fondaparinux effects because of unequal variances. A level of P 0.05 was considered statistically significant. For comparison of the fondaparinux effect at different antithrombin concentrations, the TukeyKramer honestly significant difference test for multiple comparisons was used. An level of 0.05 was considered statistically significant.
RESULTS
The normal range antithrombin plasma samples had a mean antithrombin level of 95.4% 9.2% (mean sd) (range, 75.2%115.8%). The antithrombin deficiency group had a mean antithrombin plasma level of 45.5% 13.2% (range, 15.6% 60.0%). The normal range of Heptest without addition of fondaparinux was 18 7 s (mean sd) and the normal range of Heptest HI was 8 2 s. Increasing concentrations of fondaparinux resulted in increasing clotting time in the Heptest assay, as expected (Table 1). Heptest clotting time, though, was shorter at any given fondaparinux concentration in the antithrombin-deficient samples. Roughly identical clotting times compared with samples with normal antithrombin levels were attained by doubling the fondaparinux concentration. Differences between normal range and low range antithrombin samples were significant in the entire range of fondaparinux concentrations. A scatterplot of the slopes of log (fondaparinux concentration) versus Heptest clotting time is shown in Figure 1. The dose-response to fondaparinux (increase in Heptest clotting time per fondaparinux concentration) correlates with the antithrombin concentration (regression coefficient r 0.54815). In the low range of fondaparinux concentration, addition of antithrombin concentrate results in a shift of the dose-response curve (Fig. 2, Panels A and B), with prolongation of clotting time at higher antithrombin levels. Differences between samples without antithrombin addition and 1.25 g/mL added antithrombin (final concentration), and between 1.25 and
Assay Procedures
Heptest and Heptest HI assay kits were purchased from Kappes Laborservice, Munich, Germany. Results are given as clotting time (s). The Heptest assay kit consists of a factor Xa reagent to be added to the citrated plasma sample first, and a calcium reagent also containing phospholipids and coagulation factor V, which initiates clot formation. Heptest was measured using a ball coagulometer device from Behnk, Norderstedt, Germany. Heptest HI is a modified version of the assay optimized for high heparin concentration, as used in cardiac surgery. Antithrombin was measured by chromogenic assay from DadeBehring Diagnostics. Other coagulation assays, including prothrombin time, aPTT, and fibrinogen, were measured using reagents and methods from DadeBehring Diagnostics. All of the following experiments were performed with the 25 plasma samples with normal antithrombin concentration and the 22 plasma samples with low antithrombin concentration. Plasma (950 L) was mixed with 50 L buffer (0.05 M Tris, 0.1 M NaCl, pH 7.4) (Sample A), or 25 L buffer and 25 L antithrombin concentrate (Kybernin, CLS Behring, Marburg, Germany, dissolved in buffer at a concentration of 50 U/mL) (Sample B), or with 50 L of antithrombin concentrate (Sample C). Serial dilutions of fondaparinux were prepared using the same buffer by sequentially mixing equal volumes of fondaparinux solution with buffer.
713
Table 1. Effect of Fondaparinux on Heptest Coagulation Assay Results in Plasma Samples with Normal (AT 70%) and Low (AT 60%) Antithrombin Level AT 70% (mean sd)
AT concentration (%) Heptest (s) Fondaparinux 0.02 g/mL Fondaparinux 0.04 g/mL Fondaparinux 0.08 g/mL Fondaparinux 0.16 g/mL Fondaparinux 0.32 g/mL Heptest HI (s) Fondaparinux 0.63 g/mL Fondaparinux 1.25 g/mL Fondaparinux 2.50 g/mL Fondaparinux 5.00 g/mL Fondaparinux 10.00 g/mL 95.4 9.2 19.9 9.2 43.5 6.4 51.5 5.2 62.5 5.6 75.7 9.6 18.4 1.6 20.6 2.1 22.7 2.3 24.2 3.9 25.2 5.5
AT 60% (mean sd)

45.5 13.2 24.0 12.1 38.7 8.9 45.7 6.8 54.7 6.8 63.0 6.8 16.9 1.3 18.5 1.3 19.5 1.8 20.4 2.9 21.3 5.0
P
0.0001 0.8947 0.0402 0.0014 0.0003 0.0001 0.0008 0.0001 0.0001 0.0001 0.0001
The P values refer to differences between samples with normal and low antithrombin level. Heptest HI is a version of the Heptest assay optimized for high heparin concentration. AT antithrombin.
of 10 g/mL. Differences between neighboring fondaparinux concentrations are statistically significant at the highest antithrombin concentration for 10 vs 5 g/mL and 5 vs 2.5 g/mL in the normal plasma, and for 10 vs 5 g/mL in the antithrombin deficient plasma.
DISCUSSION
Similar to unfractionated heparin, fondaparinux is dependent upon antithrombin for its anticoagulant activity. These results indicate that low antithrombin levels influence the effect of fondaparinux measured by Heptest. The anticoagulant effect can be restored either by increasing fondaparinux concentration or by adding antithrombin. These observations are relevant for fondaparinux therapy in patients with hereditary antithrombin deficiency or in intensive care patients with low antithrombin levels. Clinical trials with in vivo application of fondaparinux are needed to establish a therapeutic range for fondaparinux in Heptest and to investigate if the anticoagulant effect of fondaparinux decreases below the therapeutic range at low antithrombin concentrations. Another important observation was that in the high concentration range, increasing the fondaparinux concentration caused only a small increase in Heptest clotting time. This is related to the saturation of binding of fondaparinux to antithrombin. Increasing antithrombin to supraphysiological levels results in an extension of the linear increase in Heptest clotting time caused by addition of fondaparinux. In this respect, fondaparinux behaves similar to unfractionated heparin. Levy et al.15 found that unfractionated heparin 4.1 U/mL failed to further increase ACT values in plasma from cardiac surgery patients. In these experiments, similar to our results with fondaparinux, antithrombin supplementation led to a further prolongation of ACT at high heparin doses.
Figure 1. Correlation between the slopes of the regression

between fondaparinux concentration and Heptest coagulation time and the antithrombin levels (low range samples, 0.02 0.31 g/mL final concentration). The dose-response to fondaparinux (increase in Heptest clotting time caused by increasing concentration of Fondaparinux) is dependent upon the antithrombin concentration of the plasma. At low antithrombin concentration, a lower slope value indicating reduced anticoagulant effect is observed.
2.5 g/mL added antithrombin (final concentration) were significant for most dilutions of fondaparinux when using normal plasma (Fig. 2, Panel A), and for some of the fondaparinux dilutions when using antithrombin-deficient plasma (Fig. 2, Panel B). In the high range of fondaparinux concentrations (Fig. 2, Panels C and D), the shift of dose-response caused by addition of the antithrombin concentrate is also visible. Depending on the available antithrombin concentration, a ceiling effect indicates saturation of the antithrombin with fondaparinux with no significant increase in Heptest clotting time with increasing fondaparinux concentration. When antithrombin concentrate is added, more binding sites for fondaparinux become available, resulting in increasing Heptest HI clotting time up to a fondaparinux concentration
714
Fondaparinux and Antithrombin Level
Figure 2. Correlation between fondaparinux final concentration and Heptest coagulation time. The conventional Heptest assay was used for the low range, Heptest HI, a modified version optimized for determination of high heparin concentration for the high concentration range. Clotting time of Heptest HI is shorter than in the conventional Heptest at similar heparin concentration. Panel A: normal plasma, fondaparinux 0.020.31 g/mL, Heptest. Panel B: antithrombin-deficient plasma, fondaparinux 0.020.31 g/mL, Heptest. Panel C: normal plasma, fondaparinux 0.62510 g/mL, Heptest HI. Panel D: antithrombin-deficient plasma, fondaparinux 0.62510 g/mL, Heptest HI. Stars indicate statistically significant differences between high and intermediate and intermediate and no antithrombin addition, using TukeyKramer analysis. An value of 0.05 was considered to be significant.
In the Kybersept trial of antithrombin concentrates in patients with severe sepsis, treatment with highdose antithrombin nearly doubled the rate of bleeding complications.13 The present results indicate that, by increasing antithrombin to supernormal levels, the effect of fondaparinux concerning the prolongation of Heptest clotting time increases. A similar effect might be the reason for the increased rate of bleeding complications in patients treated with high-dose antithrombin and conventional unfractionated or low molecular weight heparin in the Kybersept trial. The experiences from the Kybersept trial led to the recommendation not to combine high-dose antithrombin therapy with heparin.16,17 Compared with unfractionated heparin, fondaparinux has a much more predictable bioavailability.10 However, its long half-life that is prolonged with renal failure is an important limitation. Laboratory monitoring is thought to be unnecessary in most patients receiving prophylactic and therapeutic dose fondaparinux, but may be important in patients with antithrombin deficiency as well as in patients receiving antithrombin concentrates in combination with fondaparinux and in patients with impaired renal function. Determination of antithrombin levels during treatment with fondaparinux may identify patients with an expected lower response, who might benefit from
antithrombin supplementation or dose adjustment of fondaparinux. REFERENCES

1. Olson ST, Chuang YJ. Heparin activates antithrombin anticoagulant function by generating new interaction sites (exosites) for blood clotting proteinases. Trends Cardiovasc Med 2002;12:331 8 2. Mayr AJ, Dunser M, Jochberger S, Fries D, Klingler A, Joannidis M, Hasibeder W, Schobersberger W. Antifactor Xa activity in intensive care patients receiving thromboembolic prophylaxis with standard doses of enoxaparin. Thromb Res 2002;105:201 4 3. Jochberger S, Mayr V, Luckner G, Fries DR, Mayr AJ, Friesenecker BE, Lorenz I, Hasibeder WR, Ulmer H, Schobersberger W, Dunser MW. Antifactor Xa activity in critically ill patients receiving antithrombotic prophylaxis with standard dosages of certoparin: a prospective, clinical study. Crit Care 2005;9:R541 8 4. Young E, Podor TJ, Venner T, Hirsh J. Induction of the acutephase reaction increases heparin-binding proteins in plasma. Arterioscler Thromb Vasc Biol 1997;17:1568 74 5. Manson L, Weitz JI, Podor TJ, Hirsh J, Young E. The variable anticoagulant response to unfractionated heparin in vivo reflects binding to plasma proteins rather than clearance. J Lab Clin Med 1997;130:649 55 6. Avidan MS, Levy JH, Scholz J, Delphin E, Rosseel PM, Howie MB, Gratz I, Bush CR, Skubas N, Aldea GS, Licina M, Bonfiglio LJ, Kajdasz DK, Ott E, Despotis GJ. A phase III, double-blind, placebo-controlled, multicenter study on the efficacy of recombinant human antithrombin in heparinresistant patients scheduled to undergo cardiac surgery necessitating cardiopulmonary bypass. Anesthesiology 2005; 102:276 84
715
7. Avidan MS, Levy JH, van Aken H, Feneck RO, Latimer RD, Ott E, Martin E, Birnbaum DE, Bonfiglio LJ, Kajdasz DK, Despotis GJ. Recombinant human antithrombin III restores heparin responsiveness and decreases activation of coagulation in heparin-resistant patients during cardiopulmonary bypass. J Thorac Cardiovasc Surg 2005;130:10713 8. Bauer KA, Hawkins DW, Peters PC, Petitou M, Herbert JM, van Boeckel CA, Meuleman DG. Fondaparinux, a synthetic pentasaccharide: the first in a new class of antithrombotic agentsthe selective factor Xa inhibitors. Cardiovasc Drug Rev 2002;20:3752 9. Choay J, Petitou M, Lormeau JC, Sinay P, Casu B, Gatti G. Structure-activity relationship in heparin: a synthetic pentasaccharide with high affinity for antithrombin III and eliciting high anti-factor Xa activity. Biochem Biophys Res Commun 1983;116:4929 10. Samama MM, Gerotziafas GT. Evaluation of the pharmacological properties and clinical results of the synthetic pentasaccharide (fondaparinux). Thromb Res 2003;109:111 11. Bauer KA, Eriksson BI, Lassen MR, Turpie AG. Fondaparinux compared with enoxaparin for the prevention of venous thromboembolism after elective major knee surgery. N Engl J Med 2001;345:130510 12. Eriksson BI, Bauer KA, Lassen MR, Turpie AG. Fondaparinux compared with enoxaparin for the prevention of venous thromboembolism after hip-fracture surgery. N Engl J Med 2001; 345:1298 304
13. Warren BL, Eid A, Singer P, Pillay SS, Carl P, Novak I, Chalupa P, Atherstone A, Penzes I, Kubler A, Knaub S, Keinecke HO, Heinrichs H, Schindel F, Juers M, Bone RC, Opal SM. Caring for the critically ill patient. High-dose antithrombin III in severe sepsis: a randomized controlled trial. JAMA 2001;286:1869 78 14. Yin ET, Wessler S, Butler JV. Plasma heparin: a unique, practical, submicrogram-sensitive assay. J Lab Clin Med 1973; 81:298 310 15. Levy JH, Montes F, Szlam F, Hillyer CD. The in vitro effects of antithrombin III on the activated coagulation time in patients on heparin therapy. Anesth Analg 2000;90:1076 9 16. Hoffmann JN, Wiedermann CJ, Juers M, Ostermann H, Kienast J, Briegel J, Strauss R, Warren BL, Opal SM. Benefit/risk profile of high-dose antithrombin in patients with severe sepsis treated with and without concomitant heparin. Thromb Haemost 2006;95:850 6 17. Kienast J, Juers M, Wiedermann CJ, Hoffmann JN, Ostermann H, Strauss R, Keinecke HO, Warren BL, Opal SM. Treatment effects of high-dose antithrombin without concomitant heparin in patients with severe sepsis with or without disseminated intravascular coagulation. J Thromb Haemost 2006;4:90 7
716
Fondaparinux and Antithrombin Level
Echo Rounds
Aorto-Pericardial Filling Without Tamponade: An Unusual Late Bentall Complication

Paula Carmona, MD Richard Bowry, MB, BS, FRCA Robert Chen, MD, FRCPC Claude Tousignant, MD, FRCPC
fter Research Ethics Board approval, we present the case of a 61-yr-old woman who was transferred to our hospital 2 yr after a Bentall procedure and with a diagnosis of right ventricular (RV) infarction and severe RV dysfunction. She presented 10 days earlier at the referring center with progressive shortness of breath, severe weakness, nausea, and vomiting. A transthoracic echocardiogram was performed that demonstrated hypokinesis of the basal and inferior walls with RV dilation. Her serum Troponin I was elevated. The initial differential diagnosis included viral myocarditis and non-ST elevation myocardial infarction possibly because of right coronary artery (RCA) pathology. The patient subsequently developed cardiogenic shock, hepato-renal failure, and disseminated intravascular coagulation. She was tracheally intubated, given inotropes, and dialyzed. At this time she was transferred to our hospital for management and further cardiothoracic opinion. On arrival, a transesophageal echocardiogram (TEE) was performed. In the upper esophageal view at 0 using color flow Doppler, a high velocity jet was seen; it appeared to originate from the anterior aspect of the aortic graft and created a large anterior collection measuring approximately 6 8 cm (Fig. 1) (see Video 1, Supplemental Digital Content 1, showing upper esophageal, long-axis view of the ascending aortic prosthesis; a high velocity jet is seen to fill the pseudoaneurysm; http://links.lww.com/A1367). The right atrium (RA)
From the Department of Anesthesia, Saint Michaels Hospital, University of Toronto, Ontario, Canada. Accepted for publication April 9, 2009. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journals Web site (www.anesthesia-analgesia.org). Address correspondence and reprint requests to Richard Bowry, MB, BS, FRCA, Department of Anesthesia, St. Michaels Hospital, 30 Bond St., Toronto, ON, Canada M5B 1W8. Address e-mail to bowryr@smh.toronto.on.ca. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc939
and RV were both dilated and no significant RV compression was seen. In the transgastric view, there was diastolic septal flattening (see Video 2, Supplemental Digital Video 2, showing transgastric short-axis view of the left and right ventricles [RV]; The RV appears dilated with a flattened septum; http://links.lww.com/A1368) as a result of right-sided volume overload. There was also severe tricuspid regurgitation (Fig. 2). The patient was transferred to the operating room with a diagnosis of RCA button dehiscence and large anterior collection. Femoral-femoral bypass and cooling were instituted before chest opening. Intraoperative TEE (midesophageal RV outflow 24 scan angle) demonstrated blood flow between the collection and the RV outflow tract (RVOT) (Fig. 3) (see Video 3, Supplemental Digital Content, showing midesophageal 4-chamber view with scan angle of 24; a communication is seen between the pericardium and the right ventricle [RV] in the region of the RV outflow tract [RVOT]; http://links.lww.com/A1369). The operative findings were complex and included large pseudoaneurysm formed by surgical adhesions, detachment of the proximal RCA graft from the synthetic aortic prosthesis, a leak at the distal graft anastomosis, a fistula between the pseudoaneurysm and the RVOT, and a second, previously unrecognized, fistula between the pseudoaneurysm and the RA. The following procedures were performed: primary repair of the distal aortic leak and dehisced RCA, debridement and over-sewing of the RA fistula, debridement and a pericardial patch to close the fistula to the RVOT. After a protracted postoperative course, the patient was discharged home on Day 31. Bentalls procedure is a surgical option for the treatment of ascending aortic aneurysm. The overall mortality for this procedure is 5% and approaches 1.5% in elective operations. Complications include thromboembolism, endocarditis, and pseudoaneurysm particularly at the coronary ostia and distal aortic suture lines. Modifications to the original procedure have occurred to reduce tension on the
717
Figure 1. Upper esophageal, short-axis view of the ascending

aorta (Ao). There is a high velocity jet (arrow) resulting from the dehiscence of the right coronary artery (RCA) anastomosis, which communicates with a large collection (*) anterior to the ascending aorta.
Figure 3. Midesophageal four-chamber view with a scan angle of 24 showing a communication (arrow) between the pericardium (*) and the right ventricle (RV) near the right ventricular outflow tract (RVOT). The left ventricle (LV) and right atrium (RA) are also demonstrated.
Figure 2. Midesophageal four-chamber view at 0 with rightward probe rotation. A pericardial fluid collection (*) with flow within on color flow Doppler is seen lateral to the right ventricle (RV). The right atrium (RA) is grossly enlarged. There is severe tricuspid regurgitation (TR).
coronary anastamosis. Of particular interest is the Cabrol procedure that includes over-sewing the native aorta around the prosthesis and creating a shunt into the RA to drain anastomotic leakage from the periprosthetic space.1 This prevents the formation of a tense hematoma and pseudoaneurysm formation. The shunt would close spontaneously but there have been reports of persistent Cabrol shunts.2,3 To our knowledge, this is the first report of a patient surviving a significant periprosthetic pseudoaneurysm with detachment of the RCA and a distal aortic graft leak after a Bentalls procedure without a Cabrol shunt. Sakano et al. reported a similar case after Bentalls procedure with a Cabrol shunt in which the pseudoaneurysm created a fistula to the RA. The Cabrol shunt was found to be completely thrombosed.4 Hoffman et al.5 reported a similar case with a fistula from the ascending aorta entering the main
718
Echo Rounds
pulmonary artery. Late coronary graft avulsion is an unusual complication after Bentalls procedure. Intraoperative biopsies in our patient were sterile invoking the possibility of an inflammatory process. Preoperative TEE was able to outline the pathology in this case. TEE also proposed an anatomic mechanism for right-sided filling in the setting of a high-pressure pseudoaneurysm. Our patients unusual presentation of an aortic pseudoaneurysm without compression of the RV or RA suggests that the pressures in the pericardium, RV, and pseudoaneurysm were very similar. An inflammatory process likely resulted in a small dehiscence at the RCA button. The resulting jet of blood created both the pseudoaneurysm and fistula to the adjacent RVOT. It is likely that this leak increased until the RCA detached completely resulting in RV dysfunction. Fortuitously, a fistula allowed considerable return of blood to the right side of the circulation. Our images lead us to suspect that the fistula to the RVOT was the primary reentry point. This would explain the dilation of the RV and the resulting tricuspid regurgitation. A second fistula to the RA found by the surgeon had not been previously identified by TEE. The identification of a large pericardial collection in the presence of hemodynamic compromise would normally trigger a diagnosis of pericardial tamponade. Routine surgical management in this case would likely have resulted in an adverse outcome. Use of TEE identified the unusual features of RCA dehiscence, RA, and RV dilation requiring closer investigation and a cautious approach to the subsequent anesthetic and surgical management. REFERENCES
1. Leverich A, Johnston C, Stiles B, Girardi L, Hartman G, Skubas NJ. Cabrol composite graft for aortic root replacement: echocardiographic imaging. Anesth Analg 2009;108:11079
2. Rosero H, Nathan PE, Rodney E, Vasavada B, Sacchi TJ. Aorta to right atrium fistula with congestive heart failure resulting from a patent Cabrol shunt after repair of aortic dissection. Am Heart J 1994;128:608 9 3. Malcolm ID. Surgical aortic perigraft to right atrial shunt: transesophageal echocardiography findings. J Am Soc Echocardiogr 1996;9:8779
4. Sakano Y, Misawa Y, Kaminishi Y, Fuse K. Aorto-right atrium fistula caused by detachment after Bentalls operation: report of a case. Surg Today 2007;37:234 6 5. Hoffman P, Lusawa T, Ro zan ski J. Detachment of the right coronary artery resulting in subsequent aneurysm of the ascending aorta and pulmonary artery fistula 7 years after a Bentall procedure. J Am Soc Echocardiogr 2005;18:e4
Clinicians Key Teaching Points
By Drs. Nikolaos Skubas, Roman Sniecinski, and Martin J. London
The Bentall procedure involves replacement of the aortic valve, ascending aorta, and reimplantation of the right and left main coronary arteries. This case illustrates a late complication of the Bentall procedure, namely coronary graft dehiscence, detected by a high velocity jet on color ow Doppler imaging in the midesophageal ascending aorta short-axis view, that led to a large pericardial uid collection. Coronary artery dehiscence would normally lead to cardiac tamponade with collapse of low-pressure chambers (right atrium and ventricle) expected on transesophageal echocardiogram (TEE) imaging. In this unusual case, tamponade was prevented by an unsuspected stula between the right ventricle and pericardium of uncertain etiology. The TEE ndings of the blood entering into the right ventricle from outside the heart helped guide the surgical approach.
719
Case Report
Perioperative Management of a Child with von Willebrand Disease Undergoing Surgical Repair of Craniosynostosis: Looking at Unusual Targets
Isabelle Maquoi, MD* Vincent Bonhomme, MD, PhD* Jacques Daniel Born, MD, PhD Marie-Franc oise Dresse, MD, PhD Elisabeth Ronge-Collard, MD Jean-Marc Minon, MD, PhD Pol Hans, MD, PhD*
We report the successful management of a craniosynostosis repair in a child with severe Type I von Willebrand disease diagnosed during the preoperative assessment and treated by coagulation factor VIII and ristocetin cofactor. Collaboration among the anesthesiologist, the neurosurgeon, the clinical pathologist, and the pediatric hematologist is important for successful management.
(Anesth Analg 2009;109:720 4)
he perioperative management of patients undergoing surgical repair of craniosynostosis is a challenge for anesthesiologists because of the risk of extensive bleeding.1,2 We report a pediatric patient with von Willebrand disease (vWD) admitted for surgical correction of craniosynostosis.
CASE DESCRIPTION
Permission was obtained from the parents to report our observations. A 10-mo, 9-kg male child was admitted to the neurosurgical department for remodeling of sagittal craniosynostosis. At 1 mo, he had undergone an uneventful repair of an inguinal hernia. He had also been anesthetized 1 wk before admission for a computed tomography scan with three-dimensional reconstruction of his malformation. Except for the dolichocephalic aspect of the skull, the physical examination was normal. During the preoperative visit, the childs father mentioned a history of vWD in his own family although he himself had no coagulation disorder. The mother reported that the child had had recurrent knee
From the *University Department of Anesthesia and ICM, Department of Neurosurgery, University Department of Pediatrics, Liege Hemophilia Treatment Center, and Department of Laboratory Medicine, Hemostasis and Thrombosis Unit, CHR Citadelle, Liege, Belgium. Accepted for publication April 15, 2009. Supported by Department of Anesthesia and ICM of Liege University Hospital, Department of Neurosurgery of CHR Citadelle, and Department of Laboratory Medicine of CHR Citadelle. Reprints will not be available from the author. Address correspondence to Pol Hans, MD, PhD, University Department of Anesthesia and ICM, Bd du 12eme de Ligne, 1, 4000 Liege, Belgium. Address e-mail to pol.hans@chu.ulg.ac.be. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aedbf9
hematomas since he began to walk. The anesthesiologist ordered a coagulation and hemostasis profile, which showed the following results: a platelet count at 353 103 mm3, a closure time of Collagen/adenosine diphosphate PFA-100 test (Siemens, see Appendix for details) at 289 s (normal range [NR]: 71111 s), a closure time of Collagen/Epinephrine PFA-100 test at 223 s (NR: 74 116 s), an activated partial thromboplastin time (aPTT) at 37 s (NR: 28 43 s), a prothrombin time (PT) at 13.4 s with an international normalized ratio at 1.0 (NR: 1.0 1.2), a coagulation factor VIII (FVIII) at 53% (NR: 50%150%), a von Willebrand factor antigen (vWF:Ag) at 17% (NR: 60%150%), a ristocetin cofactor activity (vWF:RCo) at 9% (NR: 70%132%), and an A positive blood type. As laboratory results were consistent with a severe Type I vWD, we confirmed the diagnosis by ordering a plasma vWF multimers assay. Neurosurgery was delayed to develop a perioperative care strategy with the pediatric hematologist and the clinical pathologist. Based on our plan, the patient received 250 IU of FVIII and 550 IU of vWF:RCo (Hemate P) IV 1 h before induction of anesthesia. The same dose was repeated twice daily for the first 48 h after surgery and then modified to keep the FVIII plasma level in the 80%100% range during the first postoperative week. Premedication consisted of intrarectal midazolam (0.4 mg/kg) and atropine (0.125 mg). In the operating room, two venous catheters, an arterial catheter, and standard monitoring were placed. Anesthesia was induced IV with a bolus of 0.2 g/kg sufentanil and 5 mg/kg thiopental. Tracheal intubation was facilitated by rocuronium (0.5 mg/kg). Maintenance of anesthesia was achieved with sevoflurane (2.5%3% end-tidal) vaporized in an oxygen/nitrous oxide mixture (Fio2: 0.5). The patient received 10 mg/kg tranexamic acid IV at the beginning of surgery. Baseline fluid requirement was ensured by continuous infusion of a crystalloid (Plasmalyte-A) at a 4 mL kg1 h1 rate. The surgical intervention was uneventful and lasted 212 h. At the beginning of surgery, the hematocrit value was 27.7%. The estimated blood loss was isovolumically replaced by a colloid (third generation hydroxyethyl starch,
720
Figure 1. Perioperative factor VIII (FVIII), von Willebrand factor antigen (vWF:Ag), and ristocetin cofactor (vWF:RCo) activity measured preoperatively and during the first postoperative days. Arrows represent times of IV administration of combined FVIII and ristocetin cofactor (Hemate P). Preop preoperative; D 1, 2, 3, and 6 postoperative days 1, 2, 3, and 6, respectively.
HAES 130/0.4, Voluven). After 2 h of surgery, at the time of skin closure, the hematocrit value was 18% and the patient received a 100 mL transfusion of homologous blood. The overall estimated total blood loss during the 2 h of surgery was 250 mL. This value was calculated according to the following formula: estimated blood loss estimated blood volume (Hts Hte)/Hts, where Hts and Hte represent the hematocrit value measured immediately after the arterial line insertion and at the end of this 2-h period, respectively. After tracheal extubation in the operating room, the patient was transferred to the intensive care unit for a 24-h observation period. On admission to the intensive care unit, FVIII, vWF:Ag, and vWF:RCo values were 96.2%, 62%, and 63%, respectively. A hematocrit value of 20.3% without lactic acidosis led to a further 100 mL transfusion of red blood cells from the same donor. The day after surgery, the hematocrit value was 25.8% and the child was transferred to a pediatric ward. Administration of Hemate P produced acceptable FVIII, vWF:Ag, and vWF:RCo values during the first postoperative days (Fig. 1) and was repeated once daily at half the initial dose (250 IU of FVIII and 550 IU of vWF:RCo) for 8 days. Tranexamic acid was given IV (10 mg kg1 8h1) for 24 h and then orally (20 mg kg1 8h1) for 3 wk. No antiinflammatory drugs or acetylsalicylic acid were administered. The postoperative course was unremarkable. The child was discharged from the hospital on postoperative day 6. He was reviewed by the neurosurgeon 2 wk after hospital discharge and did perfectly well.
DISCUSSION
vWD is caused by deficiency or dysfunction of vWF, which stabilizes blood coagulation FVIII and mediates platelet plug formation through the promotion of platelet-to-platelet and platelet-to-vessel wall adhesion.3 This inherited disorder has an estimated prevalence of 0.6%1.3%3 and is classified into three categories: partial quantitative deficiency (Type 1), qualitative deficiency (Type 2), and total deficiency (Type 3).4 Classification, diagnosis approach, and
therapeutic recommendations of vWD are summarized in Figure 2. Beside inherited disorders, there are rare acquired von Willebrand syndromes associated with pathological states that include lymphoproliferative and autoimmune diseases, essential thrombocythemia, cancer, and valvular heart disease. Symptoms in patients and families with vWD may vary from minimal to severe. In addition, initial coagulation tests, such as PT and aPTT, may be normal. The aPTT is sensitive to deficiencies or defects in FVIII. Because the FVIII level is not always low in individuals with vWD, the aPTT has only little screening value for that pathology. Routine coagulation tests, which should be performed in cases of clinical suspicion, vary among surgical teams. They may consist in PT, aPTT, PFA-100, or bleeding time, and eventually FVIII, vWF:Ag, and vWF:RCo. The vWF:RCo assay is a functional assay that measures the ability of vWF to interact with normal platelets. The antibiotic ristocetin allows vWF to bind to platelet membrane glycoprotein Ib, resulting in platelet clumps.5 The time required for platelet aggregation corresponds to the amount and function of vWF in the patients plasma. Our patient exhibited severe decreases in vWF:Ag (17%) and vWF:RCo activity (9%) as well as mildly low FVIII activity (53%), which characterize Type 1 vWD. Before receiving the results of the von Willebrand multimers assay, Type 2A, B, or M could not be strictly excluded. Even though these are autosomal dominant disorders, there can be a significant degree of phenotypic variability within families. The multimer profile of our patient was normal, therefore confirming the Type 1 vWD.
721
Figure 2. Diagnostic and therapeutic approach of von Willebrand disease. All Type 2s vWD are qualitative vWF deficiencies.
Type 2A is associated with decreased platelet-dependent vWF functions and a deficiency in large multimers. Type 2B is characterized by an increased vWF affinity for platelet Ib glycoproteins and a deficiency in large multimers. Type 2M corresponds to defective platelet-dependent vWF functions that are not associated with multimer defects. In Type 2N, the binding affinity of vWF for FVIII is markedly decreased. CBC count of blood cells; aPTT activated partial thromboplastin time; PT prothrombin time; TT thrombin time; vWD von Willebrand disease; vWF:Ag von Willebrand factor antigen; vWF:RCo ristocetin cofactor; RIPA test Ristocetin-induced platelet agglutination; PFA-100 platelet function assay; FVIII coagulation factor VIII.
Individuals who have blood of type O constitutively have vWF concentrations approximately 25% less than carriers of other blood types and usually are not symptomatic. However, in case of even lower vWF levels, such patients are at high risk of having a vWD, with vWF gene mutations, significant bleeding symptoms, and a strongly positive family history. Our patient was blood type A. Strategies to prevent or control bleeding in patients with vWD include the administration of desmopressin, which stimulates the release of vWF by endothelial cells, the replacement of vWF by human plasma-derived concentrates,6 and the use of hemostatic drugs which do not modify the plasma vWF:Ag.3,7,8 Transfusion of platelet concentrates is only indicated when bleeding occurs despite normal plasma values of FVIII or as an alternative therapy in Type 3 vWD. The management of patients undergoing neurosurgery with vWD is not well documented. In general surgery, studies suggest that appropriate administration of the FVIII/vWF:RCo concentrates prevents excessive bleeding in more than 90%
722
Case Report
of patients, even in severe vWD and without any serious adverse events.7,9,10 Although desmopressin is recommended as the treatment of choice for patients with Type 1 vWD, it was not administered to our patient because of frequently reported tachyphylaxis,7,11,12 risk of hyponatremia, and lower response rate in children younger than 2-yr-of-age.1315 Tranexamic acid was administered as an adjunct to FVIII/vWF:RCo concentrates. This antifibrinolytic drug inhibits the conversion of plasminogen into plasmin and stabilizes clots that have been formed.3,7,8,12 The PFA-100 device, which measures in vitro platelet adhesion and aggregation, was not used for treatment follow-up because desmopressin or FVIII/vWF factor concentrates are not only dependent on the vWF plasma concentration but also on the profile of vWF multimers, as well as on platelet vWF. In severe vWD, PFA-100 may not normalize with vWF replacement, possibly because of abnormalities in concentrate multimer profile and/or lack of intraplatelet vWF. The PFA-100 is also prolonged by
significant reductions in platelet count or hematocrit. The normalization of the PFA-100 test after administering desmopressin in Type 1 vWD can be used for testing the efficacy of desmopressin administration. The utility of this desmopressin test in other types of the disease is questionable.16 As far as our patient received factor concentrates, measuring plasma levels of those factors was obviously the best therapeutic monitoring. FVIII activity is the main predictor of surgical hemostasis. Measuring it helps avoiding supranormal levels and hence limiting the risk of venous thromboembolism.3 The initial loading and maintenance doses of vWF concentrates used in our patient are similar to those proposed by Nichols et al.3,7,17 and others for the prevention or management of bleeding in major surgery. The replacement therapy aimed at achieving plasma activity of deficient factors between 80% and 100% for the first postoperative days and higher than 50% thereafter, for a 510-day period. These recommendations are not applicable in minor surgery. In that case, one single daily infusion is recommended for 15 days to obtain target levels higher than 50% during surgery and 30% thereafter. Venous thromboembolism has been reported in association with high levels of FVIII. This is why it is advisable to measure plasma FVIII daily to avoid levels over 100%. Antithrombotic prophylaxis with low molecular weight heparin should be considered during treatment with FVIII-vWF concentrates if there are concomitant risk factors.7 Keeping vWF:RCo levels and FVIII activity below 200% should be considered. Our patient received no prophylaxis; he had no risk factor and the levels of vWF:RCo and FVIII never exceeded acceptable limits. In this report, the estimated blood loss averaged about one third of the total blood volume and was similar to published data.2,18,19 Intravascular volume replacement was achieved using a third-generation hydroxyethyl starch (Voluven), which has less impact on hemostasis than older starches. The total volume infused in our patient was 150 mL or 16.7 mL/kg. This is far below 50 mL/kg, which is considered to be safe.20 Factor VIII level is the most important determinant of surgical bleeding in vWD patients.7,21 In our patient, FVIII activity was close to normal values during surgery and the first postoperative days. Managing these patients perioperatively requires therapy to achieve acceptable plasma concentration targets of deficient coagulation and hemostasis factors. This concept is supported by previous reports on neurosurgical interventions in hemophilic patients.22,23 In conclusion, we report the successful management of a craniosynostosis repair in a child with severe Type I vWD diagnosed during the preoperative assessment. The success of the procedure basically relied on careful communication among the different physicians managing the patient and developing a perioperative strategy based on close monitoring
and appropriate replacement of deficient hemostasis factors. APPENDIX The PFA-100 device (platelet function analyzer, Siemens) measures platelet adhesion and aggregation (primary hemostasis) in vitro. This test is clearly superior to bleeding time to assess hemostasis disturbances associated with anomalies of von Willebrand disease. It mimics an artificial vessel consisting of a sample reservoir, a capillary, and a biologically active membrane with a central aperture. The aperture is coated with collagen and adenosine diphosphate or collagen and epinephrine. The application of a constant negative pressure aspirates an anticoagulated blood sample from the reservoir through the capillary, which simulates the resistance of a small artery, and through the aperture, which simulates the injured part of the vessel wall. A platelet plug forms that gradually occludes the aperture. As a consequence, the blood flow through the aperture gradually decreases and, finally, stops. The time needed for blood flow interruption (the closure time) is recorded in seconds. REFERENCES
1. Duncan C, Richardson D, May P, Thiruchelvam J, Shong DC, Potter F, Grogan J, Caswell M. Reducing blood loss in synostosis surgery: the liverpool experience. J Craniofac Surg 2008;19:1424 30 2. Haas T, Fries D, Velik-Salchner C, Oswald E, Innerhofer P. Fibrinogen in craniosynostosis surgery. Anesth Analg 2008; 106:72531 3. Nichols WL, Hultin MB, James AH, Manco-Johnson MJ, Montgomery RR, Ortel TL, Rick ME, Sadler JE, Weinstein M, Yawn BP. von Willebrand disease (VWD): evidence-based diagnosis and management guidelines, the National Heart, Lung, and Blood Institute (NHLBI) Expert Panel report (USA). Haemophilia 2008;14:171232 4. Budde U. Diagnosis of von Willebrand disease subtypes: implications for treatment. Haemophilia 2008;14(suppl 5):2738 5. Macfarlane DE, Stibbe J, Kirby EP, Zucker MB, Grant RA, McPherson J. A method for assaying von Willebrand factor (ristocetin cofactor). Thromb Diath Haemorrh 1975;34:306 8 6. Groner A. Pathogen safety of plasma-derived products Haemate P/Humate-P. Haemophilia 2008;14(suppl 5):54 71 7. Mannucci PM. Treatment of von Willebrands disease. N Engl J Med 2004;351:68394 8. Kreuz W. von Willebrands disease: from discovery to therapy milestones in the last 25 years. Haemophilia 2008;14(suppl 5):12 9. Michiels JJ, Berneman ZN, van der PM, Schroyens W, Budde U, van Vliet HH. Bleeding prophylaxis for major surgery in patients with type 2 von Willebrand disease with an intermediate purity factor VIII-von Willebrand factor concentrate (Haemate-P). Blood Coagul Fibrinolysis 2004;15:32330 10. Thompson AR, Gill JC, Ewenstein BM, Mueller-Velten G, Schwartz BA. Successful treatment for patients with von Willebrand disease undergoing urgent surgery using factor VIII/VWF concentrate (Humate-P). Haemophilia 2004;10:4251 11. Mannucci PM, Bettega D, Cattaneo M. Patterns of development of tachyphylaxis in patients with haemophilia and von Willebrand disease after repeated doses of desmopressin (DDAVP). Br J Haematol 1992;82:8793 12. Mahdy AM, Webster NR. Perioperative systemic haemostatic agents. Br J Anaesth 2004;93:84258 13. Auerswald G, Kreuz W. Haemate P/Humate-P for the treatment of von Willebrand disease: considerations for use and clinical experience. Haemophilia 2008;14(suppl 5):39 46 14. Revel-Vilk S, Schmugge M, Carcao MD, Blanchette P, Rand ML, Blanchette VS. Desmopressin (DDAVP) responsiveness in children with von Willebrand disease. J Pediatr Hematol Oncol 2003;25:874 9
723
15. Sutor AH. DDAVP is not a panacea for children with bleeding disorders. Br J Haematol 2000;108:21727 16. van Vliet HH, Kappers-Klunne MC, Leebeek FW, Michiels JJ. PFA-100 monitoring of von Willebrand factor (VWF) responses to desmopressin (DDAVP) and factor VIII/VWF concentrate substitution in von Willebrand disease type 1 and 2. Thromb Haemost 2008;100:462 8 17. Berntorp E. Prophylaxis in von Willebrand disease. Haemophilia 2008;14(suppl 5):4753 18. Hans P, Collin V, Bonhomme V, Damas F, Born JD, Lamy M. Evaluation of acute normovolemic hemodilution for surgical repair of craniosynostosis. J Neurosurg Anesthesiol 2000;12:33 6 19. Williams GD, Ellenbogen RG, Gruss JS. Abnormal coagulation during pediatric craniofacial surgery. Pediatr Neurosurg 2001; 35:512 20. Kozek-Langenecker SA, Jungheinrich C, Sauermann W, Van der LP. The effects of hydroxyethyl starch 130/0.4 (6%) on blood loss and use of blood products in major surgery: a pooled analysis of randomized clinical trials. Anesth Analg 2008;107:38290
21. Borel-Derlon A, Federici AB, Roussel-Robert V, Goudemand J, Lee CA, Scharrer I, Rothschild C, Berntorp E, Henriet C, Tellier Z, Bridey F, Mannucci PM. Treatment of severe von Willebrand disease with a high-purity von Willebrand factor concentrate (Wilfactin): a prospective study of 50 patients. J Thromb Haemost 2007;5:111524 22. Walker JA, Dixon N, Gururangan S, Thornburg C. Perioperative factor IX replacement for surgical resection of a suprasellar astrocytoma in a child with severe haemophilia B. Haemophilia 2008;14:3879 23. Cermelj M, Negro F, Schijman E, Ferro AM, Acerenza M, Pollola J. Neurosurgical intervention in a haemophilic child with a subdural and intracerebral haematoma. Haemophilia 2004;10:4057
724
Case Report
Pediatric Anesthesiology
Section Editor: Peter J. Davis
Overweight/Obesity and Gastric Fluid Characteristics in Pediatric Day Surgery: Implications for Fasting Guidelines and Pulmonary Aspiration Risk
Scott D. Cook-Sather, MD* Paul R. Gallagher, MA Lydia E. Kruge, BA* Jonathan M. Beus, BSE* Brian P. Ciampa, BS* Kevin Conor Welch, MA* Sina Shah-Hosseini, MSE* Jieun S. Choi, MD* Reshma Pachikara, BS* Kim Minger, BSN, CNOR Ronald S. Litman, DO* Mark S. Schreiner, MD*
BACKGROUND: The safety of 2-h preoperative clear liquid fasts has not been established for overweight/obese pediatric day surgical patients. Healthy children and obese adults who fasted 2 h have small residual gastric fluid volumes (GFVs), which are thought to reflect low pulmonary aspiration risk. We sought to measure the prevalence of overweight/obesity in our day surgery population. We hypothesized that neither body mass index (BMI) percentile nor fasting duration would significantly affect GFV or gastric fluid pH. In children who were allowed clear liquids up until 2 h before surgery, we hypothesized that overweight/obese subjects would not have increased GFV over lean/normal subjects and that emesis/pulmonary aspiration events would be rare. METHODS: Demographics, medical history, height, and weight were recorded for 1000 consecutive day surgery patients aged 212 yr. In addition, 1000 day surgery patients (age 212 yr) undergoing general endotracheal anesthesia were enrolled. After tracheal intubation, a 14 18F orogastric tube was inserted and gastric contents evacuated. Medications, fasting interval, GFV, pH, and emetic episodes were documented. Age- and gender-specific Center for Disease Control and Prevention growth charts (2000) were used to determine ideal body weight (IBW 50th percentile) and to classify patients as lean/normal (BMI 25th75th percentile), overweight (BMI 85th to 95th percentile), or obese (BMI 95th percentile). RESULTS: Of all day surgery patients, 14.0% were overweight and 13.3% were obese. Obese children had lower GFV per total body weight (P 0.001). When corrected for IBW, however, volumes GFV(IBW) were identical across all BMI categories (mean 0.96 mL/kg, sd 0.71; median 0.86 mL/kg, IQR 0.96). Preoperative acetaminophen and midazolam contributed to increased GFV(IBW) (P 0.025 and P 0.001). Lower GFV(IBW) was associated with ASA physical status III (P 0.024), male gender (P 0.012), gastroesophageal reflux disease (P 0.049), and proton pump inhibitor administration (P 0.018). GFV(IBW) did not correlate with fasting duration or age. Decreased gastric fluid acidity was associated with younger age (P 0.005), increased BMI percentile (P 0.036), and African American race (P 0.033). Emesis on induction occurred in eight patients (50% of whom were obese, P 0.052, and 75% of whom had obstructive sleep apnea, P 0.061). Emesis was associated with increased ASA physical status (P 0.006) but not with fasting duration. There were no pulmonary aspiration events. CONCLUSIONS: Twenty-seven percent of pediatric day surgery patients are overweight/obese. These children may be allowed clear liquids 2 h before surgery as GFV(IBW) averages 1 mL/kg regardless of BMI and fasting interval. Rare emetic episodes were not associated with shortened fasting intervals in this population.
(Anesth Analg 2009;109:72736)
he prevalence of pediatric obesity and overweight in the United States has increased from 5% in the 1960s to 16% in the early 2000s.13 Obesity affects an increasingly large proportion of adult surgical populations
From the Departments of *Anesthesiology and Critical Care Medicine, Biostatistics and Epidemiology, and Nursing, The Childrens Hospital of Philadelphia, The University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania. Accepted for publication May 8, 2009. Supported by Childrens Anesthesia Associates, Ltd. Address correspondence and reprint requests to Scott D. CookSather, MD, Department of Anesthesiology and Critical Care Medicine, The Childrens Hospital of Philadelphia, 34th St. and Civic Center Blvd., Philadelphia, PA 19104-4399. Address e-mail to sather@email.chop.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b085ff
worldwide4,5 and, with a similar trend in pediatric populations, nearly one third of children presenting for surgery are overweight or obese.6,7 There has been concern that obesity places adults at increased risk of perioperative pulmonary aspiration.8 Overweight and obese children may also be at risk, although the nature and degree of risk have not been established.9 Because pulmonary aspiration is rare,10,11 residual gastric fluid volume (GFV) has been used as a surrogate marker for pulmonary aspiration risk in studies evaluating fasting protocol safety.1214 Studies in obese adults demonstrate acceptably low GFV after 8- to 10-h fasts,15,16 and, most recently, 2 h after ingestion of clear liquid.17 Even in obese parturients, gastric emptying of 300 mL water is not delayed beyond that of normal subjects, with gastric antral volume returning to baseline in
727
1 h.18 However, perioperative GFV after a standard 2-h clear liquid fast has not been investigated in overweight/obese pediatric populations. In this study, we sought to determine the prevalences of overweight and obesity in our pediatric day surgery population and to document subject demographics and coexisting diseases. We examined a subset of patients requiring tracheal intubation to test our hypotheses that GFV and gastric fluid pH do not change with body mass index (BMI), fasting interval, or demographics. Finally, by studying a large pediatric population with significant proportions of overweight/ obese subjects, we sought to determine the incidences of and the risk factors for emesis and pulmonary aspiration on induction of anesthesia.
narrative, and young children with bronchiolitis and/or breathing problems only with upper respiratory infections were excluded. Patients with a history of gastroesophageal reflux disease (GERD) and taking medication within 1 mo before surgery were defined as having current disease. Patients with obstructive sleep apnea (OSA) had a broad range of disease, from moderate apneic pauses to severe obstruction with arterial oxyhemoglobin desaturation documented by sleep study. Data were analyzed to determine associations between BMI categories and age, gender, race, ethnicity, and common coexisting diseases and conditions.
Part B: Gastric Fluid Characteristics

From October 2005 to May 2007, 1000 evaluable day surgery subjects were enrolled, who in addition to satisfying Part A inclusion criteria also had planned tracheal intubations. Exclusion criteria were children in whom placement of an orogastric tube could interfere with the scheduled surgery (e.g., upper endoscopy), preoperative presence of an oro/nasogastric tube, contraindication to patient repositioning or to orogastric tube placement, and preoperative use of opioids or other drugs known to slow gastrointestinal transit. Verbal permission was obtained from parents/legal guardians of the participants. Children aged 7 yr provided assent. The day before surgery, patients were instructed to follow standard fasting guidelines: no solids after midnight and clear liquids up until 2 h before hospital arrival. Recent medications and clear liquid fasting intervals were recorded as well as demographic data required for Part A. As is institutional standard of practice, oral preoperative medications were administered 1530 min before induction of anesthesia: midazolam syrup (Roxane Laboratories, Columbus, OH, 2 mg/mL) 0.5 mg/kg to a maximum of 10 mg 5 mL and concentrated acetaminophen infants drops (McNeil PPC, Fort Washington, PA, 100 mg/mL) 10 15 mg/kg to a maximum of 650 mg 6.5 mL, followed by 510 mL of apple juice. After induction of anesthesia and tracheal intubation, a 14 18F multiorificed orogastric tube (Vygon, Ecouen, France) was placed and gastric fluid was aspirated using a 60-mL catheter-tip syringe (Becton Dickinson, Franklin Lakes, NJ) while the patient was placed in the supine and left and right lateral decubitus positions. GFV was measured using the graduated syringe markings and pH was measured using colorimetric paper (Micro Essential Laboratory, Brooklyn, NY). Episodes of emesis on induction were recorded, along with contextual details regarding timing, airway difficulty, and other potentially contributing factors. The attending anesthesiologist of record made an estimate of emetic volume immediately after patient intubation and then further evacuated residual stomach contents as above. To document pulmonary aspiration, we looked for evidence of bilious secretions suctioned from the tracheal tube, particulate matter at
METHODS
IRB approval was obtained for this two-part prospective study.
Part A: Epidemiology
We examined the prevalences of overweight and obesity in our day surgery population. Subjects records were reviewed with IRB waiver of the requirement for informed consent. We screened the operating room schedule at The Childrens Hospital of Philadelphia (CHOP) using OR Manager (Picis, Wakefield, MA) for 1000 consecutive day surgery patients aged 212 yr presenting from October to November 2005 at the main hospital (Philadelphia, PA) and all CHOP ambulatory surgery facilities (Chalfont, PA; Exton, PA; Voorhees, NJ). Wheelchair-bound patients and others whose height could not be measured were excluded. Patient height, weight, age, gender, race, ethnicity, medical history, and planned surgical procedure were recorded. BMI in kg/m2 was calculated for each patient, and the 2000 Center for Disease Control and Prevention (CDC) growth charts* were used to determine BMI percentiles. For categorical comparisons, subjects were divided into lean/normal (BMI 25th75th percentiles), overweight (BMI 85th to 95th percentiles), and obese (BMI 95th percentile) groups. The interquartile lean/normal group was defined with the intention of creating a robust normative subject cohort with BMIs neither adversely affected by chronic illness and/or other unknown factors limiting growth and development nor, in balance, skewed by slightly heavier subjects with BMIs between the 75th and 85th percentiles. ChartMaxx (MedPlus, Mason, OH) and CompuRecord (Phillips Medical Systems, Bothell, WA) databases were used to obtain and verify patient information. Significant medical conditions were abstracted and prevalences calculated. Reactive airway disease (RAD) had been documented in a CompuRecord checkbox with a section for supporting
*Available at: http://www.cdc.gov/nchs/about/major/nhanes/ growthcharts/clinical_charts.htm. Accessed December 4, 2008.
728
Gastric Fluid in Obese Children
or below the vocal chords, bronchospasm, and/or persistent increased supplemental oxygen requirement. BMI and BMI category were determined as in Part A, and the 2000 CDC growth charts were used to calculate ideal body weight (IBW the 50th percentile based on gender and age) in kilograms.
Sample Size Considerations and Statistical Methods

For Part A, 1000 cases were chosen to provide reliable prevalence estimates. When the sample size is 1000, a two-sided 95% confidence interval (CI) for a single proportion using the large sample normal approximation will extend 0.019 from either side of the observed proportion (range, 0.038), when the expected proportion is 10%. If the expected proportion is 15%, then the total range of 95% CI would be 0.044. Association between categorical variables and lean/normal, overweight, and obese cohorts were examined using 2 tests. Differences in continuous variables (e.g., age and BMI) among various groupings were examined using MannWhitney or KruskalWallis tests. For Part B, 1000 patients were recruited to determine the influence of BMI on GFV. With regard to GFV (in mL) and pH and how they vary with weight and BMI, we used a 0.01 two-sided Fishers z-test of the null hypothesis that the Pearson correlation coefficient is 0.00 and had 80% power to detect a correlation of 0.11 with this sample size. Within a group of 150 obese patients, there was 80% power to detect a correlation coefficient of 0.28. The KruskalWallis test was used to explore whether GFV (in mL and mL/kg IBW) and pH are similar in lean/normal, overweight, and obese subjects who fasted 2 4 h for clear liquids. Based on pairwise comparisons between groups, a sample size of 150 in each group had 80% power to detect a probability of 0.62 that an observation in, for example, the lean/normal group was less than an observation in the obese group using a Wilcoxon (MannWhitney) rank-sum test with a 0.01 two-sided significance level. The null hypothesis was that this probability was 0.50. For a multiple linear regression model which included four covariates (fasting duration, medications, medical conditions, and surgical procedure) with a squared multiple correlation R2 of 0.05, a sample size of 1000 gave 80% power to detect at the 0.01 level of significance an increase in R2 of 0.014 because of the inclusion of two additional covariates (age and BMI). Both GFV (in mL) and gastric pH were plotted against absolute body weight (ABW) and BMI, and correlation coefficients were calculated to examine the relationships between these gastric fluid characteristics and ABW and BMI. A series of univariate analyses, including Spearman correlations and MannWhitney and KruskalWallis tests, were used to examine factors associated with GFV normalized to IBW, GFV(IBW), and gastric fluid pH. Following univariate procedures, linear regression models were used to explore potential predictors of GFV(IBW) and pH.
Among the pool of potential predictors were ASA physical status; age; BMI; gender; race/ethnicity; fasting duration; midazolam, acetaminophen, and albuterol administration; and coexisting diseases/disorders such as attention deficit disorder, GERD, RAD, and OSA. Variables related to medications, medical conditions, and surgical procedures were used in the regression models by creating dichotomous variables and/or by dummy coding. Finally, patients with emetic episodes on induction of anesthesia were compared with patients who had no emesis across a series of potential risk factors, using MannWhitney tests, 2 test, and Fishers exact tests. All data analysis was conducted using SPSS for Windows, Release 15.0, 2006 (SPSS, Chicago, IL).
RESULTS
Part A
Because of exclusion criteria (most commonly, identification of wheelchair-bound patients or subjects age older than 12 yr on day of surgery), rescheduling, booking duplication, and/or incomplete data, 1177 patients were screened to enroll 1000 evaluable subjects. A variety of surgical services were represented: otolaryngology (46.6%), general surgery (10.5%), urology (10.1%), and plastic surgery (8.3%), with combined services rendered in 2.2%. Of the evaluable subjects, 42.4% were categorized as lean/normal, 14.0% as overweight, and 13.3% as obese. Table 1 displays demographic features of the day surgery population overall. Although there were more males (59.3%) than females (40.7%), there were no gender differences across BMI categories. Racial/ethnic representation was consistent with that of Philadelphia and the surrounding Delaware Valley: Caucasian (68.5%), African American (19.1%), Hispanic (2.9%), Asian (2.0%), and other race (7.4%). Subjects identified in the latter group were often of mixed racial backgrounds. Obesity appeared with more than expected frequency in Hispanic subjects (46% vs 19% overall, P 0.048). Obese children were more likely to be older (7.1 2.8 yr) than their overweight (6.7 3.1 yr) and lean/normal (6.2 3.1 yr) counterparts (P 0.002). There were no differences in weight category distributions among the four CHOP sites. There was a significant association between ASA physical status and weight category, with subjects assigned ASA physical status III more likely to be overweight or obese (P 0.030). Common coexisting diseases and conditions across the entire population sample included RAD (22.2%), history of prematurity (15.5%), OSA (13.1%), heart disease (9.9%), and GERD (7.2%). Less common significant medical conditions classified as other (i.e., cancer, autism, renal disease, and genetic disorders) were found in 16% of the patients. Patients with RAD were significantly older (P 0.004) and had higher BMI percentiles (P 0.035) than patients without RAD, although there were no differences across BMI categories. Patients with GERD
729
Table 1. Day Surgery Demographics Total n

Part A cohort Agea (yr) Heighta (cm) Weighta (kg) BMIa (kg/m2) 1000 6.4 3.0 116.6 20.3 24.9 12.5 17.3 3.3 n (%) Gender Male Female Race/ethnicity Caucasian African American Hispanic Asian Other ASA physical status I II III
a
Lean/normal
424 (42%) 6.2 3.1 114.9 20.0 22.2 9.0 16.1 1.0 n (%)
Overweight
140 (14%) 6.7 3.1 120.8 21.4 29.5 13.4 19.0 2.1 n (%)
Obese
133 (13%) 7.1 2.8 123.5 19.8 37.3 17.2 23.0 4.3 n (%)
593 (59.3) 407 (40.7) 685 (68.5) 191 (19.1) 29 (2.9) 20 (2.0) 74 (7.4) 374 (37.4) 489 (48.9) 136 (13.6)
251 (59.2) 173 (40.8) 287 (67.7) 81 (19.1) 9 (2.1) 12 (2.8) 34 (8.0) 165 (38.9) 215 (50.7) 44 (10.4)
90 (64.3) 50 (35.7) 96 (68.6) 25 (17.9) 4 (2.9) 3 (2.1) 12 (8.6) 60 (42.9) 58 (41.4) 22 (15.7)
78 (58.7) 55 (41.4) 84 (63.2) 30 (22.6) 11 (8.3) 1 (0.8) 7 (5.3) 38 (28.6) 75 (56.4) 20 (15.0)
Age, height, weight, and body mass index (BMI) displayed by mean SD.
and OSA were younger (P 0.004 and P 0.050, respectively) but were not more likely to be overweight or obese. Neither heart disease nor a history of prematurity was associated with BMI percentile. However, patients with other medical conditions were more likely to be obese (P 0.001), and there was a trend for patients with diabetes (0.4%) to be overweight or obese (P 0.067).
Part B
To enroll 1000 evaluable subjects, 1201 day surgery patients were approached, with 136 families declining participation. Sixty-five subjects were withdrawn after enrollment, most commonly after the planned intraoperative airway management changed from tracheal intubation to laryngeal mask placement. Table 2 contains descriptive statistics for the Part B cohort. Average GFV in mL/kg of ABW, GFV(ABW), was lower for the obese cohort (0.66 0.49 mL/kg) than for both the lean/normal (0.97 0.67 mL/kg) and overweight (0.92 0.66 mL/kg) cohorts (P 0.001) (Fig. 1). However, when GFV was corrected for IBW, GFV(IBW), there was no difference between BMI categories: obese (1.03 0.75 mL/kg), overweight (1.08 0.78 mL/kg), and lean/normal (0.97 0.69 mL/kg). GFV(IBW) across all subjects averaged 0.96 0.71 mL/kg, with a median of 0.86 mL/kg and an IQR of 0.96. Of 1000 patients, only 106 (10.6%) ingested clear liquids 2 4 h before surgery. In these patients, GFV(ABW) was also lower in the obese group (0.63 0.35 mL/kg) compared with that in the lean/normal (1.07 0.63 mL/kg) and overweight (0.87 0.77 mL/kg) groups (P 0.011). Once again, when corrected for IBW, GFV did not differ among weight categories: obese (1.00 0.58 mL/kg), overweight (1.05 0.90 mL/kg), lean/normal (1.06 0.65
730
mL/kg). There was no difference in GFV(IBW) for this subset of patients who fasted 2 4 h (1.04 0.68 mL/kg) when compared with all study patients (1.01 0.72 mL/kg) who fasted for up to 24 h. Including all subjects, the average fasting duration was 9.7 5.0 h with a median of 11.5 h and an IQR of 9.5. Fasting duration did not affect gastric pH. As BMI percentile increased, however, pH increased slightly (r 0.068, P 0.036). Table 3 displays univariate analyses results involving GFV(IBW) and gastric pH. Three medications (acetaminophen, midazolam, and lansoprazole), ASA physical status III, active GERD, and gender were significant predictors of GFV(IBW). African Americans had higher gastric pH (P 0.033), and increased pH was associated with younger age (P 0.005). Notably, OSA, RAD, and the category of other significant medical conditions were not associated with GFV(IBW) or gastric pH differences. Patients who took lansoprazole (Prevacid) within 12 h of anesthetic induction had dramatically lower GFV(IBW) (r 0.89, P 0.005). Gastric pH, however, was not significantly affected by the administration of any preoperative medication recorded within 12 h of surgery. Multivariable linear regression models were examined for GFV(IBW) and gastric pH, and the resulting unstandardized regression coefficients are presented in Table 4. These regression models confirmed the associations of midazolam, lansoprazole, and gender with GFV(IBW) and the associations of African American race and age with gastric pH found in univariate analysis. In addition, BMI characteristics were found to correlate with both GFV(IBW) and gastric fluid pH, though to limited degrees. Increasing BMI percentile (noncategorical) was associated with increasing
Table 2. Part B Cohort Descriptives (N 1000) Mean sd

Age (yr) Height (cm) Weight (kg) BMI (kg/m2) BMI percentile (%) GFV(ABW) (mL/kg) GFV(IBW) (mL/kg) Gastric fluid pH 5.8 2.8 114.0 19.2 24.4 14.2 17.6 4.2 61.0 31.7 0.88 0.65 0.96 0.71 2.1 0.7 n (%) Gender Male Female Race/ethnicity African American Caucasian Other BMI percentile group Lean/normal Overweight Obese ASA physical status I II III Medical conditions OSA RAD GERD Diabetes Other disease Medications Acetaminophen No acetaminophen Midazolam No midazolam Lansoprazole Metoclopramide Omeprazole Ranitidine Other medications 578 (57.8) 422 (42.2) 304 (30.4) 574 (57.4) 122 (12.2) 386 (38.6) 146 (14.6) 178 (17.8) 291 (29.1) 665 (66.5) 44 (4.4) 384 (38.4) 237 (23.7) 25 (2.5) 1 (0.1) 108 (10.8) 967 (96.7) 33 (3.3) 975 (97.5) 25 (2.5) 4 (0.4) 2 (0.2) 6 (0.6) 2 (0.2) 125 (12.5)
Figure 1. Gastric fluid volume and body mass index. Boxplots of gastric fluid volume levels by categories based on body mass index. Sample sizes: n 386; n 146; n 178. The box represents the interquartile range containing 50% of values. The whiskers are lines that extend from the box to the highest and lowest values, excluding outliers. A line across the box indicates the median. Outliers (o) are defined as cases with values between 1.5 and 3 box lengths from either end of the box. Extreme outliers (*) are cases with values more than three box lengths from either end of the box. GFV gastric fluid volume; ABW absolute body weight; IBW ideal body weight.
BMI body mass index; GFV(ABW) gastric uid volume in mL/kg of absolute body weight; GFV(IBW) gastric uid volume in mL/kg of ideal body weight; OSA obstructive sleep apnea; RAD reactive airways disease; GERD gastroesophageal reux disease.
GFV(IBW) (coefficient 0.167, P 0.018) and BMI itself was positively correlated with pH (coefficient 0.019, P 0.001). Of the patients enrolled in Part B, eight (0.8%; CI, 3.515.7/1000) vomited during induction of anesthesia (Table 5). None had evidence of pulmonary aspiration. Emesis was statistically associated with ASA physical status II or III (P 0.006) only (Table 6). Emesis was not associated with gender, age, race, or preoperative medication. Mean fasting duration was not significantly different for patients who vomited compared with those who did not: 11.2 5.9 h vs 9.7 5.0 h, respectively. Composite volume, based on clinical estimate and subsequent GFV recovery, were within the 95% CI for average GFV(IBW). There were associative trends between emetic episodes and OSA (6 of 8 subjects, P 0.061) and between emesis and
BMI category, when comparing the lean/normal or obese groups with the overweight group (P 0.052). Six patients (75%) who had an episode of emesis were scheduled for tonsillectomy and/or adenoidectomy. Specific airway difficulties preceded emesis in three cases. Three patients experienced emetic episodes shortly after exposure to N2O for IV catheter placement or very early in the induction sequence, before loss of consciousness. Two children weighing 96 and 142 kg had planned rapid sequence inductions that included propofol (200 250 mg) and succinylcholine (100 mg), and the rest had inhaled induction of anesthesia.
DISCUSSION
Following an unequivocal trend of increasing pediatric overweight and obesity in the United States13 and other developed nations,19,20 27% of our current day surgery population has a BMI 85th percentile based on year 2000 CDC growth charts. That is, more than one quarter of our day surgery caseload has a BMI at or above what was the upper 15th percentile of the pediatric population less than a decade ago. This is consistent with recent work by Nafiu et al.6 and Tait et al.7 who found 32% of two Midwestern pediatric surgical populations to be overweight/obese. Although
731
Table 3. Part B Cohort, Univariate Analysis Results for GFV(IBW) and pH GFV(IBW) Spearman r
Age (yr) Height (cm) Weight (kg) BMI (kg/m2) BMI percentile 0.035 0.066 0.075 0.044 0.061 Mean SD Gender Male Female Race/ethnicity African American Caucasian Other BMI percentile group Lean/normal Overweight Obese ASA physical status I II III Medical conditions OSA RAD GERD Diabetes Other disease Medications Acetaminophen No acetaminophen Midazolam No midazolam Lansoprazole Metoclopramide Omeprazole Ranitidine Other 0.92 0.71 1.02 0.71 0.96 0.74 0.99 0.72 0.89 0.61 0.97 0.69 1.08 0.78 1.03 0.75 0.99 0.66 0.97 0.74 0.72 0.63 0.94 0.72 0.99 0.74 0.67 0.46 0.84 0.86 0.61 0.97 0.71 0.73 0.71 0.97 0.71 0.57 0.64 0.14 0.12 0.81 0.77 0.82 0.51 1.34 0.13 1.01 0.74
pH P
0.26 0.036 0.017 0.16 0.053 P* 0.012 0.54
Spearman r
0.090 0.080 0.048 0.063 0.068 Mean SD 2.1 0.7 2.1 0.7 2.2 0.8 2.1 0.7 2.1 0.6 2.1 0.7 2.2 0.8 2.2 0.7 2.1 0.6 2.1 0.7 2.1 0.8 2.2 0.7 2.1 0.6 2.3 0.8 2.0 2.1 0.8 2.1 0.7 2.0 0.6 2.1 0.7 1.9 0.5 2.0 0.0 3.5 2.1 2.6 1.3 2.0 0.0 2.1 0.7
P
0.005 0.013 0.14 0.053 0.036 P* 0.35 0.033
0.38
0.20
0.024
0.15
0.18 0.58 0.049 0.96 0.20 0.025 0.001 0.005 0.82 0.78 0.24 0.53
0.056 0.66 0.18 0.98 0.87 0.31 0.20 0.97 0.17 0.40 0.98 0.86
GFV(IBW) gastric uid volume in mL/kg of ideal body weight; BMI body mass index; OSA obstructive sleep apnea; RAD reactive airways disease; GERD gastroesophageal reux disease. * P values are based on MannWhitney or KruskalWallis tests.
Table 4. Part B Cohort, Regression Analysis Results for GFV(IBW) and pH GFV (IBW) C
Age (yr) BMI (kg/m2) BMI percentile Gendera African Americanb Midazolamc Lansoprazolec
pH P C
0.040 0.019
SE
SE
0.009 0.006 0.049
P
0.0005 0.001 0.006
0.167 0.089 0.379 0.837
0.071 0.045 0.143 0.353
0.018 0.049 0.135 0.008 0.018
GFV(IBW) gastric uid volume in mL/kg of ideal body weight; C unstandardized regression coefcient; BMI body mass index. a Gender: 0 male, 1 female. b Race: 0 all others, 1 African American. c Medication: 0 no, 1 yes.
overweight/obesity prevalence may be lower in some rural areas when compared with that in urban areas,19 nearly uniform BMI distributions at our main hospital
732
downtown and at each of the three suburban surgical centers suggest that the pediatric obesity epidemic does not spare suburbia.
Table 5. Emesis on Induction of Anesthesia: Case Synopses

General demographic 9 yr, 48 kg, PS 2, F, C 11 yr, 142 kg, PS 3, M, C 4 yr, 14.4 kg, PS 2, M, AA 8 yr, 24.5 kg, PS 2, M, AA 4 yr, 16.9 kg, PS 2, F, AA BMI, kg/m2 (percentile) 23.9 (97) 54.0 (100) 15.0 (31) 14.7 (19) 16.1 (73) Medical history OSA, frequent gagging/ vomiting Craniopharyngioma, hypothalamic obesity, OSA Prematurity, RAD OSA OSA Surgery T&A T&A Urethral meatoplasty T&A T&A, BMT Fast (h) 11.5 21 4 12 15.5 EV (mL) 10 50 ND ND 25 GFV (mL) 56 13 20 0 7 CV (mL/kg IBW) 2.2 1.5 ND ND 2.0 pH 2 2 2 ND 3 Anesthetic issues surrounding emetic event Moderately difficult ventilation during inhalation induction despite oral airway Vomited while awake after exposure to 70% N2O for IV catheter placement During inhalation induction During inhalation induction Crying on induction with O2/ N2O and sevoflurane Emesis before loss of consciousness Multiple intubation attempts by novice N2O for IV catheter placement Vomited while awake at the start of subsequent IV induction Airway manipulation under light plane of anesthesia Laryngospasm
4 yr, 12.5 kg, PS 3, F, H 12 yr, 96 kg, PS 2, F, C 4 yr, 19.8 kg, PS 2, F, C/AA
13.7 (7) 33.7 (99)
Cerebral palsy, seizure, OSA Prematurity, RAD
T&A Tympanoplasty
14.5 7
8 120
2 0
0.6 2.7
2 ND
18.3 (96)
OSA
Adenoidectomy, BMT
4.5
20
1.3
Fast is the time interval between last clear liquid and induction of anesthesia. Residual gastric uid volume (GFV) measured with standard gastric suction technique following emetic episode. All children had received preoperative oral acetaminophen and midazolam and are presented chronologically. C Caucasian; AA African American; H Hispanic; F female; M male; PS American Society of Anesthesiologists physical status; BMI body mass index; T&A tonsillectomy and adenoidectomy; BMT bilateral myringotomy tube placement; OSA obstructive sleep apnea; RAD reactive airways disease; IV intravenous; ND no data; EV clinician-estimated volume of vomitus; CV composite volume, estimated vomitus plus GFV expressed in mL/kg ideal body weight (IBW).
Table 6. Emesis on Induction of Anesthesia: Analyses Emesis (n 8)

Agea (yr) Heighta (cm) Weighta (kg) BMIa (kg/m2) 7.5 3.6 125.2 29.6 46.8 47.5 23.7 13.9 n (%) BMI percentile Lean/normal Overweight Obese ASA physical status I II III Medical conditions OSA RAD 2 (25.0) 0 4 (50.0) 0 6 (75.0) 2 (25.0) 6 (75.0) 2 (25.0)
No emesis (n 992)
5.8 2.8 113.9 19.1 24.3 13.5 17.6 4.1 n (%) 384 (38.7) 146 (14.7) 174 (17.5) 0.006 291 (29.2) 659 (66.4) 42 (4.2) 379 (38.2) 235 (23.5) 0.061 NS P 0.052
OSA obstructive sleep apnea; RAD reactive airways disease. a Age, height, weight, and body mass index (BMI) displayed by mean SD.
Obesity is associated with several demographic factors and coexisting conditions. Racial/ethnic backgrounds have been linked to an increased prevalence of overweight/obesity, as seen in our small subset of children with Hispanic ancestry, but with a complex interplay among genetics, culture, and socioeconomic status.21,22 We found RAD to be more common with increasing BMI, though not in the categorical overweight/obese groups. This may be a result of our restrictive definition of RAD (excluding younger children who wheeze only with upper respiratory tract infections), the observed cumulative nature of obesity, and the high prevalence of RAD in our study population. That we were unable to demonstrate increased
prevalences of specific coexisting diseases, such as RAD, hypertension, OSA, and Type II diabetes in the obese subpopulation as observed by Tait et al.,7 may be a consequence of our younger age cohort (mean 6, range, 212 yr versus mean 9, range, 218 yr) with medical conditions that have yet to develop or have yet to be definitively diagnosed. Independent of subject background and coexisting conditions, we found that the average age of overweight children was older than that of lean/normal subjects, and that of obese children older still, portraying obesity as a cumulative disease beginning in early childhood, with, as Salsberry and Reagan note,22 a clear persistence across time. With overweight/obese children comprising a significant, and still growing, proportion of surgical caseloads, anesthesiologists must revisit the fundamental assumptions informing their practice and determine whether overweight/obese subpopulations deserve special precautions. Guidelines for fasting have been liberalized across several surgical populations over the last 3 decades. However, an American Society of Anesthesiologists task force cautioned that recommended fasting guidelines, such as those for clear liquids, may not apply or may need to be modified for obese subjects.11 Vaughan et al.8 suggested that obesity increased GFV (in mL) and acidity, raising concern that obese subjects were at increased risk of pulmonary aspiration. Accumulating evidence in adults1518 suggests otherwise, and our data support this evolving viewpoint as it applies to pediatrics. Growth and development complicate GFV comparisons in children. Past studies have corrected GFV (mL/kg) using ABW, but GFV must also be considered in relation to IBW. Some dimensions of gastric size, such as fasting antral area, increase with BMI,23,24
733
but overall gastric volumes may not correlate with BMI.25 Stomach capacity in obesity would seem to be variable at best, subject to both measurement technique and additional patient factors.26 Under the conservative assumptions that obese children have neither disproportionately larger lung mass (the tissue at risk in the event of aspiration) nor stomach volume (the organ of fluid containment), but might variably ingest more clear liquids than do lean/normal subjects of the same age and gender, we corrected GFV for IBW. We then make the worst case for the presence of any GFV-associated pulmonary aspiration risk in overweight and obese subjects because GFV(IBW) is magnified with increasing BMI. Although GFV(IBW) and BMI percentile were correlated to a small degree (r 0.17, P 0.018), we found a consistent average of 1 mL/kg GFV(IBW) across all weight categories, suggesting that the absolute volumes of ingested material, accompanying oropharyngeal/esophageal/gastric secretions, and gastric emptying together correlate better with age and gender (hence IBW) than with ABW and the variable contribution of body fat. With detailed medical histories available in the electronic medical records, we searched beyond overweight/obese categories to determine whether coexisting diseases/disorders contributed to unique subpopulations affecting GFV or gastric pH. Univariate regression analysis (Table 3) revealed GERD to be strongly associated with GFV, very likely due, in part, to its medical therapy. Corroborating earlier work by Nishina et al.27 and Mikawa et al.,28 we found that decreased GFV was most dramatic in patients receiving proton pump inhibitor therapy, especially in those taking lansoprazole (Prevacid). Proton pump inhibitors decrease osmotically active H and concomitant free water entering the gastric lumen, thereby potentially reducing GFV. In the final regression analysis (Table 4), four demographic factors were found to be minimally associated with gastric fluid characteristics: gender, age, race, and BMI. Differing hormonal effects on the gastrointestinal tract may play some role in females having higher GFV(IBW) than males, but the 0.1 mL/kg average difference is not of clinical importance. Younger children and African American subjects were more likely to have increased pH, and whether developmental, diet, and/or genetic factors contribute, 95% of all subjects had pH 3, suggesting minimal clinical significance. The use of pH paper as opposed to a pH meter limited the sensitivity with which we could detect differences in gastric fluid acidity. Finally, although GFV(IBW) was positively correlated with BMI percentile and gastric pH with absolute BMI, the correlations were exceedingly small and not likely useful to the clinician assessing the potential risk of aspiration. The absence of either preoperative oral acetaminophen or midazolam administration was associated with decreased GFV. Withholding midazolam reduced
734
GFV(IBW) more on average than withholding acetaminophen (Table 3) and, of the two, only midazolam was significant in multiple regression analysis (Table 4). This may be the simple consequence of larger administered volume of midazolam (0.25 mL/kg) versus acetaminophen (0.1 0.15 mL/kg) contributing differentially to GFV when children are in a weight range requiring submaximum doses. Earlier studies have shown average GFV(ABW) (not corrected for IBW) to be 0.5 mL/kg after 2- to 3-h clear liquid fasts.1214 However, these studies predated the addition of preoperative oral acetaminophen and reformulated midazolam, which in combination are standard components of our current care. These medications may increase GFV from historical averages of 0.5 to now 1 mL/kg (IBW) by adding absolute volume, by providing an osmotically active gastric load, and by stimulating salivary and gastric secretions. Several patients vomited on induction of anesthesia (Table 5), but, consistent with the known rarity of pulmonary aspiration of gastric contents in the perioperative period,11,29 none had evidence of aspiration. We found only one factor that was statistically associated with risk of emesis, ASA physical status II or III (Table 6). Although 50% (4/8) of subjects who vomited on induction were in the obese group, obesity as a risk factor did not reach statistical significance (P 0.052) in our series. Although this finding in a small cohort may be prone to Type II error, it is consistent with the recent results of Tait et al.7 in which, in 2025 patients, one obese versus six nonobese subjects vomited during induction. We found that OSA was more likely in those who vomited (75%) than those who did not (38.2%), but with the high prevalence of OSA in our surgical population, this did not reach statistical significance (P 0.061). In evaluating risk factors by univariate analysis for a variety of critical perioperative adverse respiratory events, Tait et al.7 demonstrated that OSA was most significant, with an odds ratio of 3.54. However, a much larger study population would be required to find any association between OSA and pulmonary aspiration, a rare adverse respiratory event. Historical arguments of GERD and incompetent lower esophageal sphincters being risk factors for vomiting were not borne out in our series as no patient who vomited had current or previous GERD. Three of the eight subjects who vomited did so shortly after exposure to N2O, consistent with its known emetogenic properties. Large adult clinical trials conclusively link N2O administration and postoperative emesis,30 but there are also smaller pediatric case series that document vomiting at the outset of procedures done under N2O.31,32 The mechanisms may include direct stimulation of the vomiting center, distention of airspaces in the middle ear affecting the vestibular system, or distention of the gastrointestinal tract and concomitant visceral input to the vomiting center. Given the acute time course, however, it would
appear unlikely that increased gaseous distention of the stomach and small bowel would be significant enough to overcome lower gastroesophageal sphincter tone and result in mechanical regurgitation. Furthermore, patient anxiety and/or pain on IV catheter placement may play a role in emetic episodes during this time. From our limited data it is unclear whether obesity poses an additive risk factor for vomiting with administration of N2O. Warner et al.11 found that pulmonary aspiration was more likely to occur in patients who had airway difficulties or in patients under light planes of anesthesia. Similarly, in our study, the majority of the patients with an episode of emesis during induction had a history of OSA, were about to undergo airway surgery, and several had documented airway difficulties immediately before the event. Obese pediatric patients are more likely to have OSA33 and anatomy that may make airway management more difficult during anesthesia.6,7,34 Should OSA itself contribute to aspiration risk, however, our data demonstrate that it does not do so through increased GFV. Obese subjects may have longer induction times as anesthetic induction drugs require more time to enter and achieve sufficient concentrations in effect compartments. This may be due to airways narrowed by redundant soft tissue and hindered anesthetic gas flow and/or larger compartments to be filled. Underdosing induction drugs, such as propofol and succinylcholine, the latter of which should be dosed on the basis of actual body weight, not IBW,35 may contribute risk in some cases. Most significantly, there was no difference in fasting duration between those who vomited and those who did not, suggesting that as long as patients follow the recommended 2- to 4-h clear liquid fast, their risk for emesis and pulmonary aspiration is not increased. Prolonged fasting does not completely mitigate risk: one subject in the emesis cohort fasted 21 h before surgery.
ACKNOWLEDGMENTS The authors thank several staff who helped with this project: Lynda K. Anderson, RN; A. Michael Broennle, MD; Patricia M. Browne, MD; Sabaa Dam; Lisa M. FaziDiedrich, MD; Travis C. Foster, PhD; Elynor Furlan, RN; Michael Garafolo, RN; Linda Greim, RN; Patricia Haupt, RN; Ellen C. Jantzen, MD; George Karopovich, RN; Christina Liro, MS; Marcie Peyser-Friedman, CRNP; Leslie Plona, CRNP; Liban Rodol, MD; Linda L. Thomas, RN. The authors also thank many other house staff, anesthesiologists, nurse anesthetists, surgeons, and operating room personnel who assisted us. REFERENCES
1. Ogden CL, Flegal KM, Carroll MD, Johnson CL. Prevalence and trends in overweight among US children and adolescents, 1999 2000. JAMA 2002;288:1728 32 2. Hedley AA, Ogden CL, Johnson CL, Carroll MD, Curtin LR, Flegal KM. Prevalence of overweight and obesity among US children, adolescents, and adults, 1999 2002. JAMA 2004; 291:284750 3. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999 2004. JAMA 2006;295:1549 55 4. Dindo D, Muller MK, Weber M, Clavien PA. Obesity in general elective surgery. Lancet 2003;361:20325 5. Atkins M, White J, Ahmed K. Day surgery and body mass index: results of a national survey. Anaesthesia 2002;57:169 82 6. Nafiu OO, Ndao-Brumlay KS, Bamgbade OA, Morris M, KasaVubu JZ. Prevalence of overweight and obesity in a US pediatric surgical population. J Natl Med Assoc 2007;99:46 51 7. Tait AR, Voepel-Lewis T, Burke C, Kostrzema, Lewis I. Incidence and risk factors for perioperative adverse respiratory events in children who are obese. Anesthesiology 2008; 108:375 80 8. Vaughan RW, Bauer S, Wise L. Volume and pH of gastric juice in obese patients. Anesthesiology 1975;43:686 9 9. Setzer N, Saade E. Childhood obesity and anesthetic morbidity. Paediatr Anesth 2007;17:321 6 10. Warner MA, Warner ME, Weber JG. Clinical significance of pulmonary aspiration during the perioperative period. Anesthesiology 1993;78:56 62 11. Warner MA, Warner ME, Warner DO, Warner LO, Warner EJ. Perioperative pulmonary aspiration in infants and children. Anesthesiology 1999;90:66 71 12. Splinter WM, Stewart JA, Muir JG. Large volumes of apple juice preoperatively do not affect gastric pH and volume in children. Can J Anaesth 1990;37:36 9 13. Schreiner MS, Triebwasser A, Keon TP. Ingestion of liquids compared with preoperative fasting in pediatric outpatients. Anesthesiology 1990;72:5937 14. Cook-Sather SD, Liacouras CA, Previte JP, Markakis DA, Schreiner MS. Gastric fluid measurement by blind aspiration in paediatric patients: a gastroscopic evaluation. Can J Anaesth 1997;44:168 72 15. Harter RL, Kelly WB, Kramer MG, Perez CE, Dzwonczyk RR. A comparison of the volume and pH of gastric contents of obese and lean surgical patients. Anesth Analg 1998;86:14752 16. Jurvin P, Fevre G, Merouche M, Vallot T, Desmonts JM. Gastric residue is not more copious in obese patients. Anesth Analg 2001;93:16212 17. Maltby JR, Pytka S, Watson NC, McTaggart Cowan RA, Fick GH. Drinking 300 mL of clear fluid two hours before surgery has no effect on gastric fluid volume and pH in fasting and non-fasting obese patients. Can J Anaesth 2004;51:1115 18. Wong CA, McCarthy RJ, Fitzgerald PC, Raikoff K, Avram MJ. Gastric emptying of water in obese pregnant women at term. Anesth Analg 2007;105:7515 19. Wang Y, Lobstein T. Worldwide trends in childhood overweight and obesity. Int J Pediatr Obes 2006;1:1125 20. Kipping RR, Jago R, Lawlor DA. Obesity in children. Part 1: epidemiology, measurement, risk factors, and screening. Br Med J 2008;337:9227
CONCLUSIONS
Overweight and obese children are more than one quarter of our elective day surgery caseload. Because the preoperative GFV(IBW) approximates 1 mL/kg independent of BMI category and fasting interval 2 h, we support extending the 2 h clear liquid American Society of Anesthesiologists fasting guideline to include overweight and obese children who present for day surgery. Vomiting on induction of anesthesia in this setting is uncommon and pulmonary aspiration is rare. Emetic episodes during induction may be more common in obese patients and those with OSA, however. There is no evidence to suggest that vomiting at the beginning of elective pediatric day surgery is related to clear liquid fasting intervals beyond 2 h or to increased GFVs in obesity.
735
21. Parsons TJ, Power C, Logan S, Summerbell CD. Childhood predictors of adult obesity: a systematic review. Int J Obes Relat Metab Disord 1999;23:S1S107 22. Salsberry PJ, Reagan PB. Dynamics of early childhood overweight. Pediatrics 2005;116:1329 38 23. Chiloiro M, Caroli M, Guerra V, Lodadea Piepoli A, Riezzo G. Gastric emptying in normal weight and obese childrenan ultrasound study. Int J Obes 1999;23:1303 6 24. Kim DY, Camilleri M, Murray JA, Stephens DA, Levine JA, Burton DD. Is there a role for gastric accommodation and satiety in asymptomatic obese people? Obes Res 2001;9:655 61 25. Delgado-Aros S, Cremonini F, Castillo JE, Chial HJ, Burton DD, Ferber I, Camilleri M. Independent influences of body mass and gastric volumes on satiation in humans. Gastroenterology 2004;126:432 40 26. Geliebter A. Stomach capacity in obese individuals. Obes Res 2001;9:727 8 27. Nishina K, Mikawa K, Maekawa N, Tamada M, Obara H. Omeprazole reduces preoperative gastric fluid acidity and volume in children. Can J Anaesth 1994;41:9259 28. Mikawa K, Nishina K, Maekawa N, Asano M, Obara H. Lansoprazole reduces preoperative gastric fluid acidity and volume in children. Can J Anaesth 1995;42:46772
29. Borland LM, Sereika SM, Woelfel SK, Saitz EW, Carrillo PA, Lupin JL, Motoyama EK. Pulmonary aspiration in pediatric patients during general anesthesia: incidence and outcome. J Clin Anesth 1998;10:95102 30. Myles PS, Leslie K, Chan MTV, Forbes A, Paech MJ, Peyton P, Silbert BS, Pascoe E; the ENIGMA Trial Group. Avoidance of nitrous oxide for patients undergoing major surgery. Anesthesiology 2007;107:22131 31. Babl FE, Puspitadewi A, Barnett P, Oakley E, Spicer M. Preprocedural fasting state and adverse events in children receiving nitrous oxide for procedural sedation and analgesia. Pediatr Emerg Care 2005;21:736 43 32. Denman WT, Tuason PM, Ahmed MI, Brennen LM, Cepeda MS, Carr DB. The PediSedate device, a novel approach to pediatric sedation that provides distraction and inhaled nitrous oxide: clinical evaluation in a large case series. Paediatr Anesth 2007;17:162 6 33. Slyper AH. Childhood obesity, adipose tissue distribution, and the pediatric practitioner. Pediatrics 1998;102:e4 34. Smith HL, Meldrum DJ, Brennan LJ. Childhood obesity: a challenge for the anaesthetist? Paediatr Anaesth 2002;12:750 61 35. Rose JB, Theroux MC, Katz MS. The potency of succinylcholine in obese adolescents. Anesth Analg 2000;90:576 8
736
The Perioperative Validity of the Visual Analog Anxiety Scale in Children: A Discriminant and Useful Instrument in Routine Clinical Practice to Optimize Postoperative Pain Management
Sophie Bringuier, PharmD, PhD* Christophe Dadure, MD, MSc* Olivier Raux, MD, MSc* Amandine Dubois, MSc Marie-Christine Picot, MD, PhD Xavier Capdevila, MD, PhD
BACKGROUND: Because childrens anxiety influences pain perception, perioperative anxiety should be evaluated in clinical practice with a unique, useful, and valid tool to optimize pain management. In this study, we evaluated psychometric properties of the visual analog scale (VAS)-anxiety for children and to study its perioperative relevance in clinical practice. METHODS: One hundred children scheduled for elective surgery and general anesthesia were included. VAS-anxiety was measured at four timepoints and compared with both versions of State Spielbergers questionnaires (State-Trait Anxiety Inventory for Youth [STAIY] and State-Trait Anxiety Inventory for Children STAIC) and the modified Yale Preoperative Anxiety Scale. Childrens pain, parents anxiety, and parents proxy report of childrens anxiety were evaluated using VAS. RESULTS: The correlation between STAIC and VAS-anxiety was significant on the day of discharge. Moreover, changes over time were not significant with STAIC, whereas VAS-anxiety was significantly sensitive to changes over time in the two groups of age (711 yr and 1216 yr). A receiver operating characteristic curve, using modified Yale Preoperative Anxiety Scale as reference, determined a VAS-anxiety cutoff at 30 to identify high-anxiety groups. Pain levels were significantly higher when children were anxious (VAS 30) in the postoperative period. Moreover, childrens anxiety and pain were higher when parents were anxious. CONCLUSION: VAS-anxiety is a useful and valid tool to assess perioperative anxiety in children aged 716 yr. The influence of childrens and parents anxiety on childrens postoperative pain suggests that VAS-anxiety should be recommended routinely for postoperative clinical practice to optimize anxiety and pain management.
(Anesth Analg 2009;109:73744)
erioperative anxiety is a complex combination of fear, apprehension, and worry often accompanied by physical sensations.1 Even minor surgery can be a frightening experience for children.2 4 Anxiety influences patients subjective perceptions,1 and preoperative anxiety is associated with higher level of postoperative pain.57
From the *Department of Anesthesiology and Critical Care Medicine, Lapeyronie University Hospital; Epidemiology and Clinical Research Department, Arnaud de Villeneuve University Hospital; Developmental Psychology Department, Montpellier III University; and Department of Anesthesiology and Critical Care Medicine, University Montpellier 1 and Lapeyronie University Hospital, Montpellier, France. Accepted for publication March 17, 2009. Supported by a grant from the University Hospital Center for Clinical Research, Program 2003, Montpellier, France. Address correspondence and reprint requests to Sophie Bringuier, PharmD, PhD, Department of Anesthesiology, Acute Pain and Critical Care Medicine, Lapeyronie University Hospital, Avenue du Doyen G Giraud, Montpellier 34925, France. Address e-mail to s-bringuierbranchereau@chu-montpellier.fr. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af00e4
Interest in childrens anxiety has increased in recent years, and trials have investigated the best way to measure childrens anxiety. The standard Spielberger State-trait Anxiety Inventory (STAI)8 is used most frequently in the literature. Because, this gold standard tool is not appropriate for evaluating anxiety in a busy operating room setting,9 the modified Yale Preoperative Anxiety Scale (m-YPAS) has been developed to assess anxiety for children aged 212 yr undergoing surgery. In the postoperative period, the Post-Hospital Behavioral Questionnaire10,11 has been proven to be a valid and robust instrument for the assessment of behavioral change in children, but it is not recognized as a measure of anxiety. Anxiety evaluation is essential for the clinical perioperative follow-up of children, but the multiplicity of instruments depending on the age of the child or on the period of the evaluation is limited in clinical practice. A useful and unique anxiety instrument may have multiple benefits in following anxiety throughout the hospital stay. Furthermore, in the early postoperative period, pain can be a confusion factor for a scale based on
737
Table 1. Childrens and Parents Perioperative Scales Patients

Younger (n 43) Older (n 57) Parents
The day before surgery (D-1)

VAS-anxiety VAS-pain STAIC (state) VAS-anxiety VAS-pain STAIY (state) VAS-anxiety (self and proxy)
Before induction (D0)

VAS-anxiety VAS-pain m-YPAS VAS-anxiety VAS-pain
The day after surgery (POD1)

The day of discharge (DD)

VAS visual analog scale; STAIC State-Trait Anxiety Inventory for Children; STAIY State-Trait Anxiety Inventory for Youth; m-YPAS modied Yale Preoperative Anxiety Scale; D-1 the evening before surgery; DO the day of surgery before anesthesia; POD postoperative day one; DD day of discharge.
behavioral aspects. Self-report measurements could be a more effective solution. Crandall et al.12 have demonstrated the validity of numeric 0 10 anxiety self-report scale to measure preoperative anxiety in children 713 yr. Because anxiety evaluation is also important postoperatively to optimize childrens pain management, the aims of this study were to evaluate the perioperative validity of the visual analog scale (VAS)-anxiety in children and evaluate its clinical relevance for pain management.
METHODS
This was a longitudinal observational study. Approval was obtained from the Montpellier Institutional Review Board. Children were selected from the Childrens Hospital Surgery Center, Montpellier University Hospital (France). Children of ASA status I, II, or III, aged 716 yr who were undergoing elective surgery and general anesthesia were enrolled from September 2005 to June 2006. Patients were excluded if they had a diagnosis of mental retardation or chronic pain. Outpatient surgery was not included in the study. Study information was given the day before surgery on arrival at the hospital. Parents written and childrens oral consents were obtained, and demographic information was collected. Anxiety was selfreported using a VAS at four timepoints (Table 1): the evening before surgery (D-1), the day of surgery before anesthesia (D0), postoperative day one (POD1), and the day of discharge (DD). Parents blinded to their childs responses evaluated the childs anxiety and their own anxiety using VAS. The VAS-anxiety consists of a 100-mm horizontal line with the two end points labeled no anxiety or fear and worst possible anxiety or fear. The patient is required to show the point that corresponds to their level of anxiety at that moment. VAS has been validated for adults in a previous study.13 It is a pertinent tool to assess parents anxiety.14,15 Pain intensity was also self-reported on a 100-mm VAS (VAS-pain).16 Children also completed the state version of the STAI8,17 adapted for age. This instrument is a self-administered questionnaire with two separate 20-question rating scales to evaluate trait
738
Perioperative Validity of VAS-Anxiety
and state anxiety. Only the state items were used. The pediatric version (State-Trait Anxiety Inventory for Children STAIC) was administered to children younger than 12 yr. The youth version (State-Trait Anxiety Inventory for Youth STAIY) was administered to the older children. The scores for all items were summed to create the total score. Research assistants were told to read the STAI questionnaire to young children and explain items that the children found difficult to understand. It was not administrated in the operating room before induction (D0) because it takes more than 5 min to answer. At this time, observational state anxiety was assessed using the modified m-YPAS, which is a behavioral scale developed for assessing state preoperative anxiety in children aged 712 yr.9 It consists of 27 items in five dimensions (activity, emotional, expressivity, state of arousal, vocalization, and use of parents). The item Use of parents was not included in the total score because the parents do not enter the operating room. The score is based on the sum of partial weights for each category. The psychologist who completed the m-YPAS was trained in a previous study using videotapes of children before anesthesia. The reliability levels were excellent. This research did not modify the management of the patient in any way and did not impose a protocol of sedative premedication and/or anesthesia. Children aged 711 yr were in the youngest group, and children aged 12 yr and older were in the oldest group. The sample size was based on the correlation between childrens anxiety assessed by the m-YPAS before induction in the youngest group. Given a correlation r 0.50 between VAS-anxiety and m-YPAS, a power of 90% and a two-sided level of 0.05, 37 subjects were needed. This study also included children aged 12 years and older, so we decided to increase the total number of subjects: 100 children were needed to complete this study. Continuous data are expressed as mean sd or median for non-Gaussian variables. Categorical data are expressed as frequencies (%). Continuous variables were compared with Students t-test or the MannWhitney U-test for the non-Gaussian variables. Categorical variables were compared with the 2 test.
Table 2. Childrens Characteristics 711 years old (n 43)

Boys/girls (n) Mean age (years sd) ASA, % (n) 1 2 3 Number of surgical procedures before inclusion, % (n) 0 1 2 Type of surgery, % (n) Orthopedic Plastic Abdominal Other Hospital stay (days sd)
* P 0.05 age differs signicantly.
1216 years old (n 57)

34/23 13.97 (1.42)* 77.19 (44) 19.29 (11) 3.50 (2) 35.09 (20) 36.84 (21) 28.07 (16) 75.43 (43) 10.52 (6) 12.28 (7) 1.75 (1) 3.77 (1.76)
All subjects (n 100)

58/42 11.8 (2.7) 73 (73) 25 (25) 2 (2) 35 (35) 33 (33) 32 (32) 69 (69) 11 (11) 17 (17) 3 (3) 3.9 (1.6)
24/19 9.18 (1.32)* 67.44 (29) 32.55 (14) 0 (0) 34.88 (15) 27.91 (12) 37.21 (16) 60.46 (26) 11.62 (5) 23.25 (10) 4.65 (2) 4 (1.5)
Changes over time in the two age groups were analyzed by time group interaction in a repeatedmeasures analysis of variance. The convergent validity was evaluated by the analysis of the association of VAS-anxiety with other validated measures of the same construct (m-YPAS, STAIC, and STAIY). The discriminant validity measures that association with measures of different constructs (VAS-pain). Convergent and discriminant validity were evaluated by Pearson correlations. Because m-YPAS is a valid tool to assess preoperative anxiety in children, sensitivity, specificity, and positive and negative predictive values were examined for different cutoff points of VAS-anxiety. A receiver operating characteristic (ROC) determined VAS-anxiety cutoff using m-YPAS with 30 as the reference point of high-level anxiety.9 This cutoff has been used to differentiate groups of high anxiety (30/100) of children and parents. Pearson correlation was used to investigate the relationship among parents VAS-anxiety and both self and proxy reports of childrens anxiety. The median differences were tested by a paired comparison. To study the influence of both childrens and parents anxiety on childrens postoperative pain, groups of anxiety were determined in each period using the cutoff resulting from the ROC analysis. Postoperative pain differences between groups of anxiety were tested using Students t-test. Data were analyzed using the SAS package Version 9 software (Carry, CA).
RESULTS
One hundred children were enrolled in this study. Childrens characteristics are reported in Table 2 . No statistically significant differences were reported between the age groups. All children were able to use
VAS-anxiety. In the preoperative period, VAS-anxiety did not differ significantly among history of previous surgery, gender, age, ASA status, and type of surgery. Postoperatively, VAS-anxiety did not differ with the type of surgery (POD1: P 0.42; DD: P60). The VAS-anxiety scale changed over time from the first day of hospital stay to the DD (P 0.001) (Fig. 1). The highest anxiety level was on the day of surgery before induction when 50% of the children scored higher than 30 for their anxiety. Childrens anxiety decreased in the postoperative period but was still present. The difference between VAS before induction and the day after surgery was significant (P 0.003) (Fig. 1). The analysis by time group interaction reported that VAS-anxiety did not differ significantly between the two age groups (P 0.24) from the first day of hospitalization to the DD (Fig. 2). VAS-anxiety changed over time in the two age groups (P 0.04) (Fig. 2) Table 3 describes the concurrent validity of VASanxiety. Before induction, VAS-anxiety correlated significantly with m-YPAS (r 0.67, P 0.001) in children younger than 12 yr. We found significant correlations between VAS and STAIY on the first day of hospitalization, on the POD1, and on the DD (r 0.67, P 0.001; r 0.63, P 0.001; r 0.62, P 0.002). The correlation between VAS and STAIC was only significant on the DD (r 0.66, P 0.001). For each period, the intercept between VAS-anxiety and two versions of STAI differed significantly from 0. These significant intercepts seem to be due from the range of STAI. Moreover, changes over time were not significant in the youngest age group using STAIC (P 0.05) (Fig. 3a), whereas it was significant in the oldest group using STAIY (P 0.001) (Fig. 3b). Table 4 shows the sensitivity, specificity, and positive and negative predictive values for different cutoff points
739
Figure 1. Perioperative visual analog

scale (VAS)-anxiety: values of VASanxiety in all patients in the perioperative period; the center of the box is the median, and the box represents the 25th75th percentiles. The extended bars represent the 10th90th percentiles. The points represent values higher than the 10th90th percentiles. D-1 the day before the surgery; D0 in the operating room before induction; D1 the day after surgery; and DD the day of discharge. Changes over time *P 0.05.
Figure 2. Perioperative visual analog scale (VAS)-anxiety by group of age: VAS-anxiety in the perioperative period by age group; the center of the box is the median, and the box represents the 25th75th percentiles. The extended bars represent the 10th90th percentiles. The points represent values higher than the 10th90th percentiles. D-1 the day before the surgery; D0 in the operating room before induction; D1 the day after surgery; and DD the day of discharge. Changes over time by group of age, *P 0.05; NS, P 0.05, not clinically significant of VASanxiety by age group.
Table 3. Concurrent Validity of VAS-Anxiety Pearson correlation coefcients (r) and (R2) VAS-anxiety the day before surgery
STAIC (n 43) STAIY (n 57) m-YPAS (n 43) r 0.38 R2 0.15 r 0.67* R2 0.46
VAS-anxiety before surgery
The day after surgery (POD1)

r 0.27 R2 0.07 r 0.63* R2 0.34
The day of discharge

r 0.66* R2 0.44 r 0.62* R2 0.37
r 0.67* R2 0.46
VAS visual analog scale; STAIC State-Trait Anxiety Inventory for Children; STAIY State-Trait Anxiety Inventory for Youth; m-YPAS modied Yale Preoperative Anxiety Scale. * P 0.05 clinically signicant correlations.
of VAS-anxiety. A ROC curve was generated to determine the cutoff of VAS-anxiety (Fig. 4). The area under the curve was 0.80 (CI: 0.69 0.91, P 0.001). A score of
740
30 maximized the sum of the sensitivity and specificity and was considered clinically relevant. A VASanxiety score of 30 or more was used to identify the
dependent. a, Younger children (n 43): childrens self-reported anxiety by visual analog scale (VAS)-anxiety and by State-Trait Anxiety Inventory for Children (STAIC) in the perioperative period. *P 0.05, significant changes over time of VAS-anxiety; NS. P 0.05, not significant changes over time of STAIC. b, Older children (n 57): childrens self-reported anxiety by VAS-anxiety and by State-Trait Anxiety Inventory for Youth (STAIY) in the perioperative period. *P 0.05 significant changes over time of both VASanxiety and STAIY.
Figure 3. Changes over time of anxiety is scale
anxiety score
60 50 40 30 20 10 0
EVA STAIC State
anxiety score
60 50 40
EVA STAIY State
NS
30 20
* *
10 0
D-1
POD1
DD
D-1
POD1
DD
Table 4. Characteristics of Visual Analog Scale (VAS)-Anxiety at Different Cutoff Points

VAS cutoff Sensitivity (%) Specificity (%) Positive predictive values (%) Negative predictive values (%) 20 88 51 60 85 30 78 67 67 79 40 62 75 68 71
pain and anxiety showed that the group of anxious children had a significantly higher level of pain in the postoperative period (POD1: P 0.04; DD: P 0.01) (Fig. 7a). The childrens pain was significantly higher in the group with anxious parents on the day after surgery (POD1: P 0.02) (Fig. 7b).
high-anxiety groups of children and parents. The study of parental assessment showed a significant correlation between the childrens self-report and the parents proxy report (r 0.72, P 0.001; r 0.67, P 0.001; r 0.84, P 0.001). Intercepts when parents versus children assess the childrens anxiety differed significantly from 0 on the day before surgery (P 0.0005) and the day after surgery (P 0.0284). The difference (d 1.04; P 0.002) between the self-reports and the proxy reports is significant (Fig. 5) on the day before surgery. There was no significant correlation between parents preoperative anxiety and childs preinduction anxiety. Nevertheless, the proxy report was significantly higher when the parents were anxious (VAS 30) in the three periods (P 0.008; P 0.001; and P 0.0005) (Fig. 6). The coefficient of correlation between self-reports of childrens anxiety and self-reports of parents anxiety is not significant on the day before surgery (r 0.18, P 0.19). In the postoperative period, the coefficient of correlation is significant but low (POD1: r 0.44, P 0.002; DD: r 0.53, P 0.001). The mean differences are significant in the three periods (P 0.00001; P 0.04; and P 0.03) (Fig. 5). Moreover, the childrens anxiety levels were significantly higher when the parents were anxious in the postoperative period (POD1: P 0.002; DD: P 0.002) (Fig. 6). A significant but moderate correlation coefficient was found between pain and anxiety levels in the POD1 (r 0.37, P 0.0015) and the DD (r 0.52, P 0.001). A significant difference was found between VAS-pain and VAS-anxiety (P 0.05). Postoperative pain did not differ significantly in the preoperative anxiety groups (POD1: P 0.47; DD: P 0.61). Nevertheless, simultaneous assessment of both
DISCUSSION
This study clearly demonstrates that VAS-anxiety is a useful method and has good psychometric properties to assess childrens anxiety in perioperative periods. Although VAS-pain is the universal tool to assess pain in older children, the choice of childrens anxiety scales was limited by the period of assessment or childrens age. Today, following up childrens anxiety from the first day of hospital stay to the DD has been difficult with a unique scale. In addition, childrens anxiety studies suffered from a reduced age span. This study shows that VAS-anxiety is a valid tool and sensitive to change over time, pain, and parents anxiety. Because anxiety influences pain levels, childrens anxiety must be assessed in clinical practice.57 Optimal anxiety management requires a statistically valid perioperative scale. The first step of our validation study was to compare VAS-anxiety with the fully recognized scales. Our results provide the concurrent validity of VAS-anxiety. Most importantly, we have demonstrated that the level of anxiety is dependent on the scale but not on age. In medical literature, STAI is a self-report instrument widely used to measure anxiety. The correlations between VAS-anxiety and STAI are significant in older children but only significant in children younger than 12 yr on the DD. These results might suggest that this difference is due to the childrens age. However, VAS-anxiety is able to detect statistically important changes over time in two age groups, whereas STAIC is not sensitive enough with the youngest children. Finally, the lower correlation between STAIC and VAS-anxiety did not limit the use of VAS-anxiety but demonstrated the difficulty of using STAIC in young children.
741
Figure 4. Receiver operating characteristic (ROC) curve for diagnosis of high-anxiety group.
60
80 Children VAS-anxiety 70 60 50 40 30 20 10
* * *
*
Anxious Parents Non anxious Parents *
50 VAS anxiety
children self report 40 parents proxy report 30 parents self report
20
10
0
self proxy report self report proxy report self report proxy report
0 D-1 POD1 DD
report
Figure 5. Discordance between self-reports and proxy reports: perioperative median values of visual analog scale (VAS)-anxiety: childs self-report, parents proxy report, and parents self-report. *P 0.05 the day before the surgery (D-1). This result is consistent with the study of Poma et al.,18 which showed that VAS changed significantly but the Spilebergerquestionnaire did not. Kain et al.9 also showed in their validation study that STAIC, which requires at least 510 min to complete, is not an easy tool in a busy operating room setting. Schisler et al.19 found that young childrens vocabularies may be too limited to easily understand some items of the STAIC. Furthermore, understanding the items is more difficult with premedication or after general anesthesia. The youngest age at which to use a VAS scale is also controversial. McGrath et al.20 have shown that VAS is reliable for use by children aged 5 yr and older. However, the conceptual complexity of the VAS requires the user to translate a subjective sensory experience into a linear format. Berk21 found that the
742
D-1
POD1
DD
Figure 6. Influence of state parents anxiety on childs selfreport and parents proxy report. Perioperative median values and 75th percentiles of visual analog scale (VAS)anxiety from childrens self-reports and parents proxy reports. *P 0.05, anxious parents versus nonanxious parents.
ability to seriate does not appear until 7-years old. In our study, all children were able to use VAS-anxiety from the first day of hospitalization to the DD. Although m-YPAS9 cannot be used easily in clinical practice, its use is now frequently described in the medical literature to assess young childrens anxiety before anesthetic induction. This scale has the advantage of being used from the age of 5 yr and assessing multimodal aspects of anxious behavior. The good correlation between VAS-anxiety and m-YPAS in our study confirms the validity of VAS to detect anxiety levels before surgery. The m-YPAS
60 50
60
*
Figure 7. a, Impact of children state anxiety on

postoperative pain: median values and 75th percentiles of visual analog scale (VAS)-pain of children: anxious children versus nonanxious children, *P 0.05 on postoperative day 1 (POD1) and the day of discharge (DD). b, Impact of parents state anxiety on postoperative pain: median values and 75th percentiles of VAS-pain of children: anxious parents versus nonanxious parents, *P 0.05 on POD1 and DD.
VAS Pain
40 30 20 10 0 POD1
50
VAS Pain
40 30 20 10 0
DD
POD1
DD
Anxious Children Non Anxious Children
Anxious Parents Non Anxious Parents
scale has been chosen to study the sensitivity, specificity, and positive and negative predictive values of VAS-anxiety. The maximal sum of specificity and sensitivity showed a cutoff at 30 with 78% sensitivity and 67% specificity. In an adult study, Kindler et al.13 showed a lower sensitivity at this cutoff and preferred a low-cutoff point of VAS but accepted a higher number of false-positive scores (35.1%). We chose 30 as a threshold to detect high levels of anxiety in both children and their parents. The last point of the analysis of the concurrent validity of VAS-anxiety is the comparison between self-report and proxy report. The statistical test most frequently used to validate a new method of assessment is Pearson correlation coefficient. The majority of studies, which compare different groups of patients, is limited by this test. However, a high correlation between scores does not systematically indicate high agreement.22,23 When using the VAS scale for both children and their parents, the difference can be tested. Although childrens anxiety and parents anxiety are significantly correlated, the paired comparison indicates that parents are more anxious than children in the preoperative period. Furthermore, our results show that parents overestimate the childrens own assessment. In addition, proxy-reporting scores are higher when parents are anxious. These results prove the difficulties in everyday clinical practice of perception of anxiety and demonstrate the importance of self-reporting to assess subjective measurements. Although parents proxy reports may be a useful alternative, childrens self-reporting is preferred whenever possible. This study proves that the VASanxiety detects differences between self-reporting and proxy reporting and that the self-report scale is also preferred for young children. In addition, because pain may induce bias on behavioral assessment, self-reporting of anxiety should be the primary method in the postoperative period. An important property to validate in a new scale is also the responsiveness to change.24 Our results show that VAS-anxiety is a sensible instrument to detect changes over time for the two age groups we studied.
In addition, in agreement with Caumo et al.,4 we reported that the anxiety level is significantly more important before surgery than after. This result can be explained simply by the definition of anxiety, which is a future-oriented emotion characterized by an apprehensive anticipation threat.25 The surgical act or the anesthesia constitutes a real threat for children. Postoperative anxiety is reduced, but its assessment is still important because postoperative pain interferes in this experience. The significant correlation between VAS-pain and VAS-anxiety shows the relationship between these anxiety and pain, but the moderate correlation and the significant difference proved that children are able to differently assess anxiety and pain. Because cries and agitation in the postoperative period are common behavior of pain and anxiety, the main difficulty with proxy reporting of postoperative pain is to discriminate between pain and anxiety.26 Using VAS, children were able to differently assess pain and anxiety. This result highlights the importance of self-reporting, especially postoperatively. Our study did not prove that a high score for VAS-anxiety in the preoperative period constitutes a predictor for postoperative symptoms. However, simultaneous assessment of both pain and anxiety showed that children with a high level of anxiety had a significantly higher level of pain. This result confirms that VAS-anxiety assesses state and not trait anxiety and the relevance of anxiety measures in the postoperative period. Some authors have studied the presence of the parents to facilitate induction of anesthesia. Because parents could transfer anxiety to the child, their presence rarely seems beneficial.27 Kain et al.28 clearly reported the superiority of midazolam premedication over untrained parents. In the postoperative period, we reported that the children of the anxious parents group were significantly more anxious and that those children had the highest pain levels. In agreement with Kotiniemi et al.,11 our study indicates that childrens and their parents problems are not over when
743
they leave the hospital. We demonstrate that anxiety is almost always present on the DD and that the optimal management of postoperative symptoms cannot be reduced to pain assessment. Our validation of VASanxiety constitutes an opportunity to optimize childrens pain management, and this useful instrument should be recommended for clinical practice. Several limitations should be noted in our study. First, it would be interesting to evaluate the concurrent validity of VAS-anxiety in postanesthesia care unit, but there is no gold standard to assess anxiety in the postanesthesia care unit, and it was impossible to administer 20 items in this period. The second concern is the lack of use of the testretest reliability coefficients. This important test shows the temporal stability property of a scale. Because anxiety level can change rapidly,29 this test would not be appropriate in this context. Further, randomized trials are needed to test whether VAS-anxiety is sensitive to postoperative treatment and to analyze its clinical pertinence using confounding variables. Furthermore, further studies should include outpatient surgery. In conclusion, VAS-anxiety is a useful and valid tool to assess perioperative anxiety in children 7 yr and older. Furthermore, our study demonstrated that VAS-anxiety should also be measured in the postoperative period in routine clinical practice to optimize anxiety and pain management. REFERENCES
1. Spielberger C. Anxiety as an emotional state. Anxiety: current trends in theory and research. Vol. 1. New York: Academic Press, 1972 2. Wollin SR, Plummer JL, Owen H, Hawkins RM, Materazzo F, Morrison V. Anxiety in children having elective surgery. J Pediatr Nurs 2004;19:128 32 3. Ben-Amitay G, Kosov I, Reiss A, Toren P, Yoran-Hegesh R, Kotler M, Mozes T. Is elective surgery traumatic for children and their parents? J Paediatr Child Health 2006;42:618 24 4. Caumo W, Broenstrub JC, Fialho L, Petry SM, Brathwait O, Bandeira D, Loguercio A, Ferreira MB. Risk factors for postoperative anxiety in children. Acta Anaesthesiol Scand 2000;44:7829 5. Lamontagne LL, Hepworth JT, Salisbury MH. Anxiety and postoperative pain in children who undergo major orthopedic surgery. Appl Nurs Res 2001;14:119 24 6. de Groot KI, Boeke S, van den Berge HJ, Duivenvoorden HJ, Bonke B, Passchier J. The influence of psychological variables on postoperative anxiety and physical complaints in patients undergoing lumbar surgery. Pain 1997;69:19 25 7. Kain ZN, Mayes LC, Caldwell-Andrews AA, Karas DE, McClain BC. Preoperative anxiety, postoperative pain, and behavioral recovery in young children undergoing surgery. Pediatrics 2006;118:651 8
8. Spielberger C. Manual for the state-trait anxiety inventory (form Y). Palo Alto, CA: Consulting Psychologists Press, 1983 9. Kain ZN, Mayes LC, Cicchetti DV, Bagnall AL, Finley JD, Hofstadter MB. The yale preoperative anxiety scale: how does it compare with a gold standard? Anesth Analg 1997;85:783 8 10. Vernon DT, Schulman JL, Foley JM. Changes in childrens behavior after hospitalization. Some dimensions of response and their correlates. Am J Dis Child 1966;111:58193 11. Kotiniemi LH, Ryhanen PT, Moilanen IK. Behavioural changes in children following day-case surgery: a 4-week follow-up of 551 children. Anaesthesia 1997;52:970 6 12. Crandall M, Lammers C, Senders C, Savedra M, Braun JV. Initial validation of a numeric zero to ten scale to measure childrens state anxiety. Anesth Analg 2007;105:1250 3, table of contents 13. Kindler CH, Harms C, Amsler F, Ihde-Scholl T, Scheidegger D. The visual analog scale allows effective measurement of preoperative anxiety and detection of patients anesthetic concerns. Anesth Analg 2000;90:706 12 14. Chlan LL. Relationship between two anxiety instruments in patients receiving mechanical ventilatory support. J Adv Nurs 2004;48:4939 15. Davey HM, Barratt AL, Butow PN, Deeks JJ. A one-item question with a Likert or Visual Analog Scale adequately measured current anxiety. J Clin Epidemiol 2007;60:356 60 16. Huskisson E. Measurement of pain. Lancet 1974;2:112731 17. Spielberger C. Manual for the state-trait anxiety inventory for children. Palo Alto, CA: Consulting Psychologists Press, 1973 18. Poma SZ, Milleri S, Squassante L, Nucci G, Bani M, Perini GI, Merlo-Pich E. Characterization of a 7% carbon dioxide (CO2ation paradigm to evoke anxiety symptoms in healthy subjects. J Psychopharmacol 2005;19:494 503 19. Schisler T, Lander J, Fowler-Kerry S. Assessing childrens state anxiety. J Pain Symptom Manage 1998;16:80 6 20. McGrath P, de Veber L, Hearn M. Multidimensional pain assessment in children. Adv Pain Res Ther 1985;9:38793 21. Berk L. Child development. 3rd ed. Needham Heights, MA: Allyn and Bacon, 1994 22. Theunissen NC, Vogels TG, Koopman HM, Verrips GH, Zwinderman KA, Verloove-Vanhorick SP, Wit JM. The proxy problem: child report versus parent report in health-related quality of life research. Qual Life Res 1998;7:38797 23. Cremeens J, Eiser C, Blades M. Factors influencing agreement between child self-report and parent proxy-reports on the Pediatric Quality of Life Inventory 4.0 (PedsQL) generic core scales. Health Qual Life Outcomes 2006;4:58 24. Hays RD, Hadorn D. Responsiveness to change: an aspect of validity, not a separate dimension. Qual Life Res 1992;1:735 25. Rhudy JL, Meagher MW. Negative affect: effects on an evaluative measure of human pain. Pain 2003;104:61726 26. von Baeyer CL, Spagrud LJ. Systematic review of observational (behavioral) measures of pain for children and adolescents aged 3 to 18 years. Pain 2007;127:140 50 27. Lerman J. Anxiolysis by the parent or for the parent? Anesthesiology 2000;92:9257 28. Kain ZN, Mayes LC, Wang SM, Caramico LA, Hofstadter MB. Parental presence during induction of anesthesia versus sedative premedication: which intervention is more effective? Anesthesiology 1998;89:114756 29. Kain ZN, Caldwell-Andrews AA, Mayes LC, Weinberg ME, Wang SM, MacLaren JE, Blount RL. Family-centered preparation for surgery improves perioperative outcomes in children: a randomized controlled trial. Anesthesiology 2007;106:6574
744
A Comparison of Dexmedetomidine with Propofol for Magnetic Resonance Imaging Sleep Studies in Children
Mohamed Mahmoud, MD* Joel Gunter, MD* Lane F. Donnelly, MD Yu Wang, MS Todd G. Nick, PhD Senthilkumar Sadhasivam, MD, MPH*
BACKGROUND: Magnetic resonance imaging (MRI) sleep studies can be used to guide management of children with obstructive sleep apnea (OSA) refractory to conservative therapy. Because children with OSA are sensitive to the respiratory-depressant effects of sedatives and anesthetics, provision of anesthesia for imaging studies in this patient population can be challenging. Dexmedetomidine has been shown to have pharmacological properties simulating natural sleep with minimal respiratory depression. We hypothesized that, compared with propofol, dexmedetomidine would have less effect on upper airway tone and airway collapsibility, provide more favorable conditions during dynamic MRI airway imaging in children with OSA, have fewer scan interruptions, and require less aggressive airway interventions. METHODS: In this retrospective descriptive study, we reviewed the records of 52 children receiving dexmedetomidine and 30 children receiving propofol for anesthesia during MRI sleep studies between July 2006 and March 2008. Documentation of the severity of OSA by overnight polysomnography was available for 67 of the 82 subjects, who were analyzed separately. Data analyzed included demographics, severity of OSA, comorbidities, hemodynamic changes, use of artificial airways, additional airway maneuvers, and successful completion of the MRI scan. RESULTS: Demographics, OSA severity by polysomnography, anesthetic induction, and baseline hemodynamics were comparable in both groups. An interpretable MRI sleep study was obtained for 98% of children in the dexmedetomidine group and 100% in the propofol group. Of 82 children, MRI sleep studies were successfully completed without the use of artificial airways in 46 children (88.5%) in the dexmedetomidine group versus 21 children (70%) in the propofol group (P 0.03). An artificial airway was required to complete the study in five children (12%) in the dexmedetomidine group versus nine children (35%) in the propofol group (P 0.06). Additional airway maneuvers (chin lift and shoulder roll) were required to complete the study in one child (2%) in the dexmedetomidine group and three children (10%) in the propofol group (P 0.14). Children in the dexmedetomidine group experienced reductions in heart rate, whereas those in the propofol group experienced reductions in arterial blood pressure; these reductions were statistically, but not clinically, significant. CONCLUSIONS: Dexmedetomidine provided an acceptable level of anesthesia for MRI sleep studies in children with OSA, producing a high yield of interpretable studies of the patients native airway. The need for artificial airway support during the MRI sleep study was significantly less with dexmedetomidine than with propofol. Dexmedetomidine may be the preferred drug for anesthesia during MRI sleep studies in children with a history of severe OSA and may offer benefits to children with sleep-disordered breathing requiring anesthesia or anesthesia for other diagnostic imaging studies.
(Anesth Analg 2009;109:74553)
pper airway obstruction can result from decreased airway muscle tone during anesthesia or physiologic sleep. Examination of patterns of dynamic airway collapse in children with obstructive sleep apnea (OSA) during sleep permits identification of underlying anatomic causes of airway obstruction and facilitates planning for medical and
From the Departments of *Anesthesiology and Radiology; and Division of Biostatistics and Epidemiology, Cincinnati Childrens Hospital Medical Center, Cincinnati, Ohio. Accepted for publication April 3, 2009. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journals Web site (www.anesthesia-analgesia.org). Vol. 109, No. 3, September 2009
surgical treatments required to relieve airway obstruction. Magnetic resonance imaging (MRI) sleep studies have been reported to successfully depict airway motion abnormalities that are related to OSA.19 The ideal MRI sleep study would be of the sleeping patients native airway free of artificial airways during spontaneous ventilation.
Address correspondence and reprint requests to Mohamed Mahmoud, MD, Department of Clinical Anesthesia and Pediatrics, Cincinnati Childrens Hospital Medical Center, 3333 Burnet Ave., MLC 2001, Cincinnati, OH 45229. Address e-mail to mohamed. mahmoud@cchmc.org. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc506
745
Children with OSA are sensitive to the respiratorydepressant effects of sedative and hypnotic drugs and are especially vulnerable to the development of upper airway obstruction during anesthesia and sedation.10 Providing anesthesia that mimics physiologic sleep without the use of an artificial airway in children with OSA is a challenge but is critical for accurate interpretation of the resulting scans. In contrast to other sedative drugs, dexmedetomidine has been shown to have sedative properties that parallel natural sleep, without significant respiratory depression.1113 These two advantages make dexmedetomidine an attractive drug for sedating children with OSA for MRI sleep studies. In this report, we review our clinical experiences with dexmedetomidine as the primary sedative in a series of children with OSA undergoing MRI sleep studies and compare the results with a contemporaneous series of children sedated with propofol.
METHODS
After institutional review board approval, the medical records of 82 consecutive children receiving dexmedetomidine (52 children) or propofol (30 children) anesthesia during a MRI sleep study at Cincinnati Childrens Hospital Medical Center between July 2006 and March 2008 were reviewed retrospectively. Candidate patients for review were identified from departmental anesthesia records by procedure code. Indications for MRI sleep study included persistent OSA despite surgical intervention (including previous tonsillectomy and adenoidectomy), OSA in conjunction with an underlying syndrome predisposing the patient to multilevel obstruction (such as craniofacial anomaly or trisomy 21), evaluation of supraglottic anatomy for children with OSA before anticipated complex airway surgery, and OSA associated with obesity.1,8,14 Documentation of OSA by overnight polysomnography performed at Cincinnati Childrens Hospital Medical Center was available for 41 of 52 children in the dexmedetomidine group and 26 of 30 children in the propofol group. The remaining children had either failed surgical management for OSA by clinical criteria or had polysomnography performed at another institution. Polysomnography reports were reviewed for the minimum oxygen saturation during the study, the respiratory disturbance index (the number of apneas with both central and obstructive [reduction of airflow of 90% or more] components) and hypoapneic episodes (reduction of airflow of 50% or more associated with an arousal or decrease in oxygen saturation per hour of sleep) and the obstructive index (number of obstructive apneas per hour of sleep) and for the severity of OSA (normal obstructive index 1; mild OSA obstructive index 15; moderate OSA obstructive index 510; and severe OSA obstructive index 10).
746
Dexmedetomidine Versus Propofol for Pediatric Sleep Studies
All cine MRI studies were performed on a 1.5 Tesla MRI scanner (General Electric Medical Systems, Milwaukee, WI) according to our current standard protocols. The airway was imaged from above the level of the adenoids to the level of the mid-trachea using a headneck vascular coil and 3-dimensional localization. Sagittal and axial T1-weighted spin echo images (TR 400, TE minimal, field of view 22 cm, slice thickness 4 mm, gap 1 mm, matrix 256 192, number of excitations 2), as well as axial and sagittal fast spin echo inversion recovery (TR 5000, TE 34, echo train length 12, field of view 22 cm, slice thickness 6 mm, gap 2 mm, matrix 256 192, number of excitations 2) images were obtained. MRI cine sequences, generated using a fast gradient echo sequence, were obtained in the midline sagittal location and in the axial plane at the level of the mid-portion of the tongue (flip angle 80 degrees, TR 8.2 S, TE 3.6 S, slice thickness 12 mm). Cine images were acquired from a single location approximately at 1 s intervals (about 128 consecutive images obtained during the nominal 2-min imaging time). The sagittal or axial images were then displayed in cine format, forming a real-time movie of airway motion (see Video 1, Supplemental Digital Content 1: midline sagittal cine image shows intermittent and repetitive collapse of the retroglossal airway, http://links.lww.com/A1370). An interpretable MRI sleep study was defined as a study in which all image sequences in protocol were obtained without motion artifact, and sagittal and axial cine MRIs were obtained without any artificial airway. In a majority of children in both groups, insertion of an IV catheter was facilitated by inhalation of sevoflurane and/or nitrous oxide. In the dexmedetomidine group, four patients received IM ketamine and another three children received topical EMLA to facilitate IV insertion. Three children in the propofol group received EMLA to facilitate IV insertion. Anesthesia was initiated by a bolus dose followed by an infusion of dexmedetomidine or propofol. Once an adequate level of anesthesia was achieved, nasal cannulae with a sample port for end-tidal carbon dioxide analysis were applied and the children were positioned supine in the head-and-neck vascular coil with the cervical spine in the neutral position and permitted to breathe spontaneously. With the exception of temperature, standard physiologic variables (electrocardiogram, arterial oxygen saturation, capnography, and arterial blood pressure) were monitored throughout the scanning procedure and recorded at 5-min interval on the anesthetic record. If patients moved during the study, an additional bolus dose of dexmedetomidine or propofol was administered; infusion rates for dexmedetomidine and propofol were adjusted at the discretion of the attending anesthesiologist based on subjects response. All anesthetic records were handwritten. All airway maneuvers recorded are standard practice for these
Table 1. Demographic Characteristics All subjectsa Characteristic

Age (yr) Weight (kg) Male, N (%) ASA physical status, N (%) I or II III Comorbidity, N (%) Trisomy 21, N (%) OSA severity, N (%) Mild Moderate Severe Room air Spo2 nadir (%)
Subjects with polysomnography study available Propofol (N 30) Dexmedetomidine (N 41)

10.0 (6.0, 13.0) 36.0 (23.0, 59.0) 30 (73) 31 (76) 10 (24) 31 (76) 19 (46) 16 (39) 11 (27) 14 (34) 87 (84, 91)
Dexmedetomidine (N 52)
11.0 (6.0, 15.0) 37.0 (23.0, 60.0) 34 (65) 39 (75) 13 (25) 39 (75) 23 (44) 16 (31) 11 (21) 14 (27) 87 (84, 91)
Propofol (N 26)
9.5 (5.0, 14.0) 26.0 (20.0, 47.0) 20 (77) 20 (77) 6 (23) 24 (92) 14 (54) 8 (31) 9 (35) 9 (35) 88 (84, 93)
9.5 (5.0, 14.0) 30.5 (20.0, 49.3) 23 (77) 24 (80) 6 (20) 28 (93) 17 (57) 10 (33) 8 (27) 8 (27) 88 (84, 93)
For continuous measures, 50% (25%, 75%) percentiles are reported. For categorical measures, frequencies (%) are reported. OSA obstructive sleep apnea. a Of all 82 patients, 15 subjects did not have documentation of OSA grade by polysomnography: 11 subjects were in the dexmedetomidine group, 4 subjects were in the propofol group.
specialized MRI sleep studies in our institution. We did confirm the presence of an artificial airway from two sources, the anesthesia record and the MRI sleep study report (in their report radiologists consistently comment on the presence or absence of artificial airway and the reason for placement). To minimize selection bias and the impact of other changes in management on the outcome variables and to compensate for the retrospective, nonrandomized nature of our study, the dexmedetomidine and propofol groups included all children presenting for MRI sleep studies during a defined and relatively short interval. Supplemental oxygen was administered via the nasal cannula to maintain adequate arterial oxygen saturation. Insertion of an artificial airway (either an oral airway or a nasal trumpet), placement of shoulder roll, or taping the chin to manage airway obstruction and oxygen desaturation was performed at the discretion of the attending anesthesiologist. Artificial airways were removed during the 2-min period when the sagittal cine images were acquired. Studies were terminated if the child was unable to maintain adequate arterial oxygen saturation, as determined by the clinical judgment of the attending anesthesiologist after removal of the artificial airway. At the completion of imaging, the dexmedetomidine or propofol infusion was discontinued and patients were transferred to the recovery room. Patients were discharged after meeting standard discharge criteria, including level of consciousness (awake or easy arousal with verbal commands), core temperature 36C, ability to swallow (taking oral fluids), adequacy of muscle strength (strong and close to baseline movements of extremities and head), and status consistent with the patients preoperative baseline level of function. The following data were recorded: age, weight, gender, relevant comorbidities (such as trisomy 21),
ASA physical status, severity of OSA on polysomnography (normal, mild moderate, or severe), lowest arterial oxygen saturation recorded during polysomnography, anesthetic induction drug, total dexmedetomidine or propofol bolus dose, and infusion rates of dexmedetomidine and propofol. The primary outcomes were successful completion of the MRI scan with or without an artificial airway or other airway intervention and need for an artificial airway or other airway intervention. Children requiring airway interventions (shoulder roll or chin lift) were still classified as native airways with successful completion of the MRI because radiologists comment on the presence or absence of artificial airway. This information was documented only in the anesthesia record. Secondary outcomes were hemodynamic changes during anesthesia, recovery time, and complications related to anesthesia. Continuous data are presented as mean (sd) or median (interquartile range [IQR]), depending on whether or not the data were normally distributed; categorical data are presented by frequency. Continuous measures were analyzed by two-sample t-tests or Wilcoxons ranked sum test as appropriate; categorical measures were analyzed by 2 tests or Fishers exact tests as appropriate. Changes in heart rate and arterial blood pressure were analyzed with paired t-tests. Results were considered statistically significant for P 0.05.
RESULTS
Polysomnographic sleep studies documenting the severity of OSA were available for 67 of the 82 children in the sample. Demographic characteristics including age, weight, gender, ASA physical status, comorbidities, and severity of OSA by polysomnography were comparable between the groups, both overall and for the
747
Table 2. Airway Interventions During MRI Sleep Study All subjects Dexmedetomidine (N 52)
Artificial airway N (%) None Intermittent Chin lift or shoulder roll, N (%) Any airway intervention, N (%) Interpretable MRI scan, N (%) With native airway, N (%) With artificial airway, N (%)
Fishers exact test was used to calculate the P values. MRI magnetic resonance imaging.
Subjects with polysomnography study available P

0.03
Propofol (N 30)
21 (70) 9 (30) 3 (10) 12 (40) 30 (100) 21 (70) 9 (30)
Dexmedetomidine (N 41)
36 (88) 5 (12) 1 (2) 6 (15) 40 (98) 35 (85) 5 (12)
Propofol (N 26)
17 (65) 9 (35) 3 (12) 12 (46) 26 (100) 17 (65) 9 (35)
P
0.04 0.29 0.01 1 0.06
47 (90) 5 (10) 1 (2) 6 (12) 51 (98) 46 (88) 5 (10)
0.14 0.005 1 0.03
subgroups with available polysomnographic sleep studies (Table 1). The most frequent comorbidity in both groups was trisomy 21. Additional comorbidities in the dexmedetomidine group included hypothyroidism (n 3), tracheomalacia (n 3), obesity (n 2), pulmonary hypertension (n 2), scoliosis (n 1), Angelman, Stickler, or Barnes syndrome (n 1 each), and Pierre Robin sequence (n 1). Additional comorbidities in the propofol group included obesity (n 1), scoliosis (n 1), Pierre Robin sequence (n 2), hypothyroidism (n 1), achondroplasia (n 1), Moyamoya disease (n 1), and Rubinstein-Taybi syndrome (n 1). All subjects in both groups had undergone previous adenotonsillectomy in an attempt to address OSA. Additional previous airway surgical procedures in the dexmedetomidine group included genioglossus advancement (n 2), palatoplasty (n 3), and supraglottoplasty (n 1). Additional previous airway surgical procedures in the propofol group included uvulopalatoplasty (n 1), lingual tonsillectomy (n 1), pharyngeal flap (n 3), cleft palate repair (n 2), and tongue base reduction (n 1). The median (IQR) bolus doses of dexmedetomidine and propofol were 2 (2, 2) g/kg and 1 (1, 2) mg/kg, respectively. The median (IQR) infusion rates were dexmedetomidine 2 (2, 2) g kg1 h1 and propofol 200 (150, 200) g kg1 min1. Atropine was administered to five patients (12%) in the dexmedetomidine group and one patient (4%) in the propofol group (P 0.39). Primary outcomes by treatment group are shown in Table 2. Outcomes were similar for the complete sample and for the subset with OSA confirmed by polysomnography; further discussion of primary outcomes will be restricted to the subset of subjects whose severity of OSA had been documented on polysomnography. The MRI study was completed and an interpretable MRI scan was generated in 40 of 41 children in the dexmedetomidine group (98%) and 26 of 26 children in the propofol group (100%). A larger proportion of children sedated with dexmedetomidine were able to complete the MRI scan without an
748
artificial airway compared with those sedated with propofol, although this difference achieved only borderline statistical significance (P 0.06). Children sedated with propofol were more likely to require an artificial airway during the scan compared with those sedated with dexmedetomidine (P 0.04). Overall, almost half of the children receiving propofol required an airway intervention of some sort during the MRI scan, compared with only about one sixth of those receiving dexmedetomidine (P 0.01). Subjects were stratified by the severity of OSA and the need for an artificial airway was compared by strata for both treatment groups (Table 3). Measures of the severity of OSA were comparable for each treatment group within each strata. Significantly more children with severe OSA in the propofol group (56%) required an artificial airway during the MRI sleep study compared with those in the dexmedetomidine group (7%, P 0.02). The one subject in the dexmedetomidine group who failed to complete the MRI sleep study had a history of severe OSA and experienced significant oxygen desaturation despite administration of supplemental oxygen and use of artificial airways. Two subjects in the dexmedetomidine group required additional bolus doses because of movement during the examination; none of the subjects in the propofol group required either a bolus dose or an increase in the infusion rate during the examination. No complications during recovery were noted in either group. All patients were discharged the day of the examination after they met discharge criteria; no significant complications were identified on follow-up telephone communications the day after the examination. Hemodynamic effects were observed with both dexmedetomidine and propofol (Fig. 1). Compared with their baseline heart rate (99 22 bpm), children receiving dexmedetomidine experienced a significant reduction in heart rate (78 18 bpm) after receiving the dexmedetomidine bolus (Fig. 1A; P 0.0001). Subjects in the propofol group did not demonstrate any changes in heart rate after the propofol bolus (Fig.
Table 3. Requirement for Articial Airway by Severity of OSA as Documented by Polysomnography OSA severity
Mild Obstructive index (events/h) Respiratory disturbance index (events/h) Needed artificial airway, N (%) Room air Spo2 nadir (%) Moderate Obstructive index (events/h) Respiratory disturbance index (events/h) Needed artificial airway, N (%) Room air Spo2 nadir (%) Severe Obstructive index (events/h) Respiratory disturbance index (events/h) Needed artificial airway, N (%) Room air Spo2 nadir (%)
Dexmedetomidine
N 16 2.7 1.9 3.6 1.9 2 (13) 91 (89, 94) N 11 10.2 5.8 11.0 5.8 2 (18) 86 (85, 87) N 14 21.8 11.3 23.8 11.2 1 (7) 84 (77, 88)
Propofol
N8 3.1 1.3 4.4 1.7 1 (13) 92 (88, 96) N9 8.8 3.8 10.9 4.3 3 (33) 86 (84, 89) N9 23.6 13.5 24.9 13.1 5 (56) 86 (82, 92)
P
0.53* 0.30* 1 0.54 0.54* 0.96* 0.62 0.91 0.74* 0.83* 0.02 0.45
The obstructive index is the number of obstructive apneas per hour of sleep. The respiratory disturbance index is the number of apneas including both central and obstructive components and hypoapneic episodes per hour of sleep. The severity of OSA was dened as follows normal obstructive index 1; mild OSA obstructive index 15; moderate OSA obstructive index 510; and severe OSA obstructive index 10. The obstructive index and the respiratory disturbance index values were reported as mean SD. Need for articial airway and successful MRI completions were reported by N (percentages). Room air SpO2 nadir from preprocedure polysomnography was reported as 50% (25%, 75%) percentiles. OSA obstructive sleep apnea; MRI magnetic resonance imaging. * Two-sample t-test was used. Fishers exact test was used. Wilcoxons rank sum test was used. P 0.05 was considered statistically signicant.
1A). There were no significant changes in systolic and diastolic blood pressures from baseline after dexmedetomidine administration (Fig. 1B and C). In contrast, children in the propofol group experienced significant reductions in both systolic and diastolic blood pressure after the propofol bolus; these changes were particularly dramatic for diastolic blood pressure (44 8 mm Hg after propofol bolus versus 65 18 mm Hg at baseline; Fig. 1B and C; P 0.0001). None of the hemodynamic changes were considered to be of sufficient magnitude to necessitate interrupting or terminating the imaging study. The duration of stay in the recovery room was significantly shorter in the propofol group (28 12 min) compared with the dexmedetomidine group (69 37 min, P 0.0001).
DISCUSSION
In our series, interpretable sleep study MRI scans and cine s were obtained for virtually all subjects, whether they were sedated with dexmedetomidine or propofol. However, artificial airway placement or positioning aids were required in a significantly smaller proportion of subjects sedated with dexmedetomidine (12%) than with propofol (35%). These benefits were seen primarily in children with severe OSA; there appeared to be no relative advantages for either drug in children with mild OSA symptoms. Despite the high-risk nature of the OSA population, artificial airways were required by only 12% of patients sedated with dexmedetomidine. Airway obstruction in childhood OSA can be due to a combination of structural obstruction and dynamic airway collapsibility.15,16 The decrease in airway muscular tone associated with natural sleep and sedation, as well as the supine position, lead to changes in the
anatomic positioning of structures that surround the airway.1719 These changes are more pronounced in children with OSA and often they demonstrate increased dynamic airway collapse, contributing to the clinical findings in OSA.20 Precise localization of the sites of upper airway occlusion during sleep has been attempted by a variety of techniques, including pharyngoscopy, fluoroscopy, pressure measurements, computed axial tomography, and MRI. MRI with cine sequences has become a useful tool to identify the anatomic and dynamic causes of persistent OSA.1,2123 MRI sleep studies are noninvasive and allow the whole airway to be visualized at once with high contrast resolution, demonstrating both static and dynamic abnormalities that lead to functional collapse of the airway.21,24 31 Identification of the site(s) of upper airway obstruction in patients with OSA aids in surgical planning,32 and dynamic imaging studies, such as cine MRI, have been shown to affect management decisions in more than 50% of cases of OSA.21,33 Because symptoms occur only during sleep in children with OSA, demonstration of airway collapse/ obstruction on MRI requires that the child be either asleep or sedated. Although it would be ideal to perform these studies during natural sleep, performing the studies during natural sleep is highly impractical, inefficient, and typically unsuccessful; many, perhaps most, children will not tolerate the MRI environment well enough to fall asleep and the loud noise of the gradient-echo sequences used to create the cine MRI images typically awakens those who do fall asleep.14 Upper airway collapsibility is markedly increased in both sleeping and anesthetized children.17,34 Anesthesia for children undergoing MRI sleep studies can often be a challenge because children with significant OSA are sensitive to all sedative and anesthetic drugs,
749
Figure 1. Baseline and postbolus hemodynamics. Closed circles represent dexmedetomidine group and open circles represent propofol group. In the x axis, the first data point is the mean baseline hemodynamic value and the second data point is the mean hemodynamic value after the drug bolus. (A) Comparison of baseline and postdrug heart rates: baseline mean heart rate in beats per minute (bpm) in the dexmedetomidine group (99 22 bpm) and the propofol group (95 21 bpm) were comparable (P 0.47). Compared with baseline, use of propofol did not change mean heart rate significantly but use of dexmedetomidine decreased mean heart rate (P 0.0001). Postdrug mean heart rate in the dexmedetomidine group (78 18 bpm) was significantly lower than in the propofol group (91 12 bpm; P 0.0005). (B) Comparison of baseline and postdrug systolic blood pressures: baseline mean systolic blood pressure in the dexmedetomidine group (118 15 mm Hg) and the propofol group (111 30 mm Hg) was comparable (P 0.28). Compared with baseline, use of propofol (P 0.03) decreased mean systolic blood pressure and the reduction after dexmedetomidine was not statistically significant (P 0.06). Postdrug mean systolic blood pressure in the propofol group (96 14 mm Hg) was significantly lower than the dexmedetomidine group (107 19 mm Hg; P 0.007). (C) Comparison of baseline and postdrug diastolic blood pressures: baseline mean diastolic blood pressure (in mm Hg) in the dexmedetomidine group (66 12 mm Hg) and the propofol group (65 18 mm Hg) was comparable (P 0.83). Compared with baseline, use of propofol (P 0.0001) significantly decreased mean diastolic blood pressure. Postdrug mean diastolic blood pressure in the propofol group (44 8 mm Hg) was significantly lower than the dexmedetomidine group (59 15 mm Hg; P 0.0001). and upper airway obstruction or respiratory depression can occur even at minimal levels of sedation.10 Sedatives and anesthetics commonly used in children for MRI sleep studies include pentobarbital, propofol, and benzodiazepines. Propofol and barbiturates can exacerbate upper airway obstruction and increase the risk of respiratory depression and/or apnea.10 Benzodiazepines have relaxant effects on the pharyngeal musculature, causing a reduction of the pharyngeal space.35 In contrast to the previously mentioned drugs, ketamine has been shown to preserve the hypopharyngeal size in adults.36 In an anecdotal report, a combination of ketamine 1 mg/kg and dexmedetomidine 1
750
g/kg followed by a dexmedetomidine infusion of 1 g kg1 h1 was effective in providing anesthesia without exacerbating respiratory problems during MRI in three children with trisomy 21 and OSA.37 The ideal anesthetic for MRI sleep studies would provide reliable anesthesia under near-physiologic sleep conditions, avoid airway collapse, obstruction and respiratory depression necessitating the use of artificial airways, have a predictable duration of action, and exhibit minimal cardiovascular and other side effects. Because of its sedative and anxiolytic properties, dexmedetomidine has been shown to be a useful drug for pediatric procedural sedation.38 44
Sedation with dexmedetomidine has properties that parallel natural sleep,1113 and patients clinically sedated with dexmedetomidine are still easily arousable, an effect not observed with other available sedatives.45 It, thus, seemed to us that dexmedetomidine had many of the characteristics of an ideal sedative for MRI sleep studies. After we began using dexmedetomidine to provide anesthesia during MRI sleep studies, it was our clinical impression that patients experienced fewer episodes of desaturation and airway obstruction than with pentobarbital and propofol, the drugs we had previously used, enabling us to complete the studies with fewer interruptions, especially in patients with severe OSA. Our study finding of reduced need for artificial airways in the dexmedetomidine group compared with the propofol group (Tables 2 and 3) supports our hypothesis of better preservation of upper airway tone and minimal airway collapsibility with dexmedetomidine anesthesia compared with propofol. The minimal respiratory depression associated with dexmedetomidine alone is probably not sufficient to explain the observed improved conditions during MRI sleep studies, because spontaneous respiration is also maintained in patients sedated with pentobarbital or propofol. We hypothesize that the relaxant effects of dexmedetomidine on upper airway muscle tone are similar to those seen with natural sleep and less than those of other sedatives or anesthetics for any given level of anesthesia. We are unaware of any systematic investigations of the effects of dexmedetomidine on upper airway tone or caliber and suggest that this would be an interesting subject for future investigations. The configurational changes leading to obstruction in the upper airway during anesthesia with propofol in children have been studied. Evans et al.46 showed that increasing depth of propofol anesthesia in children is associated with upper airway narrowing throughout the entire upper airway, most pronounced in the hypopharynx at the level of epiglottis. Litman et al.46,47 concluded that the dimensions of the upper airways of children change shape significantly on awakening from propofol anesthesia, and most children had the narrowest point in the pharynx at the level of the soft palate. Although the severity of obstructive apnea in children is greater in the last third of the night compared with the first third,48 the mechanisms for this are unclear. Phasic rapid-eye-movement (REM) sleep is associated with central inhibition of ventilation.49,50 Theoretically, muscle hypotonia and central inhibition of breathing may be worse during the more intense portions of REM sleep. Propofol affects REM sleep and may simulate the worst case scenario for functional airway studies. Therefore, propofol may be a preferred drug to evaluate the behavior of the pharyngeal airway in children with mild to moderate OSAS provided there is no significant accompanying oxygen desaturation. Dexmedetomidine might be the preferred
drug for children with a history of severe OSA as shown by our results. The sedative effect of dexmedetomidine is mediated via stimulation of -2 adrenoceptors in the locus coeruleus, whereas propofol acts through the inhibitory neurotransmitter -aminobutyric acid. There is some evidence that sedation induced with dexmedetomidine resembles normal sleep11,13; in rats, the pattern of c-Fos expression (a marker of activation of neurons) is qualitatively similar to that seen during normal nonREM sleep, suggesting that endogenous sleep pathways are causally involved in dexmedetomidine-induced sedation.11 Our study generates a hypothesis on finding an ideal anesthetic technique to evaluate the function of the upper airway in severe OSA. Perhaps an initial evaluation under dexmedetomidine simulating non-REM sleep followed by an infusion of propofol to simulate REM sleep would yield a comprehensive dynamic evaluation of the airway that would significantly help in surgical management decisions in these difficult children. A future prospective study to develop such a technique would address this important issue. An anesthetic challenge during dynamic MRI airway imaging is determining when to perform an airway intervention. Overnight polysomnography noting the severity of desaturations during natural sleep can be used by the anesthesiologist as a guide to determine minimal acceptable arterial oxygen saturation before intervention. If an artificial airway is required, a nasal trumpet is preferred over an oral airway because it interferes less with the evaluation of the retroglossal airway, a critical dynamic component of the MRI airway study. Despite the high affinity of dexmedetomidine for the 2 versus 1 adrenergic receptor (1620:1), significant cardiovascular effects, such as bradycardia, sinus arrest, and hypotension, have been reported.51 In our series, hemodynamic changes after dexmedetomidine were generally mild and self-limited, although the degree of bradycardia seen after the initial dose was greater in children with trisomy 21. The limitations of our report include its retrospective nature, the lack of prospectively specified thresholds for airway and hemodynamic interventions, and experience from a single center involving an uncommon, highly specialized imaging study. Even though we were unable to access an electronic anesthetic record to identify airway interventions, we feel that the use of two independent sources (the written anesthetic record and the radiologists report) provides an acceptable level of reliability and accuracy for the data. To control for any possible selection bias related to the severity of OSA, we stratified children in both groups using objective OSA criteria and excluded those children without a detailed polysomnographic report from our final analysis. Although it would obviously have been preferable to have performed a prospective, randomized study, by the time that we
751
had gained sufficient experience with dexmedetomidine to consider a formal investigation, the advantages of dexmedetomidine, especially in children with significant OSA, had led to its preference over propofol by our group for dynamic airway MRI study, which may explain the unequal numbers of subjects in this study receiving propofol and dexmedetomidine. Although MRI sleep studies to evaluate refractory OSA are a highly specialized subset of all MRI examinations, our results may have applicability to children with severe sleep-disordered breathing presenting for all types of imaging studies requiring sedation or anesthesia. In conclusion, both dexmedetomidine and propofol provided anesthesia leading to interpretable MRI sleep studies in almost all subjects. However, MRI sleep studies were completed without artificial airways in a larger proportion of children with OSA receiving dexmedetomidine than propofol. Dexmedetomidine caused significant reductions in heart rate and propofol caused significant reductions in arterial blood pressure; these hemodynamic changes did not interrupt or interfere with successful completion of the MRI sleep study. Among currently available anesthetics, dexmedetomidine may offer more favorable conditions than propofol during MRI sleep studies in children with significant OSA and may offer advantages for other imaging and diagnostic studies in children with significant sleep-disordered breathing. REFERENCES
1. Donnelly LF, Surdulescu V, Chini BA, Casper KA, Poe SA, Amin RS. Upper airway motion depicted at cine MR imaging performed during sleep: comparison between young patients with and those without obstructive sleep apnea. Radiology 2003;227:239 45 2. Donnelly LF, Casper KA, Chen B, Koch BL. Defining normal upper airway motion in asymptomatic children during sleep by means of cine MR techniques. Radiology 2002;223:176 80 3. Abbott MB, Dardzinski BJ, Donnelly LF. Using volume segmentation of cine MR data to evaluate dynamic motion of the airway in pediatric patients. AJR Am J Roentgenol 2003;181:8579 4. Donnelly LF, Casper KA, Chen B. Correlation on cine MR imaging of size of adenoid and palatine tonsils with degree of upper airway motion in asymptomatic sedated children. AJR Am J Roentgenol 2002;179:503 8 5. Suto Y, Matsuo T, Kato T, Hori I, Inoue Y, Ogawa S, Suzuki T, Yamada M, Ohta Y. Evaluation of the pharyngeal airway in patients with sleep apnea: value of ultrafast MR imaging. AJR Am J Roentgenol 1993;160:311 4 6. Shellock FG, Schatz CJ, Julien P, Steinberg F, Foo TK, Hopp ML, Westbrook PR. Occlusion and narrowing of the pharyngeal airway in obstructive sleep apnea: evaluation by ultrafast spoiled GRASS MR imaging. AJR Am J Roentgenol 1992;158: 1019 24 7. Jager L, Gunther E, Gauger J, Reiser M. Fluoroscopic MR of the pharynx in patients with obstructive sleep apnea. AJNR Am J Neuroradiol 1998;19:120514 8. Abbott MB, Donnelly LF, Dardzinski BJ, Poe SA, Chini BA, Amin RS. Obstructive sleep apnea: MR imaging volume segmentation analysis. Radiology 2004;232:889 95 9. Donnelly LF, Shott SR, LaRose CR, Chini BA, Amin RS. Causes of persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy in children with Down syndrome as depicted on static and dynamic cine MRI. AJR Am J Roentgenol 2004;183:175 81 10. Connolly LA. Anesthetic management of obstructive sleep apnea patients. J Clin Anesth 1991;3:4619
11. Nelson LE, Lu J, Guo T, Saper CB, Franks NP, Maze M. The alpha2-adrenoceptor agonist dexmedetomidine converges on an endogenous sleep-promoting pathway to exert its sedative effects. Anesthesiology 2003;98:428 36 12. Doze VA, Chen BX, Maze M. Dexmedetomidine produces a hypnotic-anesthetic action in rats via activation of central alpha-2 adrenoceptors. Anesthesiology 1989;71:759 13. Hsu YW, Cortinez LI, Robertson KM, Keifer JC, Sum-Ping ST, Moretti EW, Young CC, Wright DR, Macleod DB, Somma J. Dexmedetomidine pharmacodynamics: part I: crossover comparison of the respiratory effects of dexmedetomidine and remifentanil in healthy volunteers. Anesthesiology 2004;101: 1066 76 14. Donnelly LF. Obstructive sleep apnea in pediatric patients: evaluation with cine MR sleep studies. Radiology 2005; 236:768 78 15. Marcus CL, Katz ES, Lutz J, Black CA, Galster P, Carson KA. Upper airway dynamic responses in children with the obstructive sleep apnea syndrome. Pediatr Res 2005;57:99 107 16. Farre R, Rigau J, Montserrat JM, Buscemi L, Ballester E, Navajas D. Static and dynamic upper airway obstruction in sleep apnea: role of the breathing gas properties. Am J Respir Crit Care Med 2003;168:659 63 17. Marcus CL, McColley SA, Carroll JL, Loughlin GM, Smith PL, Schwartz AR. Upper airway collapsibility in children with obstructive sleep apnea syndrome. J Appl Physiol 1994;77:918 24 18. Cartwright RD. Effect of sleep position on sleep apnea severity. Sleep 1984;7:110 4 19. Neill AM, Angus SM, Sajkov D, McEvoy RD. Effects of sleep posture on upper airway stability in patients with obstructive sleep apnea. Am J Respir Crit Care Med 1997;155:199 204 20. Isono S, Tanaka A, Nishino T. Dynamic interaction between the tongue and soft palate during obstructive apnea in anesthetized patients with sleep-disordered breathing. J Appl Physiol 2003;95:2257 64 21. Gibson SE, Myer CM III, Strife JL, OConnor DM. Sleep fluoroscopy for localization of upper airway obstruction in children. Ann Otol Rhinol Laryngol 1996;105:678 83 22. Donnelly LF, Strife JL, Myer CM III. Is sedation safe during dynamic sleep fluoroscopy of children with obstructive sleep apnea? AJR Am J Roentgenol 2001;177:1031 4 23. Donnelly LF, Jones BV, Strife JL. Imaging of pediatric tongue abnormalities. AJR Am J Roentgenol 2000;175:489 93 24. Owen GO, Canter RJ, Robinson A. Snoring, apnoea and ENT symptoms in the paediatric community. Clin Otolaryngol Allied Sci 1996;21:130 4 25. Guilleminault C, Winkle R, Korobkin R, Simmons B. Children and nocturnal snoring: evaluation of the effects of sleep related respiratory resistive load and daytime functioning. Eur J Pediatr 1982;139:16571 26. Ali NJ, Pitson DJ, Stradling JR. Snoring, sleep disturbance, and behaviour in 4 5 year olds. Arch Dis Child 1993;68:360 6 27. Ali NJ, Pitson D, Stradling JR. Natural history of snoring and related behaviour problems between the ages of 4 and 7 years. Arch Dis Child 1994;71:74 6 28. Rosen CL. Clinical features of obstructive sleep apnea hypoventilation syndrome in otherwise healthy children. Pediatr Pulmonol 1999;27:4039 29. Marcus CL, Carroll JL, Koerner CB, Hamer A, Lutz J, Loughlin GM. Determinants of growth in children with the obstructive sleep apnea syndrome. J Pediatr 1994;125:556 62 30. Brouillette RT, Fernbach SK, Hunt CE. Obstructive sleep apnea in infants and children. J Pediatr 1982;100:31 40 31. Owens J, Opipari L, Nobile C, Spirito A. Sleep and daytime behavior in children with obstructive sleep apnea and behavioral sleep disorders. Pediatrics 1998;102:1178 84 32. Hudgel DW, Harasick T, Katz RL, Witt WJ, Abelson TI. Uvulopalatopharyngoplasty in obstructive apnea. Value of preoperative localization of site of upper airway narrowing during sleep. Am Rev Respir Dis 1991;143:942 6 33. Shott SR, Donnelly LF. Cine magnetic resonance imaging: evaluation of persistent airway obstruction after tonsil and adenoidectomy in children with Down syndrome. Laryngoscope 2004;114:1724 9 34. Isono S, Shimada A, Utsugi M, Konno A, Nishino T. Comparison of static mechanical properties of the passive pharynx between normal children and children with sleep-disordered breathing. Am J Respir Crit Care Med 1998;157:1204 12
752
35. den Herder C, Schmeck J, Appelboom DJ, de Vries N. Risks of general anaesthesia in people with obstructive sleep apnoea. BMJ 2004;329:9559 36. Drummond GB. Comparison of sedation with midazolam and ketamine: effects on airway muscle activity. Br J Anaesth 1996;76:6637 37. Luscri N, Tobias JD. Monitored anesthesia care with a combination of ketamine and dexmedetomidine during magnetic resonance imaging in three children with trisomy 21 and obstructive sleep apnea. Paediatr Anaesth 2006;16:782 6 38. Mason KP, Zgleszewski SE, Dearden JL, Dumont RS, Pirich MA, Stark CD, DAngelo P, Macpherson S, Fontaine PJ, Connor L, Zurakowski D. Dexmedetomidine for pediatric sedation for computed tomography imaging studies. Anesth Analg 2006; 103:57 62 39. Koroglu A, Teksan H, Sagir O, Yucel A, Toprak HI, Ersoy OM. A comparison of the sedative, hemodynamic, and respiratory effects of dexmedetomidine and propofol in children undergoing magnetic resonance imaging. Anesth Analg 2006;103:637 40. Munro HM, Tirotta CF, Felix DE, Lagueruela RG, Madril DR, Zahn EM, Nykanen DG. Initial experience with dexmedetomidine for diagnostic and interventional cardiac catheterization in children. Paediatr Anaesth 2007;17:109 12 41. Heard CM, Joshi P, Johnson K. Dexmedetomidine for pediatric MRI sedation: a review of a series of cases. Paediatr Anaesth 2007;17:888 92 42. Nichols DP, Berkenbosch JW, Tobias JD. Rescue sedation with dexmedetomidine for diagnostic imaging: a preliminary report. Paediatr Anaesth 2005;15:199 203
43. Berkenbosch JW, Wankum PC, Tobias JD. Prospective evaluation of dexmedetomidine for noninvasive procedural sedation in children. Pediatr Crit Care Med 2005;6:4359; quiz 40 44. Mason KP, Zurakowski D, Zgleszewski SE, Robson CD, Carrier M, Hickey PR, Dinardo JA. High dose dexmedetomidine as the sole sedative for pediatric MRI. Paediatr Anaesth 2008;18:40311 45. Venn RM, Bradshaw CJ, Spencer R, Brealey D, Caudwell E, Naughton C, Vedio A, Singer M, Feneck R, Treacher D, Willatts SM, Grounds RM. Preliminary UK experience of dexmedetomidine, a novel agent for postoperative sedation in the intensive care unit. Anaesthesia 1999;54:1136 42 46. Evans RG, Crawford MW, Noseworthy MD, Yoo SJ. Effect of increasing depth of propofol anesthesia on upper airway configuration in children. Anesthesiology 2003;99:596 602 47. Litman RS, Weissend EE, Shrier DA, Ward DS. Morphologic changes in the upper airway of children during awakening from propofol administration. Anesthesiology 2002;96:60711 48. Goh DY, Galster P, Marcus CL. Sleep architecture and respiratory disturbances in children with obstructive sleep apnea. Am J Respir Crit Care Med 2000;162:682 6 49. Millman RP, Knight H, Kline LR, Shore ET, Chung DC, Pack AI. Changes in compartmental ventilation in association with eye movements during REM sleep. J Appl Physiol 1988;65:1196 202 50. Smith CA, Henderson KS, Xi L, Chow C, Eastwood PR, Dempsey JA. Neural-mechanical coupling of breathing in REM sleep. J Appl Physiol 1997;83:192332 51. Bloor BC, Ward DS, Belleville JP, Maze M. Effects of intravenous dexmedetomidine in humans. II. Hemodynamic changes. Anesthesiology 1992;77:1134 42
753
A Novel Skin-Traction Method Is Effective for Real-Time Ultrasound-Guided Internal Jugular Vein Catheterization in Infants and Neonates Weighing Less Than 5 Kilograms
Masato Morita, MD Hiroshi Sasano, PhD, MD Takafumi Azami, PhD, MD Nobuko Sasano, PhD, MD Yoshihito Fujita, PhD, MD Shoji Ito, PhD, MD Takeshi Sugiura, PhD, MD Kazuya Sobue, PhD, MD
BACKGROUND: Internal jugular vein (IJV) catheterization in pediatric patients is sometimes difficult because of the small sizes of veins and their collapse during catheterization. To facilitate IJV catheterization, we developed a novel skin-traction method (STM), in which the point of puncture of the skin over the IJV is stretched upward with tape during catheterization. In this study, we examined whether the STM increases the cross-sectional area of the vein and thus facilitates catheterization. METHODS: This was a prospective study conducted from December 2006 to June 2008. We enrolled 28 consecutive infants and neonates weighing 5 kg who underwent surgery for congenital heart disease. The patients were randomly assigned to a group in which STM was performed (STM group) or a group in which it was not performed (non-STM group). The cross-sectional area and diameter of the right IJV in the flat position and 10 Trendelenburg position with and without applying STM were measured. We determined time from first skin puncture to the following: (a) first blood back flow, (b) insertion of guidewire, and (c) insertion of catheter. Number of punctures, success rate, complications, and degree of IJV collapse during advancement of the needle (estimated as decrease of anteroposterior diameter during advancement of the needle compared with the diameter before advancement) were also examined. RESULTS: STM significantly increased the cross-sectional area and the anteroposterior diameter of the IJV in both positions. The time required to insert the catheter was significantly shorter in the STM group, probably mainly due to a shorter guidewire insertion time. The degree of IJV collapse during advancement of the needle was much lower in the STM group. CONCLUSIONS: STM facilitates IJV catheterization in infants and neonates weighing 5 kg by enlarging the IJV and preventing vein collapse.
(Anesth Analg 2009;109:754 9)
eal-time ultrasound guidance is useful for catheterization of the internal jugular vein (IJV).13 However, IJV catheterization in pediatric patients is sometimes difficult even for skilled physicians who use ultrasound guidance. The main reasons for such a difficulty in performing catheterization are thought to be the small vein size4 and vein collapse during catheterization. A significant positive correlation has been found between IJV diameter and the success rate of IJV catheterization.5 Many reports have shown that the Trendelenburg position is effective for increasing the
From the Department of Anesthesiology and Medical Crisis Management, Nagoya City University Graduate School of Medical Sciences, Nagoya City, Aichi, Japan. Accepted for publication May 2, 2009. Address correspondence and reprint requests to Hiroshi Sasano, PhD, MD, Department of Anesthesiology and Medical Crisis Management, Nagoya City University Graduate School of Medical Sciences, 1 Azakawasumi, Mizuho-cho, Mizuho-ku, Nagoya City, Aichi 467-8601, Japan. Address e-mail to hsasano@med.nagoya-cu.ac.jp. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b01ae3
diameter of the IJV,6 8 and thereby possibly improving the success rate and decreasing the complication rate of IJV catheterization. During real-time ultrasound guidance, IJV collapse can occur when the skin is pressed with a probe or the needle is inserted into the neck. Mallory et al.8 reported that advancement of the needle decreased IJV cross-sectional area. We have observed that IJV collapse may be prevented by stretching the skin over the IJV in both the cephalad and caudad directions. We have developed a novel skin-traction method (STM), in which the puncture point of the skin over the IJV is lifted up as a result of stretching the skin in the cephalad and caudal directions with surgical tape. We have already reported that the STM could increase the cross-sectional area and anteroposterior diameter of the IJV9 and prevent IJV collapse produced by the ultrasound probe in adult volunteers.10 Thus, we hypothesized that the STM can facilitate IJV catheterization even in small children.
754
Figure 1. Differences in surface of the

neck when applying (right) and not applying (left) STM. The skin over the RIJV is stretched up, and the puncture point (circle on the skin) is lifted up by STM (double arrows). Before skin traction was performed, the angle made by two lines that between the center of the clavicle (star) and the puncture point and that between the mandibular angle (triangle) and the puncture point is kept at about 135, which makes it easy to lift up the skin around the puncture point by the STM. The skin over the RIJV is stretched until a circle drawn in the skin is changed elliptically about 1.5-fold. STM skin-traction method; RIJV right internal jugular vein.
The purpose of this study was to examine: 1) whether STM increases the cross-sectional area and diameter of the IJV and 2) whether it also facilitates real-time ultrasound-guided IJV catheterization in infants and neonates with congenital heart disease weighing 5 kg.
attached to the metal edge of the operating table so that sufficient skin traction could be maintained. The skin over the RIJV was stretched until a circle drawn on the skin was changed elliptically about 1.5-fold (Fig. 1). The time needed for taping the skin was 1 min.
METHODS
This prospective, randomized study was conducted from December 2006 to June 2008. We enrolled 28 consecutive infants and neonates weighing 5 kg (2.40 4.91 median 3.58 kg) who underwent surgery for congenital heart disease. The research protocol was reviewed and approved by the Ethics Committee of our hospital, and written informed consent was obtained from each parent of these children. Using the envelope method, the patients were randomly assigned to a group in which STM was performed (STM group) or a group in which it was not performed (non-STM group). STM was performed as follows. The skin over the right IJV (RIJV) was lifted up with several pieces of 2.6-cm-wide surgical tape (Transpore Surgical Tape, 3M, St. Paul, MN) in the cephalad and caudad directions. The skin cephalic to the RIJV was stretched cephalad, whereas the skin caudal to the RIJV was stretched caudad. Three pieces of tape were used in each direction. The other ends of the tape were firmly
Measurement of Size of IJV

Before catheterization, we measured and examined how STM could increase the cross-sectional area and diameter of the RIJV in all patients including those of the non-STM group. After the patients were anesthetized, endotracheally intubated, and mechanically ventilated, the ventilator mode was set to pressure-controlled (peak pressure 20-cm H2O, respiratory rate 20/min with application of 5-cm H2O positive end-expiratory pressure), and images with ultrasound were recorded. Real-time 2-dimensional (2D) ultrasound imaging (iLOOK25, SonoSite, Bothell, WA) with videotaping was used in this study. An ultrasound probe (L25/10 5 MHz) was placed perpendicularly on the skin surface of the neck to identify the RIJV. Care was taken not to press the neck and compress the RIJV. Patients were placed in the flat position and 10 Trendelenburg position to examine the effects of Trendelenburg position on the RIJV. A shoulder roll was placed to extend the neck, which was rotated about 30.11
755
Figure 2. Differences in the shape of RIJV on ultrasonography when applying (B) and not applying (A) STM. The cross-sectional area of the RIJV increased mainly as a result of an increase of anteroposterior diameter when applying STM. RIJV right internal jugular vein; STM skintraction method; CCA common carotid artery.
Measurements when both applying and not applying STM were done in the flat position and then in the Trendelenburg position in all 28 patients of the two groups (STM groups and non-STM groups). The point of measurement was the lesser supraclavicular fossa at the level of the cricoid cartilage.12 The images obtained using 2D ultrasound were recorded by a video tape recorder and transferred to a computer, where recorded areas and diameters of the RIJV at the end-inspiratory phase of positive ventilation were measured using ImageJ by a person blinded to the patient grouping13 (Image Processing and Analysis in Java, version 1.34 s image analysis software). The cross-sectional area, anteroposterior diameter, and transverse diameter of the RIJV were measured with and without STM (Fig. 2). All results are presented as the mean sd. Twoway repeated-measures analysis of variance with Bonferroni adjustment made for multiple comparisons was used to test for significance. P values 0.05 were considered significant.
Table 1. Baseline Demographics of the Study Groups STM group (n 14)

Age (mo) Body weight (kg) Body height (cm) Cross-sectional area (cm2) Vein depth (mm) CVP (mm Hg) Heart disease 0.64 0.74 3.57 0.85 50.4 3.99 0.17 0.09 6.45 0.97 7.92 2.61 VSD PH 6 DORV 2 TAPVC 2 TGA 2 ECD 1 HLHS 1
Non-STM group (n 14)

1.00 0.96 3.51 0.62 50.7 5.19 0.18 0.10 6.54 1.12 7.00 2.38 VSD PH 10 TOF 2 TA 1 TGA 1
Values are mean SD. There were no signicant differences between the two groups for all baseline characteristics (Welch t-test). CVP central venous pressure; VSD ventricular septal defect; PH pulmonary hypertension; DORV double outlet right ventricle; TAPVC total anomalous pulmonary venous return; TGA transposition of the great vessels; ECD endocardial cushion defect; HLHS hypoplastic left heart syndrome; TOF tetralogy of Fallot; TA tricuspid atresia; STM skin-traction method.
Efcacy of STM for Catheterization

We examined the efficacy of STM for facilitating real-time ultrasound-guided IJV catheterization in infants and neonates, who were assigned to the STM group or the non-STM group. Nine anesthesiologists, who had more than 3 yr experience with heart surgery anesthesia and were very familiar with real-time ultrasound-guided IJV catheterization, performed the catheterizations. They were each assisted by another anesthesiologist. A third person other than the operator and assistant drew one of the envelopes. If the patient was assigned to the non-STM group, the person removed several pieces of tape used to measure IJV size. An assistant sterilized the skin with povidoneiodine, cleaned off the drawn circle marker, and placed sterile drapes, while covering the tapes in the STM group to ensure that the operator was not aware of group assignment. Under general anesthesia and in the ventilator mode described earlier, catheterization was performed with real-time ultrasound 2D imaging. Another video recorder recorded actual catheterization of the IJV to determine the time taken for the procedure. With the patient in the 10 Trendelenburg
756
Skin-Traction Method for Catheterization
position, a double lumen, 17-gauge central venous catheter (Nippon-Sherwood, Tokyo, Japan) was inserted by the Seldinger technique, using a 24-gauge indwelling cannula and 0.18-in. guidewire. The following times were determined: (a) time from first puncture of skin until confirmation of the first aspiration of blood from the vein (blood back flow time), (b) time from first puncture of skin until insertion of the guidewire (total guidewire time), (c) time from first puncture of skin until insertion of the catheter (total catheter time), with the time of catheter insertion defined as aspiration of blood from the catheter, (d) B-A (pure guidewire time), (e) C-A (pure catheter time), and (f) C-B (catheter time alone). The number of puncture attempts required to achieve successful catheterization was also counted. Ultimate success or failure of catheterization and complications, such as hematoma, arterial puncture, and pneumothorax, was recorded. The degree of IJV collapse (estimated as decrease of anteroposterior diameter during advancement of the needle compared with the diameter before advancement) was measured by a person blinded to patient groupings, with the use of 2D ultrasound imaging and videotape recorder.
Table 2. Anatomical Measurement of the Right Internal Jugular Vein in 28 Patients of the Two Groups a, FP without STM
Cross-sectional area (mm ) Anteroposterior diameter (mm) Transverse diameter (mm)
2
b, FP with STM
21.1 11.5* 4.55 1.15* 5.35 1.47
c, TP without STM
19.6 9.9* 4.03 0.89* 5.61 1.61*
d, TP with STM
21.6 11.9 4.67 1.14 5.26 1.42
17.3 9.1 3.88 0.91 5.34 1.52
FP at position; TP Trendelenburg position; STM skin-traction method. * P 0.05 (versus a). P 0.05 (versus c). P 0.05 (versus b).
Table 3. Between Group Comparisons of Study End Points STM group (n 14)
Blood back-flow time (s) Total guidewire time (s) Total catheter time (s) Pure guidewire time (s) Pure catheter time (s) Catheter time alone (s) Number of punctures Success Complications Degree of collapse (%)
Values are mean SD. STM skin-traction method. * Welch t-test. a n 13.
Non-STM group (n 14)

65.0 76.4 243.7 329.4 324.6 310.4 178.7 284.6 259.6 264.1 87.1 23.4a 2.14 2.03 13/14 0/14 91.5 21.2
P value (log-rank test)

0.417 0.075 0.044 0.074 0.017 0.393 0.143*
57.9 97.5 91.5 93.7 165.9 98.4 33.6 22.6 107.9 30.4 74.4 25.5 1.29 0.61 14/14 0/14 50.4 25.4
0.01*
Using the results of our pilot study, we assumed that the expected difference in mean catheter insertion time would be 150 s with an expected standard deviation of 130 s. Assuming an of 0.05 and power 0.81, we calculated the required sample size to be 13 in each group. All results are presented as mean sd. Welch t-test was used to compare the baseline characteristics of the two groups including age, sex, body height, body weight, vein diameter in the flat position without STM, vein depth beneath the skin, and central venous pressure before the start of operation. The number of punctures and the degree of IJV collapse during advancement of the needle were also analyzed using Welch t-test. Kaplan-Meier estimates of the times described earlier (A-F) were compared using the log-rank test. P values 0.05 were considered significant.
Figure 3. Kaplan-Meier estimates were made for the following times. A, Total catheterization time (time from start of first puncture of skin until insertion of the catheter) was significantly shorter in the STM group (continuous line) than in the non-STM group (dotted line). *P 0.044 (log-rank). B, Pure guidewire time (time from first aspiration of blood from the vein until insertion of guidewire) was shorter in the STM group (continuous line) than in the non-STM group (dotted line), although not to a significant extent. P 0.074 (log-rank). STM skin-traction method.
Measurement of Size of IJV

Table 2 shows the changes in the cross-sectional area, anteroposterior diameter, and transverse diameter of the RIJV at each head tilt position when applying and not applying STM in all 28 patients in both the groups. When not applying STM, the cross-sectional area of the RIJV was significantly larger in the Trendelenburg position than in the flat position. STM significantly increased the cross-sectional area of the RIJV in the flat position from 17.3 9.1 mm2 to 21.1 11.5 mm2 and in the Trendelenburg position from 19.6 9.9 mm2 to 21.6 11.9 mm2. STM significantly increased the anteroposterior diameter of the RIJV in both the flat and Trendelenburg positions. STM significantly decreased the transverse diameter (5.61 1.61 mm vs 5.26 1.42 mm) in the Trendelenburg position, whereas STM did not change it in the flat position.
RESULTS
Of the 28 patients enrolled in this study, 14 were in the STM group and 14 were in the non-STM group. The baseline demographics of the subjects in the study are shown in Table 1, and there were no significant differences between the two groups.
757
Figure 4. Changes in shapes of right

internal jugular vein (RIJV). One of the patients in the STM group is presented in the upper figures. A, before advancement of the needle. B, during advancement. One of the patients in the non-STM group is presented in the lower figures. C, before advancement of the needle. D, during advancement. Applying STM maintained the shape of the RIJV even during advancement of the needle (B). STM skin-traction method.
Efcacy of STM for Catheterization

Table 3 shows the blood back flow time, total guidewire time, total catheter time, pure guidewire time, pure catheter time, and catheter time alone. Number of puncture attempts, ultimate success or failure of catheterization, complications, and degree of IJV collapse during advancement of the needle are also shown. Total catheter time was significantly shorter in the STM group than in the non-STM group (165.9 98.4 s vs 324.6 310.4 s, P 0.044; Fig. 3A). Pure catheter time was also significantly shorter in the STM group than in the non-STM group (107.9 30.4 s vs 259.6 264.1 s, P 0.017). Neither blood back flow time nor catheter time alone differed significantly between the groups. Pure guidewire time was shorter in the STM group, although not to a significant extent (33.6 22.6 s [STM group] vs 178.7 284.6 s [non-STM group], P 0.074; Fig. 3B). The number of puncture attempts to achieve successful catheterization was less in the STM group, although not to a significant extent. There was only one case of catheterization failure in the non-STM group, which occurred because the operator could not insert the guidewire. In this case, we counted both the total guidewire time and total catheter time as the time when the operator discontinued attempts to perform catheterization. Catheter time alone for this patient was not included as a data point. Major complications such as hematoma and arterial puncture did not occur in either group. After skin traction, the skin of some patients became temporarily pink due to the force of traction applied. However, the pink color quickly disappeared. The degree of IJV collapse (decrease of anteroposterior diameter during advancement of the needle) was less in the STM group than in the nonSTM group (50.4% 25.4% vs 91.5% 21.2%, P
758
Skin-Traction Method for Catheterization
0.001). Eleven of 14 patients had the IJV obliterated in the non-STM group (only one patient experienced IJV obliteration in the STM group; Fig. 4).
DISCUSSION
We developed a novel STM for catheterization of the IJV9 and noted that 1) STM increased the crosssectional area and the anteroposterior diameter of the IJV even in the Trendelenburg position with administration of positive end-expiratory pressure, 2) the time to insertion of catheter with real-time ultrasound guidance was significantly shorter in the STM group, and 3) the degree of IJV collapse during advancement of the needle was much lower in the STM group. These findings indicate that the STM facilitates IJV catheterization with real-time ultrasound guidance in infants and neonates. The mechanism of stretching the IJV in the anteroposterior direction with STM appears to involve lifting up the skin over the IJV and then enlarging the IJV in the anteroposterior direction. The increase in distance between skin and the IJV on ultrasonography with STM (Fig. 2) shows that this method stretches not only the IJV but also tissue, such as muscle over the IJV, releasing pressure on the IJV. The main reason for the shorter catheter insertion time is thought to be the shorter time required for guidewire insertion, which has often been a reported difficulty and the main reason for the inability to perform catheterization in pediatric patients.14,15 In the non-STM group, pure guidewire time was more than 120 s in five of 14 patients, whereas it was less for all patients in the STM group. It appears that both the increase of the anteroposterior diameter of the IJV and the prevention of vein collapse during catheterization with the
STM may facilitate catheterization and guidewire insertion. We could find no previous report on facilitation of insertion of guidewires. Thus, STM appears to be a novel, effective method for IJV catheterization. The fact that total and pure guidewire time was not significant, whereas total and pure catheter time was significant may indicate that STM is actually effective for the other process of catheterization. For instance, STM may also influence facilitating dilator insertion in very small pediatric patients and prevent guidewire kinking. Although we found no significant difference in either complications or number of attempted punctures between the two groups, a tendency toward fewer attempts at skin puncture was noted in the STM group. Fewer puncture attempts may decrease central venous catheter-associated bloodstream infection, which have been correlated with the number of needle passes.16 Furthermore, increased anteroposterior diameter and preventing collapse by applying STM may reduce the rate of inadvertent punctures of the common carotid artery, which sometimes overlaps the IJV either partially or completely.5 Further study is needed to draw appropriate conclusions regarding these complications. STM might have three disadvantages. The first is skin injury due to the force of traction applied. After skin traction, the skin of some patients temporarily became pink. However, there were no cases of skin injury. The second potential disadvantage may be a decrease of the transverse diameter in the Trendelenburg position when applying STM. The third disadvantage may be the time required for taping. However, because the time involved in taping the skin was 1 min, this does not seem to be an issue. In conclusion, with real-time ultrasound guidance, STM shortened the time for IJV catheterization significantly, while increasing cross-sectional area and anteroposterior diameter and preventing vein collapse during advancement of the needle. REFERENCES
1. Verghese ST, McGill WA, Patel RI, Sell JE, Midgley FM, Ruttimann UE. Ultrasound-guided internal jugular venous cannulation in infants: a prospective comparison with the traditional palpation method. Anesthesiology 1999;91:717
2. Hosakawa K, Shime N, Kato Y, Hoshimoto S. A randomized trial of ultrasound image-based skin surface marking versus real-time ultrasound-guided internal jugular vein catheterization in infants. Anesthesiology 2007;107:720 4 3. Wigmore TJ, Smythe JF, Hacking MB, Raobaikady R, MacCallum NS. Effects of the implementation of NICE guidelines for ultrasound guidance on the complication rates associated with central venous catheter placement in patients presenting for routine surgery in a tertiary referral center. Br J Anaesth 2007;99:6625 4. Hayashi Y, Uchida O, Takaki O, Ohnishi Y, Nakajima T, Kataoka H, Kuro M. Internal jugular vein catheterization in infants undergoing cardiovascular surgery: an analysis of the factors influencing successful catheterization. Anesth Analg 1992;74:688 93 5. Gordon AC, Saliken JC, Johns D, Owen R, Gray RR. US-guided puncture of the internal jugular vein: complications and anatomic considerations. J Vasc Interv Radiol 1998;9:333 8 6. Armstrong PJ, Sutherland R, Scott DH. The effect of position and different maneuvers on internal jugular vein diameter size. Acta Anaesthesiol Scand 1994;38:229 31 7. Parry G. Trendelenburg position, head elevation and a midline position optimize right internal jugular vein diameter. Can J Anaesth 2004;51:379 81 8. Mallory DL, Shawker T, Evans G, Mcgee W, Brenner M, Parker M, Morrison G, Mohler P, Veremakis C, Parrillo JE. Effects of clinical maneuvers on sonographically determined internal jugular vein size during venous cannulation. Crit Care Med 1990;18:1269 73 9. Morita M, Sasano H, Azami T, Sasano N, Sobue K. The skin-traction method increases the cross-sectional area of the internal jugular vein by increasing its anteroposterior diameter. J Anesth 2007;21:46171 10. Sasano H, Morita M, Azami T, Ito S, Sasano N, Kato R, Hirate H, Ito H, Takeuchi A, Sobue K. Skin-traction method prevents the collapse of the internal jugular vein caused by an ultrasound probe in real-time ultrasound-assisted guidance. J Anesth 2009;23:415 11. Sulek CA, Gravenstein N, Blackshear RH, Weiss L. Head rotation during internal jugular vein cannulation and risk of carotid artery puncture. Anesth Analg 1996;82:125 8 12. Nakayama S, Yamashita M, Osaka Y, Isobe T, Izumi H. Right internal jugular vein venography in infants and children. Anesth Analg 2001;93:331 4 13. ImageJ: Image processing and analysis in Java. Available at: http://rsb.info.nih.gov/ij/; accessed July 1, 2008 14. Milling TJ, Rose J, Briggs WM, Birkhahn R, Gaeta TJ, Bove JJ, Melniker LA. Randomized, controlled clinical trial of point-ofcare limited ultrasonography assistance of central venous cannulation: the Third Sonography Outcomes Assesment Program (SOAP-3) Trial. Crit Care Med 2005;33:1764 9 15. Alderson PJ, Burrows FA, Stemp LI, Holtby HM. Use of ultrasound to evaluate internal jugular vein anatomy and to facilitate central venous cannulation in paediatric patients. Br J Anaesth 1993;70:145 8 16. Karakitsos D, Labropoulos N, Groot ED, Patrianakos AP, Kouraklis G, Poularas J, Samonis G, Tsoutsos DA, Konstadoulakis MM, Karabinis A. Real-time ultrasound-guided catheterization of the internal jugular vein: a prospective comparison with the landmark technique in critical care patients. Crit Care 2006;10:R162
759
Case Report
Cardiac Arrest in the Neonate During Laparoscopic Surgery

Kirk Lalwani, MD, FRCA, MCR Inger Aliason, MD
We describe a case of intraoperative neonatal cardiac arrest during attempted laparoscopic surgery. Circulatory collapse occurred before peritoneal insufflation, initially obscuring the diagnosis. Emergent transthoracic echocardiography during resuscitation demonstrated intracardiac gas bubbles consistent with venous gas embolism. The site of entrainment was probably a bleeding umbilical vein transected by the umbilical trocar. Greater awareness of this complication in neonates will facilitate early diagnosis and encourage preventive measures, such as the avoidance of umbilical vessels, use of an open instead of closed access technique, and ligation of bleeding vessels after peritoneal access.
(Anesth Analg 2009;109:760 2)
irculatory collapse secondary to carbon dioxide (CO2) embolism during peritoneal insufflation for laparoscopic procedures is well known. We describe a case of neonatal cardiac arrest secondary to gas embolism during attempted laparoscopic surgery. Circulatory collapse occurred before peritoneal insufflation, initially obscuring the diagnosis. Emergency echocardiography was performed during the resuscitation and revealed the mechanism of the arrest. This report will describe the case and discuss the etiology of the air embolism, with special emphasis on the physiology of the neonatal population.
CASE REPORT
An ex-36 wk gestation 1-day-old male was scheduled for laparoscopic repair of duodenal atresia that had been diagnosed prenatally by ultrasound. He was delivered uneventfully by cesarean section after a nonreassuring fetal heart trace. He weighed 2 kg and had physical features of trisomy 21. The patients history, physical examination, and laboratory results were otherwise unremarkable. After standard monitoring and administration of oxygen, he underwent rapid-sequence IV induction with cricoid pressure, propofol, and rocuronium. His trachea was intubated uneventfully on the first attempt with a 3.0 endotracheal tube, and anesthesia was maintained with sevoflurane
From the Departments of Anesthesiology and Perioperative Medicine, and Pediatrics, Oregon Health and Science University, Portland, Oregon. Accepted for publication March 12, 2009. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journals Web site (www.anesthesia-analgesia.org). Reprints will not be available from the author. Address correspondence to Kirk Lalwani, MD, FRCA, MCR, Anesthesiology and Perioperative Medicine, BTE-2, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd., Portland, OR 97239. Address e-mail to lalwanik@ohsu.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc6f9
in oxygen and air. A small dose of fentanyl (1 g/kg) was administered IV. A second IV catheter was placed uneventfully, and the patient was prepared for surgery. Approximately 3 min after surgical incision and insertion of the umbilical trocar, the end-tidal CO2 suddenly decreased from 26 to 6 mm Hg. This was accompanied by a loss of the pulse oximetry trace, ST segment depression, and a decrease in the heart rate from 135 to 100/min while maintaining normal sinus rhythm. Pulses could not be palpated, the heart sounds were muffled, and the arterial blood pressure was approximately 40/20 mm Hg, leading to a working diagnosis of pulseless electrical activity. The differential diagnosis at this point was tension pneumothorax, venous gas embolism via the insufflator or a peripheral IV site, or hypovolemia secondary to intraabdominal vascular injury by the trocar. The surgeons were immediately asked to release the insufflation pressure, but to our surprise, they indicated that they had not yet insufflated the peritoneal cavity. Concomitant with this exchange, the sevoflurane was discontinued, the patient was hand ventilated with 100% oxygen while the integrity of the gas delivery system was checked, and the patients lungs were auscultated to exclude tension pneumothorax. All the IV lines and catheters were checked for sources of air entrainment but nothing obvious was apparent. The surgeon was asked to inspect the abdomen for trauma, which he did after partial insufflation, but there was no sign of intraabdominal trauma. The surgeons reported discoloration of the abdominal wall, and the drapes were immediately removed. The childs skin was profoundly mottled, with large purplish-black patches on the trunk, head, and all extremities. Cardiopulmonary resuscitation (CPR) was immediately commenced with chest compressions, atropine, and several epinephrine boluses. In the absence of a diagnosis, the cardiology team was contacted for emergent echocardiography; they arrived 10 min later and transthoracic echocardiography demonstrated large amounts of air in the right ventricle, pulmonary arteries, aorta, and liver tissue, sufficient to obscure cardiac anatomical details (Fig. 1) (see Video 1, Supplemental Digital Video 1: transthoracic suprasternal notch echocardiographic sweep view from left to right showing numerous air bubbles in the left and main pulmonary arteries, aortic arch, left and right atria, http://links.lww.com/A1371). Over the next 10 min of CPR with the patient in Trendelenburg position, gradual improvement in the end-tidal CO2 level and ST segment depression was accompanied by the reappearance of a
760
Figure 1. Echocardiographic image taken during CPR shows several large air bubbles () in the aorta (AO) and pulmonary arteries (PA) (main PA, left PA, right PA).
palpable pulse and the pulse oximeter wave form approximately 20 min after commencement of CPR. Ventricular systolic function was noted to be normal after CPR despite mild right heart dilation. Venous blood gas revealed hypoxemia and a profound metabolic acidosis that was treated with boluses of crystalloid and sodium bicarbonate. The surgeons, meanwhile, upon withdrawing the umbilical trocar, noticed a small amount of active bleeding from the stump of the umbilical vein. The vein was ligated and the abdomen closed. With the absence of an internal jugular venous central line, the widespread distribution of air in both sides of the heart and circulation, and the reappearance of a stable cardiovascular output, no attempt was made to place and aspirate air from an internal jugular central venous catheter. After stabilization of the vital signs with an epinephrine infusion and placement of a femoral central venous catheter, the patient was transported to the pediatric intensive care unit for elective mechanical ventilation and further management. Despite the improvement in the mottling of the patients head and trunk, all four extremities of his limbs remained purplish-black at the time of transfer to the pediatric intensive care unit. The patient had seizure-like activity overnight, but electroencephalography was negative for seizures. A brain magnetic resonance imaging was performed the next day and was normal. His extremities gradually regained color by the following morning with no tissue loss. A repeat echocardiogram the next morning demonstrated a patent ductus arteriosus with left-to-right flow and no residual air. He remained intubated and underwent successful open repair of duodenal atresia 2 days later. Karyotyping confirmed the diagnosis of trisomy 21. After tracheal extubation, the baby had no signs of neurological impairment and was discharged home uneventfully 10 days later. At follow-up in clinic 1 mo later, he was developing normally, feeding well, and had no obvious residual effects from the intraoperative cardiovascular collapse.
DISCUSSION
In our patient, massive venous air embolism (VAE) with paradoxical embolism via the foramen ovale and
the ductus arteriosus was responsible for cardiovascular collapse. The source of entrainment in this case was probably the bleeding stump of the umbilical vein noted at the site of insertion of the umbilical trocar. Typically, the offending gas is more likely to be CO2 during insufflation of the peritoneal cavity. The physiology of the neonate may make them higher risk for venous gas embolism, paradoxical embolism, and greater hemodynamic instability in the event of a gas embolism.1 Often neonates who present for surgery are mildly hypovolemic as they are nil per os, and maintenance fluids may not account for insensible and third space fluid losses. Decreased central venous pressure (and therefore right atrial pressure) can encourage venous air entrainment. In this case, there was no gradient between the surgical site and the right atrium as the patient was supine and the operating table was flat. The foramen ovale and ductus arteriosus are only functionally (and not anatomically) closed in the neonatal period, and thereby provide conduits from the venous circulation to the arterial circulation. Only 10%30% of adults have a patent foramen ovale and are therefore much less likely to have a paradoxical air embolism.2 When a VAE is small or entrainment is slow, the air can be filtered by the pulmonary vessels, which protect the systemic and coronary circulation.3 However, when the lung filter mechanism is overloaded, air can be trapped in the pulmonary capillary bed, leading to airlock, or continue through the pulmonary circulation to the systemic circulation. Neonates are less able to filter air in the pulmonary circulation because they have a significantly smaller number of gas exchanging saccules and intraacinar vessels (20 million in a term newborn versus 300 million alveoli in adults).4 A 1982 study of air embolism in children undergoing suboccipital craniotomy found no difference in the incidence of air embolism between children and adults.1 However, there was a much greater incidence of hypotension in the children. The proposed mechanism was that the volume of air entrained in the pediatric population was larger relative to their cardiac volume, which caused greater hemodynamic instability. In dogs, the lethal dose (LD100) of rapidly injected air is 7.5 mL/kg,5 but in a smaller animal, such as a rabbit, the LD50 is as low as 0.55 mL/kg.6 This study has obviously not been performed in human adults or neonates, but perhaps smaller cardiovascular systems are less able to tolerate VAE. Laparoscopic injury to enlarged paraumbilical veins as a result of portal hypertension was implicated as the cause of CO2 embolism after peritoneal insufflation in a teenager undergoing a laparoscopic cholecystectomy.7 We describe a case of intraoperative neonatal cardiac arrest because of massive gas embolism via a bleeding umbilical vein before peritoneal insufflation. The usual trocar or Veress needle placement site near the umbilicus may result in umbilical
761
vessel damage in the neonatal population and increase the likelihood of gas entrainment before or during insufflation of CO2. Therefore, identification and ligation of bleeding vessels after Veress needle or laparoscopic trocar insertion should be performed promptly to prevent air entrainment before peritoneal insufflation. Use of an open technique instead of closed access methods may also decrease the likelihood of vascular injury.8 This report also demonstrates the effectiveness of early CPR on outcome. Critical factors that influence survival outcomes from cardiac arrest include the environment in which it occurs, the duration of no flow time, the initial rhythm, the quality of CPR provided, the duration of CPR, and the preexisting condition of the child.9 Almost two thirds of inhospital pediatric cardiac arrest patients have return of spontaneous circulation, and one quarter survive to hospital discharge, of which three quarters have a good neurologic outcome. Outcomes are better in children compared with adults and for in-hospital arrests versus out-of-hospital arrests in children.9 It is also likely that operating room cardiac arrests have a better overall prognosis than in-hospital arrests in other locations as a result of immediately available experienced personnel, continuous hemodynamic monitoring, and preexisting endotracheal intubation in most cases. For pediatric patients in the perioperative period, younger children (neonates and infants), sicker children, emergency surgery, and children with congenital heart disease had a much higher incidence of cardiac arrest; therefore, when comparing risk factors and incidence of cardiac arrest in different populations of children, it is important to consider these factors.10 12 As technology and surgical skills have improved, laparoscopic procedures have become more common in the neonatal population. Despite the favorable outcome in this case, it is obvious that devastating
consequences may result from such complications in the neonate. Laparoscopic surgery entails significant physiological derangement and added potential risk in the neonate with a transitional circulation, and physicians are likely to continue to see similar complications during the perioperative care of these patients. REFERENCES
1. Cucchiara RF, Bowers B. Air embolism in children undergoing suboccipital craniotomy. Anesthesiology 1982;57:338 9 2. Fisher DC, Fisher EA, Budd JH, Rosen SE, Goldman ME. The incidence of patent foramen ovale in 1,000 consecutive patients. A contrast transesophageal echocardiography study. Chest 1995;107:1504 9 3. Butler BD, Hills BA. The lung as a filter for microbubbles. J Appl Physiol 1979;47:537 43 4. Davies G, Reid L. Growth of the alveoli and pulmonary arteries in childhood. Thorax 1970;25:669 81 5. Oppenheimer MJ, Durant TM, Lynch P. Body position in relation to venous air embolism and the associated cardiovascularrespiratory changes. Am J Med Sci 1953;225:36273 6. Munson ES, Merrick HC. Effect of nitrous oxide on venous air embolism. Anesthesiology 1966;27:7837 7. Mattei P, Tyler DC. Carbon dioxide embolism during laparoscopic cholecystectomy due to a patent paraumbilical vein. J Pediatr Surg 2007;42:570 2 8. Bonjer HJ, Hazebroek EJ, Kazemier G, Giuffrida MC, Meijer WS, Lange JF. Open versus closed establishment of pneumoperitoneum in laparoscopic surgery. Br J Surg 1997;84:599 602 9. Topjian AA, Berg RA, Nadkarni VM. Pediatric cardiopulmonary resuscitation: advances in science, techniques, and outcomes. Pediatrics 2008;122:1086 98 10. Gobbo Braz L, Braz JR, Mo dolo NS, do Nascimento P, Brushi BA, Raquel de Carvalho L. Perioperative cardiac arrest and its mortality in children. A 9-year survey in a Brazilian tertiary teaching hospital. Paediatr Anaesth 2006;16:860 6 11. Flick RP, Sprung J, Harrison TE, Gleich SJ, Schroeder DR, Hanson AC, Buenvenida SL, Warner DO. Perioperative cardiac arrests in children between 1988 and 2005 at a tertiary referral center: a study of 92,881 patients. Anesthesiology 2007;106:226 37 12. Morray JP, Geiduschek JM, Ramamoorthy C, Haberkern CM, Hackel A, Caplan RA, Domino KB, Posner K, Cheney FW. Anesthesia-related cardiac arrest in children: initial findings of the Pediatric Perioperative Cardiac Arrest (POCA) Registry. Anesthesiology 2000;93:6 14
762
Case Report
Case Report
Management of the Difcult Infant Airway with the Storz Video Laryngoscope: A Case Series
Rebecca S. Hackell, AB Lisa D. Held, DO Paul A. Stricker, MD John E. Fiadjoe, MD
The incorporation of video technology into laryngoscopes provides an additional option for the management of difficult intubations. We report the successful use of the Miller 1 Storz Video Laryngoscope in seven infants with difficult direct laryngoscopy.
(Anesth Analg 2009;109:7636)
he miniaturization of video and fiberoptic technologies has allowed the integration of video cameras into laryngoscopes of various designs. Video laryngoscopy has been shown in adults to improve glottic view, facilitate guidance by novice laryngoscopists, and improve intubation success in difficult airways.1 6 The recent availability of pediatric-sized blades has allowed the use of these laryngoscopes in pediatric patients.714 The Storz Direct Coupled Interface Video Laryngoscope (SVL) (Karl Storz GmbH, Tuttlingen, Germany) integrates a fiberoptic bundle into the light source of a standard Miller type blade with a camera in the handle of the device. The image from the tip of the device is displayed on a screen providing a magnified view during intubation. The Miller 1 video blade has been available for use in our department for difficult and routine airway management. After IRB approval, we queried our prospective difficult intubation registry to identify cases in which the SVL was used in infants identified as being difficult to intubate by direct laryngoscopy (DL). Seven cases returned from this query are presented below.
had a history of a tracheoesophageal fistula repair in the neonatal period and was reported to be difficult to intubate. Her physical examination was remarkable for micrognathia and a short neck. Difficult intubation was anticipated based on the physical findings and a history of more than three attempts at DL by an otolaryngologist at the time of her tracheoesophageal fistula repair. General anesthesia was induced with thiopental and, after confirming the ability to ventilate by face mask, vecuronium was administered. Anesthesia was maintained with oxygen and sevoflurane. Laryngoscopy was performed by a Certified Registered Nurse Anesthetist (CRNA) with the Miller 1 SVL. A Cormack and Lehane Grade 1 view was obtained on the video screen, whereas a Grade 3 view was seen by a direct line of sight view. A styletted 3.5 uncuffed endotracheal tube was advanced into the trachea under video guidance on the first attempt.
Case 2
A 9-mo-old, 6.0-kg female with trisomy 18, Cri-du-Chat syndrome, and congenital dislocation of the right hip presented for a right hip closed reduction and spica cast application. The patients trachea had previously been intubated by fiberoptic bronchoscopy without incident. Her physical examination findings included micrognathia, a short neck, and a large tongue. Anesthesia was induced with sevoflurane in nitrous oxide and oxygen, followed by maintenance with an IV propofol infusion with preservation of spontaneous respiration. The glottis was anesthetized with topical lidocaine and laryngoscopy was performed by a CRNA with the Miller 1 SVL. A Grade 1 view was seen on the video screen and the trachea was intubated successfully with a 4.0 uncuffed endotracheal tube on the first attempt. Direct line of sight view at the time of video laryngoscopy revealed a Grade 3 view.
CASE DESCRIPTIONS
Case 1
A 4-mo-old, 4.5-kg female with CHARGE syndrome presented for bilateral myringotomy with tympanostomy tube placement and bilateral nasal dilation. She
From The General Anesthesia Division, Childrens Hospital of Philadelphia, University of Pennsylvania School of Medicine, Pennsylvania, Philadelphia. Accepted for publication March 27, 2009. Supported by the Department of Anesthesiology and Critical Care Medicine of The Childrens Hospital of Philadelphia and the University of Pennsylvania School of Medicine. Reprints will not be available from the author. Address correspondence to John E. Fiadjoe, MD, Department of Anesthesiology and Critical Care Medicine, The Childrens Hospital of Philadelphia, Main Building, 9th Floor, 34th and Civic Center Blvd., Philadelphia, PA 19104. Address e-mail to fiadjoej@email.chop.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ANE.0b013e3181ad8a05
Case 3
A 10-mo-old, 7.6-kg female with Moebius syndrome, laryngotracheomalacia, micrognathia, and a
763
history of a repaired gastroschisis presented for takedown of a tongue-lip adhesion. She had a history of a previous difficult intubation at 14 days of age requiring four attempts at DL (two attempts by a pediatric anesthesiology fellow and two by the attending anesthesiologist) before successful blind tracheal intubation. On the day of the current surgery, general anesthesia was induced with sevoflurane in oxygen. An IV catheter was placed and glycopyrrolate administered. Laryngoscopy with the SVL was performed by the pediatric anesthesiology fellow. A Cormack and Lehane Grade 1 view was seen on the video screen, whereas only the epiglottis was seen by direct line of sight (Grade 3 view). This intubation attempt failed because of difficulty in directing the tube into the trachea by video guidance. A second attempt by the fellow produced a Grade 3 view on the video and was aborted. The final intubation attempt was made by the attending anesthesiologist, which resulted in a Grade 1 view on the video monitor and a Grade 4 view by direct line of sight. The patient was successfully intubated on the third attempt with a 4.0 uncuffed endotracheal tube by video laryngoscopy.
(Bonfils Pediatric Endoscope), two attempts by the pediatric anesthesia fellow, and two by the attending anesthesiologist. The fifth attempt, performed by the fellow using the Miller 1 SVL, provided a Grade 2 view on the video monitor, allowing for successful tracheal intubation with a 4.0 uncuffed endotracheal tube.
Case 6
A 10-mo-old, 9.3-kg former 33-wk triplet with translocation of chromosomes 7 and 8 and severe global developmental delay presented for a gastrostomy tube placement. The patient had a history of difficult DL (Grade 3 view with application of optimal external laryngeal manipulation) during an anesthetic 3 wk prior. Anesthesia was induced with sevoflurane in nitrous oxide and oxygen and after confirming the ability to deliver positive pressure ventilation by face mask, vecuronium was administered. The Miller 1 SVL was used by the pediatric anesthesia fellow for intubation and the vocal cords were easily visualized (Grade 1 view). A styletted 3.5 cuffed endotracheal tube was placed on the first attempt.
Case 4
A 13-mo-old, 9.1-kg male with Goldenhar Syndrome, right mandibular retrusion, limited mouth opening, cervical spine instability, repaired cleft palate, and left thumb hypoplasia presented for thumb amputation and index finger pollicization. The patient had a history of a previous difficult intubation 1 mo prior requiring fiberoptic intubation through a laryngeal mask airway after a Grade 4 view by DL. For the current case, anesthesia was induced with sevoflurane in oxygen while maintaining the head and cervical spine neutral. After IV insertion, glycopyrrolate and vecuronium were administered after confirming the ability to deliver positive pressure ventilation by face mask. An initial attempt at DL by the CRNA using a standard Miller 1 larygnoscope failed to reveal any glottic structures (Grade 4 view). A second attempt by the CRNA using the SVL revealed a Grade 2 view with optimal external laryngeal manipulation. Tracheal intubation with a 4.0 cuffed endotracheal tube was successful on the first attempt using the SVL.
Case 7
A 3-mo-old, 5.5 kg otherwise healthy male presented for excision of a left frontal scalp dermoid cyst. The patient had a difficult intubation at 6 wk of age during a laparoscopic pyloromyotomy for pyloric stenosis requiring the use of a lighted stylet after three failed attempts at DL (two by the pediatric anesthesiology fellow and one by the attending anesthesiologist) with a Grade 3 view on each attempt. General anesthesia was induced with sevoflurane in nitrous oxide and oxygen, and propofol and glycopyrrolate were administered IV. The Miller 1 SVL was used successfully by the pediatric anesthesia fellow on the first intubation attempt. A 3.5 cuffed endotracheal tube was placed without difficulty; details of direct laryngoscopic view were not documented.
DISCUSSION
The SVL integrates video technology into a Miller type blade. This allows the laryngoscopist to intubate using a direct line of sight view or a video monitor view. The deflection of the blade at the tip is approximately 12 and the built-in camera has a 60 lens. Intubation with the SVL requires a styletted endotracheal tube with a bend at the very tip of the stylet (mimicking the tip angle of the SVL). Laryngoscopy is performed with the blade in the midline or to the left in the oropharynx (to maximize space for passage of the endotracheal tube). After obtaining optimal glottic exposure on the video monitor, the laryngoscopists attention is directed to the passage of the tube along or slightly to the right of the blade shaft with careful attention paid to avoiding injury to pharyngeal and tonsillar structures. Once the styletted tube is visualized on the screen, endotracheal intubation can ensue
Case 5
A 4-mo-old, 4.8-kg former 36-wk infant with a history of dysphagia and stridor presented for bronchoscopy, endoscopy, and percutaneous endoscopic gastrostomy tube placement. At 12 h of life, the patient had a prolonged apneic episode and intubation attempts at an outside institution were reportedly difficult. Details regarding intubation attempts and laryngoscopy grade were unavailable. General anesthesia was induced with sevoflurane in nitrous oxide and oxygen. Propofol was titrated to maintain spontaneous ventilation. Intubation was unsuccessfully attempted four times with a pediatric optical stylet
764
Case Report
with withdrawal of the stylet just after the tube tip passes the vocal cords. Our case series describes the successful airway management with the Miller 1 SVL of seven infants with a history of difficult intubation by DL. This report is consistent with the published literature in adults showing that video laryngoscopy provides a glottic view equal to or better than that seen with DL, as well as a prior case report documenting successful use of the SVL in an infant after a failed DL. Wald et al.13 reported the successful tracheal intubation of a 3-wk-old, 2.1-kg neonate with the Miller 1 SVL after failed DL. The Cormack and Lehane Grade obtained with DL was 3, whereas that obtained on the video monitor with the SVL was 1. Tracheal intubation was successful on the first attempt. The majority of the published literature regarding video laryngoscopy in pediatric patients focuses on older children.710,12,14 Because of transitioning anatomy, such as the descent of the larynx as children grow, devices that perform favorably in older children may be less effective or inappropriate in infants and neonates. In all our cases in which the Cormack and Lehane view was reported for DL, there was an improvement in grade using the SVL. Furthermore, use of the SVL resulted in a successful intubation in all of these patients in whom DL would likely have been extremely difficult or failed, as shown by prior experience in each patient. Use of the SVL appears to be easily learned. All the practitioners in our series were successful at intubation despite having varying levels of experience with the device. Six of the seven patients were intubated on the first attempt. Patient 3 was intubated on the third attempt after initial laryngoscopy revealed a Grade 1 monitor view, whereas the second attempt by the same practitioner revealed a Grade 3 view. There can be variability in the view obtained from performing laryngoscopy by a single operator with the same device. This can occur as a result of changes in the airway related to the laryngoscopy (e.g., edema), variable application of optimal external laryngeal manipulation, or inadequate positioning of the device. Inadequate positioning is more likely when the operator is not very experienced with the device. We suspect that the difference in grade observed by the fellow in Case 3 was related to inadequate positioning of the device. This suspicion is supported by the fact that the final attempt made by the attending anesthesiologist was a Grade 1 view, consistent with the first attempt of the fellow. Variable application of optimal external laryngeal manipulation may also have accounted for the differences in grade; however, there was no documentation in the anesthetic record to support this. The direct line of sight views in Patients 6 and 7 were not documented in the anesthetic record, and the only evidence for difficulty with laryngoscopy was
obtained from their prior anesthetic records. As children grow, the larynx moves caudally, the mandible and maxilla enlarge creating more space for laryngoscopy. It is possible that the successful intubations with video laryngoscopy in Patients 6 and 7 may have been related to enlargement of the airway with growth rather than any unique advantages of the video laryngoscope. This is unlikely in these patients given the short time between their intubations, 3 and 6 wk, respectively. Nevertheless, video laryngoscopy is very useful in the situation in which difficulty is suspected based on history because it allows the operator to document the direct line of sight grade and then intubate using the video monitor if the direct view is suboptimal. Video laryngoscopy requires some degree of handeye coordination. Difficulty with endotracheal tube placement despite having a favorable view may be a problem with many of these devices.2,3,5 There are two main factors that contribute to difficulty with intubation using the Miller 1 SVL. First, insertion of the styletted tracheal tube while looking at the video monitor requires a different skill set from intubation with DL. Second, difficulty in advancing the tube into the trachea after passing the cords may occur because of the anterior angulation of the distal end of the styletted tube required with this indirect method, which causes the tube to strike the luminal surface of the anterior tracheal wall and impede advancement down the trachea. This can often be overcome by rotating the tube 180 (after withdrawing the stylet) as it just passes the vocal cords and orients the endotracheal tubes concavity in a posterior fashion. Some limitations of the SVL include a lack of a built-in antifog system. Warming the device before use or applying an antifog solution to the tip can mitigate this problem. Visualization is impaired with excessive secretions or blood and the distal location of the camera sometimes necessitates elevation of the epiglottis to facilitate glottic exposure. The camera attachment requires a light source and a separate video monitor connection for displaying the image, making it less portable than some of the other video devices. The cost of the SVL ranges from $18,000 to $25,000 depending on the accessories included; there is currently no disposable version. The SVL can be sterilized by high level disinfection (e.g., Cidex, STERIS, STERRAD, and Medivator) systems. Several patients in our series would likely have been tracheally intubated using a fiberoptic bronchoscope had the SVL not been available. Fiberoptic intubation in infants and neonates can be challenging. The smallest fiberoptic scopes are more difficult to control and often lack a working channel,15 and the high oxygen consumption of neonates significantly shortens the time available for intubation. Our series suggests that the Miller 1 SVL may be a useful adjunct in the management of infants who either fail DL or are anticipated to present with a difficult DL. Prospective
765
randomized studies are needed to fully evaluate this possibility. REFERENCES

1. Shippey B, Ray D, McKeown D. Use of the McGrath videolaryngoscope in the management of difficult and failed tracheal intubation. Br J Anaesth 2008;100:116 9 2. Cooper RM. The GlideScope videolaryngoscope. Anaesthesia 2005;60:1042 3. Cooper RM, Pacey JA, Bishop MJ, McCluskey SA. Early clinical experience with a new videolaryngoscope (GlideScope) in 728 patients. Can J Anaesth 2005;52:191 8 4. Hirabayashi Y, Hakozaki T, Fujisawa K, Yamada M, Suzuki H, Satoh M, Hotta K, Igarashi T, Taga N, Seo N. [GlideScope videolaryngoscope: a clinical assessment of its performance in 200 consecutive patients]. Masui 2007;56:1059 64 5. Kaplan MB, Hagberg CA, Ward DS, Brambrink A, Chhibber AK, Heidegger T, Lozada L, Ovassapian A, Parsons D, Ramsay J, Wilhelm W, Zwissler B, Gerig HJ, Hofstetter C, Karan S, Kreisler N, Pousman RM, Thierbach A, Wrobel M, Berci G. Comparison of direct and video-assisted views of the larynx during routine intubation. J Clin Anesth 2006;18:357 62 6. Kim HJ, Chung SP, Park IC, Cho J, Lee HS, Park YS. Comparison of the GlideScope video laryngoscope and Macintosh laryngoscope in simulated tracheal intubation scenarios. Emerg Med J 2008;25:279 82
7. Borland LM, Casselbrant M. The Bullard laryngoscope. A new indirect oral laryngoscope (pediatric version). Anesth Analg 1990;70:105 8 8. Weiss M, Hartmann K, Fischer JE, Gerber AC. Use of angulated video-intubation laryngoscope in children undergoing manual in-line neck stabilization. Br J Anaesth 2001;87:453 8 9. Dullenkopf A, Holzmann D, Feurer R, Gerber A, Weiss M. Tracheal intubation in children with Morquio syndrome using the angulated video-intubation laryngoscope. Can J Anaesth 2002;49:198 202 10. Milne AD, Dower AM, Hackmann T. Airway management using the pediatric GlideScope in a child with Goldenhar syndrome and atypical plasma cholinesterase. Paediatr Anaesth 2007;17:484 7 11. Trevisanuto D, Fornaro E, Verghese C. The GlideScope video laryngoscope: initial experience in five neonates. Can J Anaesth 2006;53:423 4 12. Kim JT, Na HS, Bae JY, Kim DW, Kim HS, Kim CS, Kim SD. GlideScope video laryngoscope: a randomized clinical trial in 203 paediatric patients. Br J Anaesth 2008;101:531 4 13. Wald SH, Keyes M, Brown A. Pediatric video laryngoscope rescue for a difficult neonatal intubation. Pediatr Anesth 2008;18:790 2 14. Belyamani L, Zidouh S, Kamili ND. [Intubation using a videolaryngoscope in an adolescent girl with Rubinstein-Taybi syndrome]. Can J Anaesth 2008;55:57 8 15. Wood RE. Clinical applications of ultrathin flexible bronchoscopes. Pediatr Pulmonol 1985;1:244 8
766
Case Report
Ambulatory Anesthesiology
Section Editor: Peter S. A. Glass
Routine Use of Nasogastric Tubes Does Not Reduce Postoperative Nausea and Vomiting
Karl-Heinz Kerger* Edward Mascha Britta Steinbrecher Thomas Frietsch Oliver C. Radke Katrin Stoecklein# Christian Frenkel** Georg Fritz Klaus Danner Alparslan Turan Christian C. Apfel, MD, PhD For the IMPACT Investigators
Routine use of a nasogastric (NG) tube has been suggested to prevent postoperative nausea and vomiting (PONV) despite conflicting data. Accordingly, we tested the hypothesis that routine use of a NG tube does not reduce PONV. Our work is based on data from a large trial of 4055 patients initially designed to quantify the effectiveness of combinations of antiemetic treatments for the prevention of PONV. This analysis uses propensity scores for case matching to ensure group comparability on baseline factors. Intraoperative NG tube use patients and perioperative NG tube use patients were respectively matched to nonuse patients on all available potential confounders. Matched-pairs were identified using propensity scores for 1032 patients with or without intraoperative NG tube use and 176 patients with or without perioperative NG tube use. The incidences of PONV in the intraoperative group were 44.4% vs 41.5% (P 0.35) with and without tube use, respectively, and 27.8% vs 31.3% (P 0.61) in the perioperative group. Our results provide evidence that routine use of a NG tube does not reduce the incidence of PONV.
(Anesth Analg 2009;109:768 73)
he use of a nasogastric (NG) tube to prevent postoperative nausea and vomiting (PONV) has long been suggested in the literature. Postulated mechanisms for an effect have included decompressing the stomach and decreasing acidity. Given that the experience of the person ventilating the lungs with a face mask has been described as influencing PONV1 and that use of histamine-antagonists can reduce PONV,2 the routine use of a NG tube to prevent PONV appears
plausible. The effect of a gastric tube reported in the literature is very heterogeneous,3 but individual studies may be underpowered to detect a small but still clinically relevant difference. Using a dataset of more than 1000 patients, we tested the hypothesis that routine intraoperative or perioperative use of a NG tube would not affect the incidence of PONV. The primary endpoint in this analysis was incidence of PONV during the first 24 h postoperatively.
From the *Department of Anesthesiology and Critical Care Medicine, Evangelian Deaconry Hospital, Freiburg, Germany; Department of Anesthesiology and Operative Critical Care Medicine, University Hospital Mannheim, Mannheim, Germany; Departments of Quantitative Health Sciences, and Outcomes Research, Cleveland Clinic, Cleveland, Ohio; Westpfalz-Klinikum GmbH, Kaiserslautern, Germany; Perioperative Clinical Research Core, Department of Anesthesia and Perioperative Care, University of California at San Francisco, UCSF Medical Center at Mount Zion, San Francisco, California; Klinik und Poliklinik fu r Ana sthesiologie und Intensivtherapie, University Hospital Dresden, Dresden, Germany; #Department of Anesthesiology, VU University medical center, Amsterdam, The Netherlands; **Sta dtisches Klinikum Lu neburg gemeinnu tzige GmbH, Lu neburg, Germany; Department of Anesthesiology, Intensive Care Medicine, and Pain Therapy, Heart Centre Brandenburg at Bernau, Bernau, Germany; Department of Anaesthesiology, Trakya University, Edirne, Turkey; and The Outcomes Research Consortium, Department of Anesthesiology and Perioperative Medicine, University of Louisville, Louisville, Kentucky.
Accepted for publication February 15, 2009. Supported by a grant to the Perioperative Clinical Research Core from the Department of Anesthesia and Perioperative Care at the University of California, San Francisco. The International Multicenter Protocol to Assess the Single and Combined Benefits of Antiemetic Interventions in a Controlled Clinical Trial of a 2 2 2 2 2 2 Factorial Design (IMPACT) Investigators are listed in the Appendix. Address correspondence and reprint requests to Christian C. Apfel, MD, PhD, Perioperative Clinical Research Core, Department of Anesthesia and Perioperative Care, University of California San Francisco (UCSF), UCSF Medical Center at Mount Zion, 1600 Divisadero, C-447, San Francisco, CA 94115. Address e-mail to apfel@ponv.org or apfelc@anesthesia.ucsf.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aed43b
768
Table 1. Intraoperative Nasogastric (NG) Tube Use Versus Non-Use After Matching Based on Propensity Score Intraoperative NG tube use Effect
Propensity score Years of experience of anesthesiologista Patients age (yr)a Koivurantas PONV risk scorea Apfels POV risk score Weight (kg) Body mass index (kg/m2) Gender M F ASA classificationa I II III Centre (all used)a #2 (largest centre) Operation categorya Hernia repair Cholecystectomy Hysterectomy Thyroid surgery Breast surgery Hip replacement Knee arthroscopy Upper extremity ENT & eye surgery Other gynecologic surgery Bone surgery General surgery Surgical approacha Open abdominal Laparoscopic abdominal Other Antiemetic prophylaxis Ondansetron Dexamethasone Droperidol Anesthetic regimena Inhal, Fent, N2O Prop, Fent, N2O Inhal, Remi, N2O Prop, Remi, N2O Inhal, Fent, Air Prop, Fent, Air Inhal, Remi, Air Prop, Remi, Air
No NG tube use N
516 516 516 516 514 516 516 N 62 454 247 241 28 124 19 23 149 19 9 1 5 4 26 170 18 73 138 160 218 266 252 248 41 87 48 81 45 81 48 85
N
516 516 516 516 515 515 515 N 75 441 252 221 43 126 23 28 140 21 12 2 5 5 22 169 15 74 137 160 219 258 256 240 44 83 52 78 50 84 37 88
Mean (sd)
0.43 (0.24) 8.6 (7.1) 46.9 (14.1) 0.60 (0.16) 0.40 (0.16) 70.6 (14.2) 25.3 (4.5) % 14.5 85.5 48.8 42.8 8.3 24.4 4.5 5.4 27.1 4.1 2.3 0.39 0.97 0.97 4.3 32.8 2.9 14.3 26.6 31 42.4 50.0 49.6 46.5 8.5 16.1 10.1 15.1 9.7 16.3 7.2 17.1
Mean (sd)
0.42 (0.24) 8.3 (7.8) 47.0 (14.1) 0.60 (0.16) 0.40 (0.15) 69.8 (13.4) 25.2 (4.4) % 12 88 47.9 46.7 5.4 24 3.7 4.5 28.9 3.7 1.7 0.19 0.97 0.78 5 33 3.5 14.2 26.7 31 42.3 51.6 48.8 48.1 8 16.9 9.3 15.7 8.7 15.7 9.3 16.5
P
0.90 0.43 0.94 0.73 0.34 0.32 0.58 P 0.23 0.13
0.99 0.99
0.99
0.62 0.80 0.62 0.95
Patient distribution in the intraoperative nasogastric tube use group and the no use group after matching based on propensity score. All predictors, risk factors, operations, and anesthetic regimens are distributed evenly among the groups. PONV postoperative nausea and vomiting; POV postoperative vomiting; ENT ear, nose, and throat; Inhal inhalational; Prop propofol; Fent fentanyl; Remi remifentanil. a Variable used to create propensity scores.
METHODS
Our comparative study used data from the previously published International Multicenter Protocol to Assess the Single and Combined Benefits of Antiemetic Strategies in a Controlled Clinical Trial of Factorial Design (IMPACT)4 (Appendix). In the IMPACT trial, patients were randomized in double-blind fashion and assigned to several antiemetic strategies. The insertion of the NG tube was not
randomized and left to the discretion of the anesthesiologist. In patients with an intraoperative NG tube, the tube was placed after intubation, suctioned, capped, and removed under suction immediately before extubation, whereas in patients with perioperative use, it was left in place, suctioned, and capped, with intermittent suctioning for more than 24 h after surgery. In the postanesthesia care unit, the time, severity, and characteristics of PONV were recorded on standardized
769
Table 2. Perioperative Nasogastric (NG) Tube Use Versus Non-Use After Matching Based on Propensity Score Perioperative tube use N
Propensity score Body mass index (kg/m2)a Patients age (yr)a Anxiety before induction (VAS 010)a Years of experience of anesthesiologist Koivurantas PONV risk score Apfels POV risk score Weight (kg) Gender M F ASA classificationa I II III Centre #13 (largest) BIS-level according to stratificationa No BIS 3040 5565 Operation categorya Hernia repair Cholecystectomy Hysterectomy ENT & eye surgery Other gynaecologic surgery Bone surgery General surgery Surgical approacha Open abdominal Laparoscopic abdominal Other Antiemetic prophylaxis Ondansetron Dexamethasone Droperidol Anaesthetic regimena Inhal, Fent, N2O Prop, Fent, N2O Inhal, Remi, N2O Prop, Remi, N2O Inhal, Fent, Air Prop, Fent, Air Inhal, Remi, Air Prop, Remi, Air 83 83 83 83 83 83 83 83 N 27 56 28 42 13 15 45 20 18 4 13 10 5 13 1 37 55 15 13 44 39 52 7 12 3 15 10 19 8 9
No tube use N
83 83 83 83 83 83 83 83 N 26 57 26 45 12 12 46 17 20 3 11 6 4 13 2 44 51 15 17 48 43 42 7 15 2 12 9 17 10 11
Mean (sd)
0.37 (0.26) 25.66 (4.89) 52.64 (13.17) 3.47 (2.41) 7.63 (7.89) 0.56 (0.17) 0.37 (0.18) 71.63 (13.98) % 32.5 67.5 33.7 50.6 15.7 18.1 54.2 24.1 21.7 4.8 15.7 12.1 6 15.7 1.2 44.6 66.3 18.1 15.7 53.0 47.0 62.7 8.4 14.5 3.6 18.1 12.1 22.9 9.6 10.8
Mean (sd)
0.37 (0.26) 25.58 (4.66) 52.35 (17.35) 3.70 (2.69) 7.91 (9.34) 0.59 (0.15) 0.34 (0.15) 71.45 (13.09) % 31.3 68.7 31.3 54.2 14.5 14.5 55.4 20.5 24.1 3.6 13.3 7.2 4.8 15.7 2.4 53 61.5 18.1 20.5 57.8 51.8 50.6 8.4 18.1 2.4 14.5 10.8 20.5 12.1 13.3
P
0.98 0.92 0.90 0.56 0.83 0.12 0.21 0.93 P 0.23 0.90
0.99 0.84
0.88
0.71
0.53 0.53 0.12 0.98
Patient distribution among the perioperative (24 h) nasogastric tube use group and the no use group after matching based on propensity score. All predictors, risk factors, operations, and anesthetic regimens are distributed evenly among the groups. PONV postoperative nausea and vomiting; POV postoperative vomiting; ENT ear, nose, and throat; BIS Bispectral Index; Inhal inhalational; Prop propofol; Fent fentanyl; Remi remifentanil. a Variable used to create propensity scores.
forms. PONV was defined as the occurrence of nausea (using a severity score 0 10), vomiting/retching, or both during the first 24 h after surgery.
Associations between NG tube use and three 24-h outcomes (nausea, emesis, and overall PONV) were assessed using propensity score analysis. Any baseline
770
Nasogastric Tube Use and Risk of PONV
variable even remotely predictive of NG tube use, defined as P 0.30, was included in the calculation of the propensity scores, including such factors as experience of the anesthesiologist, patient age, PONV risk score, location, surgery type, surgical approach (open versus laparoscopic), anesthetic regimen, and clinical center. Our analysis for each exposure (NG tube use versus nonuse) thus consisted of two stages. In the first stage,
all available baseline factors were used in a model to predict NG tube use (yes/no), from which each patient was assigned a predicted probability of having received a NG tube. Each patient who actually did receive a NG tube was then matched on that probability to a nonuse patient using the greedy matching algorithm5 with a matching criterion of 0.05 propensity score units. In the second stage, we compared the matched NG tube groups (yes/no) on the outcome(s) of interest, PONV, using logistic regression analyses. Multivariable models included any covariates significant at the 0.05 level, further adjusting for any remaining imbalance on available potential confounders. Note that our multivariable analysis is based on the propensitymatched patients only and is quite distinct from a traditional multivariable model using all patients in the dataset, regardless of distribution of baseline variables. The significance level for the two-tailed 2 test was 0.05. For each analysis, we performed the usual twotailed test for superiority of one treatment versus the other. We also performed a nonsuperiority analysis in which we tested the null hypothesis that NG tube use is beneficial. We defined beneficial as a reduction in the odds of having the outcome by at least 5% with NG use, corresponding to an odds ratio (OR) of 0.95 or lower. The alternative hypothesis in this one-tailed test was that the OR is 0.95, i.e., that NG tube use is either worse than nontube use (OR 1) or that it reduces the odds of the outcome no more than 5% (OR 0.95). A significant test result would thus be interpreted as NG tube use being not superior to nonuse (i.e., either equivalent or worse). The significance level for each hypothesis was 0.05. No adjustment was made for assessing the three primary outcomes. SAS statistical software (Cary, NC, version 9.1) was used for all analyses.
Figure 1. Propensity score matched comparison for patients with versus without intraoperative NG tube use. 24-h PONV incidence was 44.4% in patients with an intraoperative NG tube use versus 41.5% in controls, for a difference of 2.9% (95% CI 3.2%, 9.1%). Perioperative NG tube use propensity score results are displayed in Table 4. There was no evidence of an association between perioperative NG tube use and reduction in nausea (0.85, P 0.65), emesis (0.90, P 0.83), or overall PONV (0.84, P 0.64). The 24-h PONV incidence was 27.8% in patients with perioperative NG tube use versus 31.3% in controls, for a difference of 2.4% (95% CI 16.1%, 11.1%). In our nonsuperiority analyses, we rejected the null hypotheses that intraoperative NG tube use was more beneficial (i.e., superior) compared with non-NG tube use for two of the three outcomes of interest, PONV (multivariable P 0.033) and nausea (multivariable P 0.037), assuming that an OR between 0.95 and 1.0 represents equivalence of the two methods of care (Table 3). From these one-tailed results, we infer that the adjusted OR for perioperative NG tube use is 0.95 for PONV and nausea. Nonsuperiority was not demonstrated for perioperative NG tube use (Table 4).
DISCUSSION RESULTS
A total of 4055 patients were initially considered for analysis: 2743 patients did not receive a NG tube, 1185 received a NG tube intraoperatively, and 127 received one intra- and postoperatively for 24 h. This initial grouping demonstrated imbalance on important baseline predictors of morbidity. Propensity scores were then used to compile a subgroup of matched NG tube use and control patients for intraoperative and 24-h postoperative use. Balance was achieved for all variables used in the propensity score matching and, innate to the methodology, also for variables that influence the risk for PONV (Tables 1 and 2). Results comparing propensity-matched intraoperative NG tube use versus controls are shown in Figure 1 and with more detail in Table 3. Intraoperative use of the NG tube use was not associated with a reduction in nausea (multivariable OR of 1.23, P 0.14), vomiting (0.92, P 0.64), or PONV (1.22, P 0.16). The
This analysis of a large case-matched dataset with more than 1000 patients evaluating the effect of a NG tube on PONV shows no evidence of a reduction in incidence of PONV. This result seems surprising given that mechanistically every effort that reduces intragastric volume should decrease the incidence of vomiting. A meta-analysis performed by Cheatham et al.6 identified 26 trials with 3964 patients and found no difference in the incidence of postoperative nausea but did find a decreased risk of vomiting. However, retching, which might occur instead of vomiting in the setting of an emptied stomach, was not separately accounted for in all included studies. Additionally, the effect of a gastric tube reported in the literature is so heterogeneous that no reasonable point estimates could be calculated in a Cochrane review by Nelson et al.3 Our analysis includes a significantly larger sample size than any other previous randomized controlled
771
Table 3. Propensity Score Analysisa for Intraoperative Nasogastric (NG) Tube Use P Outcome
PONV Emesis Nausea
Model (# covariates)
Univariable Multivariable (8) Univariable Multivariable (6) Univariable Multivariable (6)
NG tube use No (%)

41.5 18.2 40.5
Yes (%)
44.4 16.9 43.2
Odds ratiob (95% CI)

1.13 (0.881.44) 1.23 (0.931.61) 0.91 (0.661.26) 0.92 (0.661.29) 1.12 (0.871.43) 1.22 (0.921.60)
Superiorityc (null: OR 1)
0.35 0.14 0.57 0.64 0.38 0.16
Nonsuperiorityd (null: OR 0.95)

0.09 0.033* 0.60 0.52 0.10 0.037*
N 1032 for univariable and N 1029 for multivariable analyses. CI condence interval; PONV postoperative nausea and vomiting; OR odds ratio. a Using greedy matching within 0.05 propensity score units. b Odds ratio for NG tube use versus nonuse (reference). c Two-tailed test of null hypothesis of no NG tube effect. d One-tailed test against the null hypothesis that the odds ratio of NG tube use is 0.95 (i.e., superior), where the alternative hypothesis and signicant result means NG tube use odds ratio 0.95 (either equivalent to or worse than non-NG tube use, given the equivalence range of 0.951.0). * Signicant at P 0.05.
Table 4. Propensity Score Analysisa for Perioperative Nasogastric (NG) Tube Use P Outcome
PONV Emesis Nausea
Model (# covariates)
Univariable Multivariable (1) Univariable Multivariable (2) Univariable Multivariable (1)
NG tube use No (%)

31.3 12.0 30.1
Yes (%)
27.8 10.8 26.5
Odds ratiob (95% CI)

0.84 (0.431.64) 0.85 (0.431.71) 0.89 (0.342.31) 0.90 (0.322.47) 0.84 (0.431.65) 0.84 (0.421.69)
Superiorityd (null: OR 1)
0.61 0.65 0.81 0.83 0.61 0.64
Nonsuperioritye (null: OR 0.95)

0.64 0.61 0.56 0.55 0.64 0.63
N 166 (83/group) for univariable and N 165 for multivariable analyses. CI condence interval; OR odds ratio. a Using greedy matching within 0.05 propensity score units. b Odds ratio for NG tube use versus nonuse (reference). c Two-tailed test of null hypothesis of no NG tube effect. d One-tailed test against the null hypothesis that the odds ratio of NG tube use is 0.95 (i.e., superior), where the alternative hypothesis and signicant result means NG tube use odds ratio 0.95 (either equivalent to or worse than non-NG tube use, given the equivalence range of 0.951.0). e Signicant at P 0.05.
trial and should thus be able to detect even small effects present. The main limitation of this analysis is that the original study was not randomized for the use of a gastric tube; however, to address this drawback patients were matched using a propensity score to yield groups balanced on potential baseline confounders.7,8 In conclusion, these results provide strong evidence that the routine use of a NG tube during surgical procedures does not reduce PONV. ACKNOWLEDGMENTS The authors thank Dr. Anuj Malhotra for his careful editorial assistance. APPENDIX The IMPACT investigators are as follows: Steering CommitteeC. C. Apfel, A. Biedler, and K. Korttila. Data Management and MonitoringC. C. Apfel, E. Kaufmann, M. Kredel, A. Schmelzer, and J. Wermelt.
772
Nasogastric Tube Use and Risk of PONV
and Data AnalysesC. C. Apfel, K.-H. Kerger, and E. Mascha. Site InvestigatorsC. C. Apfel, S. Alahuhta, F. Bach, A. Bacher, H. Bartsch, H. Bause, A. Biedler, B. Book, H. Bordon, D. Buschmann, K. Danner, O. Danzeisen, D. Detzel, L. H. J. Eberhart, H. Feierfeil, H. Forst, C. Frenkel, G. Frings, B. Fritz, G. Fritz, A. Goebel, M. Hergert, C. Heringhaus, M. Hinojosa, C. Hoehne, W. Hoeltermann, H.-B. Hopf, C. Isselhorst, R. M. Jokela, E. Kaufmann, H. Kerger, T. KangasSaarela, P. Karjaleinen, A. Kimmich, M. Koivuranta, K. Korttila, U. Koschel, P. Kranke, M. Kredel, M. Lange, F. Liebenow, W. Leidinger, M. Lucas, C. Madler, J. N. Meierhofer, F. Mertzlufft, J. Motsch, S. Mun oz, E. Palencikova, A. Paura, S. Pohl, C. Prause, R. Rincon, N. Roewer, U. Ruppert, A. Schmelzer, I. E. Schneider, R. Sneyd, Schramm, A. Soikkeli, S. Spieth, B. Steinbrecher, K. Stoecklein, M. Trick, A. Turan, S. Trenkler, I. Vedder, P. Vila, J. Wermelt, K. Werthwein, W. Wilhelm, and C. Zernak.
REFERENCES
1. Hovorka J, Korttila K, Erkola O. The experience of the person ventilating the lungs does influence postoperative nausea and vomiting. Acta Anaesthesiol Scand 1990;34:2035 2. Doenicke AW, Hoernecke R, Celik I. Premedication with H1 and H2 blocking agents reduces the incidence of postoperative nausea and vomiting. Inflamm Res 2004;53(suppl 2):S154 S158 3. Nelson R, Edwards S, Tse B. Prophylactic nasogastric decompression after abdominal surgery. Cochrane Database Syst Rev 2005:CD004929 4. Apfel CC, Korttila K, Abdalla M, Biedler A, Kranke P, Pocock SJ, Roewer N. An international multicenter protocol to assess the single and combined benefits of antiemetic interventions in a controlled clinical trial of a 2 2 2 2 2 2 factorial design (IMPACT). Control Clin Trials 2003;24:736 51
5. Parsons LS. Reducing Bias in a Propensity Score Matched-Pair Sample Using Greedy Matching Techniques. In: Proceedings of the 26th annual SAS Users Group International (SUGI) Conference, Long Beach, California, 2001:214 26 6. Cheatham ML, Chapman WC, Key SP, Sawyers JL. A metaanalysis of selective versus routine nasogastric decompression after elective laparotomy. Ann Surg 1995;221:469 76; discussion 76 8 7. Cywinski JB, Koch CG, Krajewski LP, Smedira N, Li L, Starr NJ. Increased risk associated with combined carotid endarterectomy and coronary artery bypass graft surgery: a propensitymatched comparison with isolated coronary artery bypass graft surgery. J Cardiothorac Vasc Anesth 2006;20:796 802 8. Koch CG, Khandwala F, Nussmeier N, Blackstone EH. Gender and outcomes after coronary artery bypass grafting: a propensitymatched comparison. J Thorac Cardiovasc Surg 2003;126:2032 43
773
Anesthetic Pharmacology Preclinical Pharmacology

Section Editor: Marcel E. Durieux
Clinical Pharmacology
Section Editor: Tony Gin
Automated Responsiveness Monitor to Titrate Propofol Sedation

Anthony G. Doufas, MD, PhD* Nobutada Morioka, MD Adel N. Mahgoub, MD Andrew R. Bjorksten, PhD Steven L. Shafer, MD Daniel I. Sessler, MD
BACKGROUND: In previous studies, we showed that failure to respond to automated responsiveness monitor (ARM) precedes potentially serious sedationrelated adversities associated with loss of responsiveness, and that the ARM was not susceptible to false-positive responses. It remains unknown, however, whether loss and return of response to the ARM occur at similar sedation levels. We hypothesized that loss and return of response to the ARM occur at similar sedation levels in individual subjects, independent of the propofol effect titration scheme. METHODS: Twenty-one healthy volunteers aged 20 45 yr underwent propofol sedation using an effect-site target-controlled infusion system and two different dosing protocol schemes. In all, we increased propofol effect-site concentration (Ce) until loss of response to the ARM occurred. Subsequently, the propofol Ce was decreased either by a fixed percentage (20%, 30%, 40%, 50%, 60%, and 70%; fixed percentage protocol, n 10) or by a linear deramping (0.1, 0.2, and 0.3 g mL1 min1; deramping protocol, n 11) until the ARM response returned. Consequently, the propofol Ce was maintained at the new target for a 6-min interval (Ce plateau) during which arterial samples for propofol determination were obtained, and a clinical assessment of sedation (Observers Assessment of Alertness/Sedation [OAA/S] score) performed. Each participant in the two protocols experienced each percentage or deramping rate of Ce decrease in random order. The assumption of steady state was tested by plotting the limits of agreement between the starting and ending plasma concentration (Cp) at each Ce plateau. The probability of response to the ARM as a function of propofol Ce, Bispectral Index (BIS) of the electroencephalogram, and OAA/S score was estimated, whereas the effect of the protocol type on these estimates was evaluated using the nested model approach (NONMEM). The combined effect of propofol Ce and BIS on the probability for ARM response was also evaluated using a fractional probability model (PBIS/Ce). RESULTS: The measured propofol Cp at the beginning and the end of the Ce plateau was almost identical. The Ce50 of propofol for responding to the ARM was 1.73 (95% confidence interval: 1.552.10) g/mL, whereas the corresponding BIS50 was 75 (71.377). The OAA/S50 probability for ARM response was 12.5/20 (1213.4). A fractional probability (PBIS/Ce) model for the combined effect of BIS and Ce fitted the data best, with an estimated contribution for BIS of 63%. Loss and return of ARM response occurred at similar sedation levels in individual subjects. CONCLUSIONS: Reproducible ARM dynamics in individual subjects compares favorably with clinical and electroencephalogram sedation end points and suggests that the ARM could be used as an independent instrumental guide of drug effect during propofol-only sedation.
(Anesth Analg 2009;109:778 86)
ndividual requirements for sedatives vary. For example, there is considerable interindividual variability in the propofol effect-site concentration (Ce) at loss of responsiveness.13 Consequently, standardized dosing
From the *Department of Anesthesia, Stanford University School of Medicine, Palo Alto, California; Outcomes Research Consortium, Cleveland, Ohio; Department of Anesthesia and Perioperative Care, University of California San Francisco, San Francisco, California; Department of Anaesthesia, Royal Melbourne Hospital, Parkville, Australia; and Department of Anesthesiology, Columbia University, New York City, New York. Accepted for publication May 17, 2009. Supported by Scott Laboratories, Ltd. (Lubbock, TX) and the Joseph Drown Foundation (Los Angeles, CA).
provides inadequate medication for some patients while proving excessive in others.4 Effective titration of drug effect in individual patients is necessary to avoid potential adverse events due to inappropriate dosing.5,6
Steven L. Shafer is the Editor-in-Chief of the Journal. This manuscript was handled by James G. Bovill, Guest Editor-in-Chief, and Dr. Shafer was not involved in any way with the editorial process or decision. Address correspondence and reprint requests to Anthony G. Doufas, MD, PhD, Department of Anesthesia, Stanford University School of Medicine, 300 Pasteur Dr., H3590 Stanford, CA 943055640. Address e-mail to agdoufas@stanford.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b0fd0f
778
Patients given non-opioid sedatives are unlikely to experience respiratory and/or hemodynamic complications from doses that do not cause a loss of responsiveness. We have evaluated responsiveness during propofol sedation both clinically and with an automated responsiveness monitor (ARM), a novel negative feedback system for individual titration of propofol sedation.2,7 The system incorporates a handset approximately the size of a mobile phone that consists of a conveniently located thumb switch and vibrator. It is linked to a single ear bud headphone worn by the subject. A computerized voice asks the participant to push the button at intervals. The voice request is accompanied by vibration of the handset. The voice repeats the request four times over a 10-s period, becoming louder and more insistent with each repetition. The handset vibrates during questioning, and the vibration becomes progressively more intense with each query over the 10-s period. Pressing the button in response to any of the four queries presented in this 10-s window is considered an evidence of responsiveness and stops both the queries and vibration. In previous studies, we showed that failure to respond to the ARM precedes potentially serious sedation-related adverse events associated with loss of responsiveness, such as apnea and hypotension, and that the ARM was not susceptible to false-positive responses.2,7 It remains unknown, however, whether loss and return of response to the ARM occurs at similar sedation levels, independent of the scheme used to titrate the drug effect. The issue is important because time invariance is one of the basic assumptions of pharmacokinetics.8 As a corollary, a clinical surrogate measure of sedation must be highly predictable. Predictability in this context means that the dynamics of a behavioral titration instrument like the ARM must be reliable and have a reasonably limited intra- and interindividual variability. We therefore tested the hypothesis that loss and return of response to the ARM occur at similar sedation levels in individual subjects during propofol sedation. We tested our theory by increasing the propofol Ce until loss of response to the ARM was observed. After loss of response, we compared two methods of decreasing the propofol Ce: a fixed percentage decrease and a linear deramping of the Ce until the ARM response returned. Bispectral Index (BIS) of the electroencephalogram (EEG) and Ce information were used independently, as well as in combination, to characterize ARM response dynamics.
using a fixed percentage protocol and 11 others were studied using a linear deramping protocol, as described below. Age was restricted to 20 45 yr. Volunteers fasted at least 8 h before the trial. All standard anesthetic monitors including oscillometric blood pressure, electrocardiogram, end-tidal CO2 through a sealed anesthesia mask, and pulse oximetry (Spo2) were applied to the participating volunteers. Electrodes to capture the BIS of the EEG (A-2000 monitor, BIS 3.3 algorithm, system revision 1.07, Aspect Medical Systems, Newton, MA) were applied to the forehead according to the manufacturers instructions. The BIS recording began with a 2-min period of quiet relaxation with the volunteers eyes closed. Resistance of the BIS sensors was maintained at 5 k throughout the study period. A 20-gauge catheter was inserted at the antecubital fossa on the dominant arm for the propofol infusion, whereas a 20-gauge catheter was inserted into the radial artery in the contralateral arm for blood sampling. Normothermia was maintained with forced-air warming. Volunteers breathed supplemental oxygen via a sealed anesthesia mask to maintain a Spo2 more than 92%. The ARM apparatus was strapped loosely to the dominant hand. The volunteers were trained with the ARM handset and headphone for 10 15 min before the first sedation trial. The volume of the query was adjusted to a level that they were able to hear easily. We confirmed that the volunteers responded promptly to the ARM apparatus during this prestudy period. We used a target-controlled infusion (TCI) drug delivery system according to the method of Shafer and Gregg9 to target propofol Ce using the covariateadjusted propofol kinetic model reported by Schnider et al.10 with a ke0 of 0.46/min.11 The performance of the system was previously evaluated under pseudosteady-state conditions.12 The drug delivery system consisted of a Harvard 22 (Harvard Clinical Technology, South Natick, MA) electronic syringe pump, which could be commanded by a host system (Pentium II 450 MHz microprocessor-based system) through an RS232 serial communication port. A customized software platform, written by Scott Laboratories (Lubbock, TX), was used to drive the pump.
Fixed Percentage Protocol

Ten volunteers were studied six times each during administration of propofol using an infusion ramp with a slope adjusted to change the propofol Ce with a rate of 0.6 g mL1 min1. The infusion ramp was maintained until the volunteers lost response to the ARM apparatus, because this is determined by a nonresponse to a 10-s-long query period. Subsequently, propofol Ce was decreased by one of six randomly ordered percentages: 20%, 30%, 40%, 50%, 60%, and 70%. The target propofol Ce was maintained stable at this level for 6 min (Ce plateau) before the
METHODS
With approval of the University of California, San Francisco Human Studies Committees and written informed consent, we evaluated 21 healthy volunteers of both genders between April and May of 2000. All volunteers received propofol until they stopped responding to the ARM. Ten volunteers were studied
779
Figure 1. The time course of the predicted effect-site propofol concentration (Ce) in the fixed percentage and deramping dosing protocol schemes. In the fixed percentage protocol the Ce is decreased by a certain percentage (20%, 30%, 40%, 50%, 60%, or 70%), whereas in the deramping protocol the Ce is decreased by a certain rate (0.1, 0.2, or 0.3 g mL1 min1), after the first loss of automated responsiveness monitor (ARM) (LOA) response has occurred. A 6-min-long Ce steady state (SS) completes each individual sedation trial. The third graph presents the time course for the ARM. Each ARM test lasts for 10 s and consists of four individual queries (Q). Failure to press a button in response to all four queries that are presented during a 10-s-long ARM test is considered a nonresponse. Between two successive ARM tests there is an interval of 5 s (Methods).
infusion was stopped. Sedation was evaluated at 1-min intervals during this steady-state period. After a recovery period of at least 15 min, and a reduction in the predicted Ce to 0.5 g/mL, the next trial began, again consisting of an infusion ramp followed by a decrease of a different percentage. Each participant experienced each percentage decrease (Fig. 1).
Deramping Protocol
A set of 11 different volunteers was studied three times each during administration of propofol using an infusion ramp with a slope adjusted to change propofol Ce at a rate of 0.5 g mL1 min1. The infusion was maintained until the volunteers lost response to the ARM apparatus, because this is determined by a nonresponse during a 10-s query period. Subsequently, the Ce of propofol was decreased by one of three randomly ordered slopes adjusted to change Ce by 0.1, 0.2, or 0.3 g mL1 min1. The Ce of propofol was decreased until three successive positive responses to the ARM apparatus occurred. After the third ARM response, the propofol Ce was maintained at that level for 6 min (Ce plateau) before the infusion was stopped. At the end of each steady-state period, the sedation level was evaluated. As described above, after a recovery period of at least 15 min and a reduction in the predicted Ce to 0.5 g/mL, the next trial began to test a different deramping rate. Each participant experienced each deramping rate decrease in propofol Ce (Fig. 1).
Measurements
Demographic and morphometric characteristics of the volunteers were recorded. All standard physiologic values were downloaded and recorded to an automated data acquisition system for off-line analysis. These included heart rate, arterial blood pressure, respiratory rate, end-tidal CO2, and Spo2. BIS and all
780
Automated Responsiveness Monitor
standard anesthetic monitoring data, except for blood pressure, were recorded at 15-s intervals. Each ARM test lasted for 10 s, contained four individual queries, and was repeated with a rate of four tests per minute, i.e., after each completed (four queries) ARM test there was a 5-s interval before the initiation of the next one. Failure to press the button in response to all four queries that were presented during a 10-s ARM test was considered a nonresponse. On the contrary, a positive response to any of the four queries was considered a positive response to the ARM. Both the positive and the negative responses to each ARM test were recorded approximately 5 s after the end of each test, i.e., just before the initiation of the next test. Thus, the resolution of each ARM test was approximately 15 s. A detailed diagram in Figure 1 depicts the timing characteristics of the ARM test. An arterial blood sample for propofol determination was obtained at the beginning and the end of each Ce plateau in both protocols to document the presence of steady state. The samples were analyzed using a high performance liquid chromatography assay modified from the method of Plummer.13 This method has a coefficient variation of 4.1% at a propofol plasma level of 2 g/mL. Sedation was assessed clinically using the Observers Assessment of Alertness/Sedation score (OAA/S). The OAA/S score consists of four components. As described by Chernik et al.,14 we summed the component scores. The presence of consciousness was defined as an OAA/S score higher than 10 (of 20). The score was applied every 1 min at the Ce plateau during the fixed percentage protocol, and only once at the end of each Ce plateau during the deramping protocol. An attempt was made to evaluate sedation at the end of an ARM test so as to minimize any interference with the ARM function.
Data Analysis
Demographic and morphometric data were averaged across volunteers and presented for each protocol separately. We have previously shown1 that when propofol Ce is increasing at a rate between 0.1 and 0.9 g mL1 min1, a ke0 of 0.17/min (tpeak 2.7 min) more accurately reflects the plasma-effect-site equilibration than the previously reported value of 0.46/min (tpeak 1.7 min).11 Predicted Ce reported here are thus based on a ke0 of 0.17/min. The assumption of steady-state concentration at each plateau (Ce plateau) was tested using the Bland and Altman method.15 The difference in the measured plasma concentration (Cp) between the beginning (Cp-start) and the end (Cp-end) of each concentration plateau was presented as a function of the Cp-start and Cp-end average. Differences between the two protocols were assessed using unpaired t-test. The accuracy of the TCI system was evaluated by calculating the median performance error (MDPE) and the median absolute performance (MDAPE) error for each protocol separately, as previously proposed.16 First, for each blood sample the performance error (PE) was calculated as:
protocols (7185 data points) were used for this analysis. The effect of protocol type (fixed percentage or deramping) on the Ce50, ARM was tested by permitting different Ce50, ARM values when the protocol type was added to the model as a covariate. A decrease in the objective function of the complete model more than 3.84 points indicated a significant effect of the protocol on the ARM response dynamics. The same analysis, as above, was used to estimate the probability of a response to the ARM as a function of BIS (PBIS). In this analysis, instead of the BIS value the BIS difference from the baseline, determined as BIS effect (BISeffect 100 BIS), was used. The BISeffect 50, ARM, indicating the BISeffect associated with a 50% probability for response to ARM and the BISeffect, ARM, were estimated using NONMEM. The effect of the protocol type on BISeffect 50, ARM was evaluated by incorporating the protocol type into the model as a covariate. The combined effect of propofol Ce and BISeffect on the probability for ARM response (PBIS/Ce) was also evaluated. An independent variable, IND, was calculated as:
IND 100 BIS Ce

and logistic regression was used to model the probability of response to ARM (PBIS/Ce) as a function of IND:
PE
Cm Cp Cp
100
where Cm and Cp are the measured and predicted plasma propofol concentrations, respectively. Subsequently, the MDPE and the MDAPE were calculated for each subject separately. The median (range) values for MDPE and MDAPE were reported for each protocol separately. In addition, MDPE and MDAPE values at the beginning and the end of the Ce plateau were presented in a graph for each protocol separately. Logistic regression was used to estimate the probability of response (squeezing) to the ARM device as a function of the predicted propofol Ce. Each response to the ARM was given a score of 1, and each nonresponse to the ARM was given a score of 0. The probability of responding to ARM (PCe) was then calculated as:
PBIS / Ce 1
IND BIS/Ce, ARM IND50, ARMBIS/Ce, ARM INDBIS/Ce, ARM
The combined effect of propofol Ce and BISeffect, as expressed above, was considered significant if the minimum objective function (2 log likelihood, 2LL) of the model decreased by at least 3.84 points for each parameter added to the model. Finally, in the complete model, the combined effect of propofol Ce and BISeffect was evaluated after expressing the probability of ARM response (PBIS/Ce) as the sum of the probability fractions derived from the effects of propofol Ce (PCe) and BISeffect (PBIS):
PBIS / Ce FBIS PBIS 1 FBIS PCe ,

where FBIS is the fraction of the combined probability PBIS/Ce that is determined by BISeffect. The combined effect of propofol Ce and BISeffect, expressed as above, was considered significant if it produced a decrease by 3.84 points in the 2LL of the model, for each added parameter. The probability of response to the ARM was modeled as a function of the sedation level (OAA/S score) during the Ce plateau periods in both protocols, using logistic regression. The OAA/S50, ARM and OAA/S, ARM were also estimated by NONMEM. A total of 461 data points were used for this analysis. Bootstrap resampling with replacement was used to determine 95% confidence intervals for the P50
PCe 1
Ce Ce, ARM Ce 50, ARMCe, ARM Ce Ce, ARM
where Ce50, ARM is the predicted propofol Ce associated with a 50% probability for response to the ARM and Ce, ARM is the steepness of the Ce versus probability relationship (also termed the Hill coefficient). The parameters Ce50, ARM and Ce, ARM were estimated using nonlinear mixed effects modeling (NONMEM V, GloboMax LLC, Hanover, MD). Interindividual variability was permitted and assumed to be lognormally distributed. Residual intraindividual error was assumed to be additive. All data from both
781
Figure 2. Bland and Altman analysis of the measured propofol concentrations at the beginning (plasma concentration [Cp]-start) and the end (Cp-end) of each effect-site concentration (Ce) plateau in the fixed percentage and deramping protocols. The mean and 2sd of the difference (Cp-start Cp-end) are indicated by the horizontal continuous and dotted lines, respectively. The mean difference (sd) for the fixed percentage and deramping protocols were 0.04 (0.42) and 0.08 (0.33) g/mL, respectively (unpaired t-test, P 0.672). Data obtained during the two different protocols are presented separately. Median (range) of median performance error (MDPE) and the median absolute performance (MDAPE) values are presented as a function of the time at which the samples were drawn during steady state (SS). Table 1. Performance Metrics of the Target Controlled Infusion Protocol
Fixed percentage Deramping
MDPE (%)
9 (22 to 38) 31 (6 to 97)
MDAPE (%)
17 (1038) 33 (1097)
Median perfomance error (MDPE) and median absolute performance error (MDAPE) for the two sedation protocols. A total of 120 samples obtained during the xed percentage protocol and 58 samples obtained during the deramping protocol (in one volunteer we were not able to draw samples during the last deramping trial, whereas in another volunteer the insertion of arterial line was not feasible), were used in this analysis. Values are presented as medians (range).
Table 2. NONMEM Estimates and Coefcients of Variation for the Effect of Different Parameters on the P50 for Automated Responsiveness Monitor (ARM) Response P50 for ARM response Model parameters
Ce BISeffect OAA/S score
rounding each probability to the closest integer, i.e., 1 or 0) for all the individual observations for each of the tested parameters, or combination of parameters, separately. The median (range) propofol Ce at the first loss and first recovery of ARM response for each of the individual sedation trials (i.e., fixed percentages or deramping rates) were presented for each protocol and each individual volunteer, separately. In addition, the intra- and interindividual variability, calculated as the coefficient of variation (%), in the Ce and BIS at the first loss and first recovery of ARM response were presented for each protocol separately in a tabular and graph formats.
NONMEM estimate
1.73 24.9 12.5
95 % CI
1.552.10 23.028.7 12.013.4
Gamma (se)
6.20 (0.76) 5.12 (0.52) 5.31 (0.96)
RESULTS
The volunteers participating in the fixed percentage protocol (n 10) were 32.4 7.4 yr old, weighed 69.0 7.6 kg, and were 169.7 8.1 cm tall. The volunteers participating in the deramping protocol (n 11) were 33.4 5.8 yr old, weighed 69.9 12 kg, and were 172.4 9.5 cm tall. All physiology remained within normal limits during both sedation protocols. The measured arterial propofol Cp at the beginning (Cp-start) and the end (Cp-end) of the Ce plateau differed only by 0.05 0.39 g/mL (Fig. 2). Arterial propofol Cp obtained at steady state was used to assess the performance of the TCI. The results of this analysis are presented in Table 1 and Figure 2. Timing of the sampling has not been shown to be a confounding factor regarding the performance of the TCI system. The Ce50 of propofol for responding to the ARM was 1.73 (95% confidence interval: 1.552.10) g/mL, whereas the BISeffect 50 was 24.9 (23.0 28.7), corresponding to a BIS value of approximately 75 (Table 2, Fig. 3). The OAA/S score associated with a 50%
Propofol Ce, BISeffect, and Observers Assessment of Alertness/Sedation score (OAA/S) associated with a probability of 50% (P50) for responding to the automated responsiveness monitor (ARM). Gamma () represents the steepness of the parameter versus the probability for ARM response curve. All data (7185 data points) from both protocols were included in this analysis (see also Fig. 3). Best-estimate values with their 95% condence intervals (CI) calculated using 1000 bootstrap samples with replacement from the originally estimated P50 samples are presented. Ce effect site concentration; BISeffect Bispectral Index difference from the baseline.
estimates of Ce50, BISeffect 50, and OAA/S50. One thousand bootstrap samples (simple random samples of size 21 with replacement) were created from each of the three (i.e., Ce50, BISeffect 50, and OAA/S50) samples originally estimated by NONMEM, as described above. Confidence limits for each P50 were taken as the 2.5th and 97.5th percentiles of the respective bootstrap sample distribution. The ability of each tested parameter or combination of parameters to predict the observed response to the ARM device was investigated by calculating the average of probabilities (with and without
782
Figure 3. The probability of responding to the automated responsiveness monitor (ARM) as a function of propofol effect-site concentration (Ce), Bispectral Index (BIS) (expressed as BISeffect 100 BIS), and sedation/alertness determined by the Observers Assessment of Alertness/Sedation (OAA/S) score (nonresponse to verbal and tactile stimuli correspond to scores of 10 and 9/20, respectively). Vertical dotted lines indicate the values of the above parameters that are associated with a 50% probability of responding to the ARM, whereas the horizontal lines represent the 95% confidence intervals for these values. Data obtained from both the fixed percentage and deramping protocols are included in the analysis. Table 3. Modeling the Probability of Response to Automated Responsiveness Monitor (ARM), as a Function of Propofol Effect Site Concentration (Ce), Bispectral Index (BIS), and their Combination
Model parameters (N) (A) (A1) (B) (B1) (C) (D) Ce effect (N 2) Protocol effect on (A) (N 3) BIS effect (N 2) Protocol effect on (B) (N 3) BIS/Ce effect (N 3) BIS/Ce effect (N 5) Probability of response to ARM Minimum objective function 5895.57 5892.36
2LL
3.21 (A1 A) 0.03 (B1 B) 794.08 (C A) 372.82 (C B) 305.11 (D C)
P Ce 1
Ce
Ce, ARM
Ce 50,ARMCe, ARM Ce Ce, ARM BIS effectBIS, ARM
P BIS 1
, BIS effect 100 BIS BIS effect 50, ARMBIS, ARM BIS effectBIS ARM
BIS/Ce, ARM
5474.31 5474.28
P BIS/Ce 1
IND , IND 100 BIS Ce IND 50, ARMBIS/Ce, ARM IND BIS/Ce, ARM
5101.49 4796.38
P BIS/Ce FBIS P BIS 1 FBIS P Ce, FBIS
probability for ARM response was 12.5/20 (12.0 13.4). Figure 3 presents the probability of responding to the ARM as a function of Ce, BISeffect, and OAA/S score. Propofol Ce and BIS independently predicted the observed ARM response with a probability of 0.82 and 0.84, respectively, whereas their combination increased that probability to 0.85. The model that defined the probability of the observed ARM responses as the sum of the fractional probabilities determined by propofol Ce and BIS (PBIS/Ce PCe PBIS, Table 3, row D) demonstrated the lowest minimum objective function (-2LL 4796.38), when compared with other models that used only Ce (PCe 0.82, 2LL 5895.57, Table 3, row A) or BIS (PBIS 0.84, 2LL 5474.31, Table 3, row B), as predictor variables, and the model that combined Ce and BIS in the form of an independent variable IND 100 BIS Ce (PBIS/Ce 0.84, 2LL 5101.49, Table 3, row C). In the PBIS/Ce fractional probability model, the contribution of PBIS to the overall probability was estimated to be 63%. Table 4 presents the intra- and interindividual variability, regarding the propofol Ce and BIS when the first loss and recovery of ARM response occurred. Figure 4 presents the actual Ce values and its intraindividual variability at the above end points.
DISCUSSION
Using an effect-site TCI system, we have shown that ARM can titrate propofol sedation in a reproducible manner over time even in nonsteady-state conditions. Loss and return of response to the ARM occurred at similar Ce of propofol in individual subjects and with a reasonable interindividual variability during the fixed percentage and deramping protocols. Interestingly, when BIS and propofol Ce were independently used to characterize ARM response, the latter was not influenced by the protocol titration scheme. Our TCI system performed well during the fixed percentage protocol, and its performance was similar to what we have previously demonstrated,12 using the same kinetic model for propofol, developed by Schnider et al.,10 and subsequently validated by Doufas et al.1 The system did not perform as well during the deramping protocol but was still in an acceptable range. The replacement of our original ke0 value of 0.46/min10 with the value of 0.17/min is justified by two arguments: (a) the ke0 of 0.17/min has been estimated in a study of a young healthy population1 with similar characteristics as our present participants,
783
Table 4. Intra- and Interindividual Variability of the Effect Site Concentration (Ce) and Bispectral Index (BIS) for the Loss and Recovery of Automated Responsiveness Monitor (ARM) Response End Points Loss of ARM Protocol
Fixed percentage Ce BIS Deramping Ce BIS
Recovery of ARM Intra-CV (%)

17 7 83 19 11 66
Intra-CV (%)
13 3 95 16 11 64
Inter-CV (%)
25 2 14 3 23 6 13 3
Inter-CV (%)
27 5 13 3 32 7 12 5
Intra- and interindividual variability of the Ce and BIS at the rst loss and rst recovery of the response to ARM are expressed by the coefcient variation of Ce and BIS at the respective end points for the two protocols separately. Data are presented as means SD s.
Figure 4. In the upper graph, the mean (standard deviation) propofol effectsite concentration (Ce) at which the first loss of response to the automated responsiveness monitor (ARM) and ARM recovery occurred during the different sedation trials are presented for each individual volunteer and each protocol separately. The lower graph depicts the intraindividual variability of the Ce at the above end points, expressed as the coefficient of variation (%).
using similar infusion designs as in the current trial, and (b) the ke0 value of 0.17/min was derived from a study,1 which prospectively validated the propofol kinetic set developed by Schnider et al.10 In one study, adequately characterized sedation end points like loss and recovery of consciousness (defined by the ability to respond to verbal command) occurred at similar propofol Ce in each subject, despite the presence of a large interindividual variability.3 In this study, responsiveness to the ARM apparatus demonstrated similarly predictable dynamics with a propofol Ce50, ARM of 1.73 g/mL (95% confidence interval: 1.552.10, intraindividual variability 19%). This Ce value is comparable with those we have found previously in pseudosteady-state (1.6 g/mL)2 and nonsteady-state (1.76 g/mL)1 conditions and are approximately 0.51 g/mL less than the Ce at loss of response to verbal command2,3,7,17,18 or tactile stimulation.1 Furthermore, the BIS50 for ARM response was much higher than the BIS level previously associated with loss of response to verbal7,18,19 or tactile1 stimulation. BIS is highly correlated with
784
propofol Ce,20 and it has been shown to predict clinical sedation (OAA/S score) and loss of responsiveness comparably well21 with, if not slightly better20 than, the Ce. In addition, the P50 of the OAA/S score for responding to the ARM was 12.5/20 well above the threshold for loss of consciousness (10/20). This OAA/S50, ARM value is very similar to what we demonstrated previously under pseudosteady-state conditions,2 and it was not influenced by the applied protocol scheme. Thus, this study supports our previous finding that loss of response to the ARM tends to precede loss of consciousness during propofol-only sedation and might be used as a titration instrument when the actual loss of responsiveness is not a desired end point. Although it is difficult to compare the arousing potential between an ARM and a human-based (OAA/S scale) sedation assessment, the results of the present and previous2 trials support a preponderance of the latter. Nevertheless, we cannot exclude the possibility of an interaction between these two types of stimuli during our trial.
Propofol Ce and BIS predicted the observed ARM responses with a probability of 0.82 and 0.84, respectively. This probability increased to 0.85 when information from both Ce and BIS (fractional probabilities 0.37:0.63) were simultaneously used to predict ARM response. This was highly statistically significant, in regard to model improvement (Table 3), and reflects the importance of combining BIS and Ce information when attempting to characterize ARM response as an independent sedation end point. This result of our modeling approach provides two important insights: (a) it strengthens the evidence that the ARM relays information about a real drug effect, and (b) it questions the interchangeable use of Ce and BIS, two highly correlated but different, pharmacodynamic quantities, in characterizing drug effect. Targeting the Ce, rather than the Cp propofol concentration, has been associated with fewer hemodynamic and respiratory consequences.22,23 However, a relatively rapid increase of Ce always entails the risk of concentration overshoot and oversedation when the desired end point is reached, with potential adverse physiological consequences. An almost reflexive, preferably clinical, monitoring system is necessary to prevent oversedation and enhance safety when the sedative effect is progressing quickly. Patientmaintained sedation systems that use a Ce-driven propofol TCI have managed to provide safe sedation mainly by pursuing a slow, stepwise increase in the Ce.24 Nonetheless, a certain number of patients who used patient-maintained sedation were able to deliberately oversedate themselves, reaching a potentially unsafe sedation depth.25 As expected, the use of a Ce ramp (0.5 0.6 g mL1 min1) in our study not only increased the speed of sedation induction but also resulted in a relative Ce overshoot after discontinuation of the infusion when loss of response to the ARM occurred. This Ce overshoot led to oversedation in certain instances, which was never associated with severe adverse effects, such as apnea or hypotension. We conclude that loss and return of response to the ARM occurs at similar sedation levels in individuals, even though there is considerable variability among individuals. Reproducible ARM dynamics compares favorably with clinical and EEG sedation end points and suggest that the ARM can be used as an independent instrumental guide of propofol effect. However, the wider applicability of the ARM in clinical settings, which are usually compounded by multiple stimulating events and/or drug effects, remains to be tested. ACKNOWLEDGMENTS The authors greatly appreciate the assistance of Randy Hickle, MD, Brett L. Moore, BS, and Jason Derouen, BS (all from Scott Laboratories, Inc., Lubbock, TX).
REFERENCES
1. Doufas AG, Bakhshandeh M, Bjorksten AR, Shafer SL, Sessler DI. Induction speed is not a determinant of propofol pharmacodynamics. Anesthesiology 2004;101:111221 2. Doufas AG, Bakhshandeh M, Bjorksten AR, Greif R, Sessler DI. Automated responsiveness test (ART) predicts loss of consciousness and adverse physiologic responses during propofol conscious sedation. Anesthesiology 2001;94: 58592 3. Iwakiri H, Nishihara N, Nagata O, Matsukawa T, Ozaki M, Sessler DI. Individual effect-site concentrations of propofol are similar at loss of consciousness and at awakening. Anesth Analg 2005;100:10710 4. Smith I, Monk TG, White PF, Ding Y. Propofol infusion during regional anesthesia: sedative, amnestic, and anxiolytic properties. Anesth Analg 1994;79:3139 5. Bailey PL, Pace NL, Ashburn MA, Moll JW, East KA, Stanley TH. Frequent hypoxemia and apnea after sedation with midazolam and fentanyl. Anesthesiology 1990;73:826 30 6. Malviya S, Voepel-Lewis T, Tait AR. Adverse events and risk factors associated with the sedation of children by nonanesthesiologists. Anesth Analg 1997;85:120713 7. Doufas AG, Bakhshandeh M, Haugh GS, Bjorksten AR, Greif R, Sessler DI. Automated responsiveness test and bispectral index monitoring during propofol and propofol/N2O sedation. Acta Anaesthesiol Scand 2003;47:9517 8. Schwilden H, Schuttler J. Target controlled anaesthetic drug dosing. Handb Exp Pharmacol 2008;42550 9. Shafer SL, Gregg KM. Algorithms to rapidly achieve and maintain stable drug concentrations at the site of drug effect with a computer-controlled infusion pump. J Pharmacokinet Biopharm 1992;20:147 69 10. Schnider TW, Minto CF, Gambus PL, Andresen C, Goodale DB, Shafer SL, Youngs EJ. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998;88:1170 82 11. Schnider TW, Minto CF, Shafer SL, Gambus PL, Andresen C, Goodale DB, Youngs EJ. The influence of age on propofol pharmacodynamics. Anesthesiology 1999;90:150216 12. Doufas AG, Bakhshandeh M, Bjorksten AR, Greif R, Sessler DI. A new system to target the effect-site during propofol sedation. Acta Anaesthesiol Scand 2003;47:944 50 13. Plummer GF. Improved method for the determination of propofol in blood by high-performance liquid chromatography with fluorescence detection. J Chromatogr 1987; 421:171 6 14. Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, Davidson AB, Schwam EM, Siegel JL. Validity and reliability of the Observers Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990;10:244 51 15. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:30710 16. Varvel JR, Donoho DL, Shafer SL. Measuring the predictive performance of computer-controlled infusion pumps. J Pharmacokinet Biopharm 1992;20:6394 17. Chortkoff BS, Eger EI II, Crankshaw DP, Gonsowski CT, Dutton RC, Ionescu P. Concentrations of desflurane and propofol that suppress response to command in humans. Anesth Analg 1995;81:737 43 18. Milne SE, Troy A, Irwin MG, Kenny GN. Relationship between bispectral index, auditory evoked potential index and effect-site EC50 for propofol at two clinical end-points. Br J Anaesth 2003;90:12731 19. Irwin MG, Hui TW, Milne SE, Kenny GN. Propofol effective concentration 50 and its relationship to bispectral index. Anaesthesia 2002;57:242 8 20. Struys MM, Jensen EW, Smith W, Smith NT, Rampil I, Dumortier FJ, Mestach C, Mortier EP. Performance of the ARX-derived auditory evoked potential index as an indicator of anesthetic depth: a comparison with bispectral index and hemodynamic measures during propofol administration. Anesthesiology 2002;96:80316
785
21. Struys MM, Vereecke H, Moerman A, Jensen EW, Verhaeghen D, De Neve N, Dumortier FJ, Mortier EP. Ability of the bispectral index, autoregressive modelling with exogenous input-derived auditory evoked potentials, and predicted propofol concentrations to measure patient responsiveness during anesthesia with propofol and remifentanil. Anesthesiology 2003;99:80212 22. Kazama T, Ikeda K, Morita K, Kikura M, Doi M, Ikeda T, Kurita T, Nakajima Y. Comparison of the effect-site k(eO)s of propofol for blood pressure and EEG bispectral index in elderly and younger patients. Anesthesiology 1999;90:151727
23. Struys MM, De Smet T, Depoorter B, Versichelen LF, Mortier EP, Dumortier FJ, Shafer SL, Rolly G. Comparison of plasma compartment versus two methods for effect compartmentcontrolled targetcontrolled infusion for propofol. Anesthesiology 2000;92:399406 24. Chapman RM, Anderson K, Green J, Leitch JA, Gambhir S, Kenny GN. Evaluation of a new effect-site controlled, patientmaintained sedation system in dental patients. Anaesthesia 2006;61:3459 25. Murdoch JA, Grant SA, Kenny GN. Safety of patient-maintained propofol sedation using a target-controlled system in healthy volunteers. Br J Anaesth 2000;85:299 301
786
Should Dosing of Rocuronium in Obese Patients Be Based on Ideal or Corrected Body Weight?
Christian S. Meyhoff, MD, PhD* Jrgen Lund, MD Morten T. Jenstrup, MD Casper Claudius, MD, PhD* Anne M. Srensen, MD, PhD* Jrgen Viby-Mogensen, MD, DMSc* Lars S. Rasmussen, MD, PhD, DMSc*
BACKGROUND: Pharmacokinetic studies in obese patients suggest that dosing of rocuronium should be based on ideal body weight (IBW). This may, however, result in a prolonged onset time or compromised conditions for tracheal intubation. In this study, we compared onset time, conditions for tracheal intubation, and duration of action in obese patients when the intubation dose of rocuronium was based on three different weight corrections. METHODS: Fifty-one obese patients, with a median (range) body mass index of 44 (34 72) kg/m2, scheduled for laparoscopic gastric banding or gastric bypass under propofol-remifentanil anesthesia were randomized into three groups. The patients received rocuronium (0.6 mg/kg) based on IBW (IBW group, n 17), IBW plus 20% of excess weight (corrected body weight [CBW]20% group, n 17), or IBW plus 40% of excess weight (CBW40% group, n 17). Propofol was administered as a bolus of 200 mg and an infusion at 5 mg kg1 h1 and remifentanil was administered at 1.0 g kg1 min1, both according to CBW40%. Neuromuscular function was monitored with train-of-four nerve stimulation and acceleromyography. The primary end point was duration of action, defined as time to reappearance of the fourth twitch in train-of-four. RESULTS: The median (range) duration of action was 32 (18 49), 38 (25 66), and 42 (24 66) min in the IBW, CBW20%, and CBW40% groups, respectively (P 0.001 for comparison of the IBW and CBW40% group). There were no significant differences in onset time (85 vs 84 vs 80 s) or in intubation conditions 90 s after administration of rocuronium. CONCLUSIONS: In obese patients undergoing gastric banding or gastric bypass, rocuronium dosed according to IBW provided a shorter duration of action without a significantly prolonged onset time or compromised conditions for tracheal intubation.
(Anesth Analg 2009;109:78792)
linicians may have difficulties with dosage calculations in obese patients. Dosing of neuromuscular blocking drugs according to total body weight (TBW) is simple but carries the risk of prolonged duration of action in obese patients,1,2 because there are important differences in distribution, protein binding, and elimination of drugs between obese and lean patients.3
From the *Department of Anaesthesia, Centre of Head and Orthopaedics, Copenhagen University Hospital, Rigshospitalet, Copenhagen; and Department of Anaesthesia, Hamlet Hospital, Frederiksberg, Denmark. Accepted for publication May 18, 2009. Supported by Rigshospitalets Research Foundation (to CSM), Copenhagen, Denmark and the Novo Nordisk foundation (to LSR), Copenhagen, Denmark. JV-M and CC have received consulting fees from ScheringPlough (producing Esmeron and the TOF-Watch). Reprints will not be available from the author. Address correspondence to Christian S. Meyhoff, MD, Department of Anaesthesia, Section 4231, Centre of Head and Orthopaedics, Copenhagen University Hospital, Rigshospitalet, Blegdamsvej 9, Copenhagen DK-2100, Denmark. Address e-mail to christianmeyhoff@gmail.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b0826a
Therefore, it has been suggested that dosing should be based on ideal body weight (IBW) rather than on TBW,1,2 but this could be inadequate because of prolonged onset time and poor conditions for tracheal intubation. Pharmacokinetic parameters of drugs with low lipophilicity, such as rocuronium, are not very different in obese and lean patients.4 Two studies performed using objective monitoring of neuromuscular function have, however, included few obese patients (12 in both studies1,2), and clinically relevant end points, such as time to reappearance of the fourth twitch (T4) after train-of-four (TOF) stimulation and time to TOF ratio 0.90, were not reported. These end points are important, because they represent the time when reversal of the block can be initiated5 and the patient can be safely tracheally extubated,6 respectively. In addition, dosing of rocuronium according to IBW could be inadequate because of prolonged onset time and poor conditions for tracheal intubation. Therefore, it is unclear whether rocuronium in morbidly obese patients should be dosed according to IBW or to a dosing scheme based on corrected body weight (CBW), incorporating a percentage of the difference
787
between IBW and TBW.7 We hypothesized that a dosing scheme based on IBW would be most suitable in morbidly obese patients, because duration of action would be shorter, without compromising the conditions for tracheal intubation. The aim of this study was to compare duration of action, defined as time to reappearance of T4, onset time, and tracheal intubation conditions in three groups of morbidly obese patients in which the intubation doses of rocuronium were based on three different weight corrections.
METHODS
Danish Medicines Agency and the Local Ethics Committee approved the study (NCT00540085), and written informed consent was obtained from all subjects. Patients aged 18 65 yr, scheduled for laparoscopic gastric bypass or gastric banding were eligible. Exclusion criteria were expected difficult airway, interfering neuromuscular disease, known impaired hepatic or renal function, suspected allergy to the study drugs, and medications known to influence the neuromuscular transmission. The study adhered to the International Conference on Harmonisation Good Clinical Practice standards. We recorded height, TBW, waist-hip ratio, and body fat percent, using the skinfold technique,8 after which we calculated (in kg): 1. IBW Height in cm 106 (for women) or height 102 (for men),9 2. CBW20% IBW 20% (TBW IBW), 3. CBW40% IBW 40% (TBW IBW). CBW40% was chosen because this correction is proposed for dosage of propofol in obese patients,7 and the CBW20% group was included to increase the possibility of finding an optimal dosage algorithm. Before arrival to the anesthetic room, patients were randomly allocated 1:1:1 to receive rocuronium (0.6 mg/kg) according to IBW, CBW20%, or CBW40% using a set of computer-generated random numbers kept in sealed, sequentially numbered opaque envelopes. This rocuronium intubation dose was mixed with saline to a total of 10 mL and administered by the study investigators and blinded to the patient, and the anesthetic and surgical staff. Standard monitoring consisted of noninvasive arterial blood pressure, pulse oximetry, capnography, electrocardiography, and state entropy. Patients received oxycodone (10 mg) and paracetamol (1 g) orally approximately 1 h before anesthesia. Anesthesia was induced with IV infusion of propofol (5 mg kg1 h1) and remifentanil (1.0 g kg1 min1), both according to CBW40%, and 1 min later followed by propofol (200 mg) IV. After loss of the eyelash reflex, the patients were mask ventilated while the neuromuscular monitoring equipment was calibrated. Anesthesia was maintained with infusions
788
Rocuronium in Obese Patients
of propofol and remifentanil adjusted to a state entropy between 30 and 50. Neuromuscular monitoring followed good clinical research practice (GCRP) guidelines for pharmacodynamic neuromuscular studies.10 After careful cleaning of the skin, two pediatric surface electrodes (Neotrode, ConMed Corporation, NY) were placed over the right ulnar nerve near the wrist with a distance of 3 6 cm. The forearm and four ulnar fingers were immobilized and the acceleration transducer was secured on the thumb using a Hand Adapter (Organon, Oss, the Netherlands). The response to ulnar nerve stimulation was recorded with a TOF-Watch SX (Organon), and data were collected on a laptop using the TOF-Watch SX monitor program. The monitor display was blinded to the anesthetic and surgical staff by means of an opaque sticker, and the hand was covered by the sterile surgical covering. All IV infusions were to the left arm veins. Once the patient was unconscious, a 50-Hz tetanic stimulus was applied for 5 s, and after baseline stabilization (5% variation in at least 2 min), supramaximal stimulation and calibration was ensured using the built-in calibration function (CAL 2). The allocated rocuronium dose was given over 5 s in a rapidly flowing IV line. Laryngoscopy was commenced after 80 s, and tracheal intubation conditions were evaluated after 90 s by an experienced anesthesiologist blinded to the dose of rocuronium. The evaluation was based on a standard scheme including ease of laryngoscopy, position and/or movement of the vocal cords, and reaction to intubation.10 An upper body air-warming device was used to maintain a core temperature of more than 35C and peripheral skin temperature of more than 32C, as measured on the volar side of the thenar.10 The patients lungs were ventilated to normocapnia with 60% oxygen in nitrogen. If the neuromuscular blockade was insufficient for surgery, boluses of remifentanil (0.51.0 g/kg) or propofol (20 30 mg) were given. Rocuronium (10 mg) was given if the block was still insufficient 5 min later. The surgeons requested intervention if they considered the neuromuscular blockade to be insufficient, and the anesthesiologist administered the intervention in accordance with the protocol. None of them could see the response to TOF stimulation. Neostigmine (2.5 mg) together with atropine (1 mg) or glycopyrrolate (0.5 mg) were given at end of surgery if the patient had not recovered from neuromuscular block, defined as a TOF ratio less than 0.90. The trachea was extubated when the patient was fully awake, and TOF ratio was more than 0.90. In case of supplemental rocuronium or reversal of the block, neuromuscular data after that timepoint were not included in the data analysis. To evaluate blinding, the anesthesiologist and the surgeon were asked which of the three groups they
Table 1. Characteristics for 51 Patients Who Underwent Bariatric Surgery IBW group (n 17)
Age (yr) Sex (male:female) Height (cm) Body weight (kg) Body weight in percentage of IBW Body mass index (kg/m2) Body fat in percentage Waist-hip ratio ASA physical status category (I:II:III) Type of surgery (gastric bypass:gastric banding:other)a Duration of surgery (min) Duration of anesthesia (min) 37 (2956) 7:10 173 (166196) 132 (100184) 198 (150245) 44 (3455) 45 (3554) 0.98 (0.811.15) 1:15:1 14:2:1 72 (55128) 111 (88155)
CBW20% group (n 17)

41 (2057) 1:16 167 (156178) 122 (89180) 190 (164251) 42 (3657) 46 (3849) 0.91 (0.811.12) 0:16:1 16:1:0 85 (34117) 112 (70151)
CBW40% group (n 17)

34 (2249) 2:15 169 (160196) 133 (100194) 198 (161334) 45 (3872) 46 (3255) 0.87 (0.731.00) 0:17:0 17:0:0 85 (64108) 121 (93145)
Values are median (range). CBW corrected body weight; IBW ideal body weight. a Bariatric surgery could not be completed in one patient because of poor surgical conditions and intraabdominal oozing.
believed the patients belonged to. We recorded all adverse events occurring within 24 h of surgery. TOF-Watch data were stored on a computer, and the reliability of the neuromuscular data was reviewed by two blinded assessors based on specific quality variables according to GCRP guidelines (e.g., presence of supramaximal stimulation, baseline drift, and artifacts).10 The primary end point was the duration of action, defined as time from start of rocuronium injection to reappearance of T4. The secondary end point was onset time, defined as the time from the start of rocuronium injection to 95% depression of T1. The pharmacodynamic data obtained were as follows: time to reappearance of the T1, time to T1 recovery to 25% (of the final T1), and time to TOF ratio 0.90 (with and without normalization). Because the acceleromyographic control TOF ratio before administration of a neuromuscular blocking drug most often is more than 1.00, it has been suggested to refer all TOF ratios during recovery to the control value (normalization). If, for instance, the TOF ratio is 1.20 before injection of a neuromuscular blocking drug, a recorded TOF ratio of 0.90 during recovery corresponds to a normalized value of only 0.75 (0.90/1.20).11,12 Surgical conditions were evaluated by the surgeon at the start and the end of surgery as completely satisfactory, satisfactory, slightly unsatisfactory, or unacceptable. Patient characteristics are reported with median and interquartile range. Groups are compared with Wilcoxons unpaired rank sum test using SAS for Windows, version 9.1. P 0.05 was considered statistically significant. We estimated a 10 min sd for time to reappearance of T4 based on two previous studies.1,13 We considered a difference in time to reappearance of T4 of 10 min between the IBW and CBW40% groups to be clinically relevant. We calculated that a sample size of 17 patients in each group would allow us to detect this difference, with 5% Type 1 error risk, 80% power, and 10% dropout.
RESULTS
There were 59 eligible patients in the study period, of which two refused to participate, one had an expected difficult airway, one was not included because no research staff was available, and four had surgery postponed. The characteristics of the 51 enrolled patients are presented in Table 1. Rocuronium was administered 6.5 (sd 1.4), 6.6 (sd1.4), and 6.1 (sd1.1) min after start of propofolremifentanil infusion in the IBW, CBW20%, and CBW40% groups, respectively. Onset time was not significantly different among the groups (P 0.16 for comparison of the IBW and CBW40% group; Table 2; Fig. 1). The trachea was intubated at first attempt in all but three patients (two in the IBW group and one in the CBW40% group), in whom a second or third attempt was necessary. Conditions for tracheal intubation were similar in the three groups (Table 2). Time to reappearance of T4 was significantly longer in the CBW40% group than in the IBW group (P 0.001). The surgeon requested intervention for insufficient neuromuscular blockade leading to a supplemental dose of rocuronium in three patients in the IBW group (at TOF ratio 0.18, 0.50, and 1.10) and in two patients in the CBW40% group (at TOF ratio 0.28 and 0.85; Table 2). Surgical conditions were satisfactory or completely satisfactory at start as well as during surgery in 14, 15, and 15 of the patients in the IBW, CBW20%, and CBW40% groups, respectively. One serious adverse event occurred for a patient in the CBW20% group. This patient had convulsions and transient respiratory failure 6 h after surgery and underwent second surgery because of intestinal bleeding. All other adverse events are reported in Table 2. Anesthesiologists and surgeons correctly identified 16 of the 51 patients allocation. In their answers, there were no indications of unblinding.
789
Table 2. Pharmacodynamic Parameters and Adverse Events in 51 Obese Patients Given 0.6 mg/kg of Rocuronium as Intubation Dose According to Three Different Dosing Algorithms IBW group (n 17)
Rocuronium dose (mg) Rocuronium dose (mg/kg) Onset time (s) Time to reappearance of T1 (min) Time to reappearance of T4 (min) Time to T1 recovery to 25% of final value (min) (n 14 vs 15 vs 10)a Time to TOF ratio 0.90 (min) (n 14 vs 13 vs 10)a Time to normalized TOF ratio 0.90 (min)b (n 13 vs 11 vs 8)a Conditions for intubation (excellent:good:poor) (n 15 vs 17 vs 17)c Surgical conditionsd At surgical start During surgery Supplementary propofol or remifentanil boluses in percentage Supplementary rocuronium injections Reversal of neuromuscular blockade Any adverse events in percentage Peroperative hypotension or bradycardia Postoperative nausea or vomiting Postoperative pain Other 42 (5.3) 0.31 (0.05) 85 (66146) 21 (1236) 32 (1849) 28 (2144) 63 (43107) 69 (50126) 12:3:0 14:3:0:0 12:2:0:3 3 (18) 3 (18) 1 (6) 6 (35) 1 (6) 0 4 (24) 1 (6)
CBW20% group (n 17)

44 (5.5) 0.36 (0.03) 84 (62133) 28 (1652) 38 (2566) 35 (2448) 75 (52115) 80 (56130) 14:3:0 14:3:0:0 14:1:2:0 1 (6) 0 3 (18) 6 (35) 3 (18) 2 (12) 2 (12) 1 (6)
CBW40% group (n 17)

56 (8.7) 0.42 (0.03) 80 (48124) 31 (1350) 42 (2466) 38 (2152) 76 (48105) 78 (53109) 11:6:0 14:2:0:1 13:2:0:2 1 (6) 2 (12) 5 (29) 8 (47) 2 (12) 3 (18) 2 (12) 5 (29)
Values are mean (SD) or median (range). CBW corrected body weight; IBW ideal body weight; T1 rst twitch; T4 fourth twitch; TOF train-of-four. a Time only recorded for patients that did not receive supplementary rocuronium or reversal of the neuromuscular blockade before recovery to nal TOF ratio, TOF ratio 0.90, and normalized TOF ratio 0.90 was documented with certainty. b Normalized TOF ratio 0.90 refers to the time when TOF ratio reach 90% of the control TOF before the study drug was given. c Tracheal intubation conditions10 could not be fully evaluated in two patients, because their vocal cords were not visualized. Laryngoscopy was easy and there was no reaction to insertion of the tracheal tube and cuff ination in these two patients. d Completely satisfactory: satisfactory:slightly unsatisfactory:unacceptable. One patient in the CBW40% group received supplementary rocuronium and reversal of the neuromuscular blockade.
DISCUSSION
Our results showed that in obese patients who were given propofol-remifentanil anesthesia, a dosing scheme for rocuronium (0.6 mg/kg) based on IBW was preferred to other CBWs. When dosing was based on IBW, the duration of action of rocuronium was significantly shorter, without compromising the conditions for tracheal intubation or for surgery. The mean difference in time to reappearance of T4 between the IBW and CBW40% group was 12 min (95% confidence interval (CI): 6 19 min). The primary strength of our study was the objective neuromuscular monitoring with blinded assessment of quality variables performed in accordance with the GCRP guidelines.10 We studied an obese population with little comorbidity in a highly standardized anesthetic and surgical setting, and the study groups were well balanced with respect to age and body composition. Furthermore, we report two useful end points, representing the time when reversal of the block can be initiated (reappearance of T4)5 and when the patient can be safely tracheally extubated (TOF ratio 0.90),6 respectively. However, we used a relatively high remifentanil infusion rate at induction as per our usual clinical practice, and tracheal intubation was only attempted after the neuromuscular monitoring was setup. The propofol and remifentanil would have
Figure 1. Box plot illustrating onset time (A) and time to reappearance of T4 (B) in obese patients receiving 0.6 mg/kg of rocuronium according to three different weight corrections. Median and interquartile range (box), mean (), and range (bars). IBW ideal body weight; CBW corrected body weight.
790
facilitated the conditions for tracheal intubation, which were excellent in more than two-thirds of the patients.14 The IBW group had a higher proportion of men than the CBW20% and CBW40% groups. This could be a confounding factor, because the duration of action of a neuromuscular blocking drug may be shorter in males than in females.15 All measurements of body composition were, however, comparable, and within the IBW group, time to reappearance of T4 was median 31 and 37 min for female and male patients, respectively. Thus, we do not believe that this potential confounder has contributed significantly to the results of our study. In lean patients receiving rocuronium (0.6 mg/kg) according to TBW, the time to reappearance of T1 is around 20 min,16 and time to T1 recovery to 25% of control (duration 25%) ranged from 23 to 36 min.16 18 Our data in the IBW group approach this upper range, recognizing that duration 25% and reappearance of T4 occur approximately at the same time.19 We found the 75th percentile for duration of action in the CBW40% group to be 55 min. Accordingly, reversal of the block may not be initiated in 25% of the patients for almost 1 h after an intubation dose of rocuronium based on CBW40%. In contrast, the 75th percentile for duration of action in the IBW group was only 37 min. This general difference may be clinically important in morbidly obese patients, and in addition, duration of action can be prolonged in individual patients.20 Time to TOF ratio 0.90 with and without normalization was nearly 70 min in the IBW group, and this is longer than that reported in lean patients, i.e., 43 min after receiving 0.6 mg/kg of rocuronium according to TBW, but this was assessed using mechanomyography.18 We evaluated time to TOF ratio 0.90 only in patients who did not receive supplementary rocuronium or reversal of the neuromuscular blockade before TOF ratio 0.90 was documented with certainty. It must be acknowledged that this approach tends to underestimate the median time to TOF ratio 0.90. Onset time was almost the same as in previous studies of obese patients when rocuronium was given according to IBW1 but longer than in patients given rocuronium according to TBW, in which onset time was only 60 s.2 In lean patients, onset time is reported to range from 80 to 120 s.16 18 We did not find any significant differences in onset time among the three groups, and the mean difference between the IBW and CBW40% group was only 15 s (95% CI: 3; 33 s). Thus, considering the confidence interval, we cannot exclude that onset time can be reduced by approximately 30 s by dosing rocuronium according to CBW40%. However, we do not consider this possibility to be of major clinical importance in elective patients. The conditions for tracheal intubation appeared better than those previously reported when rocuronium was administered to lean patients. Excellent
conditions were obtained only in 15% 45% of the patients in these studies.16,21 This difference may be caused primarily by the high remifentanil infusion rate used in our study, although another study found excellent conditions in 25 of 30 lean patients using alfentanil.22 The statistical power is, however, limited in these relatively small studies, and the required sample size to detect a difference of 2% in the incidence of difficult intubation conditions would be approximately 750 patients with 80% power. The rocuronium dose per actual body weight was 0.31 mg/kg in the IBW group and 0.42 mg/kg in the CBW40% group. It is interesting to note similarities between this and reports of the use of low-dose rocuronium where good conditions for tracheal intubation can still be obtained, depending on the use of adjuvant drugs and the time that tracheal intubation is attempted. For example, 0.3 mg/kg given to lean patients resulted in optimal conditions for tracheal intubation in 18 of 20 patients.23 Our data indicate that in obese patients given rocuronium according to IBW, onset time is similar to lean patients given rocuronium according to TBW, whereas duration of action may be longer. The observed onset time around 80 s was clinically acceptable, and reversal of the neuromuscular blockade (reappearance of T4) was possible after a median of 32 min in the IBW group. The observed excellent conditions for tracheal intubation may not be generalizable to obese patients in other settings where lower doses of propofol and remifentanil are used at induction. Surgical conditions in bariatric laparoscopy were not impeded by earlier reappearance of T4 in our study. Surgical conditions were good and comparable among the three groups (satisfactory or completely satisfactory in 14, 15, and 15 of the patients in the IBW, CBW20%, and CBW40% groups, respectively). Our findings concern bariatric surgery and may not be generalizable to laparotomies, surgical procedures of considerably different duration, settings using less remifentanil at induction, or patients with significant hepatic or renal dysfunction. Also, special precautions are warranted in patients with expected difficult airway. We conclude that, in our patients undergoing gastric banding or gastric bypass under propofolremifentanil anesthesia, the intubation dose of rocuronium should be calculated according to IBW. The duration of action was shorter, and this was achieved without a significantly prolonged onset time or compromised conditions for tracheal intubation or surgery. ACKNOWLEDGMENTS The authors thank Susanne Schiang, Research Nurse, Department of Anesthesia, The Juliane Marie Center, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark as well as the anesthetic and surgical staff at Hamlet
791
Hospital, Sborg, Denmark for their contribution to data collection. REFERENCES

1. Leykin Y, Pellis T, Lucca M, Lomangino G, Marzano B, Gullo A. The pharmacodynamic effects of rocuronium when dosed according to real body weight or ideal body weight in morbidly obese patients. Anesth Analg 2004;99:1086 9 2. Puhringer FK, Khuenl-Brady KS, Mitterschiffthaler G. Rocuronium bromide: time-course of action in underweight, normal weight, overweight and obese patients. Eur J Anaesthesiol Suppl 1995;11:10710 3. Adams JP, Murphy PG. Obesity in anaesthesia and intensive care. Br J Anaesth 2000;85:91108 4. Puhringer FK, Keller C, Kleinsasser A, Giesinger S, Benzer A. Pharmacokinetics of rocuronium bromide in obese female patients. Eur J Anaesthesiol 1999;16:50710 5. Kirkegaard H, Heier T, Caldwell JE. Efficacy of tactile-guided reversal from cisatracurium-induced neuromuscular block. Anesthesiology 2002;96:4550 6. Eriksson LI. The effects of residual neuromuscular blockade and volatile anesthetics on the control of ventilation. Anesth Analg 1999;89:24351 7. Servin F, Farinotti R, Haberer JP, Desmonts JM. Propofol infusion for maintenance of anesthesia in morbidly obese patients receiving nitrous oxide. A clinical and pharmacokinetic study. Anesthesiology 1993;78:657 65 8. Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:7797 9. Viby-Mogensen J, Englbaek J, Eriksson LI, Gramstad L, Jensen E, Jensen FS, Koscielniak-Nielsen Z, Skovgaard LT, Ostergaard D. Good clinical research practice (GCRP) in pharmacodynamic studies of neuromuscular blocking agents. Acta Anaesthesiol Scand 1996;40:59 74 10. Fuchs-Buder T, Claudius C, Skovgaard LT, Eriksson LI, Mirakhur RK, Viby-Mogensen J. Good clinical research practice in pharmacodynamic studies of neuromuscular blocking agents II: the Stockholm revision. Acta Anaesthesiol Scand 2007;51:789 808
11. Claudius C, Viby-Mogensen J. Acceleromyography for use in scientific and clinical practice: a systematic review of the evidence. Anesthesiology 2008;108:1117 40 12. Kopman AF, Klewicka MM, Neuman GG. The relationship between acceleromyographic train-of-four fade and single twitch depression. Anesthesiology 2002;96:5837 13. Robertson EN, Driessen JJ, Booij LH. Pharmacokinetics and pharmacodynamics of rocuronium in patients with and without renal failure. Eur J Anaesthesiol 2005;22:4 10 14. Stevens JB, Wheatley L. Tracheal intubation in ambulatory surgery patients: using remifentanil and propofol without muscle relaxants. Anesth Analg 1998;86:459 15. Adamus M, Gabrhelik T, Marek O. Influence of gender on the course of neuromuscular block following a single bolus dose of cisatracurium or rocuronium. Eur J Anaesthesiol 2008;25:589 95 16. Schultz P, Ibsen M, Ostergaard D, Skovgaard LT. Onset and duration of action of rocuroniumfrom tracheal intubation, through intense block to complete recovery. Acta Anaesthesiol Scand 2001;45:6127 17. Naguib M. Neuromuscular effects of rocuronium bromide and mivacurium chloride administered alone and in combination. Anesthesiology 1994;81:388 95 18. Dahaba AA, Bornemann H, Holst B, Wilfinger G, Metzler H. Comparison of a new neuromuscular transmission monitor compressomyograph with mechanomyograph. Br J Anaesth 2008;100:344 50 19. Lee CM. Train-of-4 quantitation of competitive neuromuscular block. Anesth Analg 1975;54:649 53 20. Claudius C, Karacan H, Viby-Mogensen J. Prolonged residual paralysis after a single intubating dose of rocuronium. Br J Anaesth 2007;99:514 7 21. Combes X, Andriamifidy L, Dufresne E, Suen P, Sauvat S, Scherrer E, Feiss P, Marty J, Duvaldestin P. Comparison of two induction regimens using or not using muscle relaxant: impact on postoperative upper airway discomfort. Br J Anaesth 2007;99:276 81 22. Kopman AF, Klewicka MM, Neuman GG. Reexamined: the recommended endotracheal intubating dose for nondepolarizing neuromuscular blockers of rapid onset. Anesth Analg 2001;93:954 9 23. Barclay K, Eggers K, Asai T. Low-dose rocuronium improves conditions for tracheal intubation after induction of anaesthesia with propofol and alfentanil. Br J Anaesth 1997;78:92 4
792
Ketamine Inhibits Maturation of Bone Marrow-Derived Dendritic Cells and Priming of the Th1-Type Immune Response
Noriyuki Ohta, MD, PhD Yoshifumi Ohashi, MD Yuji Fujino, MD, PhD
BACKGROUND: Dendritic cells (DCs) play a key role as antigen-presenting cells and growing evidence suggests that DCs influence T-cell activation and regulate the polarity of the immune response. Ketamine has been reported to have immunomodulatory properties that affect immune cells, including macrophages and natural killer cells. However, the effect of ketamine on DCs has not been characterized. We examined the immunomodulation of DCs by ketamine. METHODS: We used bone marrow-derived DCs induced by granulocytemonocytecolony stimulating factor and interleukin (IL)-4 from bone marrow and analyzed the expression of costimulatory molecules (CD40, CD80, and CD86), major histocompatibility complex class II molecules, and secretion of IL-12p40. Furthermore, we evaluated the immune response in mixed cell cultures of DCs and T cells and the contact hypersensitivity response in a whole animal. RESULTS: Ketamine suppressed the expression of CD40, CD80, and major histocompatibility complex class II molecules in DCs. DCs treated with ketamine also secreted less IL-12p40 and displayed greater endocytosis. In mixed cell cultures with CD4 T cells and DCs, ketamine-treated DCs showed less propensity to stimulate the proliferation of CD4 T cells and the secretion of interferon from CD4 T cells. Furthermore, ketamine-treated DCs impaired the induction of a cell-mediated immune response. CONCLUSION: Our findings suggest that ketamine inhibits the functional maturation of DCs and interferes with DC induction of Th1 immunity in the whole animal. These novel findings provide new insight into the immunopharmacological role of ketamine.
(Anesth Analg 2009;109:793800)
he first step in the induction of an adaptive immune response is when professional antigen-presenting cells (APCs) present antigen to nave T cells. This presentation of foreign antigen involves an antigen-major histocompatibility complex (MHC) and costimulatory signals to nave T cells. Macrophages, B lymphocytes, epithelial cells, and dendritic cells (DCs) can all act as APCs, but DCs are the most potent in the initial presentation through the MHC to nave T cells and are responsible for the subsequent T cell-specific immune response. Thus, the possibility of harnessing the power of DCs has been given prominent attention in research immunology.13 In mice, DCs develop from bone marrow (BM) stem cells. Usually, DCs exist in a functionally and phenotypically immature state and in this state, because they do not express costimulatory molecules, they do not induce an immune response. Immature DCs begin to mature when they capture and process antigens in
From the Intensive Care Unit, Osaka University Hospital, Osaka, Japan. Accepted for publication March 30, 2009. Please see supplementary material available at www.anesthesiaanalgesia.org. Address correspondence and reprint requests to Noriyuki Ohta, MD, PhD, 2-15 Yamadaoka, Suita, Osaka, Japan. Address e-mail to nohta753@hp-icu.med.osaka-u.ac.jp. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc384
peripheral tissues. Maturing DCs stimulate nave T cells through signaling both by Ag peptides, presented by MHC molecules, and by costimulatory molecules. DC maturation is accompanied by decreased Ag uptake, high levels of MHC class II and costimulatory molecule expression, and production of interleukin (IL)-12 on stimulation.1,3 Anesthetics and sedatives play an essential role in the stable and safe control of critically ill patients. Furthermore, ketamine has been often used as an anesthetic for patients who are hemodynamically unstable due, for example, to sepsis or cardiac surgery. Consequently, the effects of various anesthetics on immunity have been extensively reported. Although macrophages, lymphocytes, and natural killer cells have been investigated, little attention has been paid to the effects of anesthetics and sedatives on DCs. Among many anesthetics, ketamine is an anesthetic that affects the immunoregulatory activities on macrophages,4,5 neutrophils,6 mast cells,7 and whole blood cells.8,9 Even so, we could find no report about how anesthetics affect DCs, the most potent of the APCs. In this initial study, we designed experiments to illustrate the effects of ketamine on some of the functions of DCs. We found that ketamine inhibits the maturation of BM-derived DCs (BMDCs) and furthermore, impedes the ability of DCs to prime a Th1biased immune response.
793
METHODS
Animals
To provide cell samples, female 4 6-week-old BALB/c and C57BL/6 mice were purchased from Japan CLEA (Tokyo, Japan). They were housed with free access to chow and water in the specific pathogenfree central animal facility of Osaka University Medical School. All experimental protocols in this study were reviewed and approved by the Institutional Animal Care and Use Committee and performed according to the National Institutes of Health guidelines.
staining, 2% propidium iodide (PI) was applied to stain dead cells and the samples were analyzed using FACSCalibur and CellQuest software (BD Biosciences, Franklin Lakes, NJ). We used FITC or PE-labeled monoclonal Abs to stain for MHC class, CD40, CD80, CD86, and CD11c. Dead cells were flow-cytometorically gated out by staining with PI and only live cells were phenotypically assessed.
Interleukin-12 p40 Enzyme-linked Immunosorbent Assay

DCs were cultured in the presence or absence of ketamine for 40 h followed by stimulation with LPS (10 ng/mL) for 12 h. Following the manufacturers instructions, we analyzed culture supernatants using IL-12 p40 enzyme-linked immunosorbent assay kits (R&D Systems, Minneapolis, MN).
Reagents and Antibodies

Recombinant mouse (rm) granulocytemonocytecolony stimulating factor and rmIL-4 were purchased from R&D Systems (Minneapolis, MN). Fluoresceinconjugated dextran (40,000 molecular mass) (fluorescein isothiocyanate (FITC)-dextran) and lipolysaccharide (LPS) (from Escherichia coli 055:B5) were obtained from Sigma-Aldrich (St Louis, MO). FITC- or phycoerythrin (PE)-conjugated monoclonal antibodies (mAbs) were used to detect the expression of CD11c (HL3), MHC antigen class II (I-Ad) (M5/114.15.2), CD40 (3/23), CD80 (16-10A1), and CD86 (GL1) were purchased from BD Pharmingen (San Diego, CA). For intracellular cytokine detection, we used mAbs for IL-4 and interferon (IFN) (11B11 and XMG1.2: BD Pharmingen).
Intracellular Cytokine Assay

For intracellular cytokine assay, T cells were restimulated with ionomycin and phorbol myristate acetate in the presence of GolgiStop (BD Pharmingen). Following the makers instructions, intracellular cytokines were detected using standard Cytofix/Cytoperm kits (BD Biosciences). We used PE-labeled monoclonal Abs to stain IL-4 and IFN- and FITC-labeled mAb for staining CD4.
Annexin V and Propidium Iodide Binding Assay

For analysis of DC apoptosis, using a FACSCalibur and CellQuest software (BD Biosciences), 1 105 cells were stained with FITC-labeled anti-CD11c mAb, washed in phosphate-buffered saline, and stained with PI and allophycocyanin-labeled Annexin V (BD Pharmingen).
Isolation and Culture of DC

DCs were generated from murine BM cells using a previously described method with minor modifications.10,11 Briefly, BM was flushed from the tibiae and femurs of Balb/C mice and then depleted of red blood cells with ammonium chloride. BM cells were suspended in complete media (CM: RPMI-1640 supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U/mL penicillin, and 100 mg/mL streptomycin). Cells for each mouse were plated to two 10-cm plates and cultured for 7 days in the presence of 20 ng/mL rm granulocytemonocyte colony-stimulating factor and rmIL-4 at 37C in 5% CO2. On Day 7, nonadherent cells and loosely adherent proliferating DCs were harvested and purified on 14.5% Accudenz (Accurate Chemicals, Westbury, NJ) density gradients by centrifugation at 600g for 20 min at room temperature. Purified DCs were cultured for 40 h in the absence and presence of ketamine in concentrations of 0 to 100 M. The viability of cultured cells was assessed by the trypan-blue exclusion test. Cell viability of more than 90% was observed in all experiments in the study.
Quantitation of Ag Uptake
As described by Sallusto et al.,12 endocytosis was quantitated. In brief, 2 105 cells were equilibrated at 37C or 4C for 10 min and then pulsed with FITCdextran (Sigma-Aldrich) at a concentration of 1 mg/mL. Cold staining buffer was added to stop the reaction. The cells were washed three times and stained with PE-conjugated anti-CD11c Abs and then analyzed using a FACSCalibur. Nonspecific binding of FITCdextran to DCs was assessed by evaluating FITCdextran uptake at 4C. Mean fluorescence intensity at 37C minus mean fluorescence intensity at 4C was used as the measure of antigen uptake. To enrich CD3 T cells in the sample, splenocytes, isolated from female 5 8-week-old C57BL/6 mice (haplotype: IAb), were purified in a magnetic cell sorting and separation of biomolecules (MACS) column (Miltenyi Biotec, Bergisch Gladbach, Germany). DCs induced from Balb/C mice (haplotype: IAd) were incubated in the presence or absence of ketamine. Subsequently, DCs were stimulated by LPS (20 ng/mL) for 12 h and treated with mitomycin C (50 g/mL). Enriched CD3 T cells were cocultured with mitomycin C-treated DCs. In 96-well round-bottom
Allogeneic Mixed Cell Culture Reaction
Flow Cytometric Analysis of Surface Molecules

The expression of surface molecules on DCs was analyzed under flow cytometry. At each step of the staining, to block nonspecific binding of antibodies, 12 105 cells were incubated for 15 min on ice in staining buffer containing anti-CD16/CD32 mAb. Cells were stained with specific antibodies. After mAb
794
Inhibitory Effect of Ketamine on Dendritic Cell Function
plates, these mixed samples were cultured for 4 days in RPMI 1640 supplemented with 10% fetal bovine serum at 37C in 5% CO2 in air. Cell proliferation was estimated based on uptake of [3H]-thymidine. For this purpose, the cells were pulsed with [3H]-thymidine for the final 18 h of mixed cell culture. Radioactivity was then measured using a liquid scintillation counter (PerkinElmer, Waltham, MA). Intracellular cytokines were detected using standard Cytofix/Cytoperm kits following the manufacturers instructions (BD Bioscience).
Establishing Contact Hypersensitivity by Adoptive Transfer of DCs

DCs were prepared as described earlier. After being cultured in the presence or absence of ketamine, 5 105 DCs (in 100 L of saline) were pulsed with 100 g/mL 2,4-dinitrobenzene sulfonic acid (DNBS) and injected subcutaneously on Day 0. After 5 days, mice were challenged by the application on both sides of the right ear of 10 L of 0.2% 2,4-dinitro-1fluorobenzene (DNFB) in 4:1 acetone/olive oil solution. Negative control animals were injected with only 100 L of saline at the time of initial immunization and exposed to DNFB 5 days later. After 24 h, using an engineers spring-loaded micrometer (Mitsutoyo, Kawasaki, Japan), we measured the right (challenged) and the left (unchallenged) ear thickness. The increase in the ear thickness was evaluated by simple subtraction: thickness of the right (challenged) earthickness of the left (unchallenged) ear.
Statistics
Data for samples were compared using the Students t-test. When P 0.05 the difference was considered statistically significant. All statistical analysis was performed using JMP software (SAS Institute, Cary, NC).
RESULTS
Ketamine Inhibits Maturation of Murine DCs
To examine whether ketamine influences the maturation of DCs, we cultured immature DCs induced after 7 days of culture for another 40 h in the presence or absence of different concentrations of ketamine. The expression of surface molecules (CD11c, MHC class II) on DCs was assessed using flow cytometry. As DC maturation progressed, we observed a higher presence of MHC class II molecules. Figure 1 shows CD11c cell populations. Ketamine had no effect on the numbers of BM cells found in the wells of culture dishes after BM cells were cultured in the presence or absence of ketamine in various concentrations of up to 200 M (data not shown). Similarly, DCs cultured in the presence or absence of ketamine, at least in concentrations of up to 100 M, showed no statistically significant differences in the total number of CD11c DCs including mature MHC class IIhigh and immature MHC class IIlo DCs (Fig. 1A). To further exclude the possibility that ketamine has a toxic action on
tions of up to 100 M, ketamine has no influence and no cytotoxicity on CD11c DC. DCs were generated as described in Methods section (without or in the indicated concentration of ketamine) and analyzed at Day 9 by flow cytometry. DCs were stained for CD11c (A), and Annexin V and PI (B and C). A single typical result of Annexin V and PI staining is shown from samples for each set of three independent experiments. The percentage within each histogram represents the incidence of CD11c cells (A) and Annexin-VPI cells (C). The cytotoxic effect of ketamine was evaluated from the fraction of necrotic cells that were detected as AnnexinVPI cells. Results are means sds (n 6). DC dendritic cell; PI propidium iodide.
Figure 1. Effect of ketamine on CD11c DC. In concentra-
cultured DCs, using an Annexin V-PI binding assay, we assessed the effect of ketamine, in concentrations of up to 100 M, on apoptosis (Fig. 1B and C). Our findings showed that ketamine has no harmful effect on DCs in concentrations of up to 100 M. In subsequent protocols, we used ketamine in concentrations of 40 M. We ascertained that the total number of CD11c DCs, including mature MHC class IIhigh and immature MHC class IIlo DCs, was not affected by exposure to ketamine in concentrations of up to 100 M. As immature DCs matured, we observed higher expression of MHC class II, CD40, CD80, and CD86 molecules. We found that ketamine specifically decreased the number of MHC class IIhigh DCs. Figure 2 shows data for the expression of MHC class II, CD40, CD80, and CD86. The expression of CD40 and CD80 molecules, important for costimulating T-cell activation, was also suppressed when ketamine was present. This finding indicates that the phenotypic maturation of
795
Figure 2. Effect of ketamine on expression of costimulatory

molecules on DCs. Ketamine suppresses DC maturation. DCs were treated with ketamine (40 M). A, Flow cytometry was used to assess the expression rate for CD40, CD80, CD86, and major histocompatibility complex (MHC) class II molecules in CD11c-gated cells. Ketamine inhibits the expression on DCs of CD40, CD80, and MHC class II molecules. B, Representative histogram of the expression of each surface molecule on DC. The number in each histogram shows the percentage of DCs (CD11c cells) expressing highly each molecule. Results are means sds (n 4). DC dendritic cell.
Figure 3. Effect of ketamine on production of interleukin-12

p40 from DCs (A) and on endocytotic activity of DCs (B). Ketamine inhibits the production of interleukin-12 p40 (IL12p40) from DCs. A, DCs were treated with ketamine (40 M) for 40 h and thereafter exposed to 10 ng/mL of lipopolysaccharide (LPS) for 12 h. The expression of IL12p40 was measured by cytokine-specific ELISA in culture supernatant. B, Ketamine-treated DCs exhibit enhanced endocytotic capacity. DCs were generated as described in Methods section and harvested on Day 7. FITC-dextran was analyzed on CD11c-PEpositive cells by flow cytometry. The uptake of FITC-dextran is shown as the product of mean fluorescence intensity (MFI) at 37C minus MFI at 4C. Results are means sds (n 6). DC dendritic cell; PE phycoerythrin; FITC fluorescein isothiocyanate; ELISA enzyme-linked immunosorbent assay.
DCs is at least partially retarded by the presence of ketamine. Conversely, no differences in the expression of CD86 molecules were detected between samples that had been exposed to ketamine and control samples.
Ketamine Inhibits Secretion of Interleukin-12 p40 from DC

Mature DCs are important for the synthesis and secretion of cytokines that affect T-cell differentiation and are largely responsible for the quality of immune response. DCs produce proinflammatory cytokines: IL-12 production is a particular marker of DC maturation and can be used as method of selecting the Th1dominant adjuvant. Here, we tested the expression of IL-12p40 from LPS-stimulated DCs. As Figure 3A shows, ketamine inhibited the secretion of IL-12p40.
Ketamine Induces Immature State DCs with High Endocytotic Capacity

Evaluation of the expression of surface molecules and IL-12p40 secretions indicated that exposure to ketamine significantly suppressed the phenotypic and functional maturation of DC generated in vitro. These results did not exclude the possibility that ketamine
796
might, however, cause a general inhibition of DC functions. Consequently, we assessed whether exposure of DCs to ketamine alters the ability of DCs to capture Ag through the uptake of FITC-conjugated dextran. As Figure 3B shows, after 40 h exposure to ketamine, DCs displayed increased endocytotic capacity for FITC-dextran. All our findings, including the expression of surface molecules and IL-12p40 and endocytotic activity, strongly suggest that ketamine prevents the maturation of DCs. Ketamine inhibits the ability of DCs to stimulate T-cell proliferation and to prime the Th1-type immune response in allogeneic mixed cell culture reaction. To clarify the relevance to immune responses of the ketamine-mediated alteration of DC functions that we had so far observed, we analyzed the effect of ketamine on the mixed cell culture reaction of lymphocytes and DCs. Using samples derived from Balb/C mice (haplotype: IAd), DCs that had been incubated in
Figure 4. In the presence and absence of ketamine: in vitro dendritic cell (DC)-induced proliferation of allogeneic T cells and Th1 response. Ketamine impeded the proliferation of allogeneic T cells and Th1 response. DCs were incubated with and without ketamine (40 M) for 48 h with the administration of lipolysaccharide (LPS) for final 12 h. The DCs were washed and cocultured with T cells derived from Balb/C mice as described in Methods section. A, Mixed cell culture was performed for 4 days. For the final 18 h of mixed cell culture, [3H]-TdR was added to the culture. Cell proliferation was estimated based on uptake of [3H] thymidine, which was pulsed during the final 18 h of culture. The radioactivity of the harvested cells was measured using a liquid scintillation counter (PerkinElmer). B, After 64 h of mixed cell culture, clustering was counted under microscopic analysis. Results are means sds (n 6).
Figure 5. Ketamine inhibits the production of interferon (IFN) from CD4 T cells stimulated by allogeneic dendritic cell (DC) culture in the presence or absence of ketamine. In the presence or absence of ketamine: production of IFN- from CD4 T cells stimulated by allogeneic DC culture. Ketamine inhibited the production of IFN-. Mixed cell culture was performed with bone marrow-derived DCs (BMDCs) from Balb/C mice (IAd) and CD3 T cells from C57BL/6 mice (IAb) for 4 days. At Day 4, T cells were harvested and restimulated with ionomycin and phorbol myristate acetate (PMA) in the presence of GolgiStop (BD Pharmingen). The production of IFN- and interleukin (IL)-4 from CD4 T cells was assessed by flow cytometry with intracellular staining of cytokine using Cytofix/Cytoperm kits (BD Pharmingen). Single, typical results are shown from each set of three independent experiments. Results are means sds (n 6). (Fig. 5). No differences in the secretion of IL-4 were detected in ketamine-treated and control DC samples.
Ketamine-Treated DCs Fail to Elicit Contact Hypersensitivity Response

Contact hypersensitivity (CHS) is a typical cellmediated immune response that is induced mainly by Th1-type T cells. CHS is originally induced by epicutaneous immunization and subsequent challenge with haptens, such as DNFB. CHS can be also induced by a single subcutaneous injection at Day 0 of 5 105 DCs that have been pulsed with DNBS (the water-soluble derivative of DNFB) and subsequent epicutaneous challenge with DNFB at Day 5.13,14 We used this model to assess the in vivo effect of ketamine on DCs and DC-mediated Th1-type immune response in the whole animal. Immunization with ketamine-treated DCs elicited less CHS response than immunization with phosphate-buffered saline-treated DCs (Fig. 6).
the presence or absence of ketamine were tested for their capacity to stimulate allogeneic T cells derived from C57BL/6 (haplotype: IAb). From Day 7 cultures that had been incubated with ketamine for 40 h, DC samples were tested for their capacity to stimulate allogeneic T cells. Although coculture with LPS-stimulated DCs effectively enhanced proliferative responses, when LPSstimulated DCs were pretreated with ketamine, there was less [3H] uptake by allogeneic T cells (Fig. 4A). Another determinant of DCs potency is how well they are able to adhere to T cells to form clusters (Fig. 4B). Compared with control cells, we found that ketaminetreated DCs formed fewer clusters on T cells. We also evaluated the effects on IFN- and IL-4 synthesis when CD4 T cells were cocultured with ketamine-treated DCs. Intracellular cytokine analysis revealed a lower density of IFN-producing CD4 cells in cocultures that contained ketamine-treated BMDCs
DISCUSSION
As far as we know, this is the first report detailing the effects of ketamine on the phenotypic and functional properties of murine DCs. As such, it is also the
797
Figure 6. Ketamine inhibits Th1-type immune response

in DC-transfer model of contact hypersensitivity (CHS). Ketamine-treated DCs fail to induce a normal cell-mediated immune response. As described in Methods section, after being cultured in the presence or absence of ketamine, 5 105 DCs were pulsed with 100 g/mL DNBS and injected subcutaneously on Day 0. After 5 days, mice were challenged by the application on both sides of the right ear of 10 L 0.2% DNFB in 4:1 acetone/olive oil solution. Right (challenged) and left (unchallenged) ear thicknesses were measured after 24 h and the results are shown as simple subtraction. Results are means sds (n 6). DC dendritic cell; DNBS 2,4-dinitrobenzene sulfonic acid; DNFB 2,4-dinitro-1-fluorobenzene.
first study to show that the presence of ketamine in cultures makes DCs less able to elicit T-cell Th1-type immune response. Several previous studies have reported that ketamine affects various types of immune cells, including macrophages.4 9 Although these studies mentioned the direct effects of ketamine on particular types of cells, they did not investigate the effect of ketamine on immune responses that arise from the interaction of different immune cells. During functional maturation, DCs show increased expression of MHC and costimulatory molecules (CD40, CD80, CD86, etc), increased IL-12-secreting activity, and reduced endocytosis of antigens.13 In this study, the DC phenotype changes induced by ketamine included reduced expression of MHC class II and costimulatory molecules, less secretion of IL-12, and increased capacity to internalize antigens. These results suggest that ketamine suppresses the functional maturation of DCs. As they mature, DCs become increasingly able to stimulate T cells to proliferate and differentiate into effector T cells. We found that DCs matured by LPS acquired the ability to stimulate the proliferation of allogeneic T cells. In contrast, LPS-stimulated DCs treated with ketamine did not manifest the same ability. This reversal suggests that ketamine affects the ability of DCs to elicit T-cell response. There is accumulating evidence that the cytokine production of DCs depends on DC subsets or on stimuli received by DCs.15 IL-12 has multiple immunoregulatory functions, including the activation of the
798
Th1 T-cell subset, which plays a pivotal role in the induction of inflammation and host-defense response. Many lines of evidence indicate that the development of the Th1-type immune response is regulated by DC-derived IL-12.16 18 Our current finding that ketamine inhibits DC production of IL-12 suggests that ketamine strongly affects the differentiation of T cells. In fact, culturing CD4 T cells with ketamine-treated DCs suppressed the differentiation of CD4 T cells into IFN-producing Th1 T cells. Conversely, culturing CD4 T cells with ketamine-treated DCs did not induce the differentiation of IL-4 producing Th2 T cells. Overall, our data demonstrate that ketamine affects, at least in part, the ability of DCs to tilt the immune response toward an increased Th1 response rather than Th2 response. With practical development in the prospect of inducing Th1 responses, such as ex vivo manipulation of DCs for cell transfer,19 21 our findings indicate a candidate means of counterregulatory DC therapy. Conversely, the effect of ketamine on DCs and the Th1-type immune response suggest immunosuppression by ketamine, which possibly compromises the crucial initial protective Th1type immune response against invading pathogens. Several studies of postoperative patients have reported that ketamine suppresses the immune response.2224 It seems that just a single preoperative administration of ketamine (0.15 mg/kg) is enough to attenuate the production of proinflammatory cytokines from peripheral blood mononuclear cells and the proliferative response of mononuclear cells.23 Meanwhile, the impairment of IL-12 production from monocytes is reported to be a predictor of lethal outcome in postoperative sepsis.25,26 Moreover, a rat model of tumor metastasis has shown that ketamine promotes tumor metastasis.27 We suggest that, in view of the central role of DCs in the induction and regulation of the immune response,13 the suppressive effects of ketamine on various immune responses observed in these several reports can be mediated by the effect of ketamine on DCs.2224,27 Thus, in certain clinical settings, the immunosuppressive action of ketamine on the functions of DCs, including the suppression of IL-12 production may increase risk, so cautious use of ketamine is advised. For anesthesia in which ketamine is used as the main anesthetic reagent for major surgery, it has been reported that the plasma concentration of ketamine is about 3 g/mL.28,29 Furthermore, when administrated in combination with midazolam for sedation in emergency trauma care, the concentration of ketamine has been reported to be 2.6 2.2 g/mL.30 For pharmacological research, the concentration of ketamine used in our in vitro experiment (40 M 10.9 g/mL) was two to three times the reported concentration of ketamine in anesthesia or sedation. For an in vitro experiment, the different concentration of ketamine was reasonable. Two studies found that the administration of a single small dose of ketamine (0.15 0.50
mg/kg) in patients undergoing cardiac surgery with cardiopulmonary bypass (CPB)22 or major abdominal surgery23 reduced postoperative cytokine secretion, such as IL-6 and tumor necrosis factor-. Just one dose of ketamine resulted in a plasma concentration of 400 ng/mL. In in vivo settings, such as postoperative periods after major abdominal surgery or cardiac surgery on CPB in these studies, immunomodulatory activities induced by major surgery or the use of CPB could have a synergistic immunosuppressive effect on immune response together with ketamine. Thus, ketamine may have effects at lower concentrations than in an in vitro study. Other commentators have suggested that ketamine might have nonspecific inhibitory effects.31 One report suggested that the use of ketamine in high concentrations could be nonspecifically cytostatic to various biological activities, and the authors cautioned about the use of high concentrations in in vitro experiments. To clarify the situation, by demonstrating that the expression of CD86 was not altered in the presence of ketamine, we confirmed that ketamine did not generally suppress all DC functions. We further investigated the capacity of DCs to internalize FITC-dextran. Our assay showed that, far from being suppressed, endocytosis had increased. These findings exclude the possibility that the action of ketamine is generally and nonspecifically suppressive. The specific actions are distinctively characteristic of immature DCs.6 Furthermore, through the use of Annexin V and PI staining, we confirmed that ketamine in the range of the concentration used in this experiment was not cytotoxic. Based on our in vitro observations, we conjecture that treatment with ketamine may impair the ability of DCs to initiate a cell-mediated immune response. It has been shown that, in recipient mice transferred with DNBS-DCs, 5 105 DNBS-pulsed DCs (DNBSDC) could induce strong CHS.13,14 We checked the ability of DCs to induce the T-cellmediated immune response by sensitizing recipient mice for CHS to DNBS-DCs. Subcutaneous immunization with 5 105 DNBS-DC made the mice susceptible to CHS, but when similarly exposed to DNBS-DCs pretreated with ketamine they showed no hypersensitivity. This result is evidence that ketamine suppresses the DC-mediated induction of the Th1-type immune response in the whole animal. Furthermore, these results show that the impaired ability of ketamine-pretreated DCs to stimulate the T-cell response cannot be reversed in vivo after withdrawal of ketamine. Mostly based on in vitro experiments, it is questionable how far the findings of the current study are relevant to the in vivo conditions found in whole animals. The BMDC culture system used in most of the experiments reported here has, however, been widely used as an established means for investigating the effects of various agents on DCs.10,11 Furthermore, to assess whether the effect of ketamine found in in vitro experiments can be extrapolated into in vivo
conditions, we showed the suppressive effect of ketamine on the DC-induced immune response in a whole animal by the use of adoptive transfer model of CHS. Although we treated DCs with ketamine for 40 h and assessed the change of phenotype, we also assessed the effect of ketamine on DCs for shorter periods of exposure, including 10 h, which more closely matches ketamine administration in clinical applications, such as anesthesia. The results after 10 h incubation with ketamine show the same tendencies as results after 40 h incubation (Supplementary Figure). The findings obtained by this system are generally assumed to closely resemble what happens in whole animals. In conclusion, we have characterized a variety of effects exerted by ketamine on DCs. Resulting in a significant inhibition of Th1 development, ketamine inhibited phenotypic maturation and modulated cytokine production in these murine DCs. Exposure to ketamine may provide a nontoxic means of modulating of the immunostimulatory capacity of DCs. During sedation and analgesia in critical care, the use of ketamine can suppress the DC-mediated immune response.
Supplemental Figure. Effect of duration of dendritic cell (DC) exposure to ketamine on expression of costimulatory molecules. Ketamine suppresses DC maturation. DCs were treated with ketamine (40 M) for indicated duration (10 hr and 40 hr). A-D, flow cytometry was used to assess the expression rate for CD80 and CD86 in CD11c gated cells. Results are means sds (n 4) DC, dendritic cell. REFERENCES
1. Ueno H, Klechevsky E, Morita R Aspord C, Cao T, Matsui T, Di Pucchio T, Connolly J, Fay JW, Pascual V, Palucka AK, Banchereau J. Dendritic cell subsets in health and disease. Immunol Rev 2007;219:118 42 2. Steinman RM, Banchereau J. Taking dendritic cells into medicine. Nature 2007;449:419 26
799
3. Banchereau J, Steinman RM. Dendritic cells and the control of immunity. Nature 1998;392:24552 4. Li CY, Chou TC, Wong CS, Ho ST, Wu CC, Yen MH, Ding YA. Ketamine inhibits nitric oxide synthase in lipopolysaccharidetreated rat alveolar macrophages. Can J Anaesth 1997;44:989 95 5. Shimaoka M, Iida T, Ohara A, Taenaka N, Mashimo T, Honda T, Yoshiya I. Ketamine inhibits nitric oxide production in mouseactivated macrophage-like cells. Br J Anaesth 1996;77:238 42 6. Nishina K, Akamatsu H, Mikawa K, Shiga M, Maekawa N, Obara H, Niwa Y. The inhibitory effects of thiopental, midazolam, and ketamine on human neutrophil functions. Anesth Analg 1998;86:159 65 7. Fujimoto T, Nishiyama T, Hanaoka K. Inhibitory effects of intravenous anesthetics on mast cell function. Anesth Analg 2005;101:1054 9 8. Kawasaki C, Kawasaki T, Ogata M, Nandate K, Shigematsu A. Ketamine isomers suppress superantigen-induced proinflammatory cytokine production in human whole blood. Can J Anaesth 2001;48:819 23 9. Kawasaki T, Ogata M, Kawasaki C, Ogata J, Inoue Y, Shigematsu A. Ketamine suppresses proinflammatory cytokine production in human whole blood in vitro. Anesth Analg 1999;89:6659 10. Son YI, Egawa S, Tatsumi T, Redlinger RE Jr, Kalinski P, Kanto T. A novel bulk-culture method for generating mature dendritic cells from mouse bone marrow cells. J Immunol Methods 2002;262:14557 11. Lutz MB, Kukutsch N, Ogilvie AL, Rossner S, Koch F, Romani N, Schuler G. An advanced culture method for generating large quantities of highly pure dendritic cells from mouse bone marrow. J Immunol Methods 1999;223:7792 12. Sallusto F, Cella M, Danieli C, Lanzavecchia A. Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: downregulation by cytokines and bacterial products. J Exp Med 1995;182:389 400 13. Krasteva M, Kehren J, Horand F, Akiba H, Choquet G, Ducluzeau MT, Tedone R, Garrigue JL, Kaiserlian D, Nicolas JF. Dual role of dendritic cells in the induction and downregulation of antigen-specific cutaneous inflammation. J Immunol 1998;160:118190 14. Lappin MB, Weiss JM, Delattre V, Mai B, Dittmar H, Maier C, Manke K, Grabbe S, Martin S, Simon JC. Analysis of mouse dendritic cell migration in vivo upon subcutaneous and intravenous injection. Immunology 1999;98:181 8 15. Kadowaki N, Ho S, Antonenko S, Malefyt RW, Kastelein RA, Bazan F, Liu YJ. Subsets of human dendritic cell precursors express different toll-like receptors and respond to different microbial antigens. J Exp Med 2001;194:8639 16. Heufler C, Koch F, Stanzl U, Topar G, Wysocka M, Trinchieri G, Enk A, Steinman RM, Romani N, Schuler G. Interleukin-12 is produced by dendritic cells and mediates T helper 1 development as well as interferon-gamma production by T helper 1 cells. Eur J Immunol 1996;26:659 68 17. Macatonia SE, Hosken NA, Litton M, Vieira P, Hsieh CS, Culpepper JA, Wysocka M, Trinchieri G, Murphy KM, OGarra A. Dendritic cells produce IL-12 and direct the development of Th1 cells from naive CD4 T cells. J Immunol 1995;154:50719
18. Reis e Sousa C, Hieny S, Scharton-Kersten T, Jankovic D, Charest H, Germain RN, Sher A. In vivo microbial stimulation induces rapid CD40 ligand-independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas. J Exp Med 1997;186:1819 29 19. Kanto T, Hayashi N, Takehara T, Tatsumi T, Kuzushita N, Ito A, Sasaki Y, Kasahara A, Hori M. Impaired allostimulatory capacity of peripheral blood dendritic cells recovered from hepatitis C virus-infected individuals. J Immunol 1999;162:5584 91 20. Tatsumi T, Huang J, Gooding WE, Gambotto A, Robbins PD, Vujanovic NL, Alber SM, Watkins SC, Okada H, Storkus WJ. Intratumoral delivery of dendritic cells engineered to secrete both interleukin (IL)-12 and IL-18 effectively treats local and distant disease in association with broadly reactive Tc1-type immunity. Cancer Res 2003;63:6378 86 21. Tatsumi T, Takehara T, Kanto T, Miyagi T, Kuzushita N, Sugimoto Y, Jinushi M, Kasahara A, Sasaki Y, Hori M, Hayashi N. Administration of interleukin-12 enhances the therapeutic efficacy of dendritic cell-based tumor vaccines in mouse hepatocellular carcinoma. Cancer Res 2001;61:75637 22. Bartoc C, Frumento RJ, Jalbout M, Bennett-Guerrero E, Du E, Nishanian E. A randomized, double-blind, placebo-controlled study assessing the anti-inflammatory effects of ketamine in cardiac surgical patients. J Cardiothorac Vasc Anesth 2006;20:21722 23. Beilin B, Rusabrov Y, Shapira Y, Roytblat L, Greemberg L, Yardeni IZ, Bessler H. Low-dose ketamine affects immune responses in humans during the early postoperative period. Br J Anaesth 2007;99:5227 24. Zeyneloglu P, Donmez A, Bilezikci B, Mercan S. Effects of ketamine on serum and tracheobronchial aspirate interleukin-6 levels in infants undergoing cardiac surgery. J Cardiothorac Vasc Anesth 2005;19:329 33 25. Novotny AR, Emmanuel K, Ulm K, Bartels H, Siewert JR, Weighardt H, Holzmann B. Blood interleukin 12 as preoperative predictor of fatal postoperative sepsis after neoadjuvant radiochemotherapy. Br J Surg 2006;93:12839 26. Weighardt H, Heidecke CD, Westerholt A, Emmanuilidis K, Maier S, Veit M, Gerauer K, Matevossian E, Ulm K, Siewert JR, Holzmann B. Impaired monocyte IL-12 production before surgery as a predictive factor for the lethal outcome of postoperative sepsis. Ann Surg 2002;235:560 7 27. Melamed R, Bar-Yosef S, Shakhar G, Shakhar K, Ben-Eliyahu S. Suppression of natural killer cell activity and promotion of tumor metastasis by ketamine, thiopental, and halothane, but not by propofol: mediating mechanisms and prophylactic measures. Anesth Analg 2003;97:13319 28. Dallimore D, Anderson BJ, Short TG, Herd DW. Ketamine anesthesia in children exploring infusion regimens. Paediatr Anaesth 2008;18:708 14 29. Atallah MM, el-Mohayman HA, el-Metwally RE. Ketaminemidazolam total intravenous anaesthesia for prolonged abdominal surgery. Eur J Anaesthesiol 2001;18:29 35 30. Bourgoin A, Albane ` se J, Le one M, Sampol-Manos E, Viviand X, Martin C. Effects of sufentanil or ketamine administered in target-controlled infusion on the cerebral hemodynamics of severely brain-injured patients. Crit Care Med 2005;33:1109 13 31. Lewis E, Rogachev B, Shaked G, Douvdevani A. The in vitro effects of ketamine at large concentrations can be attributed to a nonspecific cytostatic effect. Anesth Analg 2001;92:9279
800
Increasing the Duration of Isourane Anesthesia Decreases the Minimum Alveolar Anesthetic Concentration in 7-Day-Old but Not in 60-Day-Old Rats
Greg Stratmann, MD, PhD Jeffrey W. Sall, MD, PhD Edmond I Eger, II, MD Michael J. Laster, DVM Joseph S. Bell, BA Laura D. V. May, MA Helge Eilers, MD Martin Krause, MD Frank v. d. Heusen, MD Heidi E. Gonzalez
BACKGROUND: While studying neurotoxicity in rats, we observed that the anesthetic minimum alveolar anesthetic concentration (MAC) of isoflurane decreases with increasing duration of anesthesia in 7-day-old but not in 60-day-old rats. After 15 min of anesthesia in 7-day-old rats, MAC was 3.5% compared with 1.3% at 4 h. We investigated whether kinetic or dynamic factors mediated this decrease. METHODS: In 7-day-old rats, we measured inspired and cerebral partial pressures of isoflurane at MAC as a function of duration of anesthesia. In 60-day-old rats, we measured inspired partial pressures of isoflurane at MAC as a function of duration of anesthesia. Finally, we determined the effect of administering 1 mg/kg naloxone and of delaying the initiation of the MAC determination (pinching the tail) on MAC in 7-day-old rats. RESULTS: In 7-day-old rats, both inspired and cerebral measures of MAC decreased from 1 to 4 h. The inspired MAC decreased 56%, whereas the cerebral MAC decreased 33%. At 4 h, the inspired MAC approximated the cerebral MAC (i.e., the partial pressures did not differ appreciably). Neither administration of 1 mg/kg naloxone nor delaying tail clamping until 3 h reversed the decrease in MAC. In 60-day-old rats, inspired MAC of isoflurane was stable from 1 to 4 h of anesthesia. CONCLUSIONS: MAC of isoflurane decreases over 1 4 h of anesthesia in 7-day-old but not in 60-day-old rats. Both pharmacodynamic and a pharmacokinetic components contribute to the decrease in MAC in 7-day-old rats. Neither endorphins nor sensory desensitization mediate the pharmacodynamic component.
(Anesth Analg 2009;109:8016)
nesthesia in 7-day-old rats causes neurodegeneration and delayed persistent hippocampal dysfunction.13 Whether such anesthetic neurotoxicity occurs in developing human brains remains unknown.4,5 A test of this possibility requires the development of rodent models, including a measure of clinically relevant anesthetic doses, such as anesthetic minimum alveolar anesthetic concentration (MAC) (the minimum alveolar concentration of an inhaled anesthetic required to eliminate movement in response to a supramaximal noxious stimulus in 50% of subjects). In studies of developmental anesthetic toxicity, we routinely deliver a clinically relevant anesthetic dose
From the Department of Anesthesia and Perioperative Care, University of California San Francisco, San Francisco, California. Accepted for publication April 7, 2009. Supported by a grant from the Anesthesia Patient Safety Foundation (to GS) and by NIH grant 1P01GM47818 (to EIE). Dr. Eger is a paid consultant to Baxter Healthcare Corp. who donated the isoflurane used in these studies. Address correspondence and reprint requests to Greg Stratmann, MD, PhD, Department of Anesthesia and Perioperative Care, University of California San Francisco, Box 0464, Room U286, 513 Parnassus Ave., San Francisco, CA 94143. Address e-mail to stratman@anesthesia.ucsf.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aff364
of 1 MAC by anesthetizing 10 or more rats simultaneously, tail clamping the animals every 15 min, and adjusting the inspired anesthetic concentration according to how many animals moved (see Methods). In doing so, we noticed that the isoflurane concentration required to produce 50% immobility in 7-day-old rats (hereafter referred to as inspired MAC) decreases dramatically and progressively over 4 h. Previous investigations report that duration of anesthesia does not influence MAC in 2.5-mo-old rats.6 The goals of this study were fourfold. First, we sought to determine whether the progressive decrease in inspired MAC is a real phenomenon as opposed to a consequence of the method used to anesthetize animals or to determine MAC. Here, we show that inspired MAC of isoflurane in 7-day-old but not in 60-day-old rats decreases progressively over 4 h. The second goal was to describe a model of a clinically meaningful anesthetic in 7-day-old rats. In an Anesthetic and Life Support Drugs Advisory Committee Meeting on March 29, 2007, the Food and Drug Administration attempted to interpret the relevance of preclinical findings of anesthetic toxicity in the developing brain and recognized that one of the major obstacles is the absence of a clinically relevant dosing schedule of anesthetics in animals.7 Here, we describe
801
how to deliver an inhaled anesthetic to 7-day-old rats at a clinically relevant dose. Third, we aimed to determine whether pharmacokinetic or pharmacodynamic factors govern the decrease in MAC of isoflurane in 7-day-old rats. The following observations form the rationale for this goal. Equilibration of alveolar/inspired anesthetic concentration of volatile anesthetics remains incomplete after 60 min in ventilated adult rats8 and after 3 h in spontaneously breathing tracheostomized adult rats.9 A still greater difference between alveolar and inspired isoflurane concentration might be found for spontaneously breathing 7-day-old rats because these are more susceptible to the respiratory-depressant effects of isoflurane.10 These observations suggest that the decrease in inspired MAC with increasing duration of anesthesia may be solely a pharmacokinetic phenomenon, i.e., because of incomplete equilibration of inspired and brain partial pressures of isoflurane. By determining both inspired and cerebral partial pressures at various times during the anesthetic, we show that both pharmacokinetic and pharmacodynamic factors govern the decrease in inspired MAC. The fourth goal was to test some obvious possible mechanisms mediating the decrease in MAC, namely endogenous opioids, sensory sensitization, or desensitization. Anesthesia in neonatal animals causes arterial hypotension,11 hypotension increases endogenous opioids,12 and opioids decrease MAC.13 We hypothesized that naloxone given at 3 h would reverse the decrease in MAC. Repeated supramaximal stimuli may alter the response to the same stimulus by peripheral sensitization14 or desensitization of mechanoreceptors. One concern of investigators using tail clamping for MAC determination has been a possible sensitization because of tissue injury and inflammation after repeated supramaximal stimulation.15 On the other hand, mechanoreceptors express adaptation resulting in a decreasing response to sustained stimulation.14 Furthermore, temporal summation as described by Dutton et al.16 may lead to an increase in MAC because of the stimulation pattern. All three of these possible mechanisms need to be considered in MAC testing with repeated tail clamping because they may alter the sensitivity of the animal over time independent of pharmacologic effects. If desensitization because of repeated stimulation contributes to the decrease in MAC, animals not subjected to tail clamp until late during the anesthetic should have a higher MAC on initial tail clamping than animals subjected to tail clamp throughout the course of the anesthetic. Likewise, animals undergoing experiments with delayed clamping would be expected to have a decreased MAC if sensitization or temporal summation significantly contributed to the findings. We show that neither naloxone nor delayed tail clamp changes the decrease in MAC in 7-day-old rats.
802
Anesthetic Duration Changes MAC in Rats
Table 1. Protocol for MAC Determination and Adjustment of Isourane Concentration in Response to Tail Clamping Every 15 min for 4 h Percent of rats moving in response to tail clamping
0 10 20 30 40 50 60 70 80 90 100
Subsequent percent adjustment of inspired isourane concentration

1.0 0.8 0.6 0.4 0.2 No change No change 0.1 0.2 0.3 0.5
MAC anesthetic minimum alveolar anesthetic concentration.
METHODS
With University of California, San Francisco IRB approval, 21 separate isoflurane anesthetics were conducted involving between 10 and 20 rats per anesthetic. For Aims 1 and 2 (definition of MAC as a function of anesthetic duration in 7-day-old rats (n 141) and comparison with MAC in P60 rats [n 40]), 18 anesthetics were delivered. Two anesthetics of 7-day-old rats (n 40) were used to compare inspired with brain MAC (aim 3). Of 20 rats that were simultaneously anesthetized, four animals were used for determination of brain MAC at one of the two timepoints. The remaining animals were used to determine inspired MAC. One experiment using 7-day-old rats (n 20) was conducted to test if the decrease in inspired and brain MAC observed in experiments pertaining to Aims 13 could be due to either endogenous opioids or altered peripheral mechanoreceptor sensitivity.
Rat Anesthesia
The anesthetizing chamber was a preheated, humidified glove box to which we delivered isoflurane in 6 L/min 50% oxygen/nitrogen. The chamber was part of a semiclosed anesthetic circuit incorporating a fan that recirculated gases via a canister containing soda lime and a humidifier. Gas composition within the anesthetic chamber was measured using a calibrated Datex Capnomac Ultima (Datex Instrumentarium Corp., Helsinki, Finland). A supramaximal pain stimulus was generated by application of an alligator clamp to each rats tail for 30 s or until the rat moved. Movement was defined as any movement except breathing. Tail clamping was repeated every 15 min, starting 15 min after induction of general anesthesia. Before the first tail clamping at 15 min, the inspired isoflurane concentration was set to 3.5%. Once all animals had their tails clamped, the anesthetic concentration was adjusted, if needed, according to the algorithm in Table 1. The new anesthetic concentration was recorded as the inspired MAC value for that
timepoint. For example, if 40% of rats moved with application of the tail clamp at a measured concentration of 2.5% isoflurane, MAC for that timepoint would be 2.3% (2.5% 0.2%). In addition, the isoflurane concentration for the following 15-min period would be decreased to 2.3%. Custom-made temperature probes were inserted subcutaneously over the skull to facilitate control of temperature at 36.5C 1C using computer-controlled Peltier heater/cooler plates integrated into the floor of the anesthetizing chamber.
Measurement of Inhaled and Brain Partial Pressures of Isourane

Inhaled and brain partial pressures associated with MAC were determined in 7-day-old rats but only inspired partial pressures were determined in 60-dayold rats. One and 4 h after induction of general anesthesia, anesthetic gas from the chamber was aspirated into a glass syringe and analyzed for the isoflurane concentration using gas chromatography. At each timepoint, four 7-day-old rats were decapitated and each brain quickly removed inside the anesthetic chamber and transferred to a preweighed gas-tight 20-mL syringe containing a preweighed amount of glass beads and capped with a three-way stopcock. All but 10 mL of the air in the syringe was ejected, and the syringe weighed once more to determine the weight of the brain placed in the syringe. The volume of the brain was then estimated assuming a brain density of 1.1 g/mL. The stopcock was connected to an empty 10-mL glass syringe, and the plunger of the gas-tight syringe was pushed forward to near empty, forcing the brains into the glass beads, which thereby macerated the brain. The forward movement of the plunger forced the air into the attached glass syringe. After producing maceration of the brain, the gas in the glass syringe was drawn back into the gas-tight syringe, thereby retaining any anesthetic that had transferred to the gas phase. After maceration and warming for 1 h to 37C, the plunger in the gas-tight syringe was drawn to the 20 mL mark and reequilibrated for an additional 15 min. The concentration of anesthetic in the gas phase (Cg) was measured in duplicate by gas chromatography and the two values were averaged. The concentration of anesthetic originally in the brain (Cb) was estimated from knowledge of the gas volume (Vg), the brain volume (Vb), and the separately determined brain/gas partition coefficient of the anesthetic (l). That is, the total amount of isoflurane (At) equaled the amount of agent in the gas phase (Ag Vg Cg) plus the amount retained in the brain (Ab Vb l Cg). Thus, Cb (AgAb)/Vb. The partial pressure that this represented as a percent of 1 atm thus equals 100 Cb/l.
took the opportunity to determine it concurrently for isoflurane, sevoflurane, and desflurane. Groups of four brains from 16 7-day-old rats were placed in 20 mL preweighed gas-tight syringes (capped with a three-way stopcock) containing a preweighed amount of glass beads (i.e., of known volume). The plunger of the syringe was advanced to ensure maceration of the brains. Gas containing all three volatile anesthetics then was introduced to approximately the 18 mL mark, and the syringe equilibrated for an hour in a rotameter. The plunger then was drawn to the 20 mL mark, the three-way stopcock briefly opened to allow entry of room air, and the syringe then reequilibrated at 37C for 15 min. The concentration of volatile anesthetics was determined (C1) by gas chromatography and all of the gas phase expelled through the three-way stopcock. Room air was drawn into the syringe to the 18 mL mark and the above process repeated, giving a second volatile anesthetic concentration (C2). The process again was repeated, giving C3. An exponential curve was fit to these concentrations, allowing an estimate of the decay with each dilution. l was calculated as:
l C2 /C1 C2 /Vg/ V b
The C2 and C1 values were the estimated values from the fitted curve rather than the actual determined values.
Naloxone/Delayed Tail-Clamping Experiment

In a separate experiment, application of the tail clamp began immediately in 10 (clamped) rats randomly selected from 20 simultaneously anesthetized animals and at 3 h in the other 10 (not clamped). Of those animals that were alive at 3 h (n 15), a near equal number of animals from each group was chosen randomly to receive naloxone (1 mg/kg) IP (n 4, clamped; n 3, not clamped). The others received an equal volume of normal saline. During the fourth hour of anesthesia, all rats were tail clamped and the response of the entire cohort was used to adjust the anesthetic concentration as per the algorithm in Table 1. After unblinding the investigators to group assignment, the recorded responses to tail clamping were used to determined MAC retrospectively using the algorithm in Table 1 for each of the four treatment conditions: clamped/naloxone, clamped/no naloxone, not clamped/naloxone, and not clamped/no naloxone. Consequently, MAC values of the two groups that were not clamped until 3 h of anesthesia (not clamped/naloxone and not clamped/no naloxone) are available only for the last hour of anesthesia.
Statistical Methods
Data are expressed as medians and interquartile ranges except for the brain/gas partition coefficient measurements, each of which are given as the mean and its standard deviation.
Brain/Gas Partition Coefcient Determination

Because the brain/gas partition coefficient in 7-dayold rats is not known for any volatile anesthetic, we
803
Table 2. Brain/Gas Partition Coefcients of Isourane, Sevourane, and Desurane in 7-Day-Old Rats Isourane
1.69 0.13
Data are means
SD.
Sevourane
1.10 0.10
Desurane
0.85 0.08
Figure 1. The inspired isoflurane concentration required to

produce 1 anesthetic minimum alveolar anesthetic concentration (MAC) of isoflurane decreases progressively over 4 h of anesthesia in 7-day-old rats. The MAC of 60-day-old rats is stable after 1 h. Rats were anesthetized in groups of 10 20 (14 isoflurane anesthetics in 7-day-old rats and four anesthetics in 60-day-old rats). MAC, which was determined every 15 min, was estimated as the administered isoflurane concentration adjusted by the algorithm in Table 1. Immediately after tail clamping, the administered concentration was adjusted to the isoflurane concentration given by the algorithm in Table 1. Data are medians and interquartile ranges. The data labels are median inspired isoflurane concentrations resulting in 1 MAC (inspired MAC) and are shown for 7-day-old rats only.
and 4 h of anesthesia in 7-day-old rats (n 14). The n 2 for inspired isoflurane partial pressures denotes that all rats were part of two groups of 20 simultaneously anesthetized rats each. The isoflurane partial pressure in the anesthetic chamber was determined at 1 and 4 h once (in duplicate) for each group of 20 rats. The brain partial pressures of isoflurane were determined in brains harvested at 1 h (n 6) and at 4 h (n 8). Data are medians and interquartile ranges. ***P 0.001 MannWhitney U-test.
Figure 2. Brain/inspired partial pressures of isoflurane at 1
Table 3. Inspired and Brain Partial Pressures at 1 MAC at 1 and 4 h of Isourane Anesthesia in 7-Day-Old Rats Time Inspired MAC Brain MAC Ratio brain/inspired
0.72 1.03
1h 2.75% atm 1.9% atm 4h 1.21% atm 1.27% atm Percent decrease 56 33
MAC anesthetic minimum alveolar anesthetic concentration.
Brain partial pressures of isoflurane at 1 and 4 h were compared using the MannWhitney U-test. Differences in MAC values among the four treatment conditions involving naloxone and delayed tail clamping were compared using a two-way analysis of variance. The lowest order polynomial equation resulting in an acceptable curve fit, a second order polynomial equation, was applied to the data of 7-day-old animals in the naloxone/delayed tail clamping experiment. A P value of 0.05 was considered statistically significant. Prism 4.0 for MacIntosh (GraphPad Software, San Diego, CA) was used for all analyses.
RESULTS
Brain/gas partition coefficients of isoflurane, sevoflurane, and desflurane are shown in Table 2. In 7-day-old rats, the inspired MAC of isoflurane decreased from a median of 3.5% at 15 min to 1.3% at 4 h of anesthesia (Fig. 1). In 60-day-old rats, MAC of isoflurane decreased from a median of 2.36% at 15 min to 1.65% at 60 min and was stable thereafter, reaching 1.5% at 4 h (Fig. 1). The ratio of brain/inspired partial pressures of isoflurane was determined in 7-day-old animals at 1 h (n 6) and in separate animals at 4 h (n 8) (Fig. 2).
804
At 1 h after induction of general anesthesia when the inspired partial pressure of isoflurane at MAC was 2.75%, the median brain partial pressure was 1.9% atm (interquartile range 1.7%2.3% atm) and a brain/ inspired partial pressure ratio was 0.72. This indicates that 1 h after induction of general anesthesia the inspired and brain partial pressures had not yet equilibrated. After 4 h of anesthesia, the median brain partial pressure of isoflurane at MAC was 1.27% atm (interquartile range 1.08%1.4% atm) and the inspired partial pressure of isoflurane at MAC was 1.21% atm indicating that the inspired and brain partial pressures had equilibrated (Table 3). The brain partial pressure at MAC at 4 h was 33% less than at 1 h after induction of anesthesia (P 0.001 MannWhitney U-test). Thus, from 1 to 4 h, the inspired MAC of isoflurane in 7-day-old rats decreased by 56%, whereas the brain MAC decreased by 33% (Table 3). This means that there is both a pharmacokinetic (incomplete equilibration of inspired and brain partial pressures of isoflurane) and a pharmacodynamic component (decrease in brain partial pressures from 1 to 4 h) of the decrease in MAC. The decrease in inspired MAC for isoflurane was not reversed by either naloxone administration at
Figure 3. Inspired MAC of isoflurane under four treatment

conditions: clamped naloxone, clamped no naloxone, not clamped naloxone, and not clamped no naloxone (see Methods). At the beginning of anesthesia, 10 rats were randomly assigned to be tail clamped every 15 min for the duration of anesthesia and 10 not to be tail clamped until 3 h. The curves are a second-order polynomial fit to the data from the clamped groups only. The number of surviving animals over time is given as clamped, alive and not clamped, alive underneath the graph. At 3 h, 7 of the 15 survivors (four clamped and three not clamped) were randomly assigned to receive naloxone and the others (four clamped and four not clamped) to receive normal saline. The n in the legend refers to the group size at the time of group assignment, which is 0 h for the clamped/not clamped condition and 3 h for the naloxone/no naloxone condition. Curve fits for the not clamped conditions were omitted for clarity because the 3-h timepoint in this context is crucial. If altered peripheral mechanoreceptor sensitivity had been responsible for the decrease in MAC, initial tail clamping at 3 h should have caused a greater proportion of movement in the previously not clamped rats, which was not the case. Neither naloxone injection at 3 h nor delaying tail clamping until 3 h was able to reverse the decrease in MAC in 7-day-old rats.
3 h or by delaying tail clamping until 3 h of anesthesia (Fig. 3).
DISCUSSION
We found that the inspired MAC of isoflurane decreases progressively during the first 4 h of anesthesia in 7-day-old, but not in 60-day-old adult, rats. We found an initial (15 min) inspired isoflurane MAC of 3.5%, a value that exceeds the MAC of 2.34% previously reported in 9-day-old rats.17 However, the 2.34% value was probably obtained after 12 h of anesthesia and thus is close to our 1-h value of 2.75%. Part of the decrease of MAC with increasing duration of anesthesia in 7-day-old rats results from delayed equilibration of brain and inspired partial pressures of isoflurane. However, equilibration is complete (Table 2) by 4 h of isoflurane anesthesia. Because the brain partial pressure of isoflurane that produces immobility in 50% of 7-day-old rats decreases by one third between 1 and 4 h of anesthesia, while the inspired partial pressure decreases by 56%,
there is both a pharmacokinetic and a pharmacodynamic component to the decrease in MAC. Both of these components are attenuated or absent in 60-dayold rats (i.e., MAC is the same, or nearly so, at 1 vs 4 h of anesthesia). We do not know what might underlie the decrease in the brain MAC in the immature rat, a decrease that has vanished in the mature rat. Neither endogenous opioid release nor altered peripheral mechanoreceptor sensitivity appear to contribute to this decrease in MAC. The 33%56% decreases we found in inspired, but particularly cerebral, MAC between 1 and 4 h have implications for the results of experiments of developmental neurotoxicity.1,18 20 Studies of developmental neurotoxicity should be informed by clinically meaningful anesthetic concentrations; investigators must recognize that the inhaled isoflurane concentration required to produce anesthesia (1 MAC) is initially more than twice as high as in adult rats21 and about 150% more than the published MAC for infantile rats.17 Furthermore, they should recognize that MAC for isoflurane in 7-day-old rats decreases progressively over 4 h to 60% of the published MAC for isoflurane in immature rats.17 Thus, such studies require a diminishing concentration schedule to avoid underdosing and overdosing. Based on our experience gained over the course of these studies, we recommend priming the anesthetic chamber with at least 4% isoflurane and adjusting the isoflurane concentration according to the rats responses to tail clamping as outlined in Table 1. If MAC determination is not possible or not desired, the inspired isoflurane concentrations shown in Figure 1 should provide a reasonable starting point for modeling a clinically meaningful anesthetic in 7-day-old rats. Whether these findings also apply to other animal species, other anesthetics, or to humans is unknown. In conclusion, the inspired and cerebral concentrations of isoflurane resulting in 1 MAC of anesthesia decreases progressively over 4 h as a function of both pharmacokinetic and pharmacodynamic phenomena. The pharmacodynamic component is not mediated by endorphins or by altered peripheral mechanoreceptor sensitivity. REFERENCES
1. Jevtovic-Todorovic V, Hartman RE, Izumi Y, Benshoff ND, Dikranian K, Zorumski CF, Olney JW, Wozniak DF. Early exposure to common anesthetic agents causes widespread neurodegeneration in the developing rat brain and persistent learning deficits. J Neurosci 2003;23:876 82 2. Fredriksson A, Archer T, Alm H, Gordh T, Eriksson P. Neurofunctional deficits and potentiated apoptosis by neonatal NMDA antagonist administration. Behav Brain Res 2004; 153:36776 3. Fredriksson A, Ponten E, Gordh T, Eriksson P. Neonatal exposure to a combination of N-methyl-d-aspartate and gammaaminobutyric acid type A receptor anesthetic agents potentiates apoptotic neurodegeneration and persistent behavioral deficits. Anesthesiology 2007;107:42736 4. Olney JW, Young C, Wozniak DF, Ikonomidou C, JevtovicTodorovic V. Anesthesia-induced developmental neuroapoptosis. Does it happen in humans? Anesthesiology 2004;101:2735
805
5. Anand KJ, Soriano SG. Anesthetic agents and the immature brain: are these toxic or therapeutic? Anesthesiology 2004; 101:52730 6. Eger EI II, Johnson BH. MAC of I-653 in rats, including a test of the effect of body temperature and anesthetic duration. Anesth Analg 1987;66:974 6 7. Groupe Miller C, Shafer S. Summary of Minutes of the Anesthetic and Life Support Drugs Advisory Committee Meeting on March 29, 2007. Rockville, Maryland, 2007. Available at: http:// www.fda.gov/ohrms/dockets/ac/07/minutes/2007 4285m1Final.pdf; accessed February 20, 2009 8. Wahrenbrock EA, Eger EI II, Laravuso RB, Maruschak G. Anesthetic uptake of mice and men (and whales). Anesthesiology 1974;40:19 23 9. White PF, Johnston RR, Eger EI II. Determination of anesthetic requirement in rats. Anesthesiology 1974;40:527 10. Stratmann G, Sall JW, May LD, Bell JS, Magnusson KR, Rau V, Visrodia KH, Alvi RS, Ku B, Lee MT, Dai R. Isoflurane differentially affects neurogenesis and long-term neurocognitive function in 60-day old and 7-day old rats. Anesthesiology 2009;110:834 48 11. LeDez KM, Lerman J. The minimum alveolar concentration (MAC) of isoflurane in preterm neonates. Anesthesiology 1987;67:3017 12. Molina PE. Endogenous opioid analgesia in hemorrhagic shock. J Trauma 2003;54:S126 32 13. Kurita T, Takata K, Uraoka M, Morita K, Sanjo Y, Katoh T, Sato S. The influence of hemorrhagic shock on the minimum alveolar anesthetic concentration of isoflurane in a swine model. Anesth Analg 2007;105:1639 43, table of contents
14. Reeh PW, Bayer J, Kocher L, Handwerker HO. Sensitization of nociceptive cutaneous nerve fibers from the rats tail by noxious mechanical stimulation. Exp Brain Res 1987;65:50512 15. Sobair AT, Cottrell DF, Camburn MA. A mechanical stimulator for the determination of the minimum alveolar concentration (MAC) of halothane in the rabbit. Vet Res Commun 1993; 17:375 85 16. Dutton RC, Zhang Y, Stabernack CR, Laster MJ, Sonner JM, Eger EI II. Temporal summation governs part of the minimum alveolar concentration of isoflurane anesthesia. Anesthesiology 2003;98:13727 17. Orliaguet G, Vivien B, Langeron O, Bouhemad B, Coriat P, Riou B. Minimum alveolar concentration of volatile anesthetics in rats during postnatal maturation. Anesthesiology 2001;95:734 9 18. Yon JH, Daniel-Johnson J, Carter LB, Jevtovic-Todorovic V. Anesthesia induces neuronal cell death in the developing rat brain via the intrinsic and extrinsic apoptotic pathways. Neuroscience 2005;135:81527 19. Johnson SA, Young C, Olney JW. Isoflurane-induced neuroapoptosis in the developing brain of nonhypoglycemic mice. J Neurosurg Anesthesiol 2008;20:21 8 20. Loepke AW, McCann JC, Kurth CD, McAuliffe JJ. The physiologic effects of isoflurane anesthesia in neonatal mice. Anesth Analg 2006;102:75 80 21. Holdcroft A, Bose D, Sapsed-Byrne SM, Ma D, Lockwood GG. Arterial to inspired partial pressure ratio of halothane, isoflurane, sevoflurane and desflurane in rats. Br J Anaesth 1999; 83:618 21
806
Technology, Computing, and Simulation

Section Editor: Dwayne Westenskow
Auditory Event-Related Potentials, Bispectral Index, and Entropy for the Discrimination of Different Levels of Sedation in Intensive Care Unit Patients
Matthias Haenggi, MD* Heidi Ypparila-Wolters, PhD Sarah Buerki, cand. med.* Rebekka Schlauri, cand. med.* Ilkka Korhonen, PhD Jukka Takala, MD, PhD* Stephan M. Jakob, MD, PhD*
BACKGROUND: Sedation protocols, including the use of sedation scales and regular sedation stops, help to reduce the length of mechanical ventilation and intensive care unit stay. Because clinical assessment of depth of sedation is labor-intensive, performed only intermittently, and interferes with sedation and sleep, processed electrophysiological signals from the brain have gained interest as surrogates. We hypothesized that auditory event-related potentials (ERPs), Bispectral Index (BIS), and Entropy can discriminate among clinically relevant sedation levels. METHODS: We studied 10 patients after elective thoracic or abdominal surgery with general anesthesia. Electroencephalogram, BIS, state entropy (SE), response entropy (RE), and ERPs were recorded immediately after surgery in the intensive care unit at Richmond Agitation-Sedation Scale (RASS) scores of 5 (very deep sedation), 4 (deep sedation), 3 to 1 (moderate sedation), and 0 (awake) during decreasing target-controlled sedation with propofol and remifentanil. Reference measurements for baseline levels were performed before or several days after the operation. RESULTS: At baseline, RASS 5, RASS 4, RASS 3 to 1, and RASS 0, BIS was 94 [4] (median, IQR), 47 [15], 68 [9], 75 [10], and 88 [6]; SE was 87 [3], 46 [10], 60 [22], 74 [21], and 87 [5]; and RE was 97 [4], 48 [9], 71 [25], 81 [18], and 96 [3], respectively (all P 0.05, Friedman Test). Both BIS and Entropy had high variabilities. When ERP N100 amplitudes were considered alone, ERPs did not differ significantly among sedation levels. Nevertheless, discriminant ERP analysis including two parameters of principal component analysis revealed a prediction probability PK value of 0.89 for differentiating deep sedation, moderate sedation, and awake state. The corresponding PK for RE, SE, and BIS was 0.88, 0.89, and 0.85, respectively. CONCLUSIONS: Neither ERPs nor BIS or Entropy can replace clinical sedation assessment with standard scoring systems. Discrimination among very deep, deep to moderate, and no sedation after general anesthesia can be provided by ERPs and processed electroencephalograms, with similar PKs. The high inter- and intraindividual variability of Entropy and BIS precludes defining a target range of values to predict the sedation level in critically ill patients using these parameters. The variability of ERPs is unknown.
(Anesth Analg 2009;109:80716)
ost critically ill patients receive sedative and analgesic drugs to attenuate discomfort and pain.1 The excessive use of sedatives and analgesics prolongs mechanical ventilation, and increases the incidence of nosocomial pneumonia, time spent in the intensive
From the *Department of Intensive Care Medicine, Bern University Hospital and University of Bern, Bern, Switzerland; and VTT Technical Research Centre of Finland, Tampere, Finland. Accepted for publication March 9, 2009. Supported by Datex-Ohmeda, now GE Healthcare, Helsinki, Finland. Ilkka Korhonens employer received funding from GE Healthcare to carry out research projects related to depth of anesthesia monitoring. He has been working in these projects, and part of the work reported here has resulted from these projects. The relationship does not, however, have any effect on the results reported in this study. Jukka Takala and Stephan M. Jakob, The Department of Intensive Care Medicine, Bern University Hospital, have received research grants from GE Healthcare (formerly Datex-Ohmeda), Helsinki, Finland. The authors affiliated with the department received no personal funds from GE. Vol. 109, No. 3, September 2009
care unit (ICU), and costs.2 4 Strategies to reduce the use of sedatives and analgesics may improve the outcome.2,5 Whereas under-sedation is generally easy to identify, over-sedation, with its associated problems, is more difficult to recognize but should be avoided. Although daily pauses in sedation help to avoid gross over-sedation, this is not always possible, for example, when the patients condition is unstable.5 Also, accumulation of sedatives and analgesics may
Jukka Takala is editor of Critical Care and Trauma for the Journal. This manuscript was handled by Jeffrey M. Feldman, Section Editor of Technology, Computing, and Simulation. Dr. Takala was not involved in any way with the editorial process or decision. Reprints will not be available from the author. Address correspondence to Stephan M. Jakob, MD, PhD, Department of Intensive Care Medicine, Inselspital, Freiburgstrasse, CH3010 Bern, Switzerland Address e-mail to stephan.jakob@insel.ch. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181acc85d
807
occur rapidly, especially in patients with renal and/or liver dysfunction. Monitoring the depth of sedation is difficult and is currently based on clinical assessment and the use of clinical scoring systems.4,6,7 These scoring systems cannot be applied continuously, they are subjective, and they measure arousal in response to stimulation rather than the level of consciousness in the unstimulated state, i.e., evaluating the level of sedation influences what is being assessed.6 Several methods based on the electroencephalogram (EEG) have been tested to avoid these problems, but the results have been disappointing.8,9 Entropy, a nonlinear statistic parameter which describes the order of random repetitive signals, is a relatively new method of processed EEG. In patients, Entropy translates the anesthesia-induced, more synchronized EEG into a single parameter.10 Spectral entropy can reproducibly indicate the hypnotic effects of propofol, thiopental, and different anesthetic gases.1114 Another method frequently used in the operating room to monitor depth of anesthesia is midlatency auditoryevoked potentials (MLAEPs), e.g., the Danmeter AEP Monitor/2.15 There are virtually no published studies using MLAEP in adult ICU patients16 18 and few studies are done during conscious sedation.19,20 Studies suggest that variability of the derived values impedes the use of MLAEPs to guide sedation during the light levels of sedation. The most popular method of processed EEG for assessment of sedation is the Bispectral Index (BIS).21 Although BIS has been tested and validated for use in the operating room to assess anesthetized patients,21,22 data on using BIS in the ICU to assess sedation are controversial.7,2326 The multiple concomitant medications and heterogeneity of underlying pathologies present a further challenge to the use of neuromonitoring in the ICU in general.27 We have previously shown that the time-locked cortical response to standard external auditory stimuli 100 ms (long-latency auditory-evoked potentials or event-related potentials [ERPs]) can discriminate between clinically relevant light to moderate and deep sedation levels in healthy volunteers when sedation is induced with a combination of propofol or midazolam with remifentanil.28,29 Based on our previous experience, the N100 response (the negative amplitude appearing about 100 ms after an acoustic stimulus), which represents conscious detection of changes in the acoustic environment,30,31 is the acoustic ERP best qualified to evaluate sedation, especially if presented in a sequence of four tones to see the orienting reaction coming and going, followed by a pause of 12 s to clear the sensory memory. We therefore hypothesized that these ERPs may be used to monitor the depth of sedation in ICU patients as well. As the first step to test this hypothesis, we evaluated the use of ERPs to assess the level of sedation in patients undergoing major elective surgery and admitted to the ICU for short-term postoperative mechanical ventilation.
808
Discriminating Sedation Level in ICU Patients
METHODS
This study was approved by the Ethics Committee of the Canton of Bern, and written informed consent was obtained from each patient. In a pilot series, 10 patients were recorded, but the signal quality was poor as a result of difficulties in fixing external earphones in the awakening and moving patients. For the present series of patients, a new ear canal earphone system was used. These 10 patients required an elective thoracic surgical intervention under general anesthesia and were assigned to the ICU for postoperative care. The study took place in the ICU of the Bern University Hospital. Exclusion criteria were ASA III or higher, adverse events during former surgery or anesthesia, neurological impairment in the medical history, and hearing abnormalities. The reference or baseline measurements were recorded depending on the patients availability during preoperative evaluation on the day before surgery (two patients) or several days postoperatively (eight patients) when the patients were on the regular ward and were no longer receiving narcotic or sedative drugs . The evening before surgery, the patients received their premedication according to the anesthesiologists prescription. The anesthetic regime was left to the discretion of the attending anesthesiologist. All but one patient received an epidural catheter. Postoperatively, the tracheally intubated patients were transferred to the ICU, where standard monitoring (electrocardiogram, invasive arterial blood pressure, pulse oximetry, respiratory rate, end-tidal carbon dioxide concentration) with an S/5 Critical Care Patient Monitor (Datex-Ohmeda, GE Healthcare, Helsinki, Finland) was continued. The parameters (including the processed and raw EEGs) were recorded online with S/5 Collect software (GE Healthcare) (an upgraded WinCollect version) and saved on a laptop for further analysis. Standard treatment protocols were used for hemodynamic and ventilatory management. In the ICU, four sets of measurements were taken from each patient: the first set was taken after arrival in the ICU, at Richmond Agitation-Sedation Scale (RASS)32 level 5, followed by measurements at RASS 4, then at a target of RASS 2, and after tracheal extubation at RASS 0 (Table 1). Assessments of RASS score were performed by two observers just before the recording period, but at least 15 min after the last drug adjustment to obtain a steady state, and just at the end of the sedation period. If the assessments of the observers differed from each other, consensus was sought (only needed in the lighter sedation levels). This is in accordance with the validation of the RASS provided by the original publication.7 To adjust the different sedation levels according to RASS, propofol (Disoprivan 1%, AstraZeneca, Zug, Switzerland) and remifentanil (Ultiva, GlaxoSmithKline, Scho nbu hl, Switzerland) were infused by a computer-controlled pump (Alaris Asena, ALARIS Medical Systems, Carbamed, Bern, Switzerland), driven by the Rugloop II
Table 1. Richmond Agitation-Sedation Scale (RASS) (see Ref. 32) Score

4 3 2 1 0 1 2 3 4 5
Term
Combative Very agitated Agitated Restless Alert and calm Drowsy Light sedation Moderate sedation Deep sedation Unarousable
Description
Overtly combative or violent; immediate danger to staff Pulls on or removes tube(s) or catheter(s) or has aggressive behavior toward staff Frequent nonpurposeful movement or patient-ventilator dyssynchrony Anxious or apprehensive but movements not aggressive or vigorous Not fully alert, but has sustained (more than 10 s) awakening, with eye contact to voice Briefly (less than 10 s) awakens with eye contact to voice Any movement (but no eye contact) to voice No response to voice, but any movement to physical stimulation No response to voice or physical stimulation
Procedure
1. Observe patient. Is patient alert and calm (score 0)? Does patient have behavior that is consistent with restlessness or agitation (score 1 to 4 using the criteria listed above, under description). 2. If patient is not alert, in a loud speaking voice state patients name and direct patient to open eyes and look at speaker. Repeat once if necessary. Can prompt patient to continue looking at speaker. Patient has eye opening and eye contact, which is sustained for more than 10 s (score 1). Patient has eye opening and eye contact, but this is not sustained for 10 s (score 2). Patient has any movement in response to voice, excluding eye contact (score 3). 3. If patient does not respond to voice, physically stimulate patient by shaking shoulder and then rubbing sternum if there is no response to shaking shoulder. Patient has any movement to physical stimulation (score 4). Patient has no response to voice or physical stimulation (score 5).
TCI program (BVBA Demed, Temse, Belgium), with the pharmacokinetic variable set of Schnider et al.33 for propofol and that of Minto et al.34 for remifentanil. Remifentanil target was initially set at 2 ng/mL, propofol was started at about 2 g/mL concentration, depending on the amount of anesthetics and analgesics used during the operation, and both drugs were adjusted as needed to reach the target sedation level without leaving the patient with signs indicating insufficient analgesia. A stabilization period of 15 min was maintained after each change in the doses of propofol or remifentanil. At the beginning of each step, background EEG, BIS, and Entropy were recorded for 5 min. Afterward, auditory stimulation was started and ERPs were recorded. The stimulation was applied according to a habituation paradigm. In the habituation paradigm, four equal auditory stimuli (800 Hz) were applied through earphones at intervals of 1 s, followed by a pause, and this set of stimuli was repeated after 12 s. Altogether, 40 sets of stimuli were delivered at each measurement, corresponding to a recording time of about 10 min. The tones were generated by the Presentation stimulus delivery software (Presentation Version 9.51, built 05.24.05, Neurobehavioral Systems, Albany, CA) and delivered through foam-protected in-ear earphones (ER-4S MicroPro, Ethymotic Research, Elk Grove Village, IL).
(Fz) and central (Cz), were used. Both electrodes were referred to the right mastoid, and the electrodeskin impedances were kept below 5 k[Omega]. Automatic impedance checks were performed every 15 min. The EEG signal was amplified and digitalized continuously at 200 Hz (modified EEG module, S/5 monitor, Datex-Ohmeda, GE Healthcare). This EEG module can measure a maximum of four channels, has a range of 400 V, a frequency range of 0.5 40 Hz, a resolution of 60 V, an input impedance of 10 MW, a noise level of 0.5 V root mean square (rms) from 0.5 to 40 Hz, and a common mode rejection ratio of 100 dB. BIS electrodes (Aspect Medical, S/5 BIS Module M-BIS [XP-level], Datex-Ohmeda, GE Healthcare) and Entropy electrodes (S/5 M-Entropy, Datex-Ohmeda, GE Healthcare) were placed on each side of the patients forehead according to the manufacturers recommendations. The smoothing rate of M-BIS was set to 30 s.
EEG and ERP Processing

The background EEG was first band-pass filtered with a finite impulse response type filter using cutoff frequencies of 0.5 and 32 Hz (Matlab, version 6.12, The Mathworks, Natick, MA). Five minutes of the filtered EEG signal was cut into 5-s epochs with 50% overlap. Obvious artifacts were excluded based on the maximum amplitude for each epoch. Epochs with amplitudes more than the predefined limits (100 300 V) were excluded. The appropriateness of the artifact rejection was manually confirmed. For each EEG epoch, the root mean squared total power was calculated. The power spectral density was
EEG, BIS, and Entropy Recording

The EEG signal was recorded using Ag/AgCl electrodes placed on the scalp, according to the international 10-20 system. Two electrode locations, frontal
809
estimated for each epoch using Welsh averaged periodogram method, and spectral edge frequency 95% (SEF 95%) and mean power frequency were computed from the power spectral density using a frequency range from 0.5 to 32 Hz. In addition, the following spectral powers were computed: Delta power (0.5 4 Hz), Theta power (4 8 Hz), Alpha power (8 13 Hz), Beta1 power (1320 Hz), Beta2 power (20 32 Hz), and the slow-fast ratio ([Delta Theta]/[Alpha Beta1 Beta2]). The mean parameter values of the accepted epochs were individually computed, and used in the statistical analysis. The EEG signal recorded during the auditory stimulation was first filtered using cutoff frequencies of 1 and 20 Hz and then transformed to epochs lasting from 100 to 900 ms relative to each stimulus onset. After removal of epochs with artifacts (rejection level 100 V), the individual responses to the first and second stimuli of the habituation paradigm were averaged. The responses to the third and the fourth stimuli were averaged to increase the number of stimuli for averaging, which in turn improves the signal-to-noise ratio. The N100 component was defined as a maximum negative deflection appearing 80 150 ms after the stimulus onset. The amplitude and the latency of the prominent N100 components with respect to the prestimulus baseline (50 ms to onset of the tone) were manually recorded. The first to second tone ratio of the habituation paradigm was computed by dividing the N100 amplitude of the first stimulus by the N100 amplitude of the second stimulus. An advanced signal processing technique was necessary to increase the accuracy of discrimination among the sedation levels. We therefore chose principal component (PC) analysis the averaged responses of the first, second, and the combined third and fourth stimulus of the habituation paradigm. In this analysis, each measurement can be presented as a linear sum of so-called basis vectors, which are multiplied by PCs. These can be obtained by solving the eigendecomposition problem.35 In general, the more the averaged responses have common features, the less basis vectors and PC are needed. The individual PC may then be used as features of the N100 component in the statistical analysis.
15 and 30 s, depending on the frequency, the downloaded values were cut into 30 s epochs for further processing.
The descriptive data of the N100 responses are shown as mean sd. As the Entropy and BIS are on an ordinal scale, these data are shown as median and interquartile ranges (IQR). Friedman test with respect to the sedation levels was applied. Pairwise comparisons within the time row were corrected with the StudentKeulsNewman method. A P value of 0.05 was considered significant. To find a combination of variables which was able to identify the sedation level most accurately, we applied discriminant analysis. To avoid overestimation, a maximum of four variables was tested. Discriminant analysis was first applied to the PCs of the N100 characteristics, then in a stepwise manner with addition of the variables obtained from the EEG spectrum, BIS, and Entropy values. The statistical differences among the measurement conditions (final step of the discriminant analysis) were tested using one-way analysis of variance. The analysis steps have been described in detail by Ypparila et al.36 The performance of this classification was estimated with prediction probability, calculated from Somer d.37 A value of PK 0.5 means that the parameter predicts the steps with no better than a 50:50 chance. A value of PK 1.0 means that the parameter predicts the steps correctly 100% of the time. Basic statistics were computed with the SigmaStat for Windows Version 3.1 software package (Systat Software, Point Richmond, CA). Discriminant analysis and PK were calculated with SPSS (SPSS for Windows 12.0, SPSS, Chicago, IL).
RESULTS
The 10 patients had a median age of 64 yr (range, 47 85 yr). All patients completed the full set of measurements. The data quality was very good. For the ERP recordings, 94.9% 5.3% of all individual responses of the habituation paradigm could be used, with a minimum of 72.2% in one patient during the target RASS 2 level. From the processed EEG recordings, BIS data were not transferred after selfimpedance check in two different patients, one during RASS 4 and the other at RASS 3 to 1, resulting in acceptance rates of 86% 32% and of 87% 32% of all epochs during these two sedation states. In all other sedation levels, acceptance rates were above 96% 10% for BIS and above 98% 1% for RE and SE. Stepwise maintenance of the predefined lighter sedation levels postoperatively was more difficult than anticipated: RASS 2 could be achieved in only four patients, and in two of these four patients the RASS 2 changed during the measurement period. Therefore, the lighter sedation levels (RASS 3 to 1)
BIS and Entropy Processing

Frontal muscle electromyography (EMG) can influence processed EEG because the frequency bands are overlapping. EMG is considered an artifact by the BIS device, whereas the Entropy device considers EMG (as response entropy [RE], as opposed to state entropy [SE]). Both devices have regular internal impedance checks, and both have built-in artifact recognition, which considers movements and high EMG. For data processing, we used only signals of both processed EEG devices, which were cleared by the devices built-in algorithm. As the smoothing rate of BIS is 30 s and the smoothing rate of Entropy is between
810
Figure 1. Individual event-related potential (ERP) responses of all 10 patients, recorded at central location of the electrode.
Figure 2. Amplitudes of the N100 peak (100 ms after the first auditory stimulus during the habituation paradigm) of each of the 10 patients at the different sedation levels.
were combined, and the levels RASS 5 (unarousable), RASS 4 (deep sedation), RASS 3 to 1 (light to moderate sedation), and RASS 0 (awake/after extubation) were analyzed. Drug effect-site concentrations to achieve the desired sedation levels are as follows (mean sd): remifentanil RASS 5 1.98 0.24 ng/mL, RASS 4 1.20 0.29, RASS 3 to 1 1.1 0.28; propofol 2.32 0.34 g/mL, 1.87 0.46 and 1.36 0.38. Figure 1 demonstrates the individual ERPs of all patients at the different sedation levels. The amplitude of the first N100 response (H1) at the electrode location Fz decreased from 10.6 3.3 V at baseline to
3.7 1.5 V at RASS 5 (P 0.05) and then increased to 5.3 1.6 V at RASS 0 (P 0.05, compared with RASS 5). Individual responses showed a wide variation within the measures (Fig. 2). Neither latency nor habituation was useful to discriminate between the postoperative RASS levels (data not shown). There were no differences between values recorded at electrode locations Fz and Cz (data not shown). BIS was 94 (IQR 4) at baseline, 47 (IQR 15) at level RASS 5 (P 0.05 compared with baseline), and 88 (IQR 6) after tracheal extubation (P 0.05 compared with RASS 5, Fig. 3). The corresponding values for
811
Figure 3. Box plots (median, end of box:

25th and 75th percentiles, error bars: 10th and 90th percentiles, outliers) of the processed electroencephalogram (EEG) variables Bispectral Index (BIS), response entropy (RE), and state entropy (SE) at different sedation levels. Because RE and SE are not independent measurements of Entropy, we did not perform comparisons between processed EEG variables. Friedman test was applied on sedation levels within each processed EEG variable, and pairwise comparisons within the time row were corrected with the StudentKeuls Newman method. The asterisk indicates the differences of processed EEG variables; there is no statistically significant difference between baseline and Richmond AgitationSedation Scale (RASS) 0 in any processed EEG variable.
Figure 4. Box plots of the variation of the processed electroencephalogram (EEG) parameters in the individual patients during the 10-min recordings (absolute values of Bispectral Index [BIS], response entropy [RE], and state entropy [SE]).
RE were 97 (IQR 4), 48 (IQR 9) (P 0.05 compared with baseline), and 96 (IQR 3), and for SE were 87 (IQR 5), 46 (IQR 10) (P 0.05 compared with baseline), and 87 (IQR 3), respectively (Fig. 3). Between RASS 4 and RASS 3 to 1 levels, the values frequently overlapped. Within the 10-min ERP recording time of each
812
individual patient, the processed EEG variables fluctuated frequently, with the highest variation up to an IQR of 15 (Fig. 4). Of the quantitative EEG variables (root mean square, mean power frequency, SEF 95%, power in the frequency bands), only SEF 95%, Beta 1/Beta2 power,
Table 2. Results of the Friedman Test of the Principal Components Analysis (PCA), the Processed Electroencephalogram (EEG), and the Quantitative EEG Values Variable
PC1 PC2 PC3 PC1 PC2 PC3 PC1 PC2 PC3 PC1 PC2 PC3 PC1 PC2 PC3 Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak 1 1 1 2 2 2 3 3 3 4 4 4 34 34 34
P
0.022 0.000 0.078 0.000 0.000 0.161 0.000 0.000 0.300 0.000 0.000 0.374 0.000 0.000 0.033
Variable
N100 amplitude, Peak 1, Fz N100 latency, Peak 1, Fz N100 amplitude, Peak 2, Fz N100 latency, Peak 2, Fz N100 amplitude, Peak 1, Cz N100 latency, Peak 1, Cz N100 amplitude, Peak 2, Cz N100 latency, Peak 2, Cz Response entropy State entropy BIS
P
0.000 0.000 0.000 0.000 0.000 0.000 0.003 0.000 0.000 0.000 0.000
Variable
Spectral edge frequency 95% (SEF 95%) Root mean squared (RMS) total power Total power Mean power frequency (MPF) Delta power (0.54 Hz) Theta power (48 Hz) Alpha power (813 Hz) Beta1 power (1320 Hz) Beta2 power (2032 Hz) Slow-fast ratio (Delta Theta/Alpha Beta1 Beta2)
P
0.000 0.851 0.837 0.030 0.031 0.794 0.551 0.004 0.000 0.005
Fz frontal location of the electrode; Cz central location of the electrode; BIS Bispectral Index.
Table 3. PK of the Variables Prediction probability Variable

Response entropy (RE) State entropy (SE) BIS SEF 95% H1 amplitude Principal components (PC2-Peak 1; PC1-Peak 3/4) Principal components response entropy
PCs (PC2-Peak 1; PC1-Peak 3/4) alone, and 0.93 0.06 for PCs of ERPs and RE (Table 3).
PK
0.88 0.89 0.85 0.77 0.72 0.89 0.93
Standard error
0.08 0.08 0.05 0.06 0.06 0.08 0.06
DISCUSSION
The main finding of this study was that the PC analysis of N100 responses was able to differentiate among clinically relevant sedation levels in patients. In contrast to our previous results in healthy volunteers, the absolute amplitudes of the ERPs were not able to differentiate among the sedation levels. We have no clear explanation for this difference, but it is reasonable to assume that the mixture of anesthetic and analgesic drugs, and the duration and depth of anesthesia likely contributed. The PK of PC analysis of N100 combined with quantitative EEG parameters, recorded with the same electrode, was as high as the PK of the BIS. Combining PC analysis with either BIS or Entropy did not considerably change the accuracy of sedation assessment. Although the processed EEG variables BIS, RE, and SE could discriminate among different levels of sedation, their high inter- and intraindividual variability within the clinical steady-state of sedation precludes the definition of target ranges for different sedation levels. Concerns that the ERP recordings per se might influence variability of the processed EEG are unwarranted and not supported by the literature.38 A direct comparison of the variability of ERP monitoring and processed EEG monitoring of sedation is not yet possible. The ERPs were recorded over a 10-min period, whereas BIS has a smoothing time of 30 s and RE/SE have 15/60-s smoothing times, making them prone to fluctuations. Corresponding to published data, BIS was slightly more accurate than the Entropy or EEG measures, as reflected by the higher PK.11 During the 10-min ERP recording period, the volunteers were in a stable state of sedation as estimated by the observers. Despite this, the individual EEG patterns fluctuated substantially,
Sedation levels: awake, moderate (RASS 1 to 3), and deep (RASS 4 and 5). RASS Richmond Agitation-Sedation Scale; SEF 95% spectral edge frequency 95%; BIS Bispectral Index.
and slow-fast ratio could discriminate among different sedation levels in the overall test (Table 2), but pairwise comparisons failed to differentiate further (data not shown). To improve the discriminative power of the measured parameters post hoc, we further combined clinically relevant sedation levels (deep sedation RASS 5 and 4; moderate sedation RASS 3 to 1, and awake; RASS 0 postextubation and during reference recording). The stepwise discriminant analysis of the significant variables of the PC analysis (PC2 of Peak 1, PC1 and 2 of Peak 2 and PC1 and 2 of the combined 3rd and 4th peak) resulted in 82% correctly classified cases with only the PC2 of Peak 2 and PC1 of the combined 3rd and 4th peak. Addition of any quantitative variable did not improve the classification. The same is true for the H1 and H2 amplitudes and latencies. Only the addition of RE or SE increased the size of the correct classification to 86%. Therefore the prediction probabilities PK for discrimination of the three sedation levels awake, moderate, and deep sedation were PK 0.89 (standard error 0.08) for
813
as indicated by the BIS, RE, and SE values at each sedation level (Fig. 4). This occurred despite the use of computer-controlled drug administration with validated pharmacokinetic/pharmacodynamic models. This fluctuation was more pronounced at the deeper sedation levels. A possible explanation is the decay of anesthesia, which could not be controlled for in this study setting. The high interindividual variability (Fig. 3), with substantial overlap among the different sedation levels, has been described for both BIS and Entropy.7,24 26,39 As we calculated ERPs off-line and measured only a 10-min epoch, variability of ERPs is unknown. The analysis may be modified to be applicable in real-time for on-line analysis, so the variability could be assessed in the future. In contrast to the results of previous studies in volunteers,28,29,40 measurement of only the amplitudes of the N100 component of the auditory-evoked potentials could not discriminate among the different sedation levels in this patient population. The presence of N100 responses at deep sedation levels has been described, but these responses were small41 or hardly detectable.42 This was interpreted as ongoing detection and processing of auditory information. Support of continuing cortical processing in the associative auditory cortex with reduced intensity comes from functional magnetic resonance imaging studies, in which the blood oxygen level dependency technique could find activation in areas where N100 responses were generated.43 In the discriminant analysis, the PCs of the N100 response to the first tone were not considered, and PCs of the N100 responses to the second and combined third and fourth tones were chosen. This is rather unexpected, because generally it is believed that habituation to the repetitive stimulus occurs and the amplitude of the ERP will diminish. Also unexpectedly, the low N100 amplitudes persisted at more superficial sedation levels and even after tracheal extubation. As the N100 response reflects conscious detection of any change in the auditory environment, our patients should have had easily detectable responses in the awake states, and they are to be expected in the range of the baseline recordings,30,31 independent of whether the baseline recordings were done before or days after surgery. In our prior study in volunteers, three repeated baseline measurements of N100 without any medication, separated by at least 1 wk, did not differ significantly, with baseline amplitudes being in the same range as in this study.28 Baseline recordings were done in a fully cooperative patient without any medication, whereas the RASS 0 recordings were done in a noisy ICU with a patient just recovered from anesthesia and surgery. We believe that these two states can explain the measurement differences. We cannot exclude that movements of the patients during awakening may have changed the position of the sealed in-ear headphones and modified the loudness of the stimulus. These movements caused technical
814
problems in our pilot series before we changed to the sealed headphones. This could be controlled in the future by monitoring signal delivery, e.g., bursts of midlatency or brainstem acoustic-evoked potentials or stapedius reflex potentials. In this study, only patients with obvious hearing difficulties (acoustic devices and tinnitus) were excluded. As the median age of the patients was 64 yr, the hearing threshold could have been reduced in many of the patients. An individual adaptation of the loudness, e.g., 30 dB above the hearing threshold, might have improved discrimination of sedation levels. A further reason for the discordant results compared with our previous findings is the use of other anesthetics and opiates before neurophysiological testing. The effects of combined opiates and sedative drugs on ERPs and consciousness have not been well established. High-dose opiates result in a general slowing of EEG, with a concomitant decrease in SEF 95%.44 The effect of opioids on BIS is variable: opiates alone may or may not decrease BIS.45 In studies of opioids used in combination with a hypnotic drug, BIS was decreased46 or the dose of the hypnotic could be reduced to achieve the same BIS value.47 Midlatency acoustic-evoked potentials behave the same way as the BIS, with a tendency to mirror the effect-site concentrations of the hypnotic drug rather than the sedation effect.48,49 Studies exploring opioid effects on ERPs did not show any changes, but in these studies, the doses of morphine and buprenorphine were small, and the study population consisted of chronic pain and drug-dependent patients.50,51 Our previous study28 demonstrates an opioid-independent effect of sedation on the N100 response. Therefore, it is possible that the opiates used for anesthesia blunted the N100 responses of our patients, but the fact that all patients could be tracheally extubated according to standard criteria and recovered without hypoventilation argues against substantial opiate effect-site concentrations. Another concern might arise from the fact that sedation was titrated against a sedation scale (RASS) and not against a processed EEG variable. Sedation in ICU patients is usually controlled by clinical assessment, and processed EEG parameters are not endorsed by guidelines,4 nor can they be recommended for guidance of sedation in ICU patients because of their variability.8,23 In conclusion, at present, sedation in ICU patients can only be assessed with repeated standard sedation scoring. This cannot be replaced by either ERPs or processed EEG parameters. We show that PK of the ERPs is not inferior compared with either PK of BIS or Entropy monitors, despite their persistently low amplitudes during light sedation levels and after tracheal extubation. This is conceivable because the clinical level of sedation is defined as a response to a stimulus, analogously with ERPs, whereas the processed EEG
rather reflects background activity. With on-line monitoring of ERPs and standardization of perceived loudness of the tones, the performance of ERPs might improve. REFERENCES
1. Novaes MA, Knobel E, Bork AM, Pavao OF, Nogueira-Martins LA, Ferraz MB. Stressors in ICU: perception of the patient, relatives and health care team. Intensive Care Med 1999; 25:1421 6 2. Kollef MH, Levy NT, Ahrens TS, Schaiff R, Prentice D, Sherman G. The use of continuous i.v. sedation is associated with prolongation of mechanical ventilation. Chest 1998;114:541 8 3. Rello J, Diaz E, Roque M, Valles J. Risk factors for developing pneumonia within 48 hours of intubation. Am J Respir Crit Care Med 1999;159:1742 6 4. Jacobi J, Fraser GL, Coursin DB, Riker RR, Fontaine D, Wittbrodt ET, Chalfin DB, Masica MF, Bjerke HS, Coplin WM, Crippen DW, Fuchs BD, Kelleher RM, Marik PE, Nasraway SA Jr, Murray MJ, Peruzzi WT, Lumb PD. Clinical practice guidelines for the sustained use of sedatives and analgesics in the critically ill adult. Crit Care Med 2002;30:119 41 5. Kress JP, Pohlman AS, OConnor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med 2000;342:14717 6. Kress JP, Pohlman AS, Hall JB. Sedation and analgesia in the intensive care unit. Am J Respir Crit Care Med 2002;166:1024 8 7. Ely EW, Truman B, Shintani A, Thomason JWW, Wheeler AP, Gordon S, Francis J, Speroff T, Gautam S, Margolin R, Sessler CN, Dittus RS, Bernard GR. Monitoring sedation status over time in ICU patients: reliability and validity of the Richmond Agitation-Sedation Scale (RASS). JAMA 2003;289:298391 8. Sessler C, Jo Grap M, Ramsay M. Evaluating and monitoring analgesia and sedation in the intensive care unit. Crit Care 2008;12:S2 9. Kalkman CJ, Drummond JC. Monitors of depth of anesthesia, quo vadis? Anesthesiology 2002;96:784 7 10. Viertio-Oja H, Maja V, Sarkela M, Talja P, Tenkanen N, Tolvanen-Laakso H, Paloheimo M, Vakkuri A, Yli-Hankala A, Merilainen P. Description of the Entropy algorithm as applied in the Datex-Ohmeda S/5 Entropy Module. Acta Anaesthesiol Scand 2004;48:154 61 11. Schmidt GN, Bischoff P, Standl T, Hellstern A, Teuber O, Schulte EJ. Comparative evaluation of the Datex-Ohmeda S/5 Entropy Module and the Bispectral Index monitor during propofolremifentanil anesthesia. Anesthesiology 2004;101:128390 12. Ellerkmann RK, Soehle M, Alves TM, Liermann VM, Wenningmann I, Roepcke H, Kreuer S, Hoeft A, Bruhn J. Spectral entropy and bispectral index as measures of the electroencephalographic effects of propofol. Anesth Analg 2006;102:1456 62 13. Vakkuri A, Yli-Hankala A, Talja P, Mustola S, Tolvanen-Laakso H, Sampson T, Viertio-Oja H. Time-frequency balanced spectral entropy as a measure of anesthetic drug effect in central nervous system during sevoflurane, propofol, and thiopental anesthesia. Acta Anaesthesiol Scand 2004;48:14553 14. Maksimow A, Snapir A, Sarkela M, Kentala E, Koskenvuo J, Posti J, Jaaskelainen SK, Hinkka-Yli-Salomaki S, Scheinin M, Scheinin H. Assessing the depth of dexmedetomidine-induced sedation with electroencephalogram (EEG)-based spectral entropy. Acta Anaesthesiol Scand 2007;51:2230 15. Plourde G. Auditory evoked potentials. Best Pract Res Clin Anaesthesiol 2006;20:129 39 16. Doi M, Morita K, Mantzaridis H, Sato S, Kenny GN. Prediction of responses to various stimuli during sedation: a comparison of three EEG variables. Intensive Care Med 2005;31:417 17. Rundshagen I, Schnabel K, Pothmann W, Schleich B, Schulte am Esch J. Cortical arousal in critically ill patients: an evoked response study. Intensive Care Med 2000;26:1312 8 18. Schulte-Tamburen AM, Scheier J, Briegel J, Schwender D, Peter K. Comparison of five sedation scoring systems by means of auditory evoked potentials. Intensive Care Med 1999;25:377 82 19. Oei-Lim VLB, Dijkgraaf MGW, de Smet MD, White M, Kalkman CJ. Does cerebral monitoring improve ophthalmic surgical operating conditions during propofol-induced sedation? Anesth Analg 2006;103:1189 95 Vol. 109, No. 3, September 2009
20. Huang YY, Chu YC, Chang KY, Wang YC, Chan KH, Tsou MY. Performance of AEP Monitor/2-derived composite index as an indicator for depth of sedation with midazolam and alfentanil during gastrointestinal endoscopy. Eur J Anaesthesiol 2007;24:2527 21. Johansen JW, Sebel PS. Development and clinical application of electroencephalographic bispectrum monitoring. Anesthesiology 2000;93:1336 44 22. Rampil IJ. A primer for EEG signal processing in anesthesia. Anesthesiology 1998;89:980 1002 23. LeBlanc JM, Dasta JF, Kane-Gill SL. Role of the bispectral index in sedation monitoring in the ICU. Ann Pharmacother 2006;40:490 500 24. Simmons LE, Riker RR, Prato BS, Fraser GL. Assessing sedation during intensive care unit mechanical ventilation with the Bispectral Index and the Sedation-Agitation Scale. Crit Care Med 1999;27:1499 504 25. Riker RR, Fraser GL, Simmons LE, Wilkins ML. Validating the Sedation-Agitation Scale with the Bispectral Index and Visual Analog Scale in adult ICU patients after cardiac surgery. Intensive Care Med 2001;27:853 8 26. De Deyne C, Struys M, Decruyenaere J, Creupelandt J, Hoste E, Colardyn F. Use of continuous bispectral EEG monitoring to assess depth of sedation in ICU patients. Intensive Care Med 1998;24:1294 8 27. Riker RR, Fraser GL. Sedation in the intensive care unit: refining the models and defining the questions. Crit Care Med 2002;30:16613 28. Haenggi M, Ypparila H, Takala J, Korhonen I, Luginbuhl M, Petersen-Felix S, Jakob SM. Measuring depth of sedation with auditory evoked potentials during controlled infusion of propofol and remifentanil in healthy volunteers. Anesth Analg 2004;99:1728 36, table of contents 29. Haenggi M, Ypparila H, Hauser K, Caviezel C, Korhonen I, Takala J, Jakob SM. The effects of dexmedetomidine/remifentanil and midazolam/remifentanil on auditory-evoked potentials and electroencephalogram at light-to-moderate sedation levels in healthy subjects. Anesth Analg 2006;103:11639 30. Hyde M. The N1 response and its applications. Audiol Neurootol 1997;2:281307 31. Naatanen R, Picton T. The N1 wave of the human electric and magnetic response to sound: a review and an analysis of the component structure. Psychophysiology 1987;24:375 425 32. Sessler CN, Gosnell MS, Grap MJ, Brophy GM, ONeal PV, Keane KA, Tesoro EP, Elswick RK. The Richmond AgitationSedation Scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med 2002;166:1338 44 33. Schnider TW, Minto CF, Gambus PL, Andresen C, Goodale DB, Shafer SL, Youngs EJ. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998;88:1170 82 34. Minto CF, Schnider TW, Shafer SL. Pharmacokinetics and pharmacodynamics of remifentanil. II. Model application. Anesthesiology 1997;86:24 33 35. Jolliffe IT. Principal component analysis. 2nd ed. Springer Series in Statistics. New York: Springer Verlag, 2002. 36. Ypparila H, Korhonen I, Tarvainen M, Musialowicz T, Jakob SM, Partanen J. N100 auditory potential and electroencephalogram discriminate propofol-induced sedation levels. J Clin Monit Comput 2004;18:16370 37. Smith WD, Dutton RC, Smith NT. Measuring the performance of anesthetic depth indicators. Anesthesiology 1996;84:38 51 38. Absalom AR, Sutcliffe N, Kenny GNC. Effects of the auditory stimuli of an auditory evoked potential system on levels of consciousness, and on the bispectral index. Br J Anaesth 2001;87:778 80 39. Walsh TS, Ramsay P, Lapinlampi TP, Sarkela MO, Viertio-Oja HE, Merilainen PT. An assessment of the validity of spectral entropy as a measure of sedation state in mechanically ventilated critically ill patients. Intensive Care Med 2008;34:308 15 40. Simpson TP, Manara AR, Kane NM, Barton RL, Rowlands CA, Butler SR. Effect of propofol anaesthesia on the event-related potential mismatch negativity and the auditory-evoked potential N1. Br J Anaesth 2002;89:382 8 41. Ypparila H, Karhu J, Westeren-Punnonen S, Musialowicz T, Partanen J. Evidence of auditory processing during postoperative propofol sedation. Clin Neurophysiol 2002;113: 1357 64
815
42. van Hooff JC, de Beer NA, Brunia CH, Cluitmans PJ, Korsten HH. Event-related potential measures of information processing during general anesthesia. Electroencephalogr Clin Neurophysiol 1997;103:268 81 43. Plourde G, Belin P, Chartrand D, Fiset P, Backman SB, Xie G, Zatorre RJ. Cortical processing of complex auditory stimuli during alterations of consciousness with the general anesthetic propofol. Anesthesiology 2006;104:448 57 44. Scott JC, Ponganis KV, Stanski DR. EEG quantitation of narcotic effect: the comparative pharmacodynamics of fentanyl and alfentanil. Anesthesiology 1985;62:234 41 45. Barr G, Anderson RE, Owall A, Jakobsson JG. Effects on the bispectral index during medium-high dose fentanyl induction with or without propofol supplement. Acta Anaesthesiol Scand 2000;44:80711 46. Ferreira DA, Nunes CS, Antunes LM, Santos IA, Lobo F, Casal M, Ferreira L, Amorim P. The effect of a remifentanil bolus on the bispectral index of the EEG (BIS) in anaesthetized patients independently from intubation and surgical stimuli. Eur J Anaesthesiol 2006;23:30510
47. Bouillon TW, Bruhn J, Radulescu L, Andresen C, Shafer TJ, Cohane C, Shafer SL. Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy. Anesthesiology 2004;100:135372 48. Schraag S, Flaschar J, Schleyer M, Georgieff M, Kenny GN. The contribution of remifentanil to middle latency auditory evoked potentials during induction of propofol anesthesia. Anesth Analg 2006;103:9027 49. Manyam SC, Gupta DK, Johnson KB, White JL, Pace NL, Westenskow DR, Egan TD. When is a Bispectral Index of 60 too low?: Rational processed electroencephalographic targets are dependent on the sedative-opioid ratio. Anesthesiology 2007;106:472 83 50. Lorenz J, Beck H, Bromm B. Differential changes of laser evoked potentials, late auditory evoked potentials and P300 under morphine in chronic pain patients. Electroencephalogr Clin Neurophysiol 1997;104:514 21 51. Kouri EM, Lukas SE, Mendelson JH. P300 assessment of opiate and cocaine users: effects of detoxification and buprenorphine treatment. Biol Psychiatry 1996;40:61728
816
Case Report
Evoked Potential Monitoring Identies Possible Neurological Injury During Positioning for Craniotomy
Zirka H. Anastasian, MD* Brian Ramnath, EEG/EP T, CNIM Ricardo J. Komotar, MD Jeffrey N. Bruce, MD Michael B. Sisti, MD Edward J. Gallo, EEG T, CNIM Ronald G. Emerson, MD Eric J. Heyer, MD, PhD*
Somatosensory-evoked potential (SSEP) monitoring is commonly used to detect changes in nerve conduction and prevent impending nerve injury. We present a case series of two patients who had SSEP monitoring for their surgical craniotomy procedure, and who, upon positioning supine with their head tilted 30 45, developed unilateral upper extremity SSEP changes. These SSEP changes were reversed when the patients were repositioned. These cases indicate the clinical usefulness of monitoring SSEPs while positioning the patient and adjusting position accordingly to prevent injury.
(Anesth Analg 2009;109:817821)
eurologic injury secondary to positioning is a significant perioperative problem and a common cause of patient injury in the practice of anesthesiology. Somatosensory-evoked potential (SSEP) monitoring is reproducible, reliable, and commonly used during surgical procedures to detect changes in electrophysiological conduction in peripheral nerves and central nerve pathways and, thus, to prevent nervous system damage.1 A significant change in the SSEP responses is indicated by a decrease in amplitude and/or an increase in latency. Changes in SSEP responses may be due to spinal instrumentation, hypoperfusion, hypothermia, anesthetic drugs, and positioning. SSEP monitoring has also been noted to be useful in evaluating upper extremity conduction changes related to positioning. Changes in nerve conduction are expected to occur before permanent nerve injury, thus, repositioning that reverses SSEP conduction changes should prevent perioperative peripheral nerve injury.2 4 The following case series is the first, to our knowledge, to describe two cases of asymmetric SSEP changes in the brachial plexus after positioning for craniotomy
From the Departments of *Anesthesiology, Neurology, and Neurological Surgery, New York-Presbyterian Hospital, Columbia University, New York City, New York. Accepted for publication May 18, 2009. EJH was supported in part by a grant from the NIA (RO1 AG17604). Address correspondence and reprint requests to Zirka H. Anastasian, MD, Department of Anesthesiology, Columbia University, 622 West 168th St., New York City, NY 10032. Address e-mail to zh2114@columbia.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b086bd
that resolved after repositioning. It underlines the clinical usefulness of monitoring SSEPs while positioning the patient and adjusting the position accordingly to prevent peripheral neurological injury, a common cause of patient injury and malpractice claims.
CASE DESCRIPTIONS
Case 1
A 73-yr-old woman with a history of progressive loss of coordination and left-sided hearing presented for a suboccipital craniectomy for acoustic neuroma. Her medical history was significant for hypertension and diabetes. The magnetic resonance imaging showed an acoustic neuroma with enlarged ventricles and chronic displacement of the cerebellum and brainstem. After induction of general anesthesia, the patient was positioned supine. Her head was turned approximately 45 toward the right and fixed with a Mayfield neurosurgical head holder. A sandbag was placed under the mattress on the left side of the operating table to extend the neck and left shoulder. Electrophysiologic monitoring was performed bilaterally for SSEPs in upper (median nerve) and lower (posterior tibial nerve) extremities, brainstem auditory-evoked potentials, and motor-evoked potentials. SSEPs from the median nerves were recorded using needle electrodes placed at Erbs point over the brachial plexus for assessing peripheral conduction and on the scalp at CP3 and CP4 locations, for assessing central conduction. SSEP needle electrodes were placed on the patient before positioning. Positioning required 8 min, after which, SSEPs were recorded from both upper and lower extremities. Standard filter setting for these SSEPs is 30 1500 Hz, and standard stimulation parameters were used including rate of 4.7/s, 20 mA, and 0.3 ms duration. At the start of recording, after a 1-min interval of acquiring a steady signal, the SSEP amplitude from stimulation of the left median nerve was 817
Figure 1. Somatosensory-evoked potential (SEEP) changes in the left median nerve in Case 1. SSEPs from bilateral median nerves were recorded using needle electrodes placed at Erbs point over the brachial plexus, for assessing peripheral conduction, and on the scalp at CP3 and CP4 locations, for assessing central conduction. Time in minutes is recorded on the y axis. From the start of the recordings (time 1 min), there is a unilateral decrease in the amplitude of the left median nerve recordings compared with that of the right median nerve. This difference exists at both peripheral (Erbs) and central (CP3 and CP4) recordings, suggesting a peripheral mechanism. With repositioning of the shoulder, the amplitude increases on the left side with time in central and peripheral recordings, approaching the amplitude on the right side.
noted to be reduced at Erbs point on all recording channels in comparison with contralateral recordings. When the patients left shoulder was raised to reduce extension of the neck and shoulder, the SSEPs from stimulation of the left median nerve were improved at recordings from Erbs point and subcortical and cortical electrodes and became equal in amplitude to the contralateral side in 4 min (Fig. 1). Surgical exposure was not compromised significantly with this adjustment. surgery was performed under microscope guidance. SSEPs from the median nerves were recorded using needle electrodes placed at Erbs point over the brachial plexus for assessing peripheral conduction and on the scalp at CP3 and CP4 locations, for assessing central conduction. SSEP needle electrodes were placed on the patient before positioning. Positioning required 10 min, after which, SSEPs were recorded from both upper and lower extremities. Standard filter settings for these SSEPs were the same as the previous case. At the start of recording, after a 1-min interval of acquiring a steady signal, the SSEP amplitude from stimulation of the left median nerve was reduced at Erbs point on all recording channels in comparison with contralateral recordings. When the patients left shoulder was raised to reduce extension of the neck and shoulder, the SSEPs from stimulation of the left median nerve were improved at recordings from Erbs point and subcortical and cortical electrodes and became equal in amplitude to the contralateral side in 6 min (Fig. 2). Surgical exposure was not compromised significantly in this position. Both patients gave witnessed verbal informed consent to publish any collected data.
Case 2
A 62-yr-old man with a history of transient diffuse extremity weakness presented for left temporal and suboccipital craniotomy for subtotal resection of a skull base meningioma. His medical history was significant for hypertension, gout, and diverticulitis. A magnetic resonance imaging showed the presence of a left-sided mass consistent with a meningioma, at the level of the clivus and petrous bone, with compression of the pons and ventricular dilation. After inducing general anesthesia, the patient was positioned supine. His head was turned toward the right by 30, slightly extended and fixed with a Mayfield neurosurgical head holder. A sandbag was placed under the mattress on the left side of the operating table to extend the neck and left shoulder. Electrophysiologic monitoring was performed bilaterally for SSEPs in upper (median nerve) and lower (posterior tibial nerve) extremities, and brainstem auditory-evoked potentials. Motorevoked potentials were not obtained because most of the 818
Case Report
DISCUSSION
The reversibility of upper extremity SSEP changes with repositioning in our patients establishes a cause and effect relationship between positioning and SSEP
Figure 2. Somatosensory-evoked potential (SSEP) changes in left median nerve in Case 2. SSEPs from bilateral median nerves
were recorded using needle electrodes placed at Erbs point over the brachial plexus, for assessing peripheral conduction, and on the scalp at CP3 and CP4 locations, for assessing central conduction. Time in minutes is recorded on the y axis. Upon stimulation of the left median nerve, there is no SSEP response from the start of the recordings (time 1 min). Again, the difference exists at both peripheral (Erbs) and central (CP3 and CP4) recordings, suggesting a peripheral mechanism. With repositioning of the shoulder, the amplitude increases on the left side with time in central and peripheral recordings, approaching the amplitude on the right side.
change. Although we cannot be certain that the observed SSEP changes would have been associated with a permanent deficit, if the changes indeed heralded ischemia of the nerves, clinical neuropathy would be likely after a lengthy procedure. In addition, although not routine in our center for cases not involving cervical instability, these cases identify the utility of preposition baseline SSEPs, especially because the patients will be monitored during the case. This would clearly identify positioning changes by having preand postrecordings. SSEP changes caused by positioning patients have been reported for prone2,5,6 and supine spine surgery,710 cardiac,2,4 and orthopedic surgeries.10 12 Our case report shows a change in SSEPs of the brachial plexus during positioning for craniotomy. The incidence of SSEP changes during positioning of the prone spine surgery patient ranges from 3.6% to 15%. Schwartz et al.2 found a 30% decrease in ulnar nerve SSEP amplitude in 3.6% of pediatric patients having scoliosis correction. OBrien et al.5 reported a 15% prevalence of brachial plexopathy during lumbar and spinal deformity surgery (60% decrease in amplitude or 10% increase in latency). Kamel et al.6
found that the lateral decubitus position (7.5%) and prone superman position (7.0%) had a higher incidence of position-related upper extremity SSEP changes compared with other positions including supine, arms tucked, and arms out (1.8%3.2%). The incidence of positioning changes in SSEP recordings during anterior spine surgery is lower than that of prone spine surgery.13 Schwartz et al.7 found that 1.8% of patients undergoing anterior cervical spine surgery showed intraoperative SSEP changes secondary to positioning. SSEP changes are more frequent in patients with myelopathy who were having anterior spine surgery.8 There has been only one study to describe positioning changes during SSEP monitoring for craniotomy surgery. Deinsberger et al.14 performed SSEP monitoring for positioning of patients for posterior fossa surgery in the semisitting position. Monitoring of the median and tibial nerve was recorded for 55 consecutive patients. SSEPs of the median nerve showed no changes while placing patients in the sitting position; however, tibial SSEP recordings were altered in 14 cases (25%).
819
We reviewed all cases of craniotomy for skull base tumors and acoustic neuromas in a 2-yr period to derive an incidence of SSEP changes during this category of neurosurgical procedure. At our institution, all craniotomy cases for acoustic neuromas and skull base tumors receive SSEP neuromonitoring. We found 83 cases performed in the 2-yr period between January 2007 and December 2008. Of these cases, we reviewed the neuromonitoring records and identified two cases with SSEP changes upon positioning, which are described in this case series: an incidence of 2.4%. This is approximately the incidence for SSEP changes for anterior cervical spine surgery (1.8%), but less than that found in prone and lateral decubitus positioning. The clinical correlation of SSEPs and clinical findings has been the subject of several investigations. Prielipp et al.15 studied the onset of clinical symptoms and SSEP changes in the ulnar nerve in 50 volunteers with their arm in 3090 of abduction, as well as in supination, neutral orientation, and pronation. Half of the patients who developed SSEP changes did not develop clinical symptoms, two of which manifested severe SSEP changes (60% decrease in amplitude). In contrast, Lorenzini and Poterack16 monitored SSEPs at the median and ulnar nerves in awake volunteers placed in a prone position as their arms were moved in a step-wise cephalad position. Three of seven patients developed upper extremity symptoms described as tingling, numbness, or aching in the hand, forearm, or upper arm without changes in their SSEPs, suggesting that SSEP monitoring is imperfect in detecting positioning injury. However, volunteer studies do not apply directly to anesthetized patients, who are often placed in positions that awake patients could not tolerate. These data demonstrate the false positive and negative predictive values of clinical peripheral neuropathic changes based on SSEP monitoring. There may, indeed, be instances in which SSEP changes, or the lack thereof, may be misleading in predicting postoperative clinical neuropathy. However, many studies that have evaluated SSEP monitoring with postoperative peripheral neurological changes studied SSEP changes over a short interval. In the study by Lorenzini and Poterack,16 recordings for SSEP signals were made 10 15 min after positioning the arm. The study by Prielipp et al.15 was performed for a maximum of 60 min, and on average, patients experienced clinical symptoms after 37 min, with a range of 20 59 min. In contrast, our first case took 6 h and 50 min, and the second case took 11 h and 33 min. It is unclear, but seems likely, that changes observed by SSEP monitoring would eventually lead to clinical neuropathy during these lengthy procedures. To determine if we had clinical peripheral neuropathy changes that were not predicted by SSEPs, we reviewed the case records of our department obtained from resident postoperative visits, consults, and patient phone calls after hospital discharge. From January 2007 to
820
Case Report
December 2008, 65,041 cases were reviewed. Of all types of surgical cases, 10 cases were identified in which postoperative peripheral neuropathy was reported: an incidence of 0.02%. None of these cases were for craniotomy. The major limitation of this type of review is that it is retrospective, thus underreporting of a complication can underestimate the incidence of peripheral neuropathy. However, with SSEP monitoring for cases, we may be decreasing the incidence of peripheral neuropathy due to positioning. Although the incidence of peripheral nerve injury is low, it does remain a common cause of professional liability in the practice of anesthesiology. Lee et al.17 reviewed the American Society of Anesthesiologists closed-claims analysis and found that peripheral nerve injury was the complication recorded in 21% of neurosurgical closed claims. This finding emphasizes the importance of taking measures to prevent peripheral nerve damage, including vigilance of neuromonitoring changes. These cases demonstrate that changes in SSEPs of the brachial plexus can occur during positioning for skull base craniotomy and acoustic neuromas. This case series, in addition to the literature on this topic, underlines the clinical usefulness of monitoring SSEPs during and after positioning the patient and adjusting position accordingly to prevent neurological injury. REFERENCES
1. Epstein NE, Danto J, Nardi D. Evaluation of intraoperative somatosensory-evoked potential monitoring during 100 cervical operations. Spine 1993;18:737 47 2. Schwartz DM, Drummond DS, Hahn M, Ecker ML, Dormans JP. Prevention of positional brachial plexopathy during surgical correction of scoliosis. J Spinal Disord 2000;13:178 82 3. Jellish WS, Martucci J, Blakeman B, Hudson E. Somatosensory evoked potential monitoring of the brachial plexus to predict nerve injury during internal mammary artery harvest: intraoperative comparisons of the rultract and pittman sternal retractors. J Cardiothorac Vasc Anesth 1994;8:398 403 4. Hickey C, Gugino LD, Aglio LS, Mark JB, Son SL, Maddi R. Intraoperative somatosensory evoked potential monitoring predicts peripheral nerve injury during cardiac surgery. Anesthesiology 1993;78:29 35 5. OBrien MF, Lenke LG, Bridwell KH, Padberg A, Stokes M. Evoked potential monitoring of the upper extremities during thoracic and lumbar spinal deformity surgery: a prospective study. J Spinal Disord 1994;7:277 84 6. Kamel IR, Drum ET, Koch SA, Whitten JA, Gaughan JP, Barnette RE, Wendling WW. The use of somatosensory evoked potentials to determine the relationship between patient positioning and impending upper extremity nerve injury during spine surgery: a retrospective analysis. Anesth Analg 2006;102:1538 42 7. Schwartz DM, Sestokas AK, Hilibrand AS, Vaccaro AR, Bose B, Li M, Albert TJ. Neurophysiological identification of positioninduced neurologic injury during anterior cervical spine surgery. J Clin Monit Comput 2006;20:437 44 8. Kombos T, Suess O, Da Silva C, Ciklatekerlio O, Nobis V, Brock M. Impact of somatosensory evoked potential monitoring on cervical surgery. J Clin Neurophysiol 2003;20:122 8 9. Jones SC, Fernau R, Woeltjen BL. Use of somatosensory evoked potentials to detect peripheral ischemia and potential injury resulting from positioning of the surgical patient: case reports and discussion. Spine J 2004;4:360 2 10. Swenson JD, Hutchinson DT, Bromberg M, Pace NL. Rapid onset of ulnar nerve dysfunction during transient occlusion of the brachial artery. Anesth Analg 1998;87:677 80
11. Mills WJ, Chapman JR, Robinson LR, Slimp JC. Somatosensory evoked potential monitoring during closed humeral nailing: a preliminary report. J Orthop Trauma 2000;14:16770 12. Porter SS, Black DL, Reckling FW, Mason J. Intraoperative cortical somatosensory evoked potentials for detection of sciatic neuropathy during total hip arthroplasty. J Clin Anesth 1989;1: 170 6 13. Labrom RD, Hoskins M, Reilly CW, Tredwell SJ, Wong PK. Clinical usefulness of somatosensory evoked potentials for detection of brachial plexopathy secondary to malpositioning in scoliosis surgery. Spine 2005;30:2089 93 14. Deinsberger W, Christophis P, Jodicke A, Heesen M, Boker DK. Somatosensory evoked potential monitoring during positioning of the patient for posterior fossa surgery in the semisitting position. Neurosurgery 1998;43:36 40; discussion 40 2
15. Prielipp RC, Morell RC, Walker FO, Santos CC, Bennett J, Butterworth J. Ulnar nerve pressure: influence of arm position and relationship to somatosensory evoked potentials. Anesthesiology 1999;91:34554 16. Lorenzini NA, Poterack KA. Somatosensory evoked potentials are not a sensitive indicator of potential positioning injury in the prone patient. J Clin Monit 1996;12:171 6 17. Lee LA, Posner KL, Cheney FW, Domino KB. ASA closed claims project: an analysis of claims associated with neurosurgical anesthesia. Anesthesiology 2003;99:A362
821
Patient Safety
Section Editor: Sorin J. Brull
The Diagnostic Value of the Upper Lip Bite Test Combined with Sternomental Distance, Thyromental Distance, and Interincisor Distance for Prediction of Easy Laryngoscopy and Intubation: A Prospective Study
Zahid Hussain Khan, MD* Mostafa Mohammadi, MD* Mohammad R. Rasouli, MD Fahimeh Farrokhnia, MD* Razmeh Hussain Khan
BACKGROUND: Accuracy of upper lip bite test (ULBT) has been compared with the Mallampati classification. In this study, we investigated whether the combination of the ULBT classification with sternomental distance (SMD), thyromental distance (TMD), and interincisor distance (IID) or a composite score can improve the ability to predict easy laryngoscopy and intubation compared with each test alone. METHODS: In a prospective study, 380 patients who were scheduled for elective surgery were selected randomly and enrolled in the study. Before inducing anesthesia, the airways were assessed, and ULBT class, SMD, TMD, and IID determined. Laryngoscopic view according to the Cormack and Lehane grading system was determined after induction of anesthesia and Grades 3 and 4 defined as difficult intubation. By using receiver operating characteristic analysis, the best cutoff points of the tests were calculated. Finally, sensitivity, specificity, positive and negative predictive values and accuracy of these tests and their combinations with the ULBT were calculated. RESULTS: The prevalence of difficult intubation was 5% (n 19). Class III ULBT, IID 4.5 cm, TMD 6.5 cm, and SMD 13 cm were defined as predictors of difficult intubation. There was no significant difference regarding difficult intubation based on gender (P 0.05), whereas there were significant differences between the older tests and laryngeal view (P 0.05, Mc-Nemar test). Specificity and accuracy of the ULBT were significantly higher than TMD, SMD, and IID individually (specificity was 91.69%, 82.27%, 70.64%, and 82.27%, respectively, and accuracy was 91.05%, 71.32%, 81.84%, and 76.58%, respectively). The combination of the ULBT with SMD provided the highest sensitivity. CONCLUSION: We conclude that the specificity and accuracy of the ULBT is significantly higher than the other tests and is more accurate in airway assessment. However, the ULBT in conjunction with the other tests could more reliably predict easy laryngoscopy or intubation.
(Anesth Analg 2009;109:8224)
lthough many advances have been made and many time-tested methods have been used to overcome the conundrum of an unanticipated difficult laryngoscopic tracheal intubation, most tests are not reliable.1 All preoperative airway assessment tests are characterized by low sensitivity, reasonable specificity, low positive predictive value (PPV), and significant false positives.2 4 Although our initial evaluation of the upper lip bite test (ULBT1) showed greater specificity and accuracy
From the *Department of Anesthesiology, Imam Khomeini Medical Center, Tehran University of Medical Sciences; School of Medicine, Tehran University of Medical Sciences; and Tehran University of Medical Sciences, Tehran, Iran. Accepted for publication April 18, 2009. Address correspondence and reprint requests to Zahid Hussain Khan, MD, Department of Anesthesiology, Imam Khomeini Medical Center, Tehran University of Medical Sciences, Keshavarz Blvd., Tehran 1419733141, Iran. Address e-mail to khanzh51@yahoo.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af7f0d
compared with the Mallampati classification, the ULBT has not been compared with other tests, such as measurement of sternomental distance (SMD), thyromental distance (TMD), and interincisor distance (IID). In this study, the ULBT is compared with these tests for preoperative assessment of airway and prediction of ease of tracheal intubation. We also aim to describe a composite measure that combines ULBT with each of the other measures.
METHODS
Three hundred eighty ASA I patients older than 16 yr scheduled for elective surgical procedures requiring endotracheal intubation were enrolled in the study after approval of the ethics committee of the university. Verbal informed consent was obtained from each patient before starting the study. Patients with any airway abnormality or obvious neck pathology were excluded.
822
The SMD was measured in supine position with the head fully extended and with the mouth closed. The straight distance between the upper border of the manubrium sterni and the bony point of the mentum was measured. For TMD, the straight distance between the upper border of the thyroid cartilage and the bony point of mentum was measured.5 IID was measured when the patient opened his or her mouth, and the distance between incisors was obtained. The ULBT class was determined according to the following criteria: Class I, lower incisors can bite the upper lip above the vermilion line; Class II, lower incisors can bite the upper lip below the vermilion line; and Class III, lower incisors cannot bite the upper lip.1 Anesthesia was induced with midazolam (1 mg), fentanyl (2 g/kg), thiopental (5 mg/kg), and atracurium (0.5 mg/kg). With the head in the sniffing position, laryngoscopy with a Macintosh 3 blade was attempted, and the view determined using the Cormack-Lehane (C-L) grading system.6 C-L Grades 1 and 2 were categorized as easy intubations and Grades 3 and 4 as difficult intubations. All preoperative airway assessments and measurements were performed by a third-year resident, and subsequent laryngoscopies and C-L grading were conducted and scored by one of the authors (ZHK), who was blinded to the observations made preoperatively. Statistical analysis was performed using SPSS software version 10.5 (SPSS, Chicago, IL) and MedCalc version 9.2 (MedCalc Software, Maria-kerke, Belgium). Receiver operating characteristic (ROC) analysis was used to determine the accuracy of each test and the combination of ULBT with the other tests (SMD, TMD, and IID). For this purpose, a binary variable was defined. If both the ULBT and the other test had predicted easy intubation based on obtained cutoff points, the new variable was coded as easy. If one or both of them had predicted difficult intubation, the variable value was coded as difficult intubation. Sensitivity, specificity, accuracy, PPV, and negative predictive values (NPV) were then calculated. The best cutoff points were determined by selecting a point where the sensitivity and specificity were approximately equal. Data were analyzed by using Fishers exact and Mc-Nemar tests, and a P value 0.05 was considered statistically significant.
Figure 1. The upper lip bite test (ULBT) classes versus

Cormack-Lehane (C-L) grading.
There was no significant difference regarding difficult intubation according to gender (P 0.05, 2 test). However, there were differences by gender with regard to the tests and C-L grades (P 0.05, Mc-Nemar test). A combination of each test with the ULBT showed the highest sensitivity for the combination of ULBT with SMD, whereas the highest specificity was found in the combination of ULBT with TMD.
DISCUSSION
Our results show that the ULBT has higher accuracy and specificity than the other tests and also a high NPV. The results also indicate that Class I ULBT is more likely to predict an easy intubation than the other tests. We further found that the combination of SMD and ULBT improved the sensitivity of ULBT when compared with the latter alone. This finding was the hallmark of our results when the tests were combined. However, the combination of the ULBT with the other tests did not show any superiority to the ULBT alone with regard to specificity. A combination of these tests did not enhance PPV, NPV, and accuracy compared with those obtained with the ULBT alone. One of the most important challenges in using SMD, TMD, and IID is the quantitative nature of these tests,7,8 whereas the classification of patients based on the ULBT is of a qualitative nature, making differentiation of classes easy and precise. In brief, the differences between the ULBT and the other tests are those between continuous and discrete variables. Thus, the ULBT is associated with the least interobserver variability, which adds to its advantage as an airway assessment test. These findings are consistent with a previous report by Eberhart et al.,9 which showed that the interobserver reliability of the ULBT is higher than the Mallampati classification. The prevalence of difficult intubation in our study was 5%; however, failure to intubate was not encountered. Our results corroborate those of a previous study, which reported a prevalence of 4.7% for difficult intubation.3 Wilson et al.10 suggested five risk factors in predicting difficult intubation, including
RESULTS
Three hundred nine patients (209 men) were included in the study. The mean age was 34 10 yr (mean sd). Intubation was difficult in 19 patients (5%, C-L Grades of 3 and 4 in 17 and two cases, respectively). The ULBT classes versus C-L grades are depicted in Figure 1. By using receiver operating characteristic analysis, Class III ULBT, IID 4.5 cm, TMD 6.5 cm, and SMD 13 cm were defined as cutoff points for difficult intubation (Table 1). These criteria were determined to predict difficult intubation.
823
Table 1. Area Under the ROC Curve, Sensitivity, Specicity, Positive and Negative Predictive Values, and Accuracy of Interincisor Distance (IID), Thyromental Distance (TMD), Sternomental Distance (SMD), and Upper Lip Bite Test (ULBT) Are Being Shown
Area under the ROC curve IID 4.5 cm TMD 6.5 cm SMD 13.5 cm ULBT class III IID ULBT TMD ULBT SMD ULBT 0.72 (0.60.85) 0.78 (0.660.89) 0.77 (0.670.87) 0.85 (0.740.96) 0.77 (0.730.81) 0.79 (0.740.83) 0.76 (7180) Sensitivity 68.4 (46.784.4) 73.6 (52.188) 84.2 (63.194.4) 78.9 (58.391.3) 78.9 (57.491.4) 78.9 (57.591.4) 84.2 (63.194.4) Specicity 77.0 (75.977.9)* 82.2 (81.183)* 70.6 (69.571.2)* 91.9 (90.992.6) 76.2 (7576.8) 79.8 (78.880.4) 67.9 (66.868.4) Positive predictive value 13.5 (9.316.7) 17.9 (12.721.4) 13.1 (9.814.7)* 33.3 (25.239.4) 14.9 (10.817.2) 17.0 (12.419.7) 12.1 (9.113.6) Negative predictive value 97.8 (96.499) 98.3 (9799.2) 98.8 (97.399.6) 98.8 (97.699.5) 98.6 (97.199.4) 98.6 (97.299.4) 98.8 (97.299.6) Accuracy 76.5 (74.278.2)* 81.8 (79.783.3)* 71.3 (69.272.3)* 91.0 (89.392.5) 76.3 (74.277.6) 79.7 (77.681) 68.7 (66.669.7)
Negative predictive value for all tests is high, indicating that all tests can predict easy intubation readily. ROC Receiver operating characteristic. * Signicant differences with the similar parameter in ULBT (P 0.05, Fishers exact test).
weight (P 0.05), head and neck movement (P 0.001), jaw movement (P 0.001), receding mandible (P 0.001), and protruding (buck) upper incisors (P 0.001). The ULBT when tested initially had the potential to evaluate both jaw movement and buck teeth simultaneously, providing additional support for its use as an airway assessment test. Sensitivity, specificity, and accuracy of the ULBT (78.95%, 91.96%, and 91.05%, respectively) were similar to those reported in the earlier study (76.5%, 88.7%, and 88%, respectively).1 The ULBT has high specificity and NPV, making it superior in identifying easy tracheal intubation and laryngoscopy. The sensitivity and specificity of SMD and TMD corroborate previous data.2 With regard to the combination of ULBT and other tests, we found that the combination of the SMD and the ULBT had a higher sensitivity than the ULBT alone; however, this combination had lower specificity and PPV than the ULBT alone. Our results differ from those of Richa et al.,11 which showed a low sensitivity for the combination of the ULBT with the TMD and ULBT with SMD (18.7% and 15.6%, respectively) but high specificity for these combinations (99% and 98%, respectively). The PPV of the test combinations were 60% and 50%, respectively, whereas in our study, these combinations had lower PPV.11 Because the predictive values depend on the values of sensitivity, specificity, and prevalence (which was higher in their study), this could explain the difference between our results and theirs. A test to predict difficult intubation should have high sensitivity, so that it will identify most patients in whom intubation will truly be difficult.7 It should also have a high PPV, so that only few patients with airways actually easy to intubate are subjected to the protocol for management of a difficult airway.12,13 Similarly, a test should have a high NPV to correctly predict the ease of laryngoscopy and intubation. It is being argued that both long and short TMD portend difficulty in correctly anticipating the laryngeal view,14 which challenges the validity of TMD as a useful screening test. On the other hand, the ULBT has the ability to assess jaw movement and protruding incisors simultaneously, enhancing its predictability. In conclusion, despite the fact that this study showed acceptable sensitivity and PPV of ULBT in comparison with the other tests, the ULBT is a test with
824
Easy Laryngoscopy Using ULBT
high specificity and NPV, making it a favorable test for identifying easy intubations and laryngoscopy. These findings are in agreement with those of Eberhart et al.9 However, we suggest applying a combination of assessment methods in predicting the ease of intubation. Such a combination is preferable because anatomic predictors of difficult intubation carry a low-sensitivity rate when used alone, whereas a multivariate composite risk index may achieve better results than single, independent criteria. REFERENCES
1. Khan ZH, Kashfi A, Ebrahimkhani E. A comparison of the upper lip bite test (a simple new technique) with modified Mallampati classification in predicting difficulty in endotracheal intubation: a prospective blinded study. Anesth Analg 2003;96:5959 2. Savva D. Prediction of difficult tracheal intubation. Br J Anaesth 1994;73:149 53 3. Rose DK, Cohen MM. The airway: problems and predictions in 18,500 patients. Can J Anaesth 1994;41:372 83 4. Samsoon GL, Young JR. Difficult tracheal intubation: a retrospective study. Anaesthesia 1987;42:48790 5. Karkouti K, Rose DK, Wigglesworth D, Cohen MM. Predicting difficult intubation: a multivariable analysis. Can J Anaesth 2000;47:730 9 6. Cormack RS, Lehane J. Difficult tracheal intubation in obstetrics. Anaesthesia 1984;39:110511 7. Arne J, Descoins P, Fusciardi J, Ingrand P, Ferrier B, Boudigues D, Aries J. Preoperative assessment for difficult intubation in general and ENT surgery: predictive value of a clinical multivariate risk index. Br J Anaesth 1998;80:140 6 8. Karkouti K, Rose DK, Ferris LE, Wigglesworth DF, MeisamiFard T, Lee H. Inter-observer reliability of ten tests used for predicting difficult tracheal intubation. Can J Anaesth 1996; 43:554 9 9. Eberhart LH, Arndt C, Cierpka T, Schwanekamp J, Wulf H, Putzke C. The reliability and validity of the upper lip bite test compared with the Mallampati classification to predict difficult laryngoscopy: an external prospective evaluation. Anesth Analg 2005;101:284 9 10. Wilson ME, Spiegelhalter D, Robertson JA, Lesser P. Predicting difficult intubation. Br J Anaesth 1988;61:211 6 11. Richa F, Yazbeck P, Yazigi A, Karim N, Antakly MC. Value of the association of the upper lip bite test (ULBT) with other tests in predicting difficulty of endotracheal intubation. Anesthesiology 2005:A1418 [abstract] 12. Tse JC, Rimm EB, Hussain A. Predicting difficult endotracheal intubation in surgical patients scheduled for general anesthesia: a prospective blind study. Anesth Analg 1995;81:254 8 13. Merah NA, Wong DT, Foulkes-Crabbe DJ, Kushimo OT, Bode CO. Modified Mallampati test, thyromental distance and interincisor gap are the best predictors of difficult laryngoscopy in West Africans. Can J Anaesth 2005;52:291 6 14. Brodsky JB, Lemmens HJ, Brock-Utne JG, Vierra M, Saidman LJ. Morbid obesity and tracheal intubation. Anesth Analg 2002; 94:732 6
A Macintosh Laryngoscope Blade for Videolaryngoscopy Reduces Stylet Use in Patients with Normal Airways
Andre van Zundert, MD, PhD, FRCA* Ralph Maassen, MD Ruben Lee, BE Remi Willems, MD Michel Timmerman, MD Marc Siemonsma, MD Marc Buise, MD* Marco Wiepking, MD*
BACKGROUND: Although most tracheal intubations with direct laryngoscopy are not performed with a styletted endotracheal tube, it is recommended that a stylet can be used with indirect videolaryngoscopy. Recently, there were several reports of complications associated with styletted endotracheal tubes and videolaryngoscopy. In this study, we compared three videolaryngoscopes (VLSs) in patients undergoing tracheal intubation for elective surgery: the GlideScope Ranger (GlideScope, Bothell, WA), the V-MAC Storz Berci DCI (Karl Storz, Tuttlingen, Germany), and the McGrath (McGrath series 5, Aircraft medical, Edinburgh, UK) and tested whether it is feasible to intubate the trachea of patients with indirect videolaryngoscopy without using a stylet. METHODS: Four hundred fifty consecutive adults (ASA PS III) undergoing tracheal intubation for elective surgery were randomly allocated for airway management with one of the three devices. Anesthesia induction for tracheal intubation consisted of fentanyl-propofol-rocuronium. An independent anesthesiologist used the Cormack-Lehane grading system to score an initial direct laryngoscopic view using a classic metal Macintosh blade. After subsequent positive-pressure ventilation using a face mask and an oxygen-sevoflurane mixture for 1 min, the trachea was intubated using one of the three VLSs. During intubation, the following data were collected: intubation time, number of intubation attempts, use of extra tools to facilitate intubation, and overall satisfaction score of the intubation conditions. RESULTS: The trachea of every patient was intubated using the VLSs, and none of the patients required conversion to the classic Macintosh laryngoscope. All three VLSs offered equal or better view of the glottis as assessed by the mean Cormack-Lehane grade, compared with the traditional Macintosh laryngoscopy, including a larger viewing angle of the glottic entrance. The average intubation time was 34 20 s for the GlideScope, 18 12 s for the V-MAC Storz, and 38 23 s for the McGrath VLS. Intubation with the Storz was faster (P 0.05) than the other two VLS tested and necessitated fewer additional tools (P 0.01), resulting in a higher first-pass successful intubation rate. A stylet had to be used in 7% of the patients in the Storz group versus about 50% of the patients when the other two VLS were used. CONCLUSIONS: The trachea of a large proportion of patients with normal airways can be intubated successfully with certain VLS blades without using a stylet, although the three studied VLSs clearly differ in outcome. The Storz VLS displaces soft tissues in the fashion of a classic Macintosh scope, affording room for tracheal tube insertion and limiting the need for stylet use compared with the other two scopes. Although VLSs offer several advantages, including better visualization of the glottic entrance and intubation conditions, a good laryngeal view does not guarantee easy or successful tracheal tube insertion. We recommend that the geometry of VLSs, including blade design, should be studied in more detail.
(Anesth Analg 2009;109:82531)
here is clear evidence that indirect videolaryngoscopy offers improved viewing of the glottic entrance
From the *Department of Anesthesiology, Intensive Care and Pain Therapy, Catharina Hospital, Brabant Medical School, Eindhoven, The Netherlands; Department of Anesthesiology, Ghent University Hospital, Ghent, Belgium; Department of Anesthesiology, Catharina Hospital, Eindhoven, Netherlands; Department of Anesthesiology, University Hospital, Maastricht, The Netherlands; and Department of Biomechanical Engineering, Mechanical, Materials and Maritime Engineering, Delft University of Technology, Delft, The Netherlands. Accepted for publication April 9, 2009. Supported by departmental funds of the Catharina Hospital, Eindhoven, the Netherlands. The videolaryngoscopes were made available to the study center, on a temporary basis, at no cost by the respective manufacturers: Ranger GlideScope, Verathon Medical, Bothell, WA; V-Mac Storz Macintosh, Karl Storz, Tuttlingen, Germany; and McGrath Series 5, Aircraft Medical, Edinburgh, UK. Vol. 109, No. 3, September 2009
over direct classic laryngoscopy.1,2 A better view is assumed to facilitate easier intubations, but this is not entirely confirmed,3,4 as a good laryngeal view does not guarantee easy or successful endotracheal tube
All patients in this clinical study underwent surgery and anesthesia at the Catharina Hospital, Brabant Medical School, Eindhoven, The Netherlands. Address correspondence to Andre van Zundert, MD, PhD, FRCA, Department of Anesthesiology, Intensive Care and Pain Therapy, Catharina HospitalBrabant Medical School, Michelangelolaan 2, NL-5623 EJ Eindhoven, The Netherlands. Address e-mail to zundert@iae.nl. Reprints will not be available from the author. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae39db
825
(ETT) insertion. Several manufacturers are producing videolaryngoscopes (VLSs) with differing specifications, user interface, and geometry. However, the relative performance of the different models is unknown because no comparative studies have been published. Most tracheal intubations using direct laryngoscopy are performed without instruments other than the ETT itself. However, the 60 angle of some VLS flanges limits the advancement of the ETT through the vocal cords into the trachea.57 Some manufacturers of VLSs advocate the use of stylets with the ETT to facilitate an easier insertion into the trachea.8 10 However, recent reports have shown rare but potentially serious complications associated with styletted ETT and VLSs.3,1116 In this prospective, randomized comparative study, we evaluated, in clinical circumstances, the effectiveness of three commonly used VLSs in laryngoscopy and tracheal intubation of patients with normal airways and tested whether it is feasible to intubate the trachea with indirect videolaryngoscopy without using a stylet. Our first objective was to assess the necessity of stylet use with the VLS. It is of clinical interest to determine which VLS has the most suitable geometric and ergonomic designs for tracheal placement of a standard ETT without the use of a stylet. We hypothesized that there are significant differences in the ease of insertion of the ETT given the substantially different laryngoscope blade designs of the VLSs studied. Second, we considered intubation conditions between the blades with use of a styletted ETT. Because the stylet essentially compensates for the geometrical mismatch of the VLS with the laryngeal anatomy of the patient, we expected no discernable differences among the three VLSs. Finally, we performed a comparative assessment of the glottic view among the respective VLSs and classic direct Macintosh laryngoscopy. We hypothesized that all indirect VLSs are equal or superior to classic direct laryngoscopy in terms of the glottic view and that there are no significant differences in visualization quality among the respective VLSs.
Figure 1. Photo compilation of the three videolaryngoscopes (VLSs) studied. From left to right: the GlideScope Ranger, the Storz V-MAC, and the McGrath Series-5 VLS. The preanesthetic visit of the patient (performed by an anesthesiologist not involved in this study) determined history of difficult intubation, measurement of common predictive indices for difficult intubation (BMI, thyromental distance, Mallampati grade, interdental [or intergum] distances), and evaluation of status of dentition and neck movement. When the patients arrived at the operating room, they were placed in the sniffing position with their head placed on a pillow, connected to standard monitoring devices and breathed 100% oxygen for at least 3 min. Anesthesia induction consisted of IV fentanyl 1.5 g/kg, propofol 3 mg/kg, and rocuronium 0.7 mg/kg, and the lungs were manually inflated through a face mask using sevoflurane in oxygen. After approximately 3 min, guided by objective confirmation of adequate degree of neuromuscular blockade (train-of-four monitoring), the same independent anesthesiologist not involved in this study performed an initial direct laryngoscopy using a classic metal Macintosh (Heine Optotechnik GmbH & Co. KG, Herrsching, Germany) blade; laryngoscopic view was scored according to the Cormack-Lehane (C&L) grading system, although no intubation took place. After subsequent positive-pressure ventilation using a face mask and an oxygen-sevoflurane mixture for 1 min, the trachea was intubated using one of the three VLSs available in our hospital. The anesthesiologist performing the intubation was blinded to the C&L laryngoscopy score given by the first anesthesiologist. All tracheal intubations were performed by one of five different anesthesiologists, all of whom were experienced in anesthesia, and the use of the VLS was studied (introduction of VLS course in airway skills laboratory and minimal of 30 uses with each VLS). During intubation, intraprocedural metrics of intubation difficulty (C&L grade) and our dependent variables of intubation time, number of attempts, and use of extra tools to facilitate tracheal intubation were measured. In a pilot study, the use of a stylet was
METHODS
After approval of the hospital medical ethics committee (Catharina Hospital, Eindhoven, the Netherlands) and obtaining informed consent, 450 consecutive adult patients, undergoing intubation for elective surgery, were randomly selected to receive general anesthesia and tracheal intubation using one of three VLSs: GlideScope (Ranger, Verathon Bothell, WA), V-Mac Storz Berci DCI (Storz, Karl Storz, Tuttlingen, Germany), and McGrath (McGrath Series 5, Aircraft Medical, Edinburgh, UK) (Fig. 1). Randomization was done through sealed envelope. Exclusion criteria were physical status ASA Class IIIV; age 18 yr; body mass index (BMI) 35 kg/m2; and patients with known airway pathology or cervical spine injury.
826
Intubation with Videolaryngoscopes
Table 1. Patient Characteristics and Preprocedural Intubation Conditions Videolaryngoscope

Male:female n (%) Age (yr) Weight (kg) Height (m) ASA Body mass index (kg/m2) Thyromental distance (cm) Max mouth opening (cm) Mallampati grade Mean sd (I:II:III:IV) Dentition Double denture:single:none, n Percentage
ASA American Society of Anesthesiologists.
GlideScope Ranger (n 150)

54:96 (36:64) 50.4 16.3 76.3 15.7 1.71 0.09 1.52 0.51 26.1 4.27 7.3 1.1 4.4 0.6 1.65 0.66 68:66:16:0 117:6:27 78:4:18
Storz V-Mac (n 150)

62:88 (41:59) 54.4 16.4 77.1 14.6 1.71 0.09 1.59 0.51 26.3 4.04 7.4 1.0 4.3 0.6 1.75 0.65 55:77:18:0 104:12:34 69:8:23
McGrath Series-5 (n 150)

60:90 (40:60) 56.2 15.3 76.1 14.6 1.70 0.09 1.59 0.49 26.3 4.10 7.5 1.1 4.3 0.6 1.70 0.66 61:72:17:0 101:13:36 67:9:24
favored over the gum-elastic bougie. Therefore, the choice was made to use a specific rigid stylet, formed in the shape of a hockey stick with a 90 bend, optimized for use with the VLS (GlideScope Rigid Stylet, Saturn Biomedical Systems, Burnaby, BC, Canada), as the first option if intubation was not feasible after two intubation attempts.17,18 The number of intubation attempts was counted as each approach of the ETT to the glottis entrance. Intubation time was measured (by an assistant) as the time between picking up the ETT (Hi-contour, Mallincrodt Medical, Athlone, Ireland) and the visual passage of the ETT until the vocal cords were between the two black line markings on the distal end of the ETT. Interim bag and mask ventilation time, if needed, was not included in the total intubation time. More than five attempts or 120 s were regarded as failure of intubation. If failure to secure the airway occurs with the VLS, then conventional difficult intubation protocols from the hospital were to be implemented. An overall satisfaction score of the intubation conditions was rated on a scale from 0 to 4: 0 failure, intubation not possible; 1 poor (had to use a tool other than the VLS); 2 fair (need for an extra tool plus intubation time 90 s); 3 moderate (need for extra tool to intubate the trachea, but intubation time 90 s); and 4 good (intubation successful on first or second attempt, within 90 s, and no need for extra tools to secure the airway). Attention was paid to insert and remove the VLS smoothly not to damage the oral cavity, the tongue, or the patients dentition. After removal of the VLS, the oral cavity was inspected for any bruises, lacerations, bleedings, dental damage, or other possible complications. A priori sample size testing was conducted assuming an analysis of variance (ANOVA) model for the time measurements. Using three treatments (blade), an effect size of 5 s from clinical experience, a highdesired statistical power level of 0.95 and level of
0.05, we calculated a sample size of 50 patients. We expanded the patient group to account for the three hypotheses tested and necessary correction of the sample size (i.e., Bonferroni). Data were reported as mean (sd) and incidences (both absolute and percentage). ANOVA was used to assess any differences among the groups regarding the patient-specific characteristics (i.e., age, BMI, thyromental distance, and interdental distance). Nonparametric patient metrics (i.e., gender, ASA PS, and dentition) were evaluated for differences among the groups using KruskalWallis one-way ANOVA. The differences in the dependent parameters of intubation time, attempts, use of additional tools, and overall satisfaction for the respective VLS groups was evaluated using KruskalWallis nonparametric one-way ANOVA (to forgo assumptions of normality) and Bonferroni correction for the multiple hypothesis testing. Finally, the C&L grades were compared for each of the tested VLSs and the classical Macintosh blade again using KruskalWallis one-way ANOVA. All statistical analysis was performed using MATLAB 7.2 (R2006a) (Mathworks, MA). P 0.05 was considered statistically significant.
RESULTS
Patient characteristics and preprocedural intubation conditions did not differ among patient groups (Table 1). Patients underwent a large variety of general surgery, orthopedic surgery, urology, gynecology, and plastic surgery. Peripheral oxygen saturation was maintained above 95% in all patients throughout the laryngoscopy and intubation period. All operations were completed uneventfully. We did not detect any injury of the palatoglossal arch or dental injury in any patient. Minor lip lacerations were seen in four patients. The first hypothesis concerned the differences between the VLSs studied in their effectiveness for
827
Table 2. Intubation Metrics GlideScope Ranger (n 150)

Intubation attempts Mean sd 1:2:3:4 Percentage Intubation attempts without stylet 1 attempt, n (%) 2 attempts, n (%) Total, n (% from all subjects) Intubation attempts with stylet 1 attempt, n (%) 2 attempts, n (%) Total, n (% from all subjects) Intubation Time Mean sd (s) Cormack-Lehane grade, mean sd Classic laryngoscope VLS Overall satisfaction score Good:moderate:fair:poor:failure, n Percentage 2.23 1.00* 46:40:49:15* 31:27:33:10 46 (53) 40 (47) 86 (57) 49 (76) 15 (24) 64 (43) 34 20* 1.68 0.76 1.01 0.11 87:59:4:0:0* 58:39:3:0:0
Storz V-Mac (n 150)

1.30 0.63 118:22:8:2 79:15:5:2 118 (84) 22 (16) 140 (93) 8 (80) 2 (20) 10 (7) 18 12 1.68 0.81 1.01 0.11 140:10:0:0:0 93:7:0:0:0
McGrath Series-5 (n 150)

2.52 1.00 32:30:65:23 22:20:44:15 32 (52) 30 (48) 62 (41) 65 (74) 23 (26) 88 (59) 38 23 1.77 0.83 1.01 0.08 62:83:5:0:0 42:55:3:0:0
* Statistically signicant difference over McGrath (P 0.05). Statistically signicant difference over other tested videolaryngoscopes (VLSs) (P 0.01).
three VLSs. There are less attempts necessary with the Storz VLS than with both the GlideScope and the McGrath VLS (P 0.01) and less attempts necessary with the GlideScope than with the McGrath VLS (P 0.05). The gray shading indicates attempts that were done with the addition of a stylet (i.e., more than two attempts).
Figure 2. Comparison of necessary intubation attempts using
Figure 3. Comparison of necessary intubation time for three VLSs. There is less time required to intubate with the Storz VLS than with both the GlideScope and the McGrath VLS (P 0.01), and less time necessary with the GlideScope than with the McGrath VLS (P 0.05). considering the total intubation attempts, both with and without stylet. Intubation with the Storz VLS was faster than the other two VLSs tested (KruskalWallis, 2 116, P 0.01) (Fig. 3). Concerning the second hypothesis, there were no differences using a stylet among the three VLSs with regard to the number of required intubation attempts (P 0.05). A successful intubation was achieved on the first pass with the stylet in 49 patients (76%) who used the GlideScope, eight patients (80%) who used the Storz, and 65 patients (74%) who used the McGrath. However, because we did not differentiate the time before and after using a stylet, we cannot draw any conclusions regarding differences in intubation time. Subjectively, there did not seem to be any differences concerning the intubation time among the three VLSs
intubation without a stylet. Intubation was successful on the first attempt in 46 patients (53%) with the GlideScope, 118 patients (84%) with the Storz, and 32 patients (52%) with the McGrath VLS, all without using a stylet (Table 2). A stylet had to be used to successfully intubate almost half the patients in the GlideScope group (n 64, 43%) and the McGrath group (n 88, 59%); however, it was less for the Storz group (n 10, 7%), (P 0.01) (Fig. 2). The Storz group required fewer attempts to secure the airway (KruskalWallis, 2 126, P 0.01), (Table 2) with and without stylet. The average intubation times were 34 20 s for the GlideScope, 18 12 s for the Storz, and 38 23 s for the McGrath VLS (Table 2), again
828
Figure 4. Comparison of Cormack-Lehane (C&L) grades for the pooled VLSs versus the direct Macintosh blade. There were significantly better C&L grades for each of the indirect VLS compared with the direct view (P 0.01). when using a stylet. All VLSs performed well in terms of effective intubation time necessary for securing the airway because it was not necessary to revert to standard difficult airway protocols. No dropouts were encountered, as the tracheas of all patients could be intubated using the VLSs. All the VLSs tested offered an equal or better view of the glottis as assessed by the mean C&L grade compared with traditional direct intubation techniques (P 0.01) (Fig. 4). Only very few patients had C&L grades other than Grade I with the VLS, and none of the patients included in this study had a C&L grade poorer than II. The VLSs tested were rated favorably for their larger viewing angle of the glottic entrance compared with classic laryngoscopy techniques. This proved useful for guidance of the ETT and nasogastric tube into position and avoidance of contact with soft tissues of the mouth and pharynx. In general, visualization of the glottis entrance with the VLS was not a problem, although intubation was not always straightforward. In the context of this study, overall satisfaction was greater using the Storz VLS compared with the other two VLSs (KruskalWallis, 2 95, P 0.01).
DISCUSSION
This study compares three VLS devices in a clinical setting during endotracheal intubation and confirms that a stylet is not necessary with some VLSs (Storz) but highly recommended when other VLSs (e.g., GlideScope and McGrath) are used. When using a stylet, the VLSs studied did not differ in intubation time or number of intubation attempts. This study also confirms the excellent unobstructed view of the glottic opening obtained indirectly with a VLS as opposed to the direct classic Macintosh laryngoscope. No conversion to direct laryngoscopy was necessary in any patients studied. An equal or better C&L grade could be obtained in all cases, which is in agreement with other studies.19,20 There were no significant differences in the visualization quality of the
glottis among the three VLSs except for the dimension of the device monitor. Good visualization of the glottic entrance is paramount for successful tracheal intubation. However, providing a good view of the glottis does not always correlate with successful intubation. Indeed, this study shows that devices that offer the same unhindered view of the glottis are not alike in their ease of use. This study demonstrates that stylet use is not always required with certain VLSs. The Storz VLS group was associated with a shorter intubation time, required significantly fewer attempts to secure the airway compared with the other two groups, and a stylet was required only in a minority (7%) of the patients, whereas a styletted ETT had to be used in almost half of the patients in the GlideScope and McGrath VLS groups. Presumably, this is due to the fact that the Storz VLS uses the same Macintosh laryngoscope blade as with direct laryngoscopy, providing a better view and better access, which decreases the need for stylet use. During direct laryngoscopy, a stylet is rarely used routinely at first attempt in our clinical practice. Given the fact that styletted ETT have rare but potentially significant complications, we believe it is important to reserve their use for difficult intubations.3,1116 Using a styletted ETT with the GlideScope VLS, the first pass success rates were higher in the studies performed by Sun et al.21 (94%) and Xue et al.22 (97%) compared with our study. However, the other studies used a different definition of a single pass in which a successive attempt was only recorded on retracting the ETT completely out of the mouth; in our study, each approach to the vocal cords was counted as an attempt. Similarly, Shippey et al.10 also found a first pass success rate of 93% when using the McGrath VLS and a styletted ETT. In our study, successful intubation (without using a stylet) in the Storz group was 93%, supporting our contention that a stylet is not needed at all times and, thus, preventing potential complications. Successful intubation with the Storz VLS (without stylet) is as good as with the reported success rate in the literature using the GlideScope/ McGrath VLS (with stylet).10,20 22 The direct laryngoscope and the indirect Storz VLS are generally inserted on the right side of the tongue, which is compressed and deflected laterally, whereas the indirect GlideScope and McGrath VLS are inserted in the midline and advanced over the tongue because there is no need to sweep the tongue laterally. It might be that the greater success rate using the Storz VLS without a stylet depends on the angle of its blade, which is similar to the conventional direct Macintosh laryngoscope and the fact that the tongue is displaced laterally. Presumably, use of a styletted ETT from the outset would mitigate the differences among the VLSs that we found; however, this may also increase the risk of injury. The addition of a stylet essentially
829
compensates for the geometrical shortcomings of some of the VLS designs. Since its commercial introduction in 2002, numerous studies have reported the efficacy and safety of the GlideScope VLS for tracheal intubation in patients (and simulators) with easy and difficult airways.1,5,20,23,24 However, some authors have noted that the GlideScope VLS is difficult to insert into the patients mouth, does not reach deep enough in some cases, insertion of ETT is not easy, the advancement of the ETT after removal of the stylet is difficult,22 and complications because of the use of a styletted ETT and a VLS may occur.3,1116 The arytenoid cartilages, the interarytenoid soft tissues, anterior commissure of the glottis, or the anterior wall of the cricoid cartilage sometimes interfere with advancement of the ETT into the trachea. Manipulation of the ETT orientation is often not sufficient because the curvature of the distal end of the ETT is insufficient; in such cases, an extra tool is necessary. Additionally, patient characteristics, such as dentition and mouth opening, may greatly influence the ease of insertion of the ETT. The VLS essentially positions the operators eye proximal to the larynx. Therefore, care should be taken to do the initial introduction (passing the teeth and first part of mouth) of the ETT directly, until the distal end comes into view of the VLS. Indeed, the VLS is essentially a standard laryngoscope in form and function until the critical insertion of the ETT through the vocal cords is performed. A number of techniques can be used during the intubation procedure to improve the success rate. Previously, Xue et al.22 noted that the use of a malleable stylet, preheating of the blade to body temperature, and avoiding the use of superfluous lubricant were important considerations for successful use of the GlideScope. Also, increased lifting force, withdrawal and reinsertion of the blade, and external laryngeal pressure have been proposed as helpful measures for successfully securing the airway.23 Several maneuvers may overcome the problem of inserting the ETT: relaxing the VLS; withdrawing the VLS 12 cm; use of a StyletScopeTM (Nihon Kohden Co., Tokyo, Japan) in which the operator can adjust the angle of the ETT tip between 30 and 90 by gripping the handle strongly;25 or the use of a stylet-ETT that can increase the angle between the axis of the ETT tip and the tracheal axis. We neither experience any improvement in the ease of intubation when using external laryngeal pressure nor withdrawal and reinserting the blade. In this study, it was noticed that the use of a stylet with a relatively pronounced curve (the best angle is reported to be 90)17,18 at the distal end was most helpful in advancing the tip of the ETT to the glottic opening. Further study is required to determine optimal geometrical forms for the stylet or gum-elastic bougies used for difficult intubations. More importantly, the integration of the ETT with the VLS blade is the major
830
issue for redesign in future generations of VLS, especially considering that the classical problem of visualization now seems to be resolved (all patients in our study showed a C&L I or II with all three types of VLSs). The McGrath VLS that uses a disposable blade and the recent introduction of the GlideScope Cobalt single-use disposable blades26 are promising developments, especially in busy settings in which there may not be sufficient time to sterilize the blades between uses. Portability of the VLS systems is also an issue, and there are clear advantages of the McGrath and GlideScope VLS over the Storz. The integration of an antifogging mechanism on the McGrath and the GlideScope VLS is advantageous over the Storz V-Mac VLS, which lacks this feature. Preheating the VLS with the former two is unnecessary because the light emitting diode heats a window over the video chip. If fogging does occur, it likely means that the VLS is defective. However, it is still possible to blur the view if the lubricated ETT makes contact with the imaging system. This study has several limitations: 1) The attending anesthesiologist was not blinded to the type of VLS used, which this may have introduced bias, despite being blinded to the preoperative metrics and initial C&L grade with the use of the classic Macintosh laryngoscope; 2) There are more VLSs available on the market so this review is not complete, but the three most common models available in our hospital are included; 3) There was very low patient morbidity in this study, and it remains debatable how important the metrics of intubation time, attempts, and satisfaction are with regard to patient morbidity; 4) It is clear that if the study were performed using a stylet routinely in all cases, then the second or third intubation attempt would not have been necessary; 5) The selection of patients lacking features associated with a difficult airway may have reduced the potential superiority of VLS over direct laryngoscopy and minimized the differences among the three VLS models; 6) Failure to routinely use a stylet may bias our study in favor of a device which most closely resembles a conventional Macintosh laryngoscope; 7) A study of patients with difficult airway anatomy may be needed to determine the need for the routine use of a stylet; and 8) Finally, this study deals with a specific population of elective surgical patients with normal airways and no conclusions can be made for patients in whom difficult tracheal intubation is expected.
CONCLUSIONS
The use of a styletted ETT is not ideal during tracheal intubation because it can potentially contribute to complications. Our study confirms that a large proportion of patients with normal airway anatomy can be intubated successfully with certain VLS blades without using a stylet, although there is a large difference among the types of VLSs tested. Certainly,
the problem of visualization of the glottic arch is resolved by a VLS, but this does not guarantee easy or successful endotracheal intubation. The stylet essentially compensates for the geometrical mismatch between the VLS blade and the laryngeal anatomy of the patient. The Storz VLS performed better in overall satisfaction, intubation time, and number of attempts, including attempted intubations without a stylet, most probably due to the better view and access, which limited the need for stylet use. The GlideScope and McGrath VLS are equally successful in achieving good visualization and intubation in all patients. It seems that geometry and integration of ETT with the VLS is the next question that needs to be addressed in blade design for intubation. REFERENCES
1. Cooper RM. Use of a new videolaryngoscope (GlideScope) in the management of a difficult airway. Can J Anaesth 2003; 50:6113 2. Van Zundert A, Stessel B, De Ruiter F, Giebelen D, Weber E. Video-assisted laryngoscopy: a useful adjunct in endotracheal intubation. Acta Anaesthesiol Belg 2007;58:129 31 3. Cooper RM. Complications associated with the use of the GlideScope videolaryngoscope. Can J Anaesth 2007;54:54 7 4. Kaplan MB, Ward D, Hagberg CA, Berci G, Hagiike M. Seeing is believing: the importance of video laryngoscopy in teaching and in managing the difficult airway. Surg Endosc 2006;20(suppl 2):S479 S483 5. Cooper RM, Pacey JA, Bishop MJ, McCluskey SA. Early clinical experience with a new videolaryngoscope (GlideScope) in 728 patients. Can J Anaesth 2005;52:191 8 6. Kramer DC, Osborn IP. More maneuvers to facilitate tracheal intubation with the GlideScope. Can J Anaesth 2006;53:737 7. Muallem M, Baraka A. Tracheal intubation using the GlideScope with a combined curved pipe stylet, and endotracheal tube introducer. Can J Anaesth 2007;54:77 8 8. Cuchillo JV, Rodriguez MA. Considerations aimed at facilitating the use of the new GlideScope videolaryngoscope. Can J Anaesth 2005;52:661; author reply 6612 9. Shippey B, Ray D, McKeown D. Case series: the McGrath videolaryngoscopean initial clinical evaluation. Can J Anaesth 2007;54:30713 10. Shippey B, Ray D, McKeown D. Use of the McGrath videolaryngoscope in the management of difficult and failed tracheal intubation. Br J Anaesth 2008;100:116 9
11. Vincent RD Jr, Wimberly MP, Brockwell RC, Magnuson JS. Soft palate perforation during orotracheal intubation facilitated by the GlideScope videolaryngoscope. J Clin Anesth 2007;19:619 21 12. Choo MK, Yeo VS, See JJ. Another complication associated with videolaryngoscopy. Can J Anaesth 2007;54:322 4 13. Hirabayashi Y. Pharyngeal injury related to GlideScope videolaryngoscope. Otolaryngol Head Neck Surg 2007;137:175 6 14. Hsu WT, Hsu SC, Lee YL, Huang JS, Chen CL. Penetrating injury of the soft palate during GlideScope intubation. Anesth Analg 2007;104:1609 11 15. Malik AM, Frogel JK. Anterior tonsillar pillar perforation during GlideScope video laryngoscopy. Anesth Analg 2007; 104:1610 1 16. Hsu WT, Tsao SL, Chen KY, Chou WK. Penetrating injury of the palatoglossal arch associated with use of the GlideScope videolaryngoscope in a flame burn patient. Acta Anaesthesiol Taiwan 2008;46:39 41 17. Jones PM, Turkstra TP, Armstrong KP, Armstrong PM, Cherry RA, Hoogstra J, Harle CC. Effect of stylet angulation and endotracheal tube camber on time to intubation with the GlideScope. Can J Anaesth 2007;54:217 18. Turkstra TP, Harle CC, Armstrong KP, Armstrong PM, Cherry RA, Hoogstra J, Jones PM. The GlideScope-specific rigid stylet and standard malleable stylet are equally effective for GlideScope use. Can J Anaesth 2007;54:891 6 19. Kaplan MB, Hagberg CA, Ward DS, Brambrink A, Chhibber AK, Heidegger T, Lozada L, Ovassapian A, Parsons D, Ramsay J, Wilhelm W, Zwissler B, Gerig HJ, Hofstetter C, Karan S, Kreisler N, Pousman RM, Thierbach A, Wrobel M, Berci G. Comparison of direct and video-assisted views of the larynx during routine intubation. J Clin Anesth 2006;18:357 62 20. Xue F, Zhang G, Liu J, Li X, Sun H, Wang X, Li C, Liu K, Xu Y, Liu Y. A clinical assessment of the Glidescope videolaryngoscope in nasotracheal intubation with general anesthesia. J Clin Anesth 2006;18:6115 21. Sun DA, Warriner CB, Parsons DG, Klein R, Umedaly HS, Moult M. The GlideScope video laryngoscope: randomized clinical trial in 200 patients. Br J Anaesth 2005;94:381 4 22. Xue FS, Zhang GH, Liu J, Li XY, Yang QY, Xu YC, Li CW. The clinical assessment of Glidescope in orotracheal intubation under general anesthesia. Minerva Anestesiol 2007;73:4517 23. Fun WL, Lim Y, Teoh WH. Comparison of the GlideScope video laryngoscope vs. the intubating laryngeal mask for females with normal airways. Eur J Anaesthesiol 2007;24:486 91 24. Lai HY, Chen IH, Chen A, Hwang FY, Lee Y. The use of the GlideScope for tracheal intubation in patients with ankylosing spondylitis. Br J Anaesth 2006;97:419 22 25. Hirabayashi Y. The StyletScope(R) facilitates tracheal intubation with the GlideScope. Can J Anaesth 2006;53:1263 4 26. Jones PM, Harle CC, Turkstra TP. The GlideScope cobalt videolaryngoscopea novel single-use device. Can J Anaesth 2007;54:677 8
831
Nasogastric Tube Insertion Using Different Techniques in Anesthetized Patients: A Prospective, Randomized Study
Jithesh Appukutty, MD* Prerana P. Shroff, MD
BACKGROUND: It is often difficult to correctly place nasogastric (NG) tubes under anesthesia. We hypothesized that simple modifications in technique of NG tube insertion will improve the success rate. METHODS: Two hundred patients were enrolled into the study. The patients were randomized into four groups: control, guidewire, slit endotracheal tube, and neck flexion with lateral neck pressure. The starting point of the procedure was the time when NG tube insertion was begun through the selected nostril. The end point was the time when there was either a successful insertion of the NG tube or a failure after two attempts. The success rate of the technique, duration of insertion procedure, and the occurrence of complications (bleeding, coiling, kinking, and knotting, etc.) were noted. 2, analysis of variance, and Students t-test were used to analyze the data. RESULTS: Success rates were higher in all intervention groups compared with the control group. The time necessary to insert the NG tube was significantly longer in the slit endotracheal tube group. Kinking of the NG tube and bleeding were the most common complications. CONCLUSION: The success rate of NG tube insertion can be increased by using a ureteral guidewire as stylet, a slit endotracheal tube as an introducer, or head flexion with lateral neck pressure. Head flexion with lateral neck pressure is the easiest technique that has a high success rate and fewest complications.
(Anesth Analg 2009;109:8325)
he insertion of a nasogastric (NG) tube in anesthetized, paralyzed, and intubated or unconscious patients may be difficult, with reported failure rates of nearly 50% on the first attempt with the head in neutral position.13 After a failure, subsequent attempts are usually unsuccessful due to coiling, kinking, or knotting of the NG tube as it loses stiffness due to warming to body temperature. The memory effect also contributes to subsequent failures; once kinked, the NG tube is subsequently more likely to kink at the same place. The most common sites of impaction of the NG tube are piriform sinuses and the arytenoid cartilage.4 Maneuvers to keep the NG tube along the lateral or posterior pharyngeal wall during insertion encourages the smooth passage into the esophagus.1,2,5 Common methods used to facilitate NG tube insertion include the use of a slit endotracheal tube, forward displacement of the larynx and the use of various forceps, the use of an ureteral guidewire as a
From the *Department of Anesthesiology, KJ Somaiya Medical College and Research Centre, Sion; and Department of Anaesthesiology, Seth GSMC and KEM Hospital, Parel, Mumbai, Maharashtra, India. Accepted for publication April 27, 2009. Address correspondence and reprint requests to Jithesh Appukutty, MD, 3A/501, Hema Park, Veer Savarkar Marg, Bhandup(E), Mumbai 400042, Maharashtra, India. Address e-mail to jithesh_ak1@rediffmail.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af5e1f
stylet, head flexion, lateral neck pressure, and the use of a gloved finger to steer the NG tube after impaction.2,6 8 Neck flexion, in combination with the curve of the NG tube, tends to keep the tube in close proximity to the posterior pharyngeal wall, facilitating its smooth passage into the esophagus.2 The ureteral guidewire imparts stiffness to the NG tube by acting as a stylet and preventing kinking. We hypothesized that slight modifications in NG tube insertion technique would improve the rate of successful insertion. We compared three techniques to the common method of NG insertion to determine the success rate, average time for insertion, and incidence of complications, such as bleeding, coiling, knotting, and kinking.
METHODS
Hospital Ethical committee approval was obtained, and a valid written informed consent was obtained from each patient. Patients younger than 20 yr and older than 70 yr were excluded from the study. Two hundred patients were enrolled in the study. All patients received general anesthesia and tracheal intubation for various surgical procedures that required NG tube insertion. After induction of general anesthesia and tracheal intubation, the patients were randomly allocated into four groups according to a computer-generated randomization order. In the control group (Group C), patients had a lubricated NG tube inserted gently
832
through the selected nostril, the head being maintained in the neutral position. The guidewire group (Group W) made use of a ureteral guidewire that was introduced within a 14-F NG tube until the tip of the guidewire was at the tip of the NG tube. Tube insertion was then performed in the same manner as described for the control group. In the slit tracheal tube (TT) group (Group S), the NG tube was inserted through the selected nostril and taken out through the mouth, leaving at least 10 cm of NG tube at the nostril. It was then passed through a longitudinally cut 7.0-mm internal diameter polyvinyl chloride TT, so that the tip of the NG tube was at the level of the Murphy eye of the TT. The TT was lubricated generously and was then inserted blindly into the oral cavity to a depth of 18 cm and the NG tube advanced further. The NG tube was then freed from the cut TT, and the cut TT was removed and the rest of the NG tube passed into the esophagus and was then fixed at the required length by pulling out through the nostril. In the neck flexion with lateral pressure group (Group F), a lubricated NG tube was inserted through the selected nostril to a depth of 10 cm. The patients neck was flexed, lateral neck pressure was applied, and the NG tube was advanced in a similar manner to that described for Group C. Preoperatively, the nostril to be used for NG tube insertion was chosen based on two criteria: the amount of fogging produced on a metal tongue depressor during exhalation and the relative size of the nostril. In all patient groups, a 14F, 105-cm NG tube with lead markings at the distal end was used. NG tube insertion was performed by a group of four third-year anesthesia residents (to avoid operator bias the authors did not perform NG tube insertions). These residents were all judged to be proficient in the techniques described. They were assigned patients according to a computer-generated randomization schedule. The procedure start time was defined when the NG tube insertion was begun through the selected nostril. The procedure end time was defined as the time of successful insertion of NG tube or the time after two failed attempts. The procedure duration was measured with a stopwatch. Successful NG tube insertion was confirmed when the tube passed smoothly and a gurgling sound was heard on auscultation over the epigastrium when injecting 10 cc of air through the NG tube. If the first attempt failed, the NG tube was withdrawn fully and was cleaned. Lubricating jelly was applied generously, and the procedure was repeated using the same technique. If both attempts at insertion using the selected technique were unsuccessful, then the technique was considered a failure. The NG tube was then inserted with the help of Magill forceps during a direct laryngoscopy. The following data were collected: 1. Success rate of the selected techniquefirst, second attempt, and overall. 2. Number of attempts for successful insertion.
3. Duration of insertion using the selected technique. 4. Complications during insertion kinking, knotting, and bleeding. An unpublished pilot study of 12 cases per group suggested an approximate 20% improvement (from base rate of 65% to 85%) in success rate using these techniques. Consequently, a power calculation ( 0.05 and 0.2) indicated a minimum of 46 patients for each group using an analysis of variance (ANOVA) test. Continuous data are presented as mean sd; categorical data are presented as frequency and percentage. Demographic data were analyzed by Pearsons 2 test. The time necessary to insert the NG tube in each group was compared using ANOVA test. The complication rates during insertion of NG tubes in all four groups were compared using ANOVA for multiple variables. A value of P 0.05 was considered statistically significant.
RESULTS
There were no statistically significant differences in the demographic data (age and gender) of the four patient groups. In Group C, successful NG tube insertion was achieved in 36 patients (72%) (Fig. 1) . The success rates of NG insertion were greater in Groups W, S, and F: 46 patients (92%, P 0.011), 46 patients (92%, P 0.011), and 47 patients (92%, P 0.004), respectively. In Group C, 17 patients (34%) had a NG tube placed successfully on the first attempt and 19 patients (38%) on the second attempt (Fig. 1). In Group W, 33 patients (66%) had a NG tube placed successfully on the first attempt and 13 patients (26%) on the second attempt (P 0.002 compared with Group C). In Group S, 41 patients (82%) had a NG tube placed successfully on the first attempt and five patients (10%) on the second attempt (P 0.0006 compared with Group C). In Group F, 41 patients (82%) had a NG tube placed successfully on the first attempt and six patients (12%) on the second attempt (P 0.0006 compared with Group C). Total NG tube insertion time was 56 36 s in Group C. This time was significantly longer in Group S (98 43 s) and significantly shorter in Group F (31 19 s). Group W time (42 29 s) was not statistically different from Group C. The most common complication in Group C was kinking of the NG tube, which occurred in 10 patients (20%); knotting occurred in one patient (2%) (Table 1). In Group W, the NG tube became kinked in four patients (8%) (P NS), and knotting occurred in one patient (2%) (P NS versus Group C). In Group S, 11 patients (22%) developed bleeding during NG tube insertion, significantly more frequently than in Group C; in one patient, the NG tube could not be freed from the slit TT, and the whole assembly had to be removed (complication classified as other). In Group F, four patients (8%) developed kinks during insertion of the NG tube (P NS versus Group C).
833
Figure 1. Successful nasogastric tube insertion. Group C control; Group W guidewire; Group S slit tracheal tube; and Group F neck flexion with lateral pressure.
Table 1. Duration of Nasogastric Tube Insertion (s) and Complications Group C (n 50)
Duration of insertion (s) Complication (number of cases) Kinking Knotting Bleeding Others 56 36 10 1 0 0
Group W (n 50)
42 29* 4 1 0 0
Group S (n 50)
98 43 0 0 11 1
Group F (n 50)
31 19 4 0 0 0
Group C control; Group W guidewire; Group S slit tracheal tube; Group F neck exion with lateral pressure. * P 0.166. P 0.0003. P 0.001. P 0.0005 versus control. Signicant at P 0.05.
DISCUSSION
Insertion of the NG tubes in anesthetized and intubated patients has an average failure rate of nearly 50% on the first attempt with the patients head in neutral position.2 The piriform sinuses and arytenoid cartilages are the most common sites of impaction.4 Maneuvers to avoid impaction on these structures include insertion of the NG tube along the posterior or lateral pharyngeal wall, by head flexion and lateral neck pressure, or by turning the head to one side.1,5 Other methods to facilitate NG tube insertion include the use of an ureteral guidewire or cooling the NG tube to stiffen it,9 the use of a slit-TT as a conduit,7 the use of a guitar wire as a stylet,10 endoscopic placement, or the use of various endoscopic forceps and lifting the thyroid cartilage. We observed a success rate of 34% in Group C on the first attempt, which was significantly lower compared with the success rates of the ureteral guidewire (66%), slit endotracheal tube (82%), and head flexion with lateral neck pressure (82%) groups, confirming that the latter procedures increase the success rate. A
834
Techniques for Nasogastric Tube Insertion
ureteral guidewire helps to reduce the flexibility of the NG tube, whereas a slit TT, which is resistant to kinking, directs the NG tube into the esophagus. Head flexion and lateral neck pressure help keep the NG tube along the lateral and posterior pharyngeal wall, thereby facilitating passage into the esophagus. Ratzlaff et al.11 found that the degree of NG tube flexibility significantly affected the ease with which the NG tube was inserted and also reported that the rigid tubes required fewer insertion attempts. However, as the NG tube rigidity increases, the incidence of trauma also increases, with a subsequent increase in the incidence of bleeding.10,11 We used an ureteral guidewire (6F) to decrease the flexibility of the NG tube and found that insertion was successful in 92% of patients compared with a 72% success rate in Group C. In Groups C and W, the time required for insertion was 56 36 s and 42 29 s, respectively; Group S had a longer insertion time. In Group F, the insertion time (31 19 s) was significantly shorter than Group Cs insertion time. Among the four groups, the Group F
had the shortest time to insertion whereas the Group S had the longest time. Matsuki and Zsigmond10 used guitar strings to facilitate NG tube insertion but reported a few cases of bleeding. In our study of 200 patients, 32 (16%) developed complications. The most common complications were kinking of the NG tube, knotting of the NG tube, and bleeding. We observed that of the 32 complications, 18 (56%) were due to kinking, which underscored the importance of reducing flexibility to improve the success of NG tube insertion. Decreased flexibility can be accomplished by using the ureteral guidewire as a stylet. However, bleeding was a frequent complication in the slit TT group: 11 of 50 (22%) patients experienced bleeding. The frequent incidence could be attributed to a technique that involves insertion of an additional TT into the oral cavity in an already intubated patient. This complication was evident in the patient in whom the slit TT could not be withdrawn while keeping the NG tube in place, such that the entire assembly had to be removed with great difficulty.
the slit TT was used. Overall, considering the success rate, the duration of insertion, and the complication rate, we conclude that head flexion with lateral neck pressure is the simplest technique that has the highest success rate and lowest incidence of complications. Therefore, we recommend the use of either a head flexion with lateral neck pressure or an ureteral guidewire as a stylet in all NG tube insertions. REFERENCES
1. Bong CL, Macachor JD, Hwang NC. Insertion of the nasogastric tube made easy. Anesthesiology 2004;101:266 2. Mahajan R, Gupta R. Another method to assist nasogastric tube insertion. Can J Anaesth 2005;52:6523 3. Kayo R, Kajita I, Cho S, Murakami T, Saito H. A study on insertion of a nasogastric tube in intubated patients. Masui 2005;54:1034 6 4. Parris WC. Reverse Sellick maneuver. Anesth Analg 1989;68:423 5. Ozer S, Benumof JL. Oro- and nasogastric tube passage in intubated patients: fiberoptic description of where they go at the laryngeal level and how to make them enter the esophagus. Anesthesiology. 1999;91:137 43 6. Flegar M, Ball A. Easier nasogastric tube insertion. Anaesthesia 2004;59:197 7. Sprague DH, Carter SR. An alternative method for nasogastric tube insertion. Anesthesiology 1980;53:436 8. Campbell B. A novel method of nasogastric tube insertion. Anaesthesia 1997;52:1234 9. Mahajan R, Poddar S, Grover VK. A simple and reliable method for nasogastric tube insertion. J Anaesth Clin Pharmacol 2004;20:95 6 10. Matsuki A, Zsigmond EK. Simple and reliable method of inserting a nasogastric tube during anaesthesia. Br J Anaesth 1972;44:610 11. Ratzlaff HC, Heaslip JE, Rothwell ES. Factors affecting nasogastric tube insertion. Crit Care Med 1984;12:523
CONCLUSION
The success rate of NG tube insertion can be increased by using an ureteral guidewire as a stylet, a slit TT as an introducer, or keeping the head flexed while applying lateral neck pressure. The time needed to insert a NG tube was shortest using head flexion with lateral pressure and longest with the use of a slit TT. Kinking was the most frequent complication encountered, and bleeding was the most common when
835
Case Report
Where Are My Teeth? A Case of Unnoticed Ingestion of a Dislodged Fixed Partial Denture
Gary Lau, MD* Vivek Kulkarni, MD, PhD* Gary K. Roberts, DDS John Brock-Utne, MD, PhD*
What are the dangers of swallowing foreign bodies of dental origin? How do we recognize when a patient has actually swallowed a dental appliance? How far should we pursue the retrieval of the appliance? We report a case of a patient with unnoticed ingestion of a dislodged fixed partial denture while undergoing general anesthesia and review the literature on dangers of swallowing foreign bodies of dental origin. Anesthesiologists should understand the dangers and recognize this complication when it happens, so that appropriate treatment can be pursued if necessary.
(Anesth Analg 2009;109:836 8)
ngestion of items of odontogenic origin, including dislodged teeth, crowns, fixed partial dentures (bridges), removable partial and complete dentures, as well as various other dental and orthodontic appliances by patients under general anesthesia is rare but often initially goes unnoticed, leading to potentially dangerous late complications that require invasive surgical interventions. Certain patient populations are at increased risk for unnoticed ingestion and some are at increased risk of developing perforations from the ingested bodies. The shape and dimension of the ingested odontogenic item can affect whether it will pass through the gut without incidents. Perforations occur more often in certain portions of the gastrointestinal tract (GI) than others. We report a case of a patient with unnoticed ingestion of dislodged fixed partial denture while undergoing general anesthesia and review the literature on dangers of swallowing foreign bodies of dental origin.
At the end of surgery, the patients trachea was extubated without problems and he was transported to the postoperative care unit, where he experienced a bout of coughing. Upon becoming more awake and alert, his first question was, Where are my teeth? On examination, patient was noted to have missing front incisors where his fixed partial denture had been. On questioning, the patient stated that his bridge had never come off before and did not think he had swallowed it. The operating room was thoroughly searched but his missing fixed partial denture was not located. An abdominal radiographic film revealed a radio-opaque foreign body consistent in shape with the missing fixed partial denture overlying the stomach (Fig. 1). A gastroenterologist was consulted and the patient underwent emergent upper endoscopy. However, the bridge was not visualized endoscopically. Repeat abdominal film revealed that the bridge had passed into the small intestine (Fig. 2). The patient was observed for a period of time. He neither complained of any abdominal pain, nausea nor exhibited fever or hematemesis. He was discharged home with close follow-up. On postoperative Day 3, the bridge passed into the stool without problems.
CASE DESCRIPTION
A 36-yr-old man underwent left ureteroscopic laser lithotripsy for nephrolithiasis. His surgical history was significant for a hernia repair in 2005, which was uneventful with no recorded difficulties with tracheal intubation. Examination of his airway revealed a maxillary anterior fixed partial denture that according to the patient was permanent. He stated that he had had the bridge for more than 17 yr and that it had never caused any problems. After induction of general anesthesia, the patient was easily mask ventilated. A Grade I view of the larynx was achieved with atraumatic direct laryngoscopy and a tracheal tube was placed easily. The rest of the anesthesia and surgery were uneventful.
From the Departments of *Anesthesia, and OtolaryngologyHead and Neck Surgery, Stanford University Hospital and Clinics, Stanford, California. Accepted for publication April 9, 2009. Address correspondence and reprint requests to John BrockUtne, MD, PhD, Department of Anesthesia, Stanford University School of Medicine, 300 Pasteur Dr., H3580, Stanford, CA 943055640. Address e-mail to brockutn@stanford.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae06c9
DISCUSSION
In the general population, ingestion of foreign bodies is not uncommon, especially in children, alcoholics, edentulous people, and mentally handicapped people. It has been reported that about 1500 people die annually from the ingestion of foreign bodies in the United States.1 However, the incidence of anesthetized patients ingesting dental appliances is unknown. Fortunately, the majority of foreign bodies entering the oropharynx eventually pass through the GI tract without complications. However, there is a potential risk of gut perforation, which can have serious consequences including death.2 4 Other reported late complications of unnoticed foreign body ingestion include sepsis, peritonitis, retropharyngeal and intraabdominal abscesses, esophageal impaction and stricture, ulcerative esophagitis, tracheoesophageal and enterocolonic fistulas, recurrent pneumonitis, and massive hemorrhage.35 Large objects, especially those with sharp edges, can get impacted, usually at the level of
836
Figure 1. Dental bridge seen in the stomach.
Figure 2. Dental bridge seen in the small bowel. the fourth cervical vertebra.6 Symptoms of esophageal obstruction can be nonspecific and resemble those of a mildly traumatic direct laryngoscopy, which makes it difficult for the anesthesiologists to diagnose. The symptoms include dysphagia, odynophagia, coughing, choking sensation, hematemesis, or vomiting.
Radiographic evidence that includes both anteriorposterior and lateral films is warranted. However, negative radiological findings do not exclude the possibility of a foreign body in the esophagus. Persistence of symptoms, even in the absence of positive clinical or radiological signs, warrants an endoscopic examination. Once a foreign body has reached the stomach, it has an 80%90% chance of passing along the gut spontaneously without problems, but bodies thicker than 2 cm and longer than 5 cm will not likely leave the stomach spontaneously.7,8 About 10%20% will require removal from the GI tract by endoscopy.3,8 Less than 1% of all foreign bodies will cause a perforation.6,9 However, all sharp and pointed objects should be removed before they pass from the stomach because 15%35% of this type of foreign body will cause intestinal perforation.10 Dull foreign bodies can also cause perforations by causing pressure necrosis and subsequent destruction of underlying mucosa and muscle.5 Patients with intrinsic bowel disease are at an increased risk of developing perforations from the ingested foreign body. This includes patients with bowel adhesions, inflammatory bowel disease, GI tumors, diverticulosis of large bowel, hernias, and Meckels diverticulum.4,5 The most common sites for perforation are the ileocecal junction and the sigmoid colon.5 The usual time taken for a foreign body to traverse the intestinal tract is 2 to 12 days.7 In our case, the patient did not exhibit any signs or symptoms typical of ingestion of a dental appliance other than the initial bout of coughing, which may have been related to the ingestion. This is a rare case report, in which the diagnosis of an odontogenic foreign body ingestion was made because the patient asked about the whereabouts of his own dental appliance after recovery from general anesthesia. It is fortunate that this happened in a patient who was cognitively coherent enough to express that his fixed partial denture was missing. It is easy to see how an incident like this could occur unnoticed in a mentally challenged patient and be left undiagnosed, potentially leading to further complications as described above without an obvious etiology. The uncommon presentation of this case, in the context of previous reports of ingested odontogenic items, underscores the need for a high index of suspicion in mentally challenged patients with dental restorations who are scheduled to undergo general anesthesia. Furthermore, one should keep in mind that so called permanent bridges may not be so permanent after all. It also should be noted that our review focuses on ingestion rather than pulmonary aspiration of items of odontogenic origin. Aspiration of items of odontogenic origin presents a more critical situation that has been reviewed elsewhere in the literature. In conclusion, the unnoticed swallowing of items of odontogenic origin (Fig. 3), though infrequent, can potentially be dangerous and lead to GI tract perforations. If it is suspected that an anesthetized patient has
837
Figure 3. Examples of dental appliances.
swallowed a foreign body, the appropriate medical specialist should be consulted, along with a dentist trained in the care of hospitalized and anesthetized patients, as it may be necessary to identify and remove an object with sharp edges to avoid organ perforation. Early diagnosis and treatment can avoid late complications that may require surgical intervention. Attention must be paid to patients at increased risk of unnoticed foreign body ingestion. This includes young children and patients with comorbidities that limit cognition and communication, such as stroke, dementia, Parkinsons disease, cerebral palsy, autism, and mental retardation. REFERENCES
1. DeVaneson J, Pisani A, Sharman P, Kazarian K, Mersheimer W. Metallic foreign bodies in the stomach. Arch Surg 1977;112:664 5
2. Ghori A, Dorricott NJ, Sanders DSA. A lethal ectopic denture: an unusual case of sigmoid perforation due to unnoticed swallowed dental plate. J R Coll Surg Edin 1999;44:203 6 3. Schwartz G, Polsky H. Ingested foreign bodies of the gastrointestinal tract. Am Surg 1976;42:236 8 4. Maleki M, Evans W. Foreign body perforations of the intestinal tract. Arch Surg 1970;101:4757 5. McCanse D, Kurchin A, Hinshaw J. Gastrointestinal foreign bodies. Am J Surg 1981;142:3357 6. Nandi P, Ong GB. Foreign body in the oesophagus: review of 2394 cases. Br J Surg 1978;65:59 7. Webb WA, McDaniel L, Jones L. Foreign bodies of the upper gastrointestinal tract: current management. South Med J 1978;77:1083 6 8. Perelman H. Toothpick perforation of the gastrointestinal tract. J Abdom Surg 1962;4:513 9. Gonzalez JG, Gonzalez RR, Patino JV, Garcia AT, Alvarez CP, Pedrosa CS. CT findings in gastrointestinal perforation by ingested fish bones. J Comput Assist Tomogr 1988;12: 88 90 10. Webb W. Management of foreign bodies of the upper gastrointestinal tract: update. Gastrointest Endosc 1995;41:39 51
838
Case Report
Critical Care and Trauma

Section Editor: Jukka Takala
The Practice of and Documentation on Withholding and Withdrawing Life Support: A Retrospective Study in Two Dutch Intensive Care Units
Peter E. Spronk, MD, PhD* Alexej V. Kuiper, MSc Johannes H. Rommes, MD, PhD* Joke C. Korevaar, MD Marcus J. Schultz, MD, PhD
OBJECTIVE: We determined how often life support was withheld or withdrawn in patients who died in the intensive care unit (ICU) or early after ICU discharge and evaluated documentation on decisions regarding these changes in life support orders. METHODS: This was a retrospective study in a university hospital and a general teaching hospital. Charts of patients who died during ICU stay or within 7 days after ICU discharge in 2005 were reviewed. RESULTS: Of 2578 admitted patients, 356 patients (14%) died either in the ICU or within 7 days after ICU discharge. For 9 patients data were missing, leaving 347 patients for analysis. Seventy-seven patients (22%) died with full life support, 85 (25%) died while treatment was being withheld, and 185 (53%) patients died while treatment was being withdrawn. One or more changes in life support orders were noted in 266 patients (77%). Only 8% of the patients were recorded to be incapacitated at the time of the change. Patients preferences regarding life support were documented in less than one-quarter of cases. In approximately one third of cases, it was not documented which member(s) of the ICU team were involved in an end-of-life decision. In the documented cases, end-of-life decisions were made along with the patient (7%) or with the patients representatives (59%). CONCLUSION: ICU nonsurvivors and patients who die shortly after ICU discharge predominantly die with orders to withhold or withdraw life support. Documentation on the decisions to forgo full life support is poor.
(Anesth Analg 2009;109:8416)
ntensive care unit (ICU) patients may die while treatment is being either withheld or withdrawn. Although the decision to limit or forgo further treatment may at times be made by ICU patients themselves, critically ill patients frequently are not able to make or (adequately) communicate such decisions because of sedation, cognitive dysfunction, or communication barriers like endotracheal tubes.1 With advanced care planning, patient preferences regarding discontinuation of therapy may be known by surrogate decision-makers (usually the patients spouse or life companion) and/or treating physicians. However, advanced directives/living wills frequently do not sufficiently state when and how to discontinue treatment in specific medical conditions.2 In the majority of
From the *Department of Intensive Care Medicine, Gelre Hospitals, Location Lukas, Apeldoorn; Department of Intensive Care Medicine, Academic Medical Center, University of Amsterdam; HERMES Critical Care Group; and Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands. Accepted for publication March 17, 2009. Address correspondence and reprint requests to Peter E. Spronk, MD, PhD, FCCP, Department of Intensive Care Medicine, Gelre Hospitals, Location Lukas, PO Box 9014, 7300 DS Apeldoorn, The Netherlands. Address e-mail to p.spronk@gelre.nl. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181acc64a
cases, therefore, the decision to continue (or discontinue) ICU treatment is left to the attending ICU physicians and/or other members of the ICU team.3 Documentation of medical decisions is crucial to preserve continuity inpatient care, particularly during off-hour periods. A recent study by the Dutch health inspection demonstrated that only a small percentage of necessary information was adequately recorded in the medical patient chart before surgical interventions (http://www.igz.nl/15451/475693/2007-02_Rapport_ Preoperatie1.pdf). In particular, the lack of standardization of which information should be recorded in the medical chart was prominent. Although this study did not address the issues pertaining to life support orders, in particular in the ICU setting, it is possible that documentation on this item may be worse too. In this study, we analyzed how often therapy is either withheld or withdrawn during the ICU stay in two Dutch teaching hospitals. For this analysis we collected all data of patients who died in the ICU. Because treatment and patient approach during the ICU stay could also have influenced life support orders in step-down units or normal wards, patients who died within 7 days after ICU discharge were also included in the analysis. We evaluated the preciseness of documentation on the decisions on life support orders.
841
METHODS
The clinical records of patients who were admitted to the ICU in 2005 and subsequently died during their ICU stay or within 7 days after ICU discharge were collected. As no interventions were part of the study, the need for obtaining informed consent was waived by the local ethics committee.
given or that life support was given according to the order at ICU admission. If a change in life support order was found, we searched for any documentation in the clinical records concerning the next five issues: 1. Which member(s) of the ICU team were involved in the decision to change orders on life support? 2. Were the patient and/or the patients representatives involved in the change in orders? 3. Was the change in life support preceded by psychiatric consultation? 4. Was the patient capable of making a decision regarding his medical treatment at the time the decision was made to change orders? 5. Was there knowledge of the patients preferences regarding life support?
Study Location
The study was performed in one academic hospital (Academic Medical Center (AMC), University of Amsterdam) and one university-affiliated teaching hospital (Gelre Hospital, Location Lukas). In the AMC, the ICU comprises a 28-bed closed unit in which medical/ surgical patients (including neurosurgery/neurology, cardiothoracic surgery, and cardiology patients) are under the direct care of the ICU team. The ICU team comprises 8 full-time ICU physicians, 8 subspecialty fellows, 12 residents, and occasionally 1 intern. In the Gelre Hospital, the ICU is a 10-bed closed unit with medical/surgical patients (as in AMC, with the exception of cardiothoracic surgery and neurosurgery patients). The ICU team comprises two full-time ICU physicians, five physicians who participate in evening and night shifts, and one resident.
Denitions
Full life support (order A): all life-saving interventions were to be performed when needed. Withhold life support (order B): a decision was made not to start one or more life-saving interventions, or only with certain restrictions. Interventions withheld in both ICU departments included defibrillation for ventricular fibrillation, cardiopulmonary resuscitation, treatment of arrhythmia, treatment with vasopressors and/or inotropics, tracheal intubation and mechanical ventilation, surgery, transfusion of blood products, antimicrobial treatment, or renal replacement therapy. Interventions withheld only in the ICU of Gelre Hospital included certain diagnostic procedures (which had to be documented on the form). Withdraw life support (order C): a decision was made to actively stop all life-supporting interventions that were already being performed. This meant that modalities aimed at survival were stopped, but other modalities aiming at patient comfort were continued.
Data Source
Patient data from the AMC were retrieved from the patient data management system (Metavision, iMDsoft, Sassenheim, The Netherlands). This database automatically stores patient-specific information during the complete ICU stay, including daily reports and special forms related to decisions on withholding or withdrawing life support. The reports from ICU physicians and ICU nurses, family consultations, and treatment orders were all read and documented in a separate database. In the Gelre Hospital, clinical records were only present in paper form; data similar to those mentioned above were collected. In addition, the database of the National Intensive Care Evaluation (The Netherlands, http://www.stichting-nice.nl) was used to collect baseline data of all patients.
The completed data were analyzed using SPSS version 12.0 (SPSS, Chicago, IL). Continuous data are expressed as medians with interquartile range and categorical data in percentages. Comparisons were made using MannWhitney and KruskalWallis tests for continuous data and the 2 test for categorical data. All tests were two-tailed, and differences with a P value 0.05 were considered significant. Data from the ICUs were analyzed separately but may be presented together.
Patient-Specic Data
The following data were collected for each patient: baseline data, such as gender, age, type of admission (acute medical, acute surgical, elective surgical), referring specialty (medical, surgical, neurosurgery/neurology, cardiothoracic surgery, cardiology, or other), severity of illness (Acute Physiology and Chronic Health Evaluation II score, Simplified Acute Physiology Score II), and length of stay. Orders on life support at ICU admission, subsequent changes in life support orders during the ICU stay, and life support orders at end-of-life were collected from the typically used forms and the daily reports. The date of change was also recorded. If no order on life support or no change in order on life support was mentioned, either on the form or in the daily reports, it was assumed that full life support was
842
End-of-Life Decisions in Dutch ICUs
RESULTS
Patients
Of 2578 patients admitted to the two ICUs in 2005, 293 patients (11%) died during the ICU stay and 63 patients (2%) died 7 days after ICU discharge. Of this group,
Table 1. Demographic Characteristics of All Admitted Patients Who Died in 2005 Patients who died AMC
Number of patients Died in ICU, N (%) Died 7 days after ICU discharge, N (%) Age (yr) Gender (male), N (%) Length of stay in ICU (d) APACHE II SAPS II Type of ICU admission, N (%) Acute Elective Reason for admission to the ICU, N (%) Medical Surgical Neurosurgery/neurology Cardiothoracic surgery Cardiology Other/undefined 269 237 (88) 32 (12) 66 (5266) 149 (55) 4 (27) 25 (2030) 58 (4671) 181 (67) 88 (33) 68 (25) 40 (15) 68 (25) 27 (10) 66 (25) 0 (0)
GH
78 56 (72) 22 (28) 76 (6882) 51 (65) 3 (28) 20 (1528) 51 (3864) 76 (97) 2 (3)
P value AMC versus GH

0.007* 0.028* 0.001 0.198* 0.538 0.001 0.001 0.202* 0.001*
27 (35) 42 (54) 3 (4) NA 6 (8) 0 (0)
Data are medians (interquartile range), unless stated otherwise. APACHE II Acute Physiology and Chronic Health Evaluation II; SAPS II Simplied Acute Physiology Score II; ICU intensive care unit; AMC Academic Medical Center; GH Gelre Hospital. Comparisons were made using the 2 test* for categorical data and the MannWhitney test for continuous data.
Table 2. Orders at the End of Life AMC Order A

Number of patients, N (%) Age (yr) Gender (male), N (%) Length of stay in ICU (d) Readmission, N (%) APACHE II SAPS II Order at ICU admission, N (%) Order A Order B Order C Type of ICU admission, N (%) Acute Elective Reason for ICU admission, N (%) Medical Surgical Neurosurgery/neurology Cardiopulmonary surgery Cardiology Unknown 62 60 (4571) 43 (70) 2 (13) 5 (8) 26 (1732) 58 (4474) 62 (100) 0 (0) 0 (0) 53 (86) 9 (14) 16 (26) 6 (10) 12 (19) 11 (18) 17 (24) 0 (0)
GH Order C
148 68 (5276) 80 (55) 5 (29) 14 (10) 25 (2029) 57 (5869) 140 (95) 8 (5) 0 (0) 76 (51) 72 (59) 36 (24) 24 (16) 40 (27) 9 (6) 39 (26) 0 (0)
Order B
59 67 (5377) 31 (53) 3 (27) 4 (7) 25 (2034) 57 (4672) 55 (93) 4 (7) 0 (0) 52 (88) 7 (12) 16 (27) 10 (17) 16 (27) 7 (12) 10 (17) 0 (0)
Order A
15 74 (5979) 10 (67) 2 (14) 3 (20) 20 (1225) 56 (4074) 15 (100) 0 (0) 0 (0) 15 (100) 0 (0) 6 (40) 8 (53) 0 (0) 0 (0) 1 (7) 0 (0)
Order B
16 81 (6987) 11 (69) 2 (26) 2 (13) 18 (1628) 41 (3659) 13 (81) 3 (19) 0 (0) 14 (88) 2 (12) 9 (56) 6 (38) 0 (0) 0 (0) 1 (7) 0 (0)
Order C
47 75 (6781) 30 (64) 3 (210) 7 (15) 22 (1530) 51 (3964) 36 (77) 11 (23) 0 (0) 47 (100) 0 (0) 13 (28) 28 (60) 2 (4) 0 (0) 4 (9) 0 (0)
Data are medians (interquartile range), unless stated otherwise. For denitions of orders A, B, and C, see the Methods section in the text. APACHE II Acute Physiology and Chronic Health Evaluation II; SAPS II Simplied Acute Physiology Score II; ICU intensive care unit; AMC Academic Medical Center; GH Gelre Hospital.
the clinical record files of 9 were missing, leaving 347 patients for further analysis. Demographic data are summarized in Table 1. In the AMC, relatively more patients died in the ICU as compared with the Gelre Hospital (P 0.007). Conversely, at the Gelre Hospital, a higher
percentage of patients died after ICU discharge compared with the AMC (P 0.0004). This proved to be due to a different case mix. In particular, in the AMC, 43 (86%) neurosurgical patients died in the ICU, whereas in Gelre no neurosurgery was performed. Patients who
843
Figure 1. Change and type of orders

over time. See text for details on orders.
Table 3. Order Changes in Patients Who Died in Intensive Care Unit (ICU) and in Patients Who Died Within 7 Days After ICU Discharge AMC ICU nonsurvivors
Number of evaluated patients, N Patients with changes in order, N (%) Timing of change of orders LOS in ICU before a 1st change in orders (d) Length of life after a 1st change in orders (d) Order at admission to the ICU, N (%) Order A Order B Order C Order end of life, N (%) Order A Order B Order C 237 179 (76) 4 (37) 3 (15) 228 (96) 9 (4) 0 (0) 59 (25) 50 (21) 128 (54)
GH Hospital nonsurvivors
32 27 (84) 2 (04) 1 (02) 29 (91) 3 (9) 0 (0) 3 (9) 9 (28) 20 (63)
ICU nonsurvivors
56 42 (75) 1 (06) 0 (02) 46 (82) 10 (18) 0 (0) 12 (21) 9 (16) 35 (63)
Hospital nonsurvivors
22 18 (82) 2 (18) 3 (25) 18 (82) 4 (18) 0 (0) 3 (14) 7 (32) 12 (54)
Data are medians (interquartile range), unless stated otherwise. For denitions of orders A, B, and C, see the Methods section in the text. AMC Academic Medical Center; GH Gelre Hospital; LOS length of stay.
eventually died were slightly older, had a longer ICU length of stay, and had higher Acute Physiology and Chronic Health Evaluation II and Simplified Acute Physiology Score II scores than patients who survived (data not shown).
C had a longer ICU stay than patients who died with full support.
Documentation of Changes in Life Support Orders

For 85/266 patients (32%) who had one or more changes in their life support orders, it was not clear which ICU team member was involved in the decision to change the orders (details shown in Table 4). Although 21 patients (8%) were mentioned as being capable of making a decision regarding their medical treatment at the time of a change in their life support orders, only 13 (5%) of the patients were compos mentis at the time of change and actively involved in the decision-making process. For 206 patients (78%), the patients preference regarding life support was not reported. For 94 patients (36%), it was not reported whether a patients representative existed or was involved in the decision to change life support orders.
Life Support Orders on ICU Admission and Thereafter

Of 347 patients, 321 (93%) were admitted to the ICU with an order A and 26 (7%) with an order B; there were no patients who were admitted to the ICU with an order C (Table 2). For 266 patients (76%), one or more changes in their life support orders were made before death (Fig. 1, details are shown in Table 3). Seventy-seven patients (22%) died with an order A, 85 patients (25%) with an order B, and 185 patients (53%) with an order C. The pattern of life support orders was comparable in the two ICUs. Patients who died with an order
844
Table 4. Documentation of Change in Orders AMC GH
1st code 2nd code 3rd code 1st code 2nd code (n 203) (n 84) (n 3) (n 60) (n 22)
Which people were involved in the decision to change orders on life support, N (%) Not reported Patient Physician Family Representative Was the patients representative involved in the change in orders?, N (%) Not reported Family Partner Legal representative Other None Was the patient compos mentis?, N (%) Not reported Yes No Patient was awake but considered incapablea Patient was sedated Patient was not sedated but unconscious Knowledge of patient preference by surrogate decision-makers Not reported No Yes Information given beforehand by patient Information given by family Advanced directive/living will Psychiatric consult No/not reported Before code change After code change 66 (33) 8 (4) 136 (67) 114 (56) 2 (1) 76 (37) 104 (51) 34 (17) 0 (0) 4 (2) 17 (8) 13 (6) 12 (6) 14 (7) 48 (24) 116 (57) 161 (78) 0 (0) 13 (6) 27 (13) 1 (1) 0 (0) 198 (98) 5 (2) 0 (0) 15 (18) 5 (6) 68 (81) 61 (73) 0 (0) 27 (32) 52 (62) 14 (17) 0 (0) 1(10 1 (1) 8 (10) 5 (6) 6 (7) 12 (14) 53 (63) 56 (67) 0 (0) 10 (12) 17 (20) 1 (1) 0 (0) 82 (97) 2 (3) 0 (0) 1 (25) 0 (0) 2 (50) 3 (75) 0 (0) 1 (25) 2 (67) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 1 (25) 2 (75) 4 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 4 (100) 0 (0) 0 (0) 19 (32) 11 (18) 40 (67) 38 (63) 0 (0) 18 (30) 41 (68) 8 (13) 0 (0) 3 (5) 1 (2) 1 (2) 9 (15) 14 (23) 22 (37) 14 (23) 45 (75) 0 (0) 13 (22) 2 (3) 0 (0) 0 (0) 59 (99) 1 (2) 0 (0) 0 (0) 2 (9) 22 (100) 22 (100) 0 (0) 5 (23) 6 (27) 16 (73) 0 (0) 0 (0) 0 (0) 1 (5) 2 (9) 6 (27) 8 (36) 5 (23) 16 (73) 0 (0) 2 (9) 4 (18) 0 (0) 0 (0) 22 (100) 0 (0) 0 (0)
A 3rd code change did not occur in the GH. AMC academic medical center; Code change in code (rst, second, and third time); GH Gelre Hospital. a Because of established encephalopathy, delirium, or preexisting cognitive dysfunction and/or dementia.
DISCUSSION
In this study, we found that ICU nonsurvivors predominantly die with orders for withholding or withdrawing life support. We also found that documentation on changes in life support is far from complete, and that frequently, important information was lacking. Our results regarding withholding and withdrawing life support are in line with Sprung et al.,3,4 who demonstrated that 73% of European ICU patients die while life support is being withheld or withdrawn. In contrast, Eidelman et al.5 showed that physicians in Israel do withhold, but never withdraw life-supporting interventions, which is in accordance with Jewish law. This emphasizes the fact that withholding or withdrawing life support is strongly influenced by culture. Still, in general, the incidence of withholding and withdrawal of life support is increasing, partly because more countries are legalizing this process if specific conditions are met.6 We can only hypothesize why documentation of changes in decisions on end-of-life orders was poor in
our hospitals. We may conclude this to be a reflection of bad documentation on important decisions in general. Indeed, documentation of decisions on blood transfusion, start or change of antimicrobial therapy, change of ventilatory mode, tracheal extubation, and many other important decisions in daily ICU practice is reported to be poor.7,8 However, decisions on endof-life care are considered so important, at least from a medico-judicial viewpoint, that we expected more complete and lucid documentation. In addition, although the usual manner of documentation might be different in the general as compared with the university hospital, which is probably caused by understaffing of intensivists in the general hospital, we believe that this should never be a reason for poor documentation of decisions over withholding or withdrawing therapy. Bad documentation could also be due to less continuity of care, with fewer doctors in the nonacademic setting, although the absence of differences between
845
the academic and nonacademic settings does not support this argument. Perhaps ICU physicians in our settings are too busy with direct patient care to adequately document all discussions and agreements with patients and proxies. They could also lack training or not be fully aware of the importance of strictly documenting changes in life support orders. Decisions on end-of-life orders are influenced by the professional experience of individual ICU team members making the decision (mostly ICU physicians),9 12 and by their age,11 religious beliefs,4,5 and country or culture of origin,4,11,1315 potentially resulting in differences between ICUs. Indeed, in Western countries, limitation of therapy precedes up to 90% of deaths, whereas in India the rate is reported to be no more than 22%50%.4,16 However, in Western countries the legal context of end-of-life decisions may be confusing because of the lack of specific laws.17 To support the decision-making process in the face of ethical dilemmas, several countries are developing guidelines and laws.18 In particular, patients relatives should be informed and consulted before a decision is made, the decision should be documented in the medical charts, and finally, the decision should be made collectively.6 However, these guidelines do not include statements pertaining to documentation of decisions to forego life support. Lack of documentation may mean that the decision of whether to withdraw or withhold life support treatment is made only by the physician, without consulting the relatives. Noninvolvement was shown to be as high as 50% for decisions to change end-of-life orders in the study by Ferrand et al.19 Our findings regarding poor documentation of decisions to change life support orders do not necessarily relate to an insufficiently careful decision-making process. However, all patients life support orders should be adequately documented for evaluation and legal purposes. There are several limitations to our study. First, our study design does not allow generalization of our findings to all ICU patients. Indeed, we only included ICU nonsurvivors and patients who died shortly after ICU discharge in The Netherlands. Second, this study was restricted to only two ICUs in The Netherlands. Results may not be generalized to other institutions, particularly those with a patient population substantially different from that of the study hospitals. We showed that proportionally more patients died in the ICU in the academic AMC setting than in the Gelre ICU, which proved to be due to differences in case mix. Nevertheless, the similarities between these two centers with regard to practices on withholding or withdrawing life support are striking, particularly in view of the fact that they represented an academic and a nonacademic, albeit university-affiliated, teaching setting. This at least suggests that the situations in other hospitals in The Netherlands may be similar to those in our centers.
846
In conclusion, ICU nonsurvivors in The Netherlands die predominantly with orders to withhold or withdraw life support. Documentation of these decisions is poor and inconsistent, however, and deserves more attention from attending physicians for evaluation and legal purposes. REFERENCES
1. Luce JM. Ethical principles in critical care. JAMA 1990; 263:696 700 2. Ditto PH, Danks JH, Smucker WD, Bookwala J, Coppola KM, Dresser R, Fagerlin A, Gready RM, Houts RM, Lockhart LK, Zyzanski S. Advance directives as acts of communication: a randomized controlled trial. Arch Intern Med 2001;161:42130 3. Sprung CL, Cohen SL, Sjokvist P, Baras M, Bulow HH, Hovilehto S, Ledoux D, Lippert A, Maia P, Phelan D, Schobersberger W, Wennberg E, Woodcock T. End-of-life practices in European intensive care units: the Ethicus Study. JAMA 2003;290:790 7 4. Vincent JL. Forgoing life support in western European intensive care units: the results of an ethical questionnaire. Crit Care Med 1999;27:1626 33 5. Eidelman LA, Jakobson DJ, Pizov R, Geber D, Leibovitz L, Sprung CL. Foregoing life-sustaining treatment in an Israeli ICU. Intensive Care Med 1998;24:162 6 6. Ferrand E, Marty J. Prehospital withholding and withdrawal of life-sustaining treatments. The French LATASAMU survey. Intensive Care Med 2006;32:1498 1505 7. Wynn A, Wise M, Wright MJ, Rafaat A, Wang YZ, Steeb G, McSwain N, Beuchter KJ, Hunt JP. Accuracy of administrative and trauma registry databases. J Trauma 2001;51:464 8 8. Golob JF Jr, Fadlalla AM, Kan JA, Patel NP, Yowler CJ, Claridge JA. Validation of Surgical Intensive Care-Infection Registry: a medical informatics system for intensive care unit research, quality of care improvement, and daily patient care. J Am Coll Surg 2008;207:164 73 9. Cook DJ, Guyatt GH, Jaeschke R, Reeve J, Spanier A, King D, Molloy DW, Willan A, Streiner DL. Determinants in Canadian health care workers of the decision to withdraw life support from the critically ill. Canadian Critical Care Trials Group. JAMA 1995;273:703 8 10. Walter SD, Cook DJ, Guyatt GH, Spanier A, Jaeschke R, Todd TR, Streiner DL. Confidence in life-support decisions in the intensive care unit: a survey of healthcare workers. Canadian Critical Care Trials Group. Crit Care Med 1998;26:44 9 11. Cuttini M, Nadai M, Kaminski M, Hansen G, de Leeuw R, Lenoir S, Persson J, Rebagliato M, Reid M, de Vonderweid U, Lenard HG, Orzalesi M, Saracci R. End-of-life decisions in neonatal intensive care: physicians self-reported practices in seven European countries. EURONIC Study Group. Lancet 2000;355:2112 8 12. Burns JP, Mitchell C, Griffith JL, Truog RD. End-of-life care in the pediatric intensive care unit: attitudes and practices of pediatric critical care physicians and nurses. Crit Care Med 2001;29:658 64 13. Randolph AG, Zollo MB, Egger MJ, Guyatt GH, Nelson RM, Stidham GL. Variability in physician opinion on limiting pediatric life support. Pediatrics 1999;103:e46 14. Nyman DJ, Sprung CL. End-of-life decision making in the intensive care unit. Intensive Care Med 2000;26:1414 20 15. Vincent JL. Cultural differences in end-of-life care. Crit Care Med 2001;29:N525 16. Prendergast TJ, Claessens MT, Luce JM. A national survey of end-of-life care for critically ill patients. Am J Respir Crit Care Med 1998;158:11637 17. Zamperetti N, Proietti R. End of life in the ICU: laws, rules and practices: the situation in Italy. Intensive Care Med 2006;32:1620 2 18. Steinberg A, Sprung CL. The dying patient: new Israeli legislation. Intensive Care Med 2006;32:1234 7 19. Ferrand E, Robert R, Ingrand P, Lemaire F. Withholding and withdrawal of life support in intensive-care units in France: a prospective survey. French LATAREA Group. Lancet 2001; 357:9 14
Low Tidal Volume Ventilation in a Porcine Model of Acute Lung Injury Improves Cerebral Tissue Oxygenation
Johannes Bickenbach, MD* Norbert Zoremba, MD Michael Fries, MD Rolf Dembinski, MD, PhD* Robert Doering Eileen Ogawa, MD* Rolf Rossaint, MD, PhD* Ralf Kuhlen, MD, PhD
BACKGROUND: In study, we investigated the effects of different tidal volumes on cerebral tissue oxygenation and cerebral metabolism in a porcine model of acute lung injury (ALI). We hypothesized that mechanical ventilation with low tidal (LT) volumes improves cerebral tissue oxygenation and metabolism after experimentally induced ALI. METHODS: After inducing experimental ALI by surfactant depletion, we studied two conditions in 10 female pigs: 1) LT volume ventilation with 6 mL/kg body weight, and 2) high tidal (HT) volume ventilation with 12 mL/kg body weight. Variables of gas exchange, hemodynamic, continuous cerebral tissue oxygen tension (ptiO2), cerebral microdialysis, and systemic cytokines were analyzed. After induction of ALI, data were collected at 2, 4, and 8 h. The primary end point was the change in ptiO2. For group comparisons, a t-test was used. A value of 0.05 was considered to indicate statistical significance. RESULTS: At baseline and after induction of ALI, no differences between groups were found in ptiO2; however, ptiO2 was significantly lower in the HT group after 4 and 8 h. Pao2 and Paco2 showed no significant differences between the groups at all timepoints. Regarding cerebral microdialysis, a significantly higher level of extracellular lactate could be demonstrated after 2, 4, and 8 h in the HT group. The release of cytokines resulted in higher values for interleukin-6 and interleukin-8 in the HT group. CONCLUSION: Protective ventilation with LT yielded a significant improvement in cerebral tissue oxygenation and metabolism compared to HT ventilation in a porcine model of ALI. There was dissociation between arterial and cerebral tissue oxygenation. Cerebral oxygenation and metabolism might have possibly been impaired by a more distinctive inflammatory response in the HT group.
(Anesth Analg 2009;109:84755)
cute respiratory distress syndrome (ARDS) is still a life-threatening disease with high mortality.1,2 However, partly, because of less invasive mechanical ventilation, survival rates have improved over the past decade,3,4 and hence various studies have focused on long-term effects after ARDS. Seventy percent of ARDS survivors show distinctive neurological impairment, including memory, language, and cognitive decline.5 Moreover, 90% of critical care patients who undergo long-term mechanical ventilation (i.e., 5 days) display these findings.6 Still, the mechanisms that are of importance for this poor neurological outcome have not yet been fully explained. Memory impairment is associated with hypoxic damage to hippocampal structures.79 However, refractory hypoxemia inherent to ARDS cannot be seen
From the *Department of Surgical Intensive Care; Department of Anesthesiology, RWTH University Hospital, Aachen, Germany; and Department of Intensive Care Medicine, Helios Hospital Berlin Buch, Berlin, Germany. Accepted for publication April 27, 2009. Address correspondence and reprint requests to Johannes Bickenbach, Department of Surgical Intensive Care, RWTH University Hospital, Pauwelsstr. 30, D-52074 Aachen, Germany. Address e-mail to jbickenbach@ukaachen.de. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ad5769
as the single mechanism contributing to neurological impairment, as we have previously demonstrated.10 Ventilation strategy may affect the central nervous system (CNS) by modifying the inflammatory response.1114 We, therefore, designed a laboratory animal study to investigate effects of ventilation with different tidal volumes on cerebral tissue using fluorescence quenching (ptiO2) as a local measure of oxygenation and intracerebral microdialysis as a measure of local metabolism. We hypothesized that, after experimentally induced acute lung injury (ALI), mechanical ventilation with low tidal (LT) volumes would result in improved cerebral tissue oxygenation and brain metabolism by a more attenuated inflammatory response when compared to high tidal (HT) volume ventilation.
METHODS
The experimental protocol was approved by the appropriate governmental committee for the use and care of laboratory animals.
Animal Preparation
Ten female pigs weighing 30.2 2.0 kg (mean sd) were included in the study. Preexisting diseases were excluded after examination by a veterinarian.
847
IM azaperone (6 mg/kg) and atropine (0.01 mg/kg) were administered 45 min before anesthesia. Anesthesia was performed by continuous infusion of thiopental (510 mg/kg/h) and fentanyl (8 12 g/kg/h). Animals were then orally intubated and placed in a supine position. Volume-controlled ventilation was used with a respiratory rate of 15 to 25 breaths/min to yield a Paco2 of 35 to 40 mm Hg, a positive endexpiratory pressure of 0 mm Hg and an inspiration:expiration ratio of 1:1. Initially, tidal volume was adjusted to 10 mL/kg body weight. The inspiratory oxygen fraction was kept at 1.0 throughout the experiment. An arterial catheter (Vygon, Ecouen, France) and an 8.5-F venous sheath (Arrow Deutschland GmbH, Erding, Germany) were percutaneously positioned after ultrasound-guided puncture of the femoral vessels. A Swan-Ganz catheter (Arrow Deutschland GmbH, Erding, Germany) was positioned into a pulmonary artery under transducer pressure guidance. The blood temperature, as determined by the pulmonary artery catheter, was maintained around 38C using a convective heating blanket (Warm Touch 5200, Tyco Healthcare, Pleasanton, CA). Fluid replacement was provided by administration of balanced electrolyte solution in accordance with hemodynamics. After positioning of the head in midline and fixation with a custom-made holding device, a burr hole with a diameter of 2 mm was created 1 cm lateral to the sagittal suture and 1 cm rostral to the coronal suture into the right hemisphere. A double-lumen BOLT Catheter (Licox IM3, Integra Neurosciences, Plainsboro, NJ) was carefully screwed into the skull, and the dura was carefully punctured with a needle before advancing the OxyLite sensor and the microdialysis catheter.
Oxygen delivery (DO2) and oxygen consumption (VO2) were calculated using standard formula:
[DO2 CaO2 CO; VO2 avDO2 CO]

Cerebral microdialysis was performed using a probe with a membrane length of 10 mm, an outer diameter of 0.6 mm, and a cutoff at 20 kDa (CMA 70, CMA-Microdialysis, Sweden). A precision infusion pump (CMA 102, CMA-Microdialysis, Sweden) was used at a flow rate of 2 L/min (Perfusion fluid CNS, CMA-Microdialysis, Sweden).1517 Before and at the end of the experiment, the in vitro recovery rates for each probe were determined in a calibration solution equilibrated to 37C by continuing the perfusion with the settings used during the experiment. In all experiments, no significant changes in the in vitro recovery rate were found. The measured experimental values were adjusted according to the relative recovery rate to estimate the in vivo extracellular concentration of the substances in the immediate environment of the probes. Lactate and glucose were collected for every measurement timepoint, stored in aliquots at 10C, and later analyzed enzymatically with a CMA 600 Microdialysis Analyzer. The lactate/pyruvate ratio (L/P ratio) was also determined as a marker for disturbed cellular energy metabolism in hypoxia.18
Cerebral ptiO2 was Measured with an OxyLab System (Oxford Optronics, UK)
Blood samples for the determination of S-100 protein and proinflammatory cytokines (interleukin [IL]-6 and IL-8) were taken from the arterial line and allowed to clot for 30 min at room temperature. After centrifugation with 3000g for 10 min, the supernatant was stored in aliquots at 80C until analysis. Serum levels of IL-6 and IL-8 were quantitated using commercially available ELISAs (R&D Systems, Minneapolis, MN). S-100 levels were measured using an automated immunoluminometric assay (DiaSorin, Dietzenbach, Germany).
Measurements
Mean arterial blood pressure (MAP), mean pulmonary arterial pressure (MPAP), central venous pressure (CVP), and pulmonary arterial occlusion pressure were measured (pvb Medizintechnik, Kirchseeon, Germany; AS/3 Compact; Datex Ohmeda, Helsinki, Finland). Cardiac output (CO) and mixed venous saturation (SvO2) were continuously recorded (Vigilance, Edwards Lifesciences Irvine, CA). Heart rate was derived from the blood pressure curve. Systemic vascular resistance (SVR) and pulmonary vascular resistance (PVR) were calculated using standard formula:
Experimental Protocol
Induction of experimental ALI was performed by repeated surfactant washout.19,20 The experimental protocol was started when an arterial oxygen tension (PaO2) 100 mm Hg persisted for 60 min without additional lavages. After induction of ALI, the animals were randomly assigned to two groups: (a) the tidal volume was set to 6 mL/kg body weight (LT group); and (b) 12 mL/kg body weight (HT group). A positive end-expiratory pressure of 5 mm Hg was adjusted in both groups. No further interventions were performed. At baseline, PacO2 levels were allowed to range from 35 to 45 mm Hg. Later, the respiratory rate was adjusted to minimize hypercapnia in the LT group. MAP was allowed to range between 60 and 150 mm Hg.21
[SVR ( MAP-CVP)/CO 80; PVR (MPAP-pulmonary capillary wedge pressure)/ CO 80].

Arterial and mixed venous blood gases, hemoglobin, electrolytes, and lactate were analyzed (ABL 510 and OSM 3; Radiometer, Copenhagen, Denmark).
848
Neurological Consequences of Mechanical Ventilation
Table 1. Hemodynamic Variables BL

HR (bpm) HT LT MAP (mm Hg) HT LT CVP (mm Hg) HT LT CO (L/min) HT LT MPAP (mm Hg) HT LT PCWP (mm Hg) HT LT SVR (dyn sec cm5) HT LT PVR (dyn sec cm5) HT LT 89.4 9.4 104.6 16.2 129.2 18.6 128.6 16.7 6.0 1.8 6.6 1.9 3.8 0.8 3.2 0.5 17.8 2.8 17.4 6.6 7.6 2.3 7.4 0.9 2677 852 2630 150 218 53 200 21
ALI
118.6 19.8 92 17.0 86.4 9.7 95.4 7.1 6.8 1.3 7.4 1.1 4.6 0.9 3.3 0.3 27.2 7.2 29.0 2.0 8.0 1.8 9.6 1.8 1421 306 2018 466 349 149 454 59
2h
132.6 43.8 90.8 13.8 79.2 15.8 95.4 3.8 7.2 1.5 8.6 1.3 3.1 0.6 2.9 0.4 31.4 3.4 33.4 3.8 9.8 1.5 9.6 1.8 1902 732 2376 250 562 143 687 159
4h
136.4 42.9 97.6 11.5 75.6 9.4 89.8 8.7* 8.6 2.5 9.2 0.8 3.1 0.8 2.7 0.4 29.8 6.7 34.4 5.2 11.4 2.7 10.0 1.6 1850 775 2309 223 502 226 137 97
8h
188.3 84.2 116.6 47.6 63.6 23.8 63.2 10.6* 11.7 14.5 12.9 12.2 3.0 0.8 2.7 0.2 30.6 7.5 27.8 5.7 10.2 4.0 10.0 1.4 1678 699 2619 856 547 173 556 62
Data are presented as mean SD. BL baseline; ALI acute lung injury; HR heart rate; MAP mean arterial blood pressure; CVP central venous pressure; CO cardiac output; MPAP mean pulmonary artery pressure; PCWP pulmonary capillary wedge pressure; SVR systemic vascular resistance; PVR pulmonary vascular resistance. * P 0.05 between groups.
After equilibration of 1 h, baseline measurements were obtained. Hemodynamics, pulmonary gas exchange, respiratory mechanics, brain tissue oxygen tension, intracerebral microdialysis, and systemic cytokines were measured at baseline, after induction of ALI, and 2, 4, and 8 h afterward.
Hemodynamic Variables
Overall, hemodynamic variables were relatively unchanged in both groups throughout the study period. Fluid replacement was equal in the groups, with a mean of 5 mL/kg/h. The MAP was significantly lower in the HT group at 4 and 8 h (P 0.05 for both). Other hemodynamic variables are summarized in Table 1.
All data are expressed as mean sd. Normal distribution of the data was confirmed using the KolmogorovSmirnov test. Group comparisons at the given timepoints were performed using the t-test for independent measurements. To detect significant changes within groups over time, analysis of variance was performed and followed by post hoc tests for multiple comparisons (Bonferroni). In case of significant differences, exact P values are reported. In general, a P value of 0.05 was considered to indicate statistical significance.
Respiratory Mechanics
There were no differences between the groups at baseline and at induction of ALI. Significantly higher mean and peak inspiratory pressures were observed in the HT group at 2, 4, and 8 h (Table 2).
Gas Exchange and Cerebral Tissue Oxygenation

No differences between the groups could be shown for the pH value at any timepoint (Table 3). Regarding PaO2, induction of ALI resulted in a remarkable decrease in arterial oxygenation. However, at no time were significant differences between the groups observed. PacO2 increased by approximately 50% after induction of ALI. No differences were noted between groups. As depicted in Figure 1, ptiO2 was not significantly different between the groups at baseline. In both groups, induction of ALI resulted in a dramatic decrease in ptiO2. There was already a strong tendency toward improved ptiO2 values after 2 h of ventilation in animals ventilated with LT volumes, in comparison to animals in the HT groups, reaching statistical significance after 4 and 8 h
RESULTS
All animals survived the entire study period. No differences in baseline variables were observed between groups. A mean of 7 1 lavages were performed to obtain a stable ALI. In both groups, induction of ALI resulted in comparable significant reductions of oxygen-derived variables (PaO2, ptiO2, and VO2), hemodynamic (MAP), and metabolic variables (pH). Concurrently, mean and peak inspiratory pressures increased significantly.
849
Table 2. Respiratory Mechanics BL

RR HT LT PEEP (mbar) HT LT MIP (mbar) HT LT PIP (mbar) HT LT 17 2 18 2 0 0 15 1.5 14 0.5 20 2.1 19 1.3
ALI
22 5 27 5 0 0 29 2 27 1 37 3 35 2.4
2h
34 2 37 6 5 5 37 4.3 23 2* 56 8 28 2.4*
4h
37 4 37 6 5 5 38 5.9 22 2.4* 57 6.7 28 2.4*
8h
37 4 37 6 5 5 39 5.8 22 2.2* 61 4.6 28 2.2*
Data are presented as mean SD. BL baseline; ALI acute lung injury; RR respiratory rate; PEEP positive end-expiratory pressure; MIP mean inspiratory pressure; PIP peak inspiratory pressure. * P 0.05 between groups. P 0.05 vs. ALI.
Table 3. Variables of Gas Exchange and Tissue Oxygenation BL

Pao2 (mm Hg) HT LT PaCo2 (mm Hg) HT LT pH HT LT Do2 (mL/min) HT LT Vo2 (mL/min) HT LT 482.6 36.7 474.8 95.2 36.2 3.8 39.6 1.6 7.5 0.0 7.5 0.04 5454 971 4730 868 731 305 735 303
ALI
43.3 4.4 50.8 11.3 61.2 8.3 60.2 8.8 7.26 0.05 7.32 0.08 3968 1356 3607 903 1804 498 1461 378
2h
118.2 142.2 84.120.3 63.0 21.7 68.2 12.2 7.28 0.13 7.24 0.06 3699 844 3884 1050 1359 359 1560 280
4h
135.0 162.9 142.4 114.2 69.2 30.1 69.8 12.6 7.32 0.13 7.26 0.09 3904 923 3754 449 1677 490 1297 286
8h
217.5 188.3 219.2 112.9 63.5 21.9 64.0 11.7 7.3 0.14 7.3 0.04 4013 804 3770 119 1349 744 796 455
Data are presented as mean SD. BL baseline; ALI acute lung injury; pH pH value; PaCO2 arterial carbon dioxide partial pressure; PaO2 arterial oxygen partial pressure; DO2 oxygen delivery; VO2 oxygen consumption. * P 0.05 between groups. P 0.05 vs. ALI.
(P 0.05). Ventilation with LT volumes resulted in a significant increase of ptiO2 8 h after ALI. Concerning arterial oxygen saturation, a decrease could be observed after ALI in both groups (Figure 2), followed by an increase thereafter. No significant differences between the groups could be shown.
a trend toward lower extracellular glucose levels in the HT group after 4 and 8 h. The L/P ratio did not achieve statistical significance, but there was a strong tendency toward a higher level in the HT group (Figure 3).
Venous Oxygen Saturation was Signicantly Lower in the HT Group After 8 h

No significant differences between the groups were found for DO2 and VO2, as shown in Table 2. VO2 values were clearly higher in the HT group after 4 and 8 h (Figure 2).
S-100 and Cytokines

Serum concentrations of S-100 protein were not significantly different between the groups at baseline. After induction of ALI, no noticeable increase was seen in either group. After 2 h, a decrease of S-100 was noted in the LT group, whereas the levels in the HT group remained unchanged. Continuing higher levels for S-100 were found in the HT group after 4 h (0.3 0.1 vs 0.1 0.02 g/L, P 0.052) and 8 h (0.3 0.17 vs 0.1 0.03 g/L, P 0.05). The measurement of IL-6 showed no differences between the groups at baseline and after ALI. After 2 h, significantly higher levels could be demonstrated in the HT group (279.7 187.1 vs 45.4 101.6 pg/mL,
Cerebral Microdialysis
For lactate, no difference could be shown at baseline and after induction of ALI (Figure 3). It was significantly higher in the HT group after 2 h (3.4 1.4 vs 1.1 0.5 mmol/L), 4 h (2.4 0.8 vs 1.0 0.7 mmol/L), and 8 h (1.9 0.6 vs 1.0 0.4 mmol/L). No significant differences in glucose were seen. There was
850
Figure 1. ptiO2. Data are presented as mean sd *P 0.05 between groups. $P 0.05 vs acute lung injury (ALI) (exact P value is given in brackets).
Figure 3. Extracellular metabolites of cerebral microdialysis.

Data are presented as mean sd. *P 0.05 between groups.
Systemic Lactate, Hemoglobin, and Electrolytes

For lactate, significantly higher levels in the HT group were observed after ALI and at 2 and 8 h (Figure 5). Ventilation with LT volumes ameliorated the increases observed in the HT group and led to a significant decrease at 2, 4, and 8 h after ALI. There were no differences in hemoglobin and electrolyte levels between the groups at any timepoint (Table 4).
DISCUSSION
presented as mean sd. *P 0.05 between groups.
Figure 2. Arterial and venous oxygen saturation. Data are
P 0.05). After 4 h, a clear trend toward higher IL-6 release was seen in the HT group (591.6 480.2 vs 66.7 137.7, P 0.07). The analysis of IL-8 revealed higher levels after 2 h (2785.8 2246.1 vs 349.1 510.4, P 0.1) and 8 h (2752.5 3597.9 vs 2.2 2.7, P 0.2), although statistical significance was not reached (Figure 4).
The purpose of this study was to determine the effects of different modes of mechanical ventilation on the CNS in a porcine model of ALI. Our major finding was that cerebral ptiO2 improved only during ventilation with LT volumes, although global variables of oxygenation were comparable. Importantly, measurement of ptiO2 is a technique to reveal oxygen metabolism and therefore the vitality of organs. Concerning brain tissue, previous experiments have demonstrated physiological brain ptiO2 values of 25 to 40 mm Hg.22 Thresholds of 10 mm Hg are
851
Figure 4. Systemic cytokines. Data are presented as mean

sd *P 0.05 between groups. #P 0.05 vs acute lung injury (ALI) (exact P value is given in brackets).
Figure 5. Systemic lactate. Data are presented as mean sd. *P 0.05 between groups. #$P 0.05 vs acute lung injury (ALI) (exact P values are given in brackets). indicative of ischemia. In a study evaluating hypoxic events, ptiO2 values were 11 3 mm Hg at an arterial PO2 of 40 mm Hg.23 These results are partly consistent with ours. We found even lower ptiO2 levels at an arterial PO2 around 50 mm Hg (Figure 6). This is the first study evaluating ptiO2 levels in a model of ALI. However, with regard to the reproducibility of arterial PO2 values in both groups, the
852
course of this variable alone cannot explain the development of ptiO2 levels in this study. In ARDS patients, numerous factors have an impact on the CNS, such as long-term sedation,24 hyperglycemia,6 and inflammation. Concerning the impact of mechanical ventilation, it is well known that ventilation with LT volumes is associated with an attenuated inflammatory response.1114 Conversely, there is growing evidence that more injurious, HT volume ventilation results in a massive inflammatory response due to strain on noninjured, non-atelectatic lung regions in ALI.25 HT volumes lead to increased levels of peripheral organ dysfunction.26 However, the mechanisms by which mechanical ventilation culminates in organ dysfunction and how lung-protective ventilatory strategies would reduce these effects are unknown. Imai et al. demonstrated epithelial cell apoptosis of peripheral organs during injurious ventilation with HT volumes. It was suggested that increased levels of chemokines and inflammatory mediators enhance these effects leading to multiple organ dysfunction.27 Imai et al.s study did not examine the CNS, and the extent to which these mechanisms could also contribute to neuronal damage is unclear. However, inflammation is also involved in neurodegenerative processes, particularly in hippocampal neurons.28 In our study, although only reaching significance for IL-6 after 2 h, the more distinctive cytokine levels in the HT group might have been a consequence of mechanical stress in the lung due to strain and higher inspiratory pressures. This systemic inflammatory response might have affected the CNS in our model as well, leading to compromised cerebral oxygenation and metabolism in the HT group. According to Imai et al.s results, apoptosis might be one explanation for the observed CNS compromise. Within apoptotic processes, mitochondrial apoptotic alterations are initiated. Mitochondria are relevant for cellular metabolism, cellular oxygen use, and highenergy phosphate production (ATP).29 Mitochondrial damage leads to increased reliance on anaerobic metabolism, which favors lactic acid production. As observed in our study, the phenomena of altered tissue oxygen use at the mitochondrial level may contribute to dysfunction of the integrity of the CNS. The tendency toward higher VO2 in the HT group could provide an indication for these metabolic changes. Furthermore, inflammation may cause cerebral microcirculatory dysfunction.30 Such alterations in microcirculation, triggered by a more pronounced inflammatory response in the HT group, could be a further mechanism explaining lower cerebral ptiO2 levels. Moreover, the use of cerebral microdialysis allowed examination of brain metabolism directly in the region of interest.3133 The extracellular measurement of energy-related metabolites reflects metabolic events in
Table 4. Hemoglobin and Electrolytes BL

Hb (g/dL) HT LT K (mmol/L) HT LT Na (mmol/L) HT LT Ca (mmol/L) HT LT 10.2 0.7 10.9 0.4 4.1 0.2 4.0 0.1 139 1 141 2 1.4 0.1 1.4 0.0
ALI
10.6 0.6 10.5 0.6 4.0 0.2 4.1 0.2 141 1 142 2 1.4 0.5 1.4 0.0
2h
11.9 1.4 10.8 0.5 4.3 0.3 4.4 0.2 141 2 141 1 1.3 0.1 1.4 0.0
4h
11.9 1.0 10.9 1.0 4.6 0.1 4.5 0.2 142 2 141 1 1.3 0.1 1.4 0.0
8h
11.3 1.1 10.9 1.0 4.7 0.2 4.4 0.2 142 2 141 1 1.3 0.0 1.3 0.0
Data are presented as mean SD. BL baseline; ALI acute lung injury; Hb arterial hemoglobin; K sodium; Na potassium; Ca calcium. * P 0.05 between groups.
Figure 6. Spearman correlation coefficient between PaO2 and ptiO2.
the intracellular compartment. Although data suggest high cerebral lactate levels for ischemia,21 indicative as a marker for secondary injury after brain damage,34,35 significant increases in cerebral lactate concentrations can also be found after global hypoxia.36 We also found significantly higher cerebral lactate levels in the HT group. Interestingly, lactate levels further increased in the HT group after normoxic conditions had been restored. Again, a systemic and regional inflammatory response triggered by more injurious mechanical ventilation could explain these mechanisms in the HT group. We suggest that the accumulation of anaerobic metabolites, as observed by cerebral microdialysis, additionally underlines the compromised cellular oxygen use in the CNS because of a more pronounced inflammatory response in the HT group. The course of lactate seems comparable when measured systemically and intracerebrally, as demonstrated in previous studies.37 This may underline the reliability of the techniques used to indicate hypoxic phenomena.
Although significant differences were not seen in the L/P ratios, at least the strong tendency in these results may further demonstrate anaerobic conditions due to mechanical ventilation with HT volumes. Hemodynamic results showed significantly lower MAP levels in the HT group after 4 and 8 h. However, in both groups, MAP levels were above the lower pressure limit for autoregulation, which is one of the most important mechanisms regulating cerebral blood flow (CBF).37 The cerebral circulation plays a major role in maintaining a constant chemical microenvironment in the brain. It is of prime importance in autoregulation, and therefore, although significantly different, MAP levels were set between the defined limits at all timepoints in both groups. Another mechanism influencing CBF is the PacO2.38 The vasodilatory effects have direct impact on CBF, and consequently on brain tissue oxygenation and metabolism.37,39 41 Because PacO2 levels in our study
853
were comparable between groups, it can be assumed that CBF was within a normal range. In addition, apart from a global measurement like CBF, regulation of cerebral microcirculation is dependent on numerous factors.42 Different microcirculatory effects due to the different tidal volumes could have played a role in the observed changes of cerebral tissue oxygenation and metabolism. The release of specific markers can be used to identify neuronal damage. S-100 protein, prevalent in astroglial cells of the CNS, is a sensitive and specific marker representing brain tissue damage and disintegration of the blood brain barrier. It has been shown to correlate well with the extent and prognosis of brain injury.10,43 46 The peak serum concentrations at baseline and after induction of ALI are comparable to those obtained in other studies concerning brain damage.10,47,48 However, it seems justified to assume that the early peak after baseline represents neuronal damage due to insertion of the microprobe. The constantly higher levels of S-100 in the HT group may reflect persistent neuronal damage that was not detectable in the LT group. Although statistical significance was only reached after 8 h, this may be considered as improved neuronal recovery in the LT group. However, these results may be limited, as mechanical ventilation per se may lead to a release of S100 protein.49 Our study has limitations. First, we cannot show evidence but can only speculate on the potential mechanisms playing a role in the observed results. In retrospect, the analysis of further organ dysfunction markers, such as monocyte-chemotactic protein 1 or growth-regulated oncogene, as well as quantitation of apoptosis could have provided further insights into the pathomechanisms. Therefore, elaborate in vivo and in vitro analysis is required. However, this is the first study evaluating neuronal compromise in experimental ALI and detecting relevant phenomena in conjunction with injurious mechanical ventilation. The specific aim was to directly monitor its neurological consequences. Second, concerning the length of the study, the effect on neuronal damage might have occurred more clearly in a chronic trial. Any long-term effects like neurocognitive decline after ALI cannot be demonstrated here. Nevertheless, even after this short-term observation, significant findings could be demonstrated, allowing new insights into the pathophysiology of brain damage after ALI. Within these acutely instrumented animals, the assessment of cerebral tissue oxygenation and metabolism in a model of ALI may be a valuable tool for identifying early neurological compromise. Third, the CBF was not measured in our study. However, both major variables influencing CBF, namely PacO2 and MAP, showed a comparable course in both groups. Besides, it is more likely that microcirculatory effects influenced by an inflammatory
854
response play a role in the differences observed. Therefore, it might have been helpful to also monitor microcirculation. Intracranial pressure (ICP) was also not monitored. As ICP is partly affected by the CBF and PacO2, we believe that ICP did not increase. Other factors influencing ICP, such as edema or swelling, seem unlikely. It is also possible that a further potential pattern of injury could have been caused by inserting a further probe. We wanted to focus on the measurement of oxygenation and resulting metabolism. In conclusion, we demonstrate for the first time that, in an established porcine model of ALI, LT volume ventilation leads to significantly improved cerebral tissue oxygenation when compared to HT volume ventilation. One explanation could be a more pronounced inflammatory response during mechanical ventilation with HT volumes. This might have injurious effects on cerebral oxygenation and, consequently, on cerebral metabolism, because activation of complex inflammatory cascades can lead to early impairment of neuronal metabolism.50 Conversely, we suggest that there may be a link between an attenuated release of inflammatory mediators during lungprotective ventilation and the impact of neuroinflammatory processes. These data have potential implications for the clinical effects of mechanical ventilation in ARDS. The mechanism by which the mode of ventilation affects brain function should be investigated as a basis for developing possible treatment options for this scenario in clinical circumstances. REFERENCES
1. Kraft P, Friedrich P, Pernerstorfer T, Fitzgerald RD, Koc D, Schneider B, Hammerle AF, Steltzer H. The Acute Respiratory Distress Syndrome: definitions, severity, and clinical outcome: an analysis of 101 clinical investigations. Intensive Care Med 1996;22:519 29 2. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke KJ, Hudson L, Lamy M, Legall JR, Morris A, Spragg R. The Am-Eur Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes and clinical trial coordination. Am J Respir Crit Care 1994;149:818 24 3. The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301 8 4. Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schettino GP, Lorenzi-Filho G, Kairalla RA, Deheinzelin D, Munoz C, Oliveira R, Takagaki TY, Carvalho CR. Effect of a protectiveventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:34754 5. Hopkins RO, Weaver LK, Collingridge D, Parkinson BR, Chan KJ, Orme JF. Two-year cognitive, emotional, and quality-of-life outcomes in acute respiratory distress syndrome. Am J Respir Crit Care Med 2005;171:340 7 6. Hopkins RO, Jackson JC. Long-term neurocognitive function after critical illness. Chest 2006;30:869 78 7. Manns JR, Hopkins RO, Squire LR. Semantic memory and the human hippocampus. Neuron 2003;38:2733 8. Hopkins RO, Kesner RP, Goldstein M. Item and order recognition memory in subjects with hypoxic brain injury. Brain Cogn 1995;27:180 201 9. Hopkins RO, Kesner RP, Goldstein M. Memory for novel and familiar spatial and linguistic temporal distance information in hypoxic subjects. J Int Neuropsychol Soc 1995;1:454 68
10. Fries M, Bickenbach J, Henzler D, Beckers S, Dembinski R, Sellhaus B, Rossaint R, Kuhlen R. S-100 protein and neurohistopathologic changes in a porcine model of acute lung injury. Anesthesiology 2005;102:7617 11. Parsons PE, Eisner MD, Thompson BT, Matthay MA, Ancukiewicz M, Bernard GR, Wheeler AP; NHLBI Acute Respiratory Distress Syndrome Clinical Trials Network. Lower tidal volume ventilation and plasma cytokine markers of inflammation in patients with acute lung injury. Crit Care Med 2005;33:1 6 12. Frank JA, Parsons PE, Matthay MA. Pathogenetic significance of biological markers of ventilator-associated lung injury in experimental and clinical studies. Chest 2006;130:1906 14 13. Meduri GU, Headley S, Kohler G, Stentz F, Tolley E, Umberger R, Leeper K. Persistent elevation of inflammatory cytokines predicts a poor outcome in ARDS. Plasma IL-1 beta and IL-6 levels are consistent and efficient predictors of outcome over time. Chest 1995;107:106273 14. Wolthuis EK, Choi G, Dessing MC, Bresser P, Lutter R, Dzoljic M, van der Poll T, Vroom MB, Hollmann M, Schultz MJ. Mechanical ventilation with lower tidal volumes and positive end-expiratory pressure prevents pulmonary inflammation in patients without preexisting lung injury. Anesthesiology 2008;108:46 54 15. Ungerstedt U. Microdialysisprinciples and applications for studies in animal and man. J Intern Med 1991;230:36573 16. Arner P, Bolinder J. Microdialysis of adipose tissue. J Intern Med 1991;230:381 6 17. Muller M. Science, medicine, and the future: microdialysis. BMJ 2002;324:588 91 18. Magnoni S, Ghisoni L, Locatelli M, Caimi M, Colombo A, Valeriani V, Stochetti N. Lack of improvement in cerebral metabolism after hyperoxia in severe head injury: a microdialysis study. J Neurosurg 2003;98:952 8 19. Lachmann B, Robertson B, Vogel J. In vivo lung lavage as an experimental model of the respiratory distress syndrome. Acta Anaesthesiol Scand 1980;24:231 6 20. Dembinski R, Max M, Lopez F, Kuhlen R, Kurth R, Rossaint R. Effect of inhaled prostacyclin in combination with almitrine on ventilation-perfusion distributions in experimental lung injury. Anesthesiology 2001;94:461 8 21. Schulz MK, Wang LP, Tange M, Bjerre P. Cerebral microdialysis monitoring: determination of normal and ischemic cerebral metabolisms in patients with aneurysmal subarachnoid hemorrhage. J Neurosurg 2000;93:808 14 22. Hoffman WE. Measurement of intracerebral oxygen pressure: practicalities and pitfalls. Curr Opin Anaesthesiol 1999;12: 497502 23. Martinez-Tica JF, Berbarie R, Davenport P, Zornow MH. Monitoring PO2, PCO2, and pH during graded levels of hypoxemia in rabbits. J Neurosurg Anesthesiol 1999;11:260 3 24. Starr JL, Whalley LJ. Drug induced dementia. Drug Saf 1994;11:310 7 25. Gattinoni L, Caironi P, Carlesso E. How to ventilate patients with acute lung injury and acute respiratory distress syndrome. Curr Opin Crit Care 2005;11:69 76 26. Ranieri VM, Giunta F, Suter PM, Slutsky AS. Mechanical ventilation as a mediator of multisystem organ failure in acute respiratory distress syndrome. JAMA 2000;284:43 4 27. Imai Y, Parodo J, Kajikawa O, de Perrot M, Fischer S, Edwards V, Cutz E, Liu M, Keshavjee S, Martin TR, Marshall JC, Ranieri VM, Slutsky AS. Injurious mechanical ventilation and endorgan epithelial cell apoptosis and organ dysfunction in an experimental model of acute respiratory distress syndrome. JAMA 2003;289:2104 12 28. Zassler B, Weis C, Humpel C. Tumor necrosis factor-alpha triggers cell death of sensitized potassium chloride-stimulated cholinergic neurons. Brain Res Mol Brain Res 2003;113:78 85 29. Rolfe DF, Brown GC. Cellular energy utilization and molecular origin of standard metabolic rate in mammals. Physiol Rev 1997;77:73158 30. Rosengarten B, Hecht M, Auch D, Ghofrani HA, Schermuly RT, Grimminger F, Kaps M. Microcirculatory dysfunction in the brain precedes changes in evoked potentials in endotoxininduced sepsis syndrome in rats. Cerebrovasc Dis 2007;23:140 7 Vol. 109, No. 3, September 2009
31. Valtysson J, Persson L, Hillered L. Extracellular ischemia markers in repeated global ischemia and secondary hypoxaemia monitored by microdialysis in rat brain. Acta Neurochir 1998;140:38795 32. Ronne-Englstrom E, Carlson H, Liu Y, Ungerstedt U, Hillered L. Influence of perfusate glucose concentration on dialysate lactate, pyruvate, aspartate, and glutamate levels under basal and hypoxic conditions: a microdialysis study in rat brain. J Neurochem 1995;65:257 62 33. Landolt H, Langemann H. Cerebral microdialysis as a diagnostic tool in acute brain injury. Eur J Anaesthesiol 1996;13:269 78 34. Hillered L, Persson L, Nilsson P, Ronne-Englstrom E, Enblad P. Continuous monitoring of cerebral metabolism in traumatic brain injury: a focus on cerebral microdialysis. Curr Opin Crit Care 2006;12:112 8 35. Ungerstedt U, Rostami E. Microdialysis in neurointensive care. Curr Pharm Des 2004;10:214552 36. Zoremba N, Homola A, Rossaint R, Sykova E. Brain metabolism and extracellular space diffusion parameters during and after transient global hypoxia in the rat cortex. Exp Neurol 2007;203:34 41 37. Hutchinson PJ, Gupta AK, Fryer TF, Al-Rawi PG, Chatfield DA, Coles JP, OConnell MT, Kett-White R, Minhas PS, Aigbirhio FI, Clark JC, Kirkpatrick PJ, Menon DK, Pickard JD. Correlation between cerebral blood flow, substrate delivery, and metabolism in head injury: a combined microdialysis and triple oxygen positron emission tomography study. J Cereb Blood Flow Metab 2002;22:735 45 38. Hemphill JC III, Knudson MM, Derugin N, Morabito D, Manley GT. Carbon dioxide reactivity and pressure autoregulation of brain tissue oxygen. Neurosurgery 2001;48:377 83 39. Akca O, Doufas A, Morioka N, Iscoe S, Fisher J, Sessler DI. Hypercapnia improves tissue oxygenation. Anesthesiology 2002;97:801 6 40. OHara JA, Hou H, Demidenko E, Springett RJ, Khan N, Swartz HM. Simultaneous measurement of rat brain cortex PtO2 using EPR oximetry and a fluorescence fiber-optic sensor during normoxia and hyperoxia. Physiol Meas 2005;26:20313 41. Akca O, Liem E, Suleman M, Doufas AG, Galandiuk S, Sessler DI. Effect of intra-operative end-tidal carbon dioxide partial pressure on tissue oxygenation. Anaesthesia 2003;58:536 42 42. Iadecola C, Nedergaard M. Glial regulation of the cerebral microvasculature. Nat Neurosci 2007;10:1369 76 43. Bo ttiger BW, Mo bes S, Gla tzer R, Bauer H, Gries A, Bertsch P, Motsch H, Martin E. Astroglial protein S-100 is an early and sensitive marker of hypoxic brain damage and outcome after cardiac arrest in humans. Circulation 2001;103:2694 8 44. Fassbender K, Schmidt R, Schreiner A, Fatar M, Muhlhauser F, Daffertshofer M, Hennerici M. Leakage of brain-originated proteins in peripheral blood: temporal profile and diagnostic value in early ischemic stroke. J Neurol Sci 1997;148:1015 45. Romner B, Ingebrigtsen T, Kongstad P, Borgesen SE. Traumatic brain damage: serum S-100 protein measurements related to neuroradiological findings. J Neurotrauma 2000;17:6417 46. Fries M, Kunz D, Gressner AM, Rossaint R, Kuhlen R. Procalcitonin serum levels after out-of-hospital cardiac arrest. Resuscitation 2003;59:1059 47. Rosen H, Rosengren L, Herlitz J, Blomstrand C. Increased serum levels of the S-100 protein are associated with hypoxic brain damage after cardiac arrest. Stroke 1998;29:4737 48. Krieter H, Denz C, Janke C, Bertsch T, Luiz T, Ellinger K, Van Ackern K. Hypertonic-hyperoncotic solutions reduce the release of cardiac troponin I and S-100 after successful cardiopulmonary resuscitation in pigs. Anesth Analg 2002;95:1031 6 49. Routsi C, Stamataki E, Nanas S, Psachoulia C, Stathopoulos A, Koroneos A, Zervou M, Jullien G, Roussos C. Increased levels of serum S100B protein in critically ill patients without brain injury. Shock 2006;26:20 4 50. Semmler A, Hermann S, Mormann F, Weberpals M, Paxian SA, Okulla T, Scha fers M, Kummer MP, Klockgether T, Heneka MT. Sepsis causes neuroinflammation and concomitant decrease of cerebral metabolism. J Neuroinflammation 2008;5:38
855
Pressure Support Ventilation and Biphasic Positive Airway Pressure Improve Oxygenation by Redistribution of Pulmonary Blood Flow
Alysson R. Carvalho, PhD* Peter M. Spieth, MD* Paolo Pelosi, MD, PhD Alessandro Beda, PhD* Agnaldo J. Lopes, MD, DSc Boriana Neykova, MD Axel R. Heller, MD, PhD* Thea Koch, MD, PhD* Marcelo Gama de Abreu, MD, MSc, PhD, DEAA*
BACKGROUND: Spontaneous breathing (SB) activity may improve gas exchange during mechanical ventilation mainly by the recruitment of previously collapsed regions. Pressure support ventilation (PSV) and biphasic positive airway pressure (BIPAP) are frequently used modes of SB, but little is known about the mechanisms of improvement of lung function during these modes of assisted mechanical ventilation. We evaluated the mechanisms behind the improvement of gas exchange with PSV and BIPAP. METHODS: Five pigs (2529.3 kg) were mechanically ventilated in supine position, and acute lung injury (ALI) was induced by surfactant depletion. After stabilization, BIPAP was initiated with lower continuous positive airway pressure equal to 5 cm H2O and the higher continuous positive airway pressure titrated to achieve a tidal volume between 6 and 8 mL/kg. The depth of anesthesia was reduced, and when SB represented 20% of total minute ventilation, PSV and BIPAP SB were each performed for 1 h (random sequence). Whole chest helical computed tomography was performed during end-expiratory pauses and functional variables were obtained. Pulmonary blood flow (PBF) was marked with IV administered fluorescent microspheres, and spatial cluster analysis was used to determine the effects of each ventilatory mode on the distribution of PBF. RESULTS: ALI led to impairment of lung function and increase of poorly and nonaerated areas in dependent lung regions (P 0.05). PSV and BIPAP SB similarly improved oxygenation and reduced venous admixture compared with controlled mechanical ventilation (P 0.05). Despite that, a significant increase of nonaerated areas in dependent regions with a concomitant decrease of normally aerated areas was observed during SB. In five of six lung clusters, redistribution of PBF from dependent to nondependent, better aerated lung regions were observed during PSV and BIPAP SB. CONCLUSIONS: In this model of ALI, the improvements of oxygenation and venous admixture obtained during assisted mechanical ventilation with PSV and BIPAP SB were explained by the redistribution of PBF toward nondependent lung regions rather than recruitment of dependent zones.
(Anesth Analg 2009;109:856 65)
ontrolled mechanical ventilation may be necessary to achieve adequate gas exchange and reduce the work of breathing in patients with acute lung injury (ALI).13 However, clinical and experimental studies have shown that spontaneous breathing (SB) activity
From the *Clinic of Anesthesiology and Intensive Care Medicine, University Clinic Carl Gustav Carus, Dresden, Germany; University Augusto Motta, Rio de Janeiro, Brazil; Department of Ambient, Health and Safety, University of Insubria, Varese, Italy; and Institute of Radiology, University Clinic Carl Gustav Carus, Technical University Dresden, Dresden, Germany. Accepted for publication May 4, 2009. Supported, in part, by a research grant from the European Society of Anaesthesiology, Brussels, Belgium. Address for correspondence and reprint requests to Dr. Marcelo Gama de Abreu, Clinic of Anesthesiology and Intensive Care Medicine, University Clinic Carl Gustav Carus, Technical University Dresden, Fetscherstr. 74, 01307 Dresden, Germany. Address e-mail to mgabreu@uniklinikum-dresden.de. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aff245
may improve gas exchange and lung function during mechanical ventilation, as well as reduce the need for sedation, cardio-circulatory drug support,4 6 and even mitigate atrophy of diaphragm myofibers.7 Controlled mechanical ventilation with deep sedation and/or muscle paralysis also modifies the displacement pattern of the diaphragm.8,9 The unopposed increase of intraabdominal pressure may reduce the transpulmonary pressure in dependent lung regions, promoting lung collapse5,8 and redistributing tidal ventilation toward nondependent regions.10,11 In this context, SB activity may contribute to restoring the physiological displacement of the diaphragm, improving regional ventilation of dependent lung regions12 and enhancing hemodynamics through decreased intrathoracic pressure.2,1315 The recruitment of dorsal and usually better perfused lung regions through inspiratory efforts is considered the main mechanism behind the improvement in gas exchange during
856
Figure 1. Time course of interventions. Therapy 1 and 2 correspond to the sequence of spontaneous breathing (pressure support
ventilation [PSV] or biphasic positive airway pressure [BIPAP] spontaneous breathing [SB]) after the randomization.
assisted ventilation,5,8,12 but this concept was challenged recently.16 Pressure support ventilation (PSV) and biphasic positive airway pressure with SB (BIPAP SB), both characterized by decelerating inspiratory flow patterns, are frequently used modes of mechanical ventilation in the clinical arena. Although PSV supports every triggered breath with positive pressure, BIPAP allows SB without support at two different airway pressure levels. This may lead to enhanced ventilation and perfusion of dependent lung regions, making BIPAP SB superior to PSV.3,17 However, more recent works have suggested that PSV and BIPAP SB similarly improve gas exchange when compared with controlled mechanical ventilation in experimental ALI.16,18 The main objective of this study was to identify the mechanisms behind the improvement in lung function during PSV and BIPAP SB. For this purpose, we assessed the distribution of lung aeration and the spatial distribution of changes in pulmonary perfusion, so-called clusters, during PSV and BIPAP SB in experimental ALI. We hypothesized that redistribution of pulmonary blood flow (PBF) toward better aerated regions plays an important role in the improvement of oxygenation during PSV and BIPAP SB.
during inspiration, taking the first value at the beginning of the inspiratory cycle as offset. Respiratory drive (P0.1) was assessed as the difference between Paw at the beginning of inspiration and 100 ms thereafter.16 Mean systemic and pulmonary arterial pressures as well as central venous and pulmonary artery wedge pressures were measured. Cardiac output, venous admixture, and oxygen delivered and consumption were calculated using standard formulas.20
Computed Tomography
Helical computed tomography (CT) scans of the whole lung were obtained with a Somatom Sensation 16 (Siemens, Erlangen, Germany) during controlled mechanical ventilation, before and after the induction of ALI, as well as during assisted mechanical ventilation with PSV and BIPAP SB. Scans were obtained during breath-hold maneuvers at end-expiratory pressure and with simultaneous clamping of the endotracheal tube. The CT scanner was set as follows: collimation, 16 0.75 mm; pitch, 1.35; bed speed, 38.6 mm/s; voltage, 120 kV; and tube current-time product, 120 mAs. Images were reconstructed with slices of 5 mm thickness, without gaps between slices, yielding images with 512 512 pixels with a surface of 0.443 0.443 mm2 (voxel size 0.98 mm3). The radiograph attenuation of each pixel, expressed in Hounsfield units (HU), was primarily determined by the density (mass/volume ratio) of the tissue and expressed as the CT number, i.e., CT/(1000) (volume of gas/[volume of gas volume of tissue]). The attenuation scale arbitrarily assigns to bone a value of 1000 HU (complete absorption), to air a value of 1000 HU (no absorption), and to water a value of 0 HU; blood and lung tissue have a density ranging between 20 and 40 HU.21,22 After manual segmentation of the region of interest, images were analyzed for calculation of total lung and total gas volumes as well as total lung mass and percentages of hyperaerated (1000 to 900 HU), normally aerated (900 to 500 HU), poorly aerated (500 to 100 HU), and nonaerated (100 to 100 HU) compartments in total lung volume, as suggested elsewhere.21,22 The lung volume (i.e., the sum of gas plus tissue volume) was calculated as follows: ([size of the pixel]2 slice thickness total number of pixels of the region of interest for the whole lung). Weight of
METHODS
The protocol of this study was approved by the local animal care committee and the Government of the State of Saxony, Germany. Figure 1 shows the sequence of interventions performed, which are described in detail in this section. Five female pigs were anesthetized and mechanically ventilated with volume-controlled mode with inspiratory square flow waveform using a tidal volume (VT) 12 mL/kg, I:E 1:1, inspired oxygen fraction (Fio2) 0.5, positive end-expiratory pressure 5 cm H2O, and respiratory rate (RR) set to achieve an arterial pressure of carbon dioxide (Paco2) between 30 and 45 mm Hg.
Measurements of Functional Variables

Airway (Paw) and esophageal (Pes) pressures as ) were continuously recorded.19 The well as airflow (V product of Pes versus time (PTP) was calculated
857
lungs was calculated as: ([1 CT/1000] [size of the pixel]2 slice thickness total number of pixels of the region of interest for the whole lung).22 To assess the regional distribution of aeration in noninjured and injured lungs as well as in controlled and assisted ventilation, the whole lung was divided into 10 zones in the cranio-caudal and also the ventrodorsal axes. The median CT-scan attenuation and the volume of each zone were calculated. Three-dimensional volume meshes were created by masking the lung boundary of each compartment with a routine written in Matlab (MathWorks) by one of the authors (ARC). Color mapping was used to represent lung aeration compartments.
Distribution of PBF
Regional PBF was marked with IV administered fluorescent, color-labeled 15-m diameter microspheres before and after induction of lung injury in controlled mechanical ventilation as well as during assisted mechanical ventilation with PSV and BIPAP SB. The colors used were blue-green, carmine, crimson, red, orange, scarlet, and yellow-green. A different color was assigned randomly and administered at each timepoint to mark regional perfusion under each experimental condition. Immediately before injection, the microspheres were vortexed, sonicated for 90 s, and drawn into a 2-mL syringe. All injections were performed over 60 s to average blood flow over several cardiac cycles and VTs. During injection, approximately 1.5 106 microspheres were administered. Postmortem processing of lungs was performed as previously described.23 Briefly, lungs were flushed with 50 mL/kg of a hydroxyethyl starch 130/0.4 solution (Voluven, Fresenius Kabi, Bad Homburg, Germany) and air-dried by continuous tracheal airflow for 7 days (continuous pressure of 25 cm H2O). The lungs were then coated with a one-component polyurethane foam (BTI Befestigungstechnik, Ingelfingen, Germany), suspended vertically in a square box, and embedded in rapidly setting urethane foam (polyol and isocyanate, Elastogran, Lemfo rde, Germany). The foam block was cut into cubes of 1.3 cm3. Each cube was weighed and assigned a threedimensional coordinate. The samples were then soaked for 2 days in 2 mL of 2-ethoxyethyl acetate (Aldrich Chemical, Milwaukee, WI) to retrieve the fluorescent dye. The fluorescence was read in a luminescence spectrophotometer (LS-50B; Perkin-Elmer, Beaconsfield, UK). The measured intensity of fluorescence in each probe was normalized according to its own weight using Eq. 1:
probes. The mean normalized relative flow was therefore 1.0. The distribution of PBF along the cranio-caudal and ventro-dorsal axes under each experimental condition was assessed by means of linear regression. Additionally, a three-dimensional reconstruction of the lung was performed, considering the spatial coordinates of each lung piece and the PBF at each of the x, y, and z coordinates. Color mapping was used to identify the rel,i. The color regional distribution of PBF based on Q rel under map was then normalized by the maximum Q each experimental condition, resulting in a color scale ranging from dark blue (0.0, lowest perfusion) to dark red (1.0, highest perfusion). Thereafter, the lungs were divided into 10 zones of equal heights along the cranio-caudal and ventro-dorsal axes, as described, for rel CT analysis. The relative blood flow content (Q content) of each zone was calculated taking the sum of rel,i in that zone divided by the sum of Q rel in the Q whole lung. The volume of each zone was also computed by the simple arithmetical sum of each lung piece in that zone. The spatial representation of the lung volume meshes and the distribution of PBF were obtained using a routine written in Matlab (MathWorks) by one of the authors (ARC).
After instrumentation, animals were allowed to stabilize for 15 min (baseline). ALI was induced by repetitive lung lavage until the Pao2/Fio2 ratio decreased to 200 mm Hg and did not spontaneously recover during a 30-min period.24 The endotracheal tube was disconnected from the ventilator and warmed isotonic saline solution (30 mL/kg, 37C 39C) was instilled (height of approximately 40 cm above the endotracheal tube). After that, the fluid was retrieved passively by gravity drainage. Further lavages were performed if Pao2/Fio2 ratio exceeded 200 mm Hg. The countdown of 30 min was restarted in this case. After injury stabilization, the ventilator was switched to BIPAP with lower continuous positive airway pressure (CPAPlow) 5 cm H2O, higher CPAP (CPAPhigh) titrated to obtain VT between 6 and 8 mL/kg, and RR to achieve Paco2 between 50 and 60 mm Hg without SB. The depth of anesthesia was reduced (0.51.5 mg kg1 h1, midazolam; 4 6 mg kg1 h1, ketamine; and 0.1 0.3 g kg1 h1, remifentanil), and when SB represented more than 20% of total minute ventilation, the ventilatory mode was switched to PSV or BIPAP SB. During PSV, pressure support was set to obtain VT between 6 and 8 mL/kg, the flow trigger was 2.0 L/min, and the cycling-off criteria was 25% of peak flow. During BIPAP SB, the I:E ratio was adjusted to obtain mean Paw in the range of 8 10 cm H2O to permit comparability with the PSV mode. The animals lungs were then ventilated for a period of 1 h with each mode (random sequence).
rel,i x i/ Q
x / n
i
(1)
rel,i is the weight-normalized relative PBF of where Q the probe i; xi is the obtained fluorescence divided by the weight of the probe i, and n is the number of
858
PSV and BiPAP Spontaneous Breathing Redistribute Pulmonary Blood Flow
Table 1. Respiratory Variables Controlled ventilation (square inspiratory ow pattern) Baseline

MV Total (L/min) Controlled (L/min) Spontaneous (L/min) VT Total (mL) Controlled (mL) Spontaneous (mL) RR Total (/min) Controlled (/min) Spontaneous (/min) Ppeak Total (cm H2O) Controlled (cm H2O) Paw mean (cm H2O) P0.1 (cm H2O) PTP (cm H2O s min1) 4.5 (3.36.8) 4.5 (3.36.8) 324.7 (320.7333.4) 324.7 (320.7333.4) 14(1020) 14 (1020) 18.9 (18.819.2) 18.9 (18.819.2) 10.7 (10.510.8)
Assisted ventilation (decelerating inspiratory ow pattern) PSV

6.8 (5.87.1) 160.1 (130.4165.1) 35 (3152) 22.1 (21.822.5) 7.8 (7.68.9) 0.43 (0.40.6) 30.8 (15.339.7)
Injury
3.6 (3.34.6) 3.6 (3.34.6) 332.0 (313.5333.2) 332.0 (313.5333.2) 12 (1013) 12 (1013) 31.7* (30.132.6) 31.7* (30.132.6) 14.9* (13.315.3)
BIPAP SB
7.1 (4.97.3) 1.5 (1.31.6) 5.3 (3.65.6) 109.9 (91.0113.4) 157.8 (141.4169.2) 93.2 (86.5109.4) 55 (4966) 10 (612) 45 (3960) 10.5 (9.811.6) 22.3 (22.022.5) 9.4 (9.19.8) 4.9 (2.25.2) 168.6 (106.4192.2)
Values provided as median (interquartile ranges). PSV pressure support ventilation; BIPAP SB biphasic positive airway pressure spontaneous breathing; MV minute ventilation; VT tidal volume; RR respiratory rate; Ppeak peak airway pressure; Paw mean mean airway pressure; P0.1 airway pressure gradient measured at 100 ms after start of inspiration; PTP inspiratory pressure time product of esophageal pressure; , not applicable. * P 0.05 versus baseline; P 0.05 versus injury; P 0.05 versus PSV.
CT-scan images, fluorescent microspheres injection, and functional variables measurements were performed under each experimental condition (Fig. 1). Animals were then killed by IV administration of 2 g of thiopental and 50 mL of KCl (1 M), and lungs were extracted for determination of PBF distribution.
Postmortem Lung Processing

Lungs were air-dried at 25 cm H2O for 7 days and cut into pieces of approximately 1.33 mm3. The intensity of fluorescence in each piece was measured and normalized for its own weight.23 A three-dimensional reconstruction of the lung was performed, considering the spatial coordinates of each lung piece and the PBF at each of the x, y, and z coordinates. Thereafter, lungs were divided into 10 zones along the cranio-caudal and ventro-dorsal axes and the relative blood flow content of each zone was calculated.
whether the pattern of change is common across animals. To verify if a single dominant animal influenced the clustering processing in the meta-cluster, the amount of each animals pieces in each respective cluster was calculated. Because five animals were studied, we expected that the ideal cluster from the meta-cluster procedure should be composed of a fraction of 20% of pieces per animal. However, values between 20% 10% were accepted. The clusters were created without reference to the spatial location of pieces within the lung. To display and assess spatial clustering using data from metaclusters, a meta-lung was created by linear transformations (stretching and compressing) along the x, y, and z axes of the coordinates for each animal, as previously proposed.25
Statistical Analysis Cluster Analysis

For each animal, lung pieces sharing similar characteristics of change in PBF were grouped into clusters. In addition, meta-cluster analysis was used to identify stereotypical changes in PBF that were common to all animals.25 For this purpose, the pieces from all animals were merged into one dataset and the nonhierarchical clustering method was used to identify clusters that have the same pattern across all animals. Accordingly, clustering of pieces from individual animals shows whether changes in flow over each experimental condition occur in different parts of the lung, and the meta-clustering demonstrates
Values are presented as median and interquartiles (1st quartile3rd quartile). Comparisons were performed using Wilcoxons test for paired data (Software package SPSS, version 12.0, Chicago, IL), and the Bonferroni-Holm procedure was used to adjust for multiple tests. Statistical significance was set at P 0.05.
RESULTS
Surfactant depletion increased Paw (Table 1), impaired oxygenation, and worsened venous admixture (Table 2). Furthermore, increases in total lung volume,
859
Table 2. Gas Exchange, Hemodynamics, and Oxygen Transport/Consumption Controlled ventilation (square ow waveform) Baseline
Gas exchange Pao2/Fio2 (mm Hg) VA/ Q t (%) Q Paco2 (mm Hg) Hemodynamics CO (L/min) HF (/min) MAP (mm Hg) MPAP (mm Hg) CVP (mm Hg) PCWP (mm Hg) PVR (dyn s cm5) SVR (dyn s cm5) Oxygen transport and consumption O2 (mL/min) D O2 (mL/min) V 533.2 (517.6536.8) 5.5 (5.16.9) 33.1 (30.936.7) 2.5 (2.23.2) 72 (7276) 77 (6990) 22 (1623) 10 (712) 13 (1214) 260.5 (125.8288) 1994.2 (1711.62414.4) 356.8 (281.6391.4) 105.3 (86.4138.2)
Assisted ventilation (decelerating ow pattern) PSV

262.8 (243.2290.4) 16.7 (9.817.4) 55.0 (49.655.8) 2.8 (2.73.3) 86 (79103) 76 (7183) 30 (2935) 8 (710) 12 (1114) 407.2 (388.1690.6) 1628.7 (1552.22303.0) 419.0 (326.9465.1) 153.4 (124.5158.5)
Injury
126.0* (102.0141.4) 26.1* (25.233.1) 37.2 (35.437.8) 2.0* (2.03.0) 67 (5779) 69 (6974) 30* (2832) 10 (1011) 15 (1316) 627.5* (333.3800.0) 1594.6 (1547.62313.7) 274.4* (237.3287.3) 135.6 (106.7142.7)
BIPAP SB
218.8 (198.2229.4) 19.2 (13.519.7) 66.2 (65.968.0) 3.3 (3.23.9) 96 (91113) 69 (6876) 35 (3036) 7 (78) 12 (812) 509.1 (456.0573.3) 1285.0 (1236.41719.9) 450.6 (381.1468.5) 104.0 (99.7107.7)
Values provided as median (interquartile ranges). PSV pressure support ventilation; BIPAP SB biphasic positive airway pressure spontaneous breathing; PaO2/FIO2 ratio between arterial pressure of oxygen and inspired oxygen VA/ Q t venous admixture; PaCO2 arterial pressure of carbon dioxide; CO cardiac output; HF heart frequency; MAP mean arterial blood pressure; MPAP mean pulmonary fraction; Q O2 oxygen delivery; arterial pressure; CVP central venous pressure; PCWP pulmonary artery wedge pressure; PVR pulmonary vascular resistance; SVR systemic vascular resistance; D O2 oxygen consumption. V * P 0.05 versus baseline; P 0.05 versus injury; P 0.05 versus PSV.
Table 3. Computed Tomography Data Controlled ventilation (square inspiratory ow pattern) Baseline
Total lung volume (mL) Total gas volume (mL) Total lung mass (g) Hyperinflated (% Vol) Normally aerated (% Vol) Poorly aerated (% Vol) Nonaerated (% Vol) 1015.1 (809.81161.9) 718.4 (518.9800.5) 296.7 (290.9361.5) 3.0 (1.43.5) 81.9 (77.083.1) 11.5 (10.814.0) 2.6 (2.43.3)
Assisted ventilation (decelerating inspiratory ow pattern) PSV

797.7 (706.6974.1) 320.5 (237.0324.4) 560.7 (382.2566.9) 0.9 (0.52.4) 42.4 (33.049.0) 29.5 (24.329.8) 32.7 (25.234.6)
Injury
1025.9* (947.21295.5) 621.0 (524.7753.2) 495.5* (422.4542.3) 2.3 (1.73.1) 63.3* (61.269.0) 20.0 (17.522.8) 11.5* (10.414.6)
BIPAP SB
832.2 (825.1948.2) 310.7 (290.7436.0) 522.8 (389.1541.5) 0.9 (0.82.1) 43.5 (34.954.6) 24.6 (20.226.1) 28.7 (25.329.9)
Values provided as median (interquartile ranges). Hyperaerated, normally aerated, poorly aerated, and nonaerated compartments were computed according to references.21,22 PSV pressure support ventilation; BIPAP SB biphasic positive airway pressure spontaneous breathing. * P 0.05 versus baseline; P 0.05 versus injury; P 0.05 versus PSV.
total lung mass, and volume of nonaerated areas were observed (Table 3). As shown in Table 1, total minute ventilation increased with both modes of assisted mechanical ventilation (P 0.05). The reduction of VT during PSV and BIPAP SB was accompanied by an increase in the total RR (P 0.05). Additionally, a significant reduction of total gas volume and the volume of normally aerated areas were observed (P 0.05, Table 3). Despite this, oxygenation increased and venous admixture decreased with both modes of assisted compared with controlled mechanical ventilation (P 0.05), whereas P0.1 and PTP were higher during BIPAP SB than PSV. No significant differences between PSV and BIPAP SB were observed with regard to gas exchange and distribution of aeration (Table 3).
860
Regional Distribution of Aeration and PBF

Figure 2 shows the three-dimensional representation of the distribution of lung aeration and PBF for one representative animal. At baseline, nonaerated areas were constrained to the most caudal regions. After induction of ALI, nonaerated areas increased in the dorsal lung zones, whereas normally aerated areas could be observed mainly in ventral and cranial regions. During PSV and BIPAP SB, the amount of nonaerated areas in dorsal parts of the lungs increased further (Figs. 2 and 3A, left panels). Similarly, after ALI, PBF shifted from caudal and dorsal to cranial and ventral lung regions, respectively (P 0.05). During PSV and BIPAP SB, PBF further decreased in dorsal and caudal areas, but no differences were observed between modes (Figs. 2 and 3A, right panels).
Figure 2. Three-dimensional representation of the distribution of aeration assessed by static computed tomography (CT), and the spatial distribution of weight-normalized relative pulmonary blood flow (PBF) in one representative animal. CT images were obtained at the end of expiration. The x, y, and z axes represent the spatial orientation of lungs. Two different projections are shown: the upper row presents a frontal plan projection and the lower row presents the same lungs rotated by 120 around the vertical axis (z). Red represents hyperaerated (1000 to 900 Hounsfield units [HU]); blue, normally aerated (900 to 500 HU); gray, poorly aerated (500 to 100 HU); and pink, nonaerated areas (100 to 100 HU). Color mapping was used to illustrate PBF, with the color intensity normalized by the maximum PBF at each experimental condition. The normalized color bar is presented with dark blue and red, representing the lowest and highest perfusion levels, respectively. Note the increase of poorly and nonaerated areas from the caudal to cranial and dorsal to ventral areas after induction of acute lung injury, as well as during assisted spontaneous breathing with pressure support ventilation (PSV) and with biphasic positive airway pressure and spontaneous breathing (BIPAP SB). Also note the redistribution of PBF toward better aerated cranial and ventral lung regions.
Cluster Analysis
Figure 4 shows the results of the cluster analysis for the same animal. Lung pieces could be grouped into six main clusters in that animal (Fig. 4A, left panels). Cluster A presented an almost constant distribution of PBF. Cluster B presented a pattern of increase of PBF through all experimental interventions, whereas Cluster C presented the opposite pattern. Cluster D presented a decrease of PBF from controlled to assisted ventilation with PSV and BIPAP SB, whereas Cluster F presented the opposite pattern. Cluster E presented a pattern of increase of PBF from baseline to injury, with a decrease of PBF from injury to PSV and BIPAP SB. The spatial distribution of clusters (Fig. 4A, right panel) evidenced a reduction of PBF in caudal and dorsal regions between baseline and injury (Cluster C), with a concomitant increase in the cranial and ventral regions (Clusters B and E). During PSV and BIPAP SB, a further reduction of PBF in caudal and dorsal regions was observed (Clusters C, D, and E), as well as an increase in PBF to cranial and ventral regions (Clusters B and F). Figure 4B shows the six clusters of the meta-cluster analysis (Fig. 4B, left panel). Cluster A presented an almost constant pattern of PBF under all experimental conditions. Cluster B presented an increase of PBF after SB was resumed, whereas Cluster F presented the opposite pattern. Cluster C presented a decrease of PBF throughout the experimental conditions, whereas Cluster D presented the reverse pattern. Cluster F presented an increase of PBF from baseline to injury and a decrease of PBF from controlled
ventilation to SB. Figure 4B (middle panels) shows the percentages of the number of pieces of each animal related to the total amount of pieces of the meta-lung. Note that the clustering process was quite representative of the overall behavior. Figure 4B (right panels) shows the spatial distribution of meta-clusters. As can be observed, the clusters that presented a reduction in PBF were located in dorsal and caudal regions, whereas the clusters that presented an increase in PBF were located in ventral and cranial regions of lungs.
DISCUSSION
The main findings of this study were that: 1) in a surfactant depletion model of ALI, PSV and BIPAP SB led to similar improvement in oxygenation and reduction in venous admixture compared with controlled mechanical ventilation; 2) P0.1 and PTP were higher with BIPAP SB than with PSV; 3) neither PSV nor BIPAP SB resulted in increased aeration of dependent lung regions compared with controlled mechanical ventilation; 4) during PSV and BIPAP SB, pulmonary perfusion shifted from dependent to nondependent lung regions.
Regional Distribution of Aeration and PBF During Controlled Mechanical Ventilation

Controlled mechanical ventilation with muscle paralysis modifies the displacement pattern of the diaphragm.8,9 Accordingly, normally aerated and hyperaerated areas are usually located in nondependent regions, whereas poorly aerated and nonaerated areas
861
Figure 3. Left column: Regional distribution of computed tomography (CT) attenuation, expressed in Hounsfield units (HU),
in 10 zones of lungs along the cranio-caudal (upper panels) and ventro-dorsal (lower panels) axes at Baseline and Injury (A) as well as at institution of pressure support ventilation (PSV) and biphasic positive airway pressure (BIPAP), respectively (B). The horizontal lines marked the ranges for each compartment in the CT attenuation plot. Symbols represent the median at each measurement condition, with circles, triangles, stars, and squares representing Baseline, Injury, assisted ventilation with PSV, and assisted ventilation with BIPAP, respectively. Right column: Regional distribution of the weight-normalized relative pulmonary blood flow content (PBF) in 10 zones along the cranio-caudal (upper panels) and ventro-dorsal (lower panels) axes. Vertical bars represent 1st and 3rd quartiles. *P 0.05, Baseline versus Injury (controlled ventilation).
are seen in dependent lung regions.10 Our results are in agreement with these observations. The meta-cluster analysis showed a consistent pattern of changes in regional distribution of PBF across animals. Although PBF decreased in dorsal regions, an increase of blood flow in ventral regions occurred. However, no differences in the regional distribution of PBF were observed in more central parts of the lungs after induction of ALI, even though these regions presented large amounts of poorly or nonaerated lung tissue. The reduced but still present blood flow through the dorsal and caudal regions likely explains the increased venous admixture and the reduced Pao2/Fio2 ratio under controlled ventilation after induction of ALI.
Regional Distribution of Aeration and PBF During PSV and BIPAP SB

SB activity has been proposed as an alternative to improve respiratory function and reduce sedation and circulatory drug support during ALI.3 Although it
862
cannot be considered a gold standard, PSV is the most frequently used form of assisted mechanical ventilation in the clinical setting.26 During PSV, every flow- or pressure-triggered breath is assisted with positive pressure, helping to unload the respiratory muscles and to reduce the work of breathing, as well as to enhance the synchrony between subject and mechanical ventilator. However, depending on the settings of the mechanical ventilator, excessive unloading of respiratory muscles can also lead to loss of movement of the diaphragm, especially in the posterior muscular sections, which would reduce the transpulmonary pressure in dependent lung regions and mimic-controlled ventilation.5,8,11 On the other hand, BIPAP SB supports inspiration only if it begins simultaneously with the change from CPAPlow to CPAPhigh, but nonsupported breathing is possible in both levels. Although nonsupported breaths are associated with increased work of breathing, they may be useful to improve ventilation of
1, at each experimental intervention (B baseline; I injury; PSV pressure support ventilation; Bi biphasic positive airway pressure, left panels). Changes in flow for a piece at each specified experimental condition was calculated as the rel at that condition and the mean Q rel throughout all experimental interventions. Six patterns of difference between the Q changes in pulmonary blood flow (clusters) were identified by the nonhierarchical analysis. Each cluster is depicted with different colors (left panels, Clusters AF). The number of pieces in each cluster and the percentage of each cluster in the total lung are also exhibited above each panel. Right: Three-dimensional representation of the spatial distribution of the six color-coded clusters. B, Clusters identified through the nonhierarchical clustering method throughout all experimental conditions in all animals (meta-cluster analysis). The number of pieces in each cluster and the percentage of each cluster in the total meta-lung are also exhibited above each panel. The bold lines represent the mean value of the respective color-coded cluster. Middle panels: Percentage (number of occurrences) of pieces from each animal (A1A5) in the respective clusters; horizontal lines mark the range of 20% 10%. Right: Three-dimensional representation of the spatial distribution of the six color-coded clusters (meta-lung). Two different projections are shown. The upper right panel presents a frontal plan projection and the lower right panel presents the same picture rotated by 120 along the vertical axis (z). The x, y, and z axes represent the spatial orientation of the meta-lung.
Figure 4. A, Patterns of changes in the residuals of the weight-normalized pulmonary blood flow of the pig presented in Figure
dependent lung areas by means of contraction of the posterior muscular sections of the diaphragm and also increase lung perfusion through reduction of intrathoracic pressure. Accordingly, different authors have suggested that BIPAP SB leads to an improvement in lung function when compared with PSV, as a consequence of the better matching of ventilation/ perfusion ratio in dependent regions.17,18 However, the distribution of PBF is influenced by different factors, such as gravity, the fractal structure of pulmonary capillaries, trans-capillary pressure gradients, and, most importantly, hypoxic pulmonary vasoconstriction.2730 Our findings are in line with previous studies showing that PSV and BIPAP SB improve oxygenation and reduce venous admixture compared with controlled ventilation.16 18 Additionally, our results
showed a significant increase in the respiratory drive and in PTP, which reflects the respiratory effort and the oxygen consumption of the respiratory muscles,31 during BIPAP SB compared with PSV. Thus, we expected that the respiratory muscles, and more particularly the diaphragm, would generate higher transpulmonary pressures in dependent zones during BIPAP SB, resulting in recruitment and improved ventilation/perfusion matching in those zones. However, the improvement in pulmonary function could not be attributed to increased end-expiratory lung gas volumes or normally aerated areas. This result is somewhat surprising, given that the surfactant depletion model is believed to be a very recruitable model of ALI.32,33 However, we cannot exclude the possibility that SB may have influenced the distribution of
863
regional ventilation or that intratidal lung recruitment occurred. It is worth noting that the redistribution of intrapulmonary gas with PSV and BIPAP SB was followed by the same pattern of the redistribution of PBF. Two clusters representing 33% of the total lung tissue and situated in ventral areas revealed an increase of PBF with PSV and BIPAP SB. In contrast, two other clusters representing 14% of lung tissue and situated in dorsal areas showed a decrease of PBF. These findings support the hypothesis that improved regional aeration/perfusion matching occurred during assisted ventilation with PSV and BIPAP SB. On the other hand, the regional distribution of aeration and PBF did not differ between PSV and BIPAP SB, as we expected. The most likely explanation for redistribution of PBF toward nondependent lung regions is that peak and mean Paw were lower during PSV and BIPAP SB than during controlled mechanical ventilation. Thus, lung capillaries in nondependent lung zones probably had lower impedance to perfusion, allowing more efficient hypoxic pulmonary vasoconstriction in dependent areas. Because both PSV and BIPAP SB used decelerating inspiratory flow, whereas controlled ventilation used square inspiratory flow, we also cannot exclude that redistribution of airway-alveolar flow improved ventilation in nondependent zones during assisted compared with controlled mechanical ventilation. In addition, it is possible that the rapid shallow breathing pattern and hypercapnia observed during assisted mechanical ventilation further contributed to a shift of ventilation to nondependent zones. In contrast to our study, other authors have reported increased oxygenation and reduced venous admixture with BIPAP SB compared with PSV.17 This is possibly explained by the fact that those authors did not match assisted ventilation modes for both mean Paw and minute ventilation simultaneously, as performed in this work. Our findings also differ from the clinical reports by Cereda et al.34 and Putensen et al.17 One possible explanation is that patients investigated in those studies may have presented more severe lung injury. Another possible reason is that patients with ALI/acute respiratory distress syndrome may exhibit blunted hypoxic pulmonary vasoconstriction35 because of the inhibitory effects of proinflammatory mediators, nitric oxide, and endotoxin itself.36 39
hypoxic pulmonary vasoconstriction, may play a pivotal role in improvement of oxygenation during assisted mechanical ventilation. Consequently, lack of improvement of Pao2 after resuming SB in mechanically ventilated patients may suggest impairment of hypoxic pulmonary vasoconstriction and/or diffuse loss of lung aeration.
Limitations
This study has several limitations. First, we used a relatively low positive end-expiratory pressure level (5 cm H2O), which possibly precluded the more dependent alveolar units from being kept open at the end of expiration, even if they may have opened during inspiration (intratidal recruitment). Second, we did not use a crossover design for controlled and assisted ventilation, which may have somewhat biased our analysis. We decided for the fixed sequence of controlled followed by assisted mechanical ventilation because it more closely reflects clinical practice. Moreover, if we had included controlled ventilation in the randomization, redistribution of PBF during PSV and BIPAP SB could have contaminated controlled ventilation (carryover effect). Third, we evaluated only immediate physiological effects of PSV and BIPAP SB. Thus, we cannot exclude that these modes may lead to recruitment of dependent lung zones over the long term. Fourth, depth of sedation was increased during controlled compared with assisted mechanical ventilation. However, the drugs we used, namely midazolam, ketamine and remifentanil, seem not to influence the tonus of pulmonary vasculature.40,41 Fifth, it must be kept in mind that our evaluation was performed in a model of ALI that does not reproduce all features of the more complex human ALI/acute respiratory distress syndrome. Because PSV and BIPAP SB seem to be more likely to result in improved lung function when the hypoxic vasoconstriction reflex is preserved, extrapolation of our results to the clinical scenario must consider this fact. We conclude that in a surfactant depletion model of ALI, PSV and BIPAP SB improve oxygenation and reduce venous admixture to similar extents compared with controlled ventilation. Such effects are better explained by redistribution of lung perfusion toward nondependent lung zones than recruitment of dependent lung regions. REFERENCES
Possible Clinical Implications

Our findings contribute to further understand the physiological mechanisms leading to improvement of Pao2 and reduction of venous admixture when switching from controlled to assisted spontaneous mechanical ventilation. According to our data, redistribution of PBF from dependent toward nondependent, better aerated regions, which is closely related to preserved
864
1. Hering R, Zinserling J, Wrigge H, Varelmann D, Berg A, Kreyer S, Putensen C. Effects of spontaneous breathing during airway pressure release ventilation on respiratory work and muscle blood flow in experimental lung injury. Chest 2005;128:2991 8 2. Putensen C, Hering R, Muders T, Wrigge H. Assisted breathing is better in acute respiratory failure. Curr Opin Crit Care 2005;11:63 8 3. Putensen C, Muders T, Varelmann D, Wrigge H. The impact of spontaneous breathing during mechanical ventilation. Curr Opin Crit Care 2006;12:13 8
4. Putensen C, Zech S, Wrigge H, Zinserling J, Stuber F, Von ST, Mutz N. Long-term effects of spontaneous breathing during ventilatory support in patients with acute lung injury. Am J Respir Crit Care Med 2001;164:439 5. Putensen C, Hering R, Wrigge H. Controlled versus assisted mechanical ventilation. Curr Opin Crit Care 2002;8:517 6. Wrigge H, Zinserling J, Neumann P, Defosse J, Magnusson A, Putensen C, Hedenstierna G. Spontaneous breathing improves lung aeration in oleic acid-induced lung injury. Anesthesiology 2003;99:376 84 7. Levine S, Nguyen T, Taylor N, Friscia ME, Budak MT, Rothenberg P, Zhu J, Sachdeva R, Sonnad S, Kaiser LR, Rubinstein NA, Powers SK, Shrager JB. Rapid disuse atrophy of diaphragm fibers in mechanically ventilated humans. N Engl J Med 2008;358:132735 8. Froese AB, Bryan AC. Effects of anesthesia and paralysis on diaphragmatic mechanics in man. Anesthesiology 1974;41:24255 9. Froese AB. Anesthesia-paralysis and the diaphragm: in pursuit of an elusive muscle. Anesthesiology 1989;70:88790 10. Puybasset L, Cluzel P, Gusman P, Grenier P, Preteux F, Rouby JJ. Regional distribution of gas and tissue in acute respiratory distress syndrome. I. Consequences for lung morphology. CT Scan ARDS Study Group. Intensive Care Med 2000;26:857 69 11. Duggan M, Kavanagh BP. Pulmonary atelectasis: a pathogenic perioperative entity. Anesthesiology 2005;102:838 54 12. Wrigge H, Zinserling J, Neumann P, Muders T, Magnusson A, Putensen C, Hedenstierna G. Spontaneous breathing with airway pressure release ventilation favors ventilation in dependent lung regions and counters cyclic alveolar collapse in oleic-acidinduced lung injury: a randomized controlled computed tomography trial. Crit Care 2005;9:R780 9 13. Putensen C, von ST, Hering R, Stuber F, Zinserling J. Effect of different ventilatory support modalities on the ventilation to perfusion distributions. Acta Anaesthesiol Scand Suppl 1997; 111:119 22 14. Hering R, Viehofer A, Zinserling J, Wrigge H, Kreyer S, Berg A, Minor T, Putensen C. Effects of spontaneous breathing during airway pressure release ventilation on intestinal blood flow in experimental lung injury. Anesthesiology 2003;99:1137 44 15. Varelmann D, Wrigge H, Zinserling J, Muders T, Hering R, Putensen C. Proportional assist versus pressure support ventilation in patients with acute respiratory failure: cardiorespiratory responses to artificially increased ventilatory demand. Crit Care Med 2005;33:1968 75 16. Gama de Abreu M, Spieth PM, Pelosi P, Carvalho AR, Walter C, Schreiber-Ferstl A, Aikele P, Neykova B, Hubler M, Koch T. Noisy pressure support ventilation: a pilot study on a new assisted ventilation mode in experimental lung injury. Crit Care Med 2008;36:818 27 17. Putensen C, Mutz NJ, Putensen-Himmer G, Zinserling J. Spontaneous breathing during ventilatory support improves ventilationperfusion distributions in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 1999;159:1241 8 18. Henzler D, Pelosi P, Bensberg R, Dembinski R, Quintel M, Pielen V, Rossaint R, Kuhlen R. Effects of partial ventilatory support modalities on respiratory function in severe hypoxemic lung injury. Crit Care Med 2006;34:1738 45 19. Gama de Abreu M, Quelhas AD, Spieth P, Brauer G, Knels L, Kasper M, Pino AV, Bleyl JU, Hubler M, Bozza F, Salluh J, Kuhlisch E, Giannella-Neto A, Koch T. Comparative effects of vaporized perfluorohexane and partial liquid ventilation in oleic acid-induced lung injury. Anesthesiology 2006;104:278 89 20. Bleyl JU, Ragaller M, Tscho U, Regner M, Kanzow M, Hubler M, Rasche S, Albrecht M. Vaporized perfluorocarbon improves oxygenation and pulmonary function in an ovine model of acute respiratory distress syndrome. Anesthesiology 1999;91:4619 21. Vieira SR, Puybasset L, Richecoeur J, Lu Q, Cluzel P, Gusman PB, Coriat P, Rouby JJ. A lung computed tomographic assessment of positive end-expiratory pressure-induced lung overdistension. Am J Respir Crit Care Med 1998;158:15717 22. Gattinoni L, Caironi P, Pelosi P, Goodman LR. What has computed tomography taught us about the acute respiratory distress syndrome? Am J Respir Crit Care Med 2001;164:170111
23. Hubler M, Souders JE, Shade ED, Polissar NL, Schimmel C, Hlastala MP. Effects of vaporized perfluorocarbon on pulmonary blood flow and ventilation/perfusion distribution in a model of acute respiratory distress syndrome. Anesthesiology 2001;95:1414 21 24. Lachmann B, Robertson B, Vogel J. In vivo lung lavage as an experimental model of the respiratory distress syndrome. Acta Anaesthesiol Scand 1980;24:231 6 25. Hlastala MP, Lamm WJ, Karp A, Polissar NL, Starr IR, Glenny RW. Spatial distribution of hypoxic pulmonary vasoconstriction in the supine pig. J Appl Physiol 2004;96:1589 99 26. Esteban A, Ferguson ND, Meade MO, Frutos-Vivar F, Apezteguia C, Brochard L, Raymondos K, Nin N, Hurtado J, Tomicic V, Gonzalez M, Elizalde J, Nightingale P, Abroug F, Pelosi P, Arabi Y, Moreno R, Jibaja M, DEmpaire G, Sandi F, Matamis D, Montanez AM, Anzueto A. Evolution of mechanical ventilation in response to clinical research. Am J Respir Crit Care Med 2008;177:170 7 27. Glenny RW, Robertson HT. Fractal modeling of pulmonary blood flow heterogeneity. J Appl Physiol 1994;70:1024 30 28. Glenny RW, Lamm WJ, Bernard SL, An D, Chornuk M, Pool SL, Wagner WW Jr, Hlastala MP, Robertson HT. Selected contribution: redistribution of pulmonary perfusion during weightlessness and increased gravity. J Appl Physiol 2000;89:1239 48 29. Altemeier WA, McKinney S, Glenny RW. Fractal nature of regional ventilation distribution. J Appl Physiol 2000;88:15517 30. Starr IR, Lamm WJ, Neradilek B, Polissar N, Glenny RW, Hlastala MP. Regional hypoxic pulmonary vasoconstriction in prone pigs. J Appl Physiol 2005;99:36370 31. McGregor M, Becklake MR. The relationship of oxygen cost of breathing to respiratory mechanical work and respiratory force. J Clin Invest 1961;40:971 80 32. Van der Kloot TE, Blanch L, Youngblood AM, Weinert C, Adams AB, Marini JJ, Shapiro RS, Nahum A. Recruitment maneuvers in three experimental models of acute lung injury. Effect on lung volume and gas exchange. Am J Respir Crit Care Med 2000;161:148594 33. Luecke T, Meinhardt JP, Herrmann P, Weiss A, Quintel M, Pelosi P. Oleic acid vs saline solution lung lavage-induced acute lung injury: effects on lung morphology, pressure-volume relationships, and response to positive end-expiratory pressure. Chest 2006;130:392 401 34. Cereda M, Foti G, Marcora B, Gili M, Giacomini M, Sparacino ME, Pesenti A. Pressure support ventilation in patients with acute lung injury. Crit Care Med 2000;28:1269 75 35. Schuster DP, Anderson C, Kozlowski J, Lange N. Regional pulmonary perfusion in patients with acute pulmonary edema. J Nucl Med 2002;43:86370 36. Schuster DP, Haller J. Regional pulmonary blood flow during acute pulmonary edema: a PET study. J Appl Physiol 1990; 69:353 61 37. Spohr F, Cornelissen AJ, Busch C, Gebhard MM, Motsch J, Martin EO, Weimann J. Role of endogenous nitric oxide in endotoxininduced alteration of hypoxic pulmonary vasoconstriction in mice. Am J Physiol Heart Circ Physiol 2005;289:H82331 38. Ichinose F, Zapol WM, Sapirstein A, Ullrich R, Tager AM, Coggins K, Jones R, Bloch KD. Attenuation of hypoxic pulmonary vasoconstriction by endotoxemia requires 5-lipoxygenase in mice. Circ Res 2001;88:832 8 39. Caironi P, Ichinose F, Liu R, Jones RC, Bloch KD, Zapol WM. 5-Lipoxygenase deficiency prevents respiratory failure during ventilator-induced lung injury. Am J Respir Crit Care Med 2005;172:334 43 40. Nakayama M, Murray PA. Ketamine preserves and propofol potentiates hypoxic pulmonary vasoconstriction compared with the conscious state in chronically instrumented dogs. Anesthesiology 1999;91:760 71 41. Bjertnaes L, Hauge A, Kriz M. Hypoxia-induced pulmonary vasoconstriction: effects of fentanyl following different routes of administration. Acta Anaesthesiol Scand 1980;24:537
865
The Success of Emergency Endotracheal Intubation in Trauma Patients: A 10-Year Experience at a Major Adult Trauma Referral Center
Christopher T. Stephens, MD Stephanie Kahntroff, MD Richard P. Dutton, MD, MBA
BACKGROUND: Emergency airway management is a required skill for many anesthesiologists. We studied 10 yr of experience at a Level 1 trauma center to determine the outcomes of tracheal intubation attempts within the first 24 h of admission. METHODS: We examined Trauma Registry, quality management, and billing system records from July 1996 to June 2006 to determine the number of patients requiring intubation within 1 h of hospital arrival and to estimate the number requiring intubation with the first 24 h. We reviewed the medical record of each patient in either cohort who underwent a surgical airway access procedure (tracheotomy or cricothyrotomy) to determine the presenting characteristics of the patients and the reason they could not be orally or nasally intubated. RESULTS: All intubation attempts were supervised by an anesthesiologist experienced in trauma patient care. Rapid sequence intubation with direct laryngoscopy was the standard approach throughout the study period. During the first hour after admission, 6088 patients required intubation, of whom 21 (0.3%) received a surgical airway. During the first 24 h, 10 more patients, for a total of 31, received a surgical airway, during approximately 32,000 attempts (0.1%). Unanticipated difficult upper airway anatomy was the leading reason for a surgical airway. Four of the 31 patients died of their injuries but none as the result of failed intubation. CONCLUSIONS: In the hands of experienced anesthesiologists, rapid sequence intubation followed by direct laryngoscopy is a remarkably effective approach to emergency airway management. An algorithm designed around this approach can achieve very high levels of success.
(Anesth Analg 2009;109:866 72)
emodynamic instability, time pressure, lack of patient cooperation, risk of aspiration, the need for cervical spine protection, and facial injuries all contribute to the difficulty of airway management in trauma patients. The past decades have seen the development and promulgation of standard techniques (especially the American Society of Anesthesiologists difficult airway algorithm1), new rescue devices, such as the laryngeal mask airway (LMA), and a gradual shift in the United States away from anesthesiologists and toward emergency medicine physicians as the primary airway managers in the emergency department (ED). Published reports suggest a high level of success with emergency intubation, but few studies have examined a large series of patients and none in recent years.
From the Division of Trauma Anesthesiology, Department of Anesthesiology, University of Maryland School of Medicine, Baltimore, Maryland. Accepted for publication March 30, 2009. Address correspondence and reprint requests to Richard P. Dutton, MD, MBA, Division of Trauma Anesthesiology, University of Maryland Medical System, 22 South Greene St., Baltimore, MD 21201. Address e-mail to rdutton@anes.umm.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ad87b0
Endotracheal intubation is considered definitive airway management in the trauma patient,1 because it allows for deep sedation and analgesia, controlled mechanical ventilation, and protection of the airway from aspiration. A review of the literature suggests that anesthetized rapid sequence intubation (RSI) is the most common method of securing the airway.2 8 This method has been shown to result in higher rates of success in first-pass attempts, speed of placement, and fewer complications. Before 1990, trauma patients were intubated by medics in the field and in transport, by surgeons and generalist physicians in the ED, or by anesthesia providers summoned to the ED. Pharmacologic options and technology for rescuing a lost airway were limited. Rapid sequence approaches were used but led to a relatively high surgical airway rate. Harrison et al.9 reported the need for a surgical airway in 2% (6 of 302) of patients intubated during prehospital transport. In the past 2 decades, there have been substantial changes in airway management equipment. The use of intubating stylets has become more common, and rescue devices, such as the Combitube (Tyco-Kendall, Mansfield, MA) and the LMA (Vitaid, Toronto, Ontario, Canada), are now available. The development of Emergency Medicine as a recognized specialty has meant that most emergency
866
Figure 1. Emergency airway management algorithm at the R Adams Cowley Shock Trauma Center. It is assumed that an airway is absolutely required and that patients cannot be reawakened electively. LMA laryngeal mask airway. airway management in the United States now occurs without the involvement of an anesthesiologist. A recent study evaluated the need for a surgical airway with the use of a defined airway algorithm in a Level 1 trauma center, where emergency physicians are primarily responsible for airway management.10 This study reported a surgical airway rate of 2.6% and a complication rate of 9.8%. A European study of anesthesia-trained prehospital providers documented a failure rate (surgical airway) of 3.9% in a population of trauma patients.10,11 We undertook a retrospective review of emergency airway management at our center to establish our current success rate and identify risk factors associated with difficult airway management. We describe the management algorithm used, the overall success rate, and the factors associated with the need for a surgical airway. within the first 24 h after admission, on the assumption that, although less time pressured, these cases would still require the elements of an emergent RSI, including cricoid pressure and manual in-line cervical stabilization. The R Adams Cowley Shock Trauma Center (STC) of the University of Maryland Medical System, located in downtown Baltimore, is the primary adult resource center for trauma in the state of Maryland and annually receives more than 5500 primary admissions and another 1000 patients transferred from other hospitals with critical injuries. Patients arriving at the STC are assessed by a multidisciplinary team. Initial airway assessment and management is the responsibility of the attending anesthesiologist, a trauma specialist, who is typically supervising one or more residents (anesthesia or emergency medicine). The emergency airway management algorithm in use at the STC for the period of study is shown in Figure 1.12 This is a simplification of the standard American Society of Anesthesiologists algorithm1 adopted for use in this specialized setting, where the option to reverse the anesthetic and awaken the patient is seldom practical. This algorithm is followed at the discretion of the attending anesthesiologist, in accordance with the circumstances of the case. Intermittent positive pressure ventilation via facemask is used to support the patient if desaturation occurs, and cricoid pressure and in-line cervical stabilization may
METHODS
With the approval of the IRB, we conducted a retrospective review of all patients who underwent attempted intubation within 24 h of arrival between July 1, 1996, and June 30, 2006, a period of 10 yr. We chose to focus on the first hour of care to capture those patients intubated on an emergent basis. We excluded patients transferred from other hospitals, many of whom were already intubated. We also examined outcomes in the broader group of patients intubated
867
be relaxed if they are preventing successful intubation. Failure in three attempts at direct laryngoscopy leads to preparations for a surgical airway and an attempt to manage the situation with placement of a LMA, with continued cricoid pressure. If LMA placement allows acceptable oxygenation and ventilation, then a variety of options are available at the discretion of the anesthesiologist. These include renewed attempts at intubation with different equipment or positioning, fiberoptic intubation through the LMA, or a surgical airway under semicontrolled conditions (cricothyrotomy or tracheotomy). If LMA placement is not successful, then a cricothyrotomy or tracheotomy is performed immediately by the surgical team. Trauma anesthesiology billing records, quality management (QM) records, and the Trauma Registry database were used to identify patients who underwent emergency airway management during the study period. Maintenance of a Trauma Registry is a requirement for certification as a trauma center in the United States. Data are entered during daily review of each patients records by a specialty-trained and certified group of registrars (who also enter billing information), and information is periodically updated and reviewed by supervisors, before our reporting it to the state Trauma System. The quality of Registry data is audited internally, and again by the state certifying authority, as part of the trauma center designation process. Billing records are similarly scrutinized by both insurers and the state rate-setting commission. Billing records were used to approximate the number of patients intubated within 24 h of admission. Because electronic billing records could not identify patients presenting to the operating room (OR) for general anesthesia who had a definitive airway already in place, this number is necessarily an estimate. Trauma Registry and anesthesia QM records were then used to identify all patients who had a cricothyrotomy or tracheotomy performed within 24 h of admission. These medical records were reviewed in detail. Figures 2a c show the distribution of patients under study. Nine patients were intubated successfully (three within 1 h and an additional six within 24 h), but then underwent tracheotomy for long-term management of severe facial injuries or brain trauma. These patients were considered successful airway management cases and were excluded from further analysis. Computerized medical records, paper records, and QM files were reviewed to elicit more detailed information regarding each case. These data included the mechanism of injury, the indication for intubation, the cause of difficulties, and the patients ultimate outcome. Need for a surgical airway was a designated QM indicator throughout the study period, meaning that each of these cases was reviewed by an uninvolved anesthesia provider and discussed by the Divisional Quality Management (QM) Committee.
868
Experience with Emergency Intubations
Figure 2. a c, Trauma admission and outcomes of urgent and emergency airway management.
Minutes of these discussions were available for review. All surgical airway cases identified had original medical records and most had available QM documentation. We also examined all QM records during this decade for evidence of morbidities associated with emergency intubation that did not result in a surgical airway: specifically, worsened neurologic injury or evidence of anoxic injury.
RESULTS
Approximately 32,000 trauma patients required intubation within 24 h of admission during the study period. Of the 6088 patients in whom intubation was attempted within 1 h, 6008 received orotracheal airways, 59 received nasotracheal airways, and 21 received surgical airways. Of the 21 emergency surgical airways, 17 were cricothyrotomies and 4 were tracheotomies (Fig. 2b). The rate of surgical airway access in emergencies was 0.3%. For patients requiring airway management between 1 and 24 h of admission, there were 10 additional surgical airways placed in approximately 26,000 attempts (estimated from OR billing records as described above), for a surgical airway rate of approximately 0.04% in these less urgent cases. The 31 patients receiving surgical airways within 24 h of admission are enumerated in Table 1. The study methodology did not permit us to identify how often an intubating stylet (gum elastic bougie) was used to facilitate intubation (anecdotally, this is common) or
Table 1. Patients Requiring Surgical Airway Management Within 24 h of Trauma Center Admission
Emergency (within 1 h of admission) Yes Yes Yes Yes Yes
N 31 1 2 3 4 5
Sex M M F M M
Age 63 19 24 25 29
Cric or trach C C C C T
Reason Self-inflicted GSW to the mid and upper face. GSW to chest and back. Arrived in full cardiac arrest. MVC head-on collision. Cardiac arrest at the scene. Multiple GSW to chest, neck, back. Cardiac arrest in the field. Copious emesis from mouth. GSW with injury to trachea and thyroid gland. Cricothyrotomy unsuccessful by resident. Tracheotomy placed by surgical attending. GSW to mandible. Copious bloody secretions in airway. Fall down stairs. Blood noted in airway. Bradycardia and desaturations during attempt at intubation. MVC with resulting facial trauma. Fell at nursing home. Edema of true vocal cords and polyps. Trismus noted. Boulder to head. Sedated for fracture reduction, with bradycardia and desaturations requiring emergency intubation. LMA was placed. MVC. Unable to visualize vocal cords. Chemical explosion to face. Unable to visualize vocal cords secondary to blood and anterior anatomy. MCC. Broken teeth. Unable to visualize vocal cords. Pedestrian struck by a motor vehicle, with multiple facial injuries. MCC. Unable to visualize vocal cords. MVC. Blood in oropharynx, which made visualization difficult. MVC. Failed awake fiberoptic intubation secondary to airway edema. Intraoral GSW. Failed airway secondary to tongue edema. Patient found down. Subdural hematoma. Unable to visualize vocal cords secondary to large epiglottis. MVC. Patient was seizing. Unable to visualize vocal cords. MVC. Supraglottic edema. Ecchymosis around posterior pharynx, base of tongue and floor of mouth. Tongue tip necrosis. Self-inflicted GSW to neck and cheek. Deep penetrating injuries to submental area. Free segment of mandible missing. Difficult visualization. There was significant bleeding, broken teeth, and a dental plate, which made the intubation difficult. MVC patient to operating room for orthopedic procedure. Undiagnosed adenocarcinoma with tracheal compression. Rigid neck secondary to metastatic disease. Fall with brachial artery injury. Unable to visualize vocal cords. MCC. Intubation failed with desaturation and vomiting. Fall with C3-C7 fractures. Failed awake fiberoptic in OR with desaturation. Fall with pelvic fracture. Vocal cords were not visualized. Pedestrian struck, with multiple facial bone fractures. Vocal cords were not visualized. MVC. Difficult laryngeal visualization. GSW to leg. Difficult laryngeal visualization. Hit by a jet ski in the mandible. Initial intubation successful. Patient extubated after surgery, developed laryngospasm, and could not be reintubated.
GCS 7 3 3 3 15
Disposition Died Died Died Died Lived
6 7 8 9 10
M M M F M
18 59 41 86 47
Yes Yes Yes Yes Yes
C C C C T
13 3 15 13
Lived Lived Lived Lived Lived
11 12 13 14 15 16 17 18 19 20 21
F M F M M M M M M M M
79 47 44 48 32 49 73 33 66 70 63
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
C C C C T C C C T C C
15 15 15 11 13 11 15 7 12 Unknown 15
Lived Lived Lived Lived Lived Lived Lived Lived Lived Lived Lived
22
36
No
11
Lived
23
69
No
15
Lived
24 25 26 27 28 29 30 31
M M M M M M M M
50 32 66 60 58 41 24 40
No No No No No No No No
T C C T T T T C
15 15
Lived Lived Lived Lived Lived Lived Lived Lived
15
15 15
Cric cricothyrotomy; Trach tracheotomy; GCS Glasgow Coma Scale score; GSW gunshot wound; MVC motor vehicle collision; MCC motorcycle collision; LMA laryngeal mask airway; OR operating room.
an LMA was placed and then followed by successful oral intubation (a rare event). Overall this cohort included 26 men and 5 women, which is consistent with the demographics of trauma patients. Seven different mechanisms of injury were identified in this cohort: gunshot wound, motorcycle collision, motor vehicle collision, chemical explosion, pedestrian struck, fall from a height, and struck by a falling object. The average age was 44.7 yr (range, 18 86 yr). In the 25 patients for whom an admission Glasgow Coma Scale score was available, the median
score was 13 (range, 315). The majority of patients were suffering from severe, multisystem injury. Five different causes for failed endotracheal intubation were identified: foreign material in the pharynx or larynx, direct injury to the head or neck with loss of normal upper airway anatomy and airway edema, a pharyngeal mass, laryngospasm, and difficult premorbid anatomy. The distribution of causes is shown in Figure 3. Emesis, blood, and broken teeth were the identified foreign materials preventing intubation. Three patients had disruption of airway anatomy
869
DISCUSSION
Risk factors for poor outcome from airway management include the potential for aspiration of gastric contents, exacerbation of occult cervical spine injuries, hemodynamic instability, and traumatic brain injury. This retrospective study describes outcomes of emergency airway management in a group of 32,000 adult trauma patients during the decade from July, 1996, to June, 2006. All patients were managed using a standard protocol adapted directly from the American Society of Anesthesiologists Difficult Airway algorithm,1,12 and all airway management occurred under the direction of an anesthesiologist with specialty expertise in trauma. Our aim was to review our experience with emergency airway management in trauma patients, to assess the effectiveness of the protocol used, and to categorize our failed efforts in an attempt to better predict potentially difficult airways. One limitation of our study is its retrospective nature. Although we were able to examine medical records of the patients who required surgical airways, we could not individually review all 57,000 admissions during this time period. We were thus forced to estimate the number of patients requiring intubation within 24 h of admission from OR billing records. Emergency intubations performed in the Trauma Resuscitation Unit are billed separately, with a specific time recorded in the Trauma Registry, and thus could be exactly counted. It was not possible for us to retrospectively determine intrinsic details of the successful intubations, including number of attempts, number of operators, equipment used, and lowest oxygen saturation. This detail would have allowed a closer analysis. It is possible that some of the patients in this series should have received a surgical airway sooner, for example, to mitigate the effects of hypoxia or hypercarbia on traumatic brain injury. It was also not possible to assess the contribution of micro- or macroaspiration on the development of subsequent respiratory distress in a population of patients with many risk factors for respiratory failure, including premorbid conditions, prehospital aspiration, traumatic brain injury, acute and chronic intoxication, hemorrhagic shock, transfusion, long bone fractures, and direct pulmonary injury. Another limitation is the data collected on the need for a surgical airway. Although we can identify the causes with reasonable accuracy, because of the contemporaneous QM review of these high-profile cases, we cannot identify how often similar patients were managed successfully. We are unable to assess the specific impact of attending discretion in avoiding or mitigating difficult airway management. It is likely that some patients with observable risk factors for difficult intubation were managed differently than routine patients: earlier hands-on involvement of the attending anesthesiologist is one possibility; another
Figure 3. Causes of the need for surgical airway access.
because of gunshot wounds to the face or neck; one patient who was a pedestrian struck by a motor vehicle had massive facial distortion. Edema preventing intubation was at the supraglottic, laryngeal, or subglottic level and was the result of airway burns, facial fractures, or chemical toxicity. Anatomic variation was noted as the primary cause of difficulty in patients who were found to have difficult laryngoscopic views without obvious injury to the head or neck. This category included variations such as obesity, limited mouth opening, short thyromental distance, limited neck mobility, and anterior larynx. In five of these cases, no specific variation was noted, and the inability to visualize the larynx was described as surprising by the attending anesthesiologist. Remarkably, 87% (27/31) of the patients requiring a surgical airway survived to hospital discharge, and none of the four deaths appeared to be the primary result of failed airway management. Two of the four deaths had already suffered cardiac arrest at the scene of injury, with subsequent transient recovery of circulation but ongoing hemodynamic instability, and one other was brought to the trauma resuscitation unit in full arrest. These three patients were all judged by the QM review process to have died of exsanguinating hemorrhage that was not preventable, although an exacerbating effect of hypoxia cannot be excluded. The fourth death occurred 3 days after admission in a patient who sustained a gunshot wound to the face resulting in a carotid artery dissection and middle cerebral artery stroke. A review of QM records revealed only one patient with the suggestion of exacerbation of an occult cervical spine injury during intubation, although focused review of this case suggested that the neurologic deficit identified after intubation may have been present earlier, but inadequately documented. No new changes in neurologic status as the result of hypoxia during intubation efforts were identified, although the impact of transient hypoxia on patients with traumatic brain injury cannot be excluded. No patients died of a hypoxic cardiac arrest during airway management efforts.
870
might be the decision to use an awake fiberoptic approach in cooperative patients. This is one area in which the involvement of anesthesiologists, as opposed to emergency medicine physicians, may have played an important role. Our success in emergent airway management (99.7%) compares favorably with previously published series. In a consensus paper from the Eastern Association for the Surgery of Trauma, Dunham et al.13 describe an overall failure to intubate with RSI in trauma centers to be 1.7% in a population of 943 patients. A review of other literature reports surgical airway rates ranging from 0.3% to 5.6% at major Level 1 trauma centers. Of these studies, the majority of institutions report a failed intubation rate of 1%2%.14 24 Although premorbid anatomic variations were the leading cause of failed oral intubation in our series, head and neck injuries were also an important risk factor. Thirteen of our 31 patients who required surgical airways had injuries involving the head, face, neck, upper chest, or a combination. This is consistent with prior studies of emergency airway management in which many of the difficulties encountered were in patients with trauma to the face or neck. The cited report by Dunham et al.13 describes the evidence base supporting techniques for emergency intubation in trauma patients. The group reported that many patients with severe neck injury required immediate airway intervention secondary to cervical hematomas and laryngotracheal injury. Fourteen percent (3 of 21) of our failed emergency intubation group had a significant neck injury. This retrospective survey, the largest reported, confirms an overall high rate of success in emergency airway management. Although secondary effects of hypoxia could not be assessed in this review, the fact that no patients died of acute hypoxia is encouraging. Having intubations supervised by a small group of specialist anesthesiologists is likely beneficial, as is the immediate presence of surgical support when difficulties arise. Others have reported that training is important to increase the percentage of successful airway placements.25,26 One previous study reported on senior Emergency Medicine residents rotating as the trauma airway manager for 2 mo in addition to OR training throughout their residency, with a combined experience of 70 80 intubations.6 This is comparable with the 50 80 intubations, many urgent, that an average resident performs in 1 mo in our practice. We believe that this level of experience and training contributes strongly to good outcomes. Using a small selection of adjunctive technologies, only the intubating stylet (bougie) and the LMA keeps our protocol simple and allows for practice and familiarity for all providers. Although many other adjunctive devices have been recommended for emergency airway management, it is likely that practice and experience are important when using the device in
an emergency.27 It is possible that the video laryngoscope (Glidescope, Verathon, Bothell, WA), which has entered our practice in the past year, will further improve our experience, or that simulator training, now routine at our site, will improve performance in emergencies. In summary, it is possible to achieve a high rate of success in emergency airway management. The need for a surgical airway in our practice is primarily the result of premorbid anatomic variation, although trauma to the face or neck also contributes. The cornerstones of effective airway management in our series were a simple protocol based on rapid sequence induction of anesthesia, judicious use of selected adjunctive devices, and an experienced anesthesia faculty present at each admission. REFERENCES
1. American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Practice guidelines for management of the difficult airway: an updated report by the American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Anesthesiology 2003;98:1269 77 2. Committee on Trauma, American College of Surgeons. Advanced trauma life support for doctors. Chicago: American College of Surgeons, 2004 3. Graham CA, Beard D, Henry JM, McKeown DW. Rapid sequence intubation of trauma patients in Scotland. J Trauma 2004;56:1123 6 4. Walls RM. Management of the difficult airway in the trauma patient. Emerg Med Clin North Am 1998;16:45 61 5. Sakles JC, Laurin EG, Rantapaa AA, Panacek EA. Airway management in the emergency department: a one-year study of 610 tracheal intubations. Ann Emerg Med 1998;31:32532 6. Omert L, Yeaney W, Mizikowski S, Protetch J. Role of the emergency medicine physician in airway management of the trauma patient. J Trauma 2001;51:1065 8 7. Bushra JS, McNeil B, Wald DA, Schwell A. A comparison of trauma intubations managed by anesthesiologists and emergency physicians. Acad Emerg Med 2002;9:404 5 8. McBrien ME, Pollok AJ, Steedman DJ. Advanced airway control in trauma resuscitation. Arch Emerg Med 1992;9:177 80 9. Harrison T, Thomas SH, Wedel SK. In-flight oral endotracheal intubation. Am J Emerg Med 1997;6:558 61 10. Casey ZC, Smally AJ, Grant RJ, McQuay J. Trauma intubations: can a protocol-driven approach be successful? J Trauma 2007;63:955 60 11. Timmermann A, Eich C, Russo SG, Natge U, Bra uer A, Rosenblatt WH, Braun U. Prehospital airway management: a prospective evaluation of anaesthesia trained emergency physicians. Resuscitation 2006;70:179 85 12. Dutton RP, McCunn M. Anesthesia for trauma. In: Miller RD, ed. Millers anesthesia. 6th ed. Philadelphia: Elsevier Churchill Livingstone, 2005: 245195 13. Dunham CM, Barraco RD, Clark DE, Daley BJ, Davis FE, Gibbs MA, Knuth T, Letarte PB, Luchette FA, Omert L, Weireter LJ, Wiles CE. Guidelines for emergency tracheal intubation immediately after traumatic injury. J Trauma 2003;55:16279 14. Mulder DS, Marelli D. The 1991 Fraser Gurd lecture: evolution of airway control in the management of injured patients. J Trauma 1992;33:856 62 15. Talucci RC, Shaikh KA, Schwab CW. Rapid sequence induction with oral endotracheal intubation in the multiply injured patient. Am Surg 1988;54:1857 16. Bogetz M, Katz J. Airway management of the trauma patient. Semin Anesth 1985;4:114 23 17. Mandavia DP, Qualls S, Rokos I. Emergency airway management in penetrating neck injury. Ann Emerg Med 2000;35:2215 18. Pierre EJ, McNeer RR, Shamir MY. Early management of the traumatized airway. Anesthesiol Clin 2007;25:111
871
19. Salvino CK, Dries D, Gamelli R, Murphy-Macabobby M, Marshall W. Emergency cricothyroidotomy in trauma victims. J Trauma 1993;34:5035 20. DeLaurier GA, Hawkins ML, Treat RC, Mansberger AR. Acute airway management: role of cricothyroidotomy. Am Surg 1990;56:125 21. Wright MJ, Greenberg DE, Hunt JP, Madan AK, McSwain NE. Surgical cricothyroidotomy in trauma patients. South Med J 2003;96:4657 22. Bair AE, Panacek EA, Wisner DH, Bales R, Sakles JC. Cricothyroidotomy: a 5-year experience at one institution. J Emerg Med 2003;24:151 6 23. Levitan RM, Rosenblatt B, Meiner EM, Reilly PM, Hollander JE. Alternating day emergency medicine and anesthesia resident responsibility for management of the trauma airway: a study of laryngoscopy performance and intubation success. Ann Emerg Med 2004;43:48 53
24. Levitan RM, Everett WW, Ochroch EA. Limitations of difficult airway prediction in patients intubated in the emergency department. Ann Emerg Med 2004;44:30713 25. Hawkins ML, Shapiro MB, Cue JI, Wiggins SS. Emergency cricothyrotomy: a reassessment. Am Surg 1995;61:525 26. Shearer V. Modern airway management for the trauma patient. Curr Opin Anaesthesiol 2000;13:1359 27. Smith CE, DeJoy SJ. New equipment and techniques for airway management in trauma. Curr Opin Anaesthesiol 2001;14:197209
872
The Effects of Endotracheal Suctioning on the Accuracy of Oxygen Consumption and Carbon Dioxide Production Measurements and Pulmonary Mechanics Calculated by a Compact Metabolic Monitor
George Briassoulis, MD, PhD* Panagiotis Briassoulis, MD Evi Michaeloudi, MD* Diana-Michaela Fitrolaki, MD* Anna-Maria Spanaki, MD* Efrossini Briassouli, MD
BACKGROUND: Open endotracheal suctioning (ETS), which is performed regularly in mechanically ventilated patients to remove obstructive secretions, can cause an immediate decrease in dynamic compliance and expired tidal volume and result in inadequate or inaccurate sidestream respiratory monitoring, necessitating prolonged periods of stabilization of connected metabolic monitors. We investigated the immediate effect of open ETS on the accuracy of oxygen consumption (VO2) and carbon dioxide production (VCO2) measurements and calculated lung mechanics, respiratory quotient, and resting energy expenditure in mechanically ventilated children without severe lung pathology, when using a compact modular metabolic monitor (E-COVX) continuously recording patient spirometry and gas exchange measurements. METHODS: Open ETS was performed when clinically indicated in 11 children mechanically ventilated for sepsis or head injury. A total of 2800 pulmonary 1-min gas exchange measurements were recorded in 28 ETS instances for 50 consecutive minutes before and 50 min after the standardized procedure. RESULTS: Pulmonary mechanics and indirect calorimetry did not differ between preand postsuction sets of measurements. Pre- and postsuction VO2, VCO2, dynamic airway resistance, dynamic compliance, and expiratory minute ventilation remained stable from 5 to 55 min after tracheal suctioning and did not differ among different ventilatory modes. Average paired differences of sequential pre- and postsuction VO2, VCO2, respiratory quotient, and resting energy expenditure were 0.6%, 1%, 0.1%, and 0.3%. Ratio differences between the first and the second periods of measurements (125 vs 26 50 sets of 1-min measurements) did not differ in the two groups. CONCLUSIONS: Pulmonary mechanics and indirect calorimetry measurements are not influenced after uneventful open ETS in well-sedated patients. The E-COVX is able to reliably record spirometry and metabolic indices as early as 5 min after suctioning at different ventilator modes.
(Anesth Analg 2009;109:8739)
he most accurate method for determining resting energy expenditure (REE) in hospitalized patients is indirect calorimetry.1,2 Metabolic monitoring devices used in the critical care setting, however, are subject to a range of conditions that may compromise their accuracy.3 More specifically, metabolic monitors errors were shown to be significantly affected by oxygen concentration and minute ventilation4 and when used
From the *Pediatric Intensive Care Unit, University Hospital, University of Crete, Heraklion, Greece; Department of Anaesthesiology, School of Medicine, University of Athens; and The 1st Department of Internal Medicine-Propaedeutic, University of Athens, Athens, Greece. Accepted for publication May 4, 2009. Address correspondence and reprint requests to George Briassoulis, MD, Pediatric Intensive Care Unit, University Hospital of Heraklion, 71110 Heraklion, Crete, Greece. Address e-mail to ggbriass@otenet.gr. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b018ee
during inhaled anesthesia.5 Additionally, older systems like Deltatrac II (Datex Ohmeda 2000, Helsinki, Finland), which measure gas volume in a mixing chamber, are relatively expensive, require a high level of technical expertise, and are time consuming to calibrate.6 New compact modular metabolic monitors like the E-COVX (formerly M-COVX, GE Healthcare/ Datex-Ohmeda), which use a breath-by-breath method to analyze oxygen consumption (VO2) and carbon dioxide production (VCO2), are less expensive and simpler to use, perform calibration automatically, and are much smaller in size.7 Using such a simple monitor in certain ventilation modes and in nonsedated patients, however, may not provide measurements within a clinically accepted range.8 Open endotracheal suctioning (ETS) is performed regularly in mechanically ventilated children to remove obstructive secretions. It was shown that ETS can cause an immediate decrease in dynamic compliance and expired tidal volume in ventilated children intubated
873
with small endotracheal tubes (ETs), probably indicating loss of lung volume caused by the suctioning procedure.9 Such a change, which might be followed by marked decrease in SaO2, can be prevented by positive pressure oxygen administration or by using a closed system of ETS to maintain lung volume and minute ventilation.10 Especially in patients with lung injury, suctioninginduced lung derecruitment could be prevented by performing recruitment maneuvers during suctioning and minimized by avoiding disconnection.9 However, the effect of the frequently needed ETS on accuracy of lung mechanics and VO2 and VCO2 measurements using a simple compact modular metabolic monitor is not known. We aimed to determine the immediate effect of open ETS on the accuracy of VO2 and VCO2 measurements and calculated dynamic lung compliance, airway resistance, minute ventilation, respiratory quotient (RQ), and REE, when using the E-COVX by means of a prospective observational clinical study in critically ill children without severe lung pathology.
METHODS
Patient Population
Eleven pediatric intensive care unit patients without severe lung pathology, mechanically ventilated for sepsis or head injury, were studied. Patients with primary (asthma, bronchiolitis, and pneumonia) or severe (acute respiratory distress syndrome) lung pathology were excluded. Patients with sepsis or trauma and any mild-to-moderate pulmonary complication, such as atelectasis, ventilator-associated pneumonia, contusion (excluding air leaks), or acute lung injury, were only included if 1) not desaturated more than 10% 13 min after any ET disconnection and 2) not meeting any of the other exclusion criteria. The investigation was approved by and conducted according to the local ethics committee guidelines, and informed consent was obtained from each patients relative. All patients were hemodynamically stable and in the supine position with a 30 head lift inclination. Mechanical ventilation was performed through a cuffed ET, and only patients with a fractional inspired oxygen (Fio2) 60%, a respiratory rate below 35 bpm, and an ET leak below 10% (inspiratory tidal volume expiratory tidal volume/inspiratory tidal volume 100) were included in the study. Patients ventilated with Heliox/nitric oxide mixture, on the oscillator ventilator, receiving renal replacement therapy, and those receiving bolus enteral feeds or expected to be extubated within 24 h of admission were excluded from the study. The make of the ventilator (Servo i, Marquet, Solna, Sweden) was documented, in addition to the ventilation mode and settings at the time of the measurement.
and were able to breathe spontaneously. The suctioning procedure was performed as follows. The patients were administered 100% inspired oxygen for 30 s before suctioning. They were disconnected from the ventilator, a suction catheter was passed down 2 cm below the distal end of the ET tip, continuous suction was applied, and the catheter withdrawn while rotating slightly. The trachea was suctioned two times for 10 s with an interval of 10 s between each suctioning, during which the catheter was changed. Brief hand ventilation was used for lung recruitment between and after suctioning. The patient was then immediately reconnected to the ventilator circuit. This resulted in a 30-s period of disconnection from the ventilator. Any adverse events were documented. We used a range of catheter sizes, according to availability and the ET size. The suction apparatus was set on medium, corresponding to a vacuum pressure of approximately 250 mm Hg (33 kPa or 4.8 psi), measured at the source with tubing clamped. After suctioning was completed, the Fio2 was immediately changed to presuction settings unless desaturation occurred, in which case Fio2 was gradually turned down as SaO2 improved. The nurse in charge performed all the suctioning procedures in each patient. Throughout the observation and suctioning periods, there was continuous electrocardiological and pulse oximetry monitoring. During and for 90 min before this process, no changes were made to the ventilator settings and physiotherapy was not allowed.
Apparatus
The new compact modular metabolic monitor E-COVX (GE Healthcare/Datex-Ohmeda, also known as M-COVX) has been marketed as the replacement for the Deltatrac II (Datex Ohmeda 2000) for complete sidestream respiratory monitoring with patient spirometry and gas exchange measurement. The mode of action of the system has been described in detail.8 This device relates flow measurements made at the mouth by pneumotachograph to measurements of inspired and expired gas composition by matching the two wave forms, thereby allowing continuous, breath-bybreath monitoring of an intubated patients VO2 and VCO2. The pneumotachograph derives the tidal volume from the pressure difference across a fixed orifice and is therefore potentially influenced by acute changes of resistance in the spirometry tubing and undetected leaks in the system. In the P-Lite (15300 mL) or D-Lite (300 mL) flow sensor, selected according to the patients recorded tidal volume and located proximal to the Y-piece to the patients ET, the flow measurement is based on the pressure decrease across a special proprietary turbulent flow restrictor. It uses mathematical integration of flow and time-synchronized continuous gas sampling to provide data. The gas sample is continuously drawn from the connector to the gas analyzer unit of the E-COVX. Both O2 and CO2
Suctioning Procedure
Suctioning for study purposes coincided with the nursing staffs planned time of ETS as clinically indicated, so that patients experienced no additional discomfort. The patients were well sedated but not paralyzed
874
Suction and Gas Exchange Measurements
measures are based on the sidestream principle. Detection through the D-Lite or Pedi-Lite flow sensor and gas sampler is achieved at respiration rates of 4 35 bpm and 4 50 bpm, respectively.
Table 1. Clinical Characteristics of Patients Enrolled in the Study (n 11) Frequency (%)
Age (yr) Body weight (kg) Sex Male Female Diagnosis Sepsis Head injury PeLoad PRISM TISS Ventilatory mode PRVC SIMV PS Fio2 (%) SaO2 (%) Heart rate (bpm) Blood pressure (mm Hg) Temperature (C) Energy intake (kcal/d) Midazolam (mg kg1 h1) Fentanyl (g kg1 h1) Length of stay (d) Duration of mechanical ventilation (d)
Mean sd (range)
9.5 4.6 (617) 30 18 (1580)
Lung Mechanics and Metabolic Gas Exchange Measurements

Breath-by-breath values were averaged over each minute of recording and used for analysis. Pulmonary 1-min gas exchange measurements were recorded for 50 consecutive minutes before (presuction) and 50 min after the standardized suctioning procedure, starting 5 min postsuction. (A steady-state condition must be present to ensure that the gas exchange measurement is equivalent to the tissue gas exchange. According to the manufacturer, a steady state can be defined as a period of time after the patient has stabilized from any changes [about 5 min] and will not incur further changes in his/her treatment that may affect gas exchange or increase metabolism.) Measurements from the E-COVX were collected after a 5-min warm-up period which included automatic calibration. Steady state was defined by five consecutive 1-min measurements of VO2 and VCO2 having 10% variation. Throughout the study, measurements took place only when the patient was in steady state. We consistently used a heat- and moisture-exchange filter alone, avoiding heated water bath humidification, followed by regular checks on the spirometry tubing and checks for tidal volume consistency between the E-COVX module and the ventilator. Measurements of ventilation (dynamic lung compliance, dynamic expiratory airway resistance, mechanical and spontaneous expired tidal volume, and total respiratory rate) and gas exchange (VO2, VCO2, RQ) were simultaneously recorded pre- and post-ETS. Expired tidal volume recordings were used rather than inspired tidal volume to minimize errors due to any ET leaks. The machine uses inspiratory volumes, as these are the more reliable measurements; expiratory volumes are dependent upon assumptions of expired temperature (assumed to be 35C) and humidity (assumed to be 100%), and the modified Weir formula for converting VO2 and VCO2. The monitor displays a 5-min average for REE, but we obtained the 1-min averages with the S5 software. Because not all 1-min variables could be collected through the S5 software, the 5-min recorded means (of the 1-min measurements and calculations) were ultimately analyzed and compared over the 50-min periods before and after suctioning.
73 27 73 27
9 4.2 (213) 11.5 6.2 (627) 37 7.5 (2248)
64 18 18
45 10 (3060) 99 1 (97100) 99 31 (58159) 103 8 (82116) 37.3 0.65 (36.138) 676 240 (2401000) 0.57 0.35 (01) 1.82 1.5 (04) 26.7 29.5 (498) 20.7 26.5 (395)
PeLoad pediatric logistic organ dysfunction score; PRISM pediatric risk of mortality; TISS therapeutic intervention scoring system; PRVC pressure-regulated controlled ventilation; SIMV synchronized intermittent mandatory ventilation; PS pressure support ventilation; FIO2 fractional inspired oxygen; SaO2 saturation of arterial oxygen.
The mean difference between the two sets of measurements represents the performance bias. On average, the bias indicates how far off an individual VCO2 or VO2 measurement or calculated RQ or REE will be after suctioning compared with a presuction recorded value. Furthermore, to examine whether the ETS might have influenced the accuracy of the measurements closer to the intervention (after 5 min) more than the later ones (after 50 min), we compared the within-groups ratio differences between the first and the second periods of measurements (125 vs 26 50 sets of 1-min measurements). Statistical significance was defined as P 0.05. SPSS version 15.0 (SPSS, Chicago, IL) was used for all data entry and statistical procedures except the Bland-Altman plot, which was constructed using the MedCalc version 9.3.1 (MedCalc Software, Belgium).
RESULTS
A total of 2800 pulmonary 1-min gas exchange measurements were recorded in 28 ETS instances in 11 patients, producing 560 records of means of five successive 1-min measurements (5-min means). Of those, 1400 pulmonary sets of 1-min gas exchange measurements (or 280 5-min means) were recorded before and the same (1400 1-min or 280 5-min means)
Data were tested for normality using the KolmogorovSmirnov and Lilliefors tests. Normally distributed data were analyzed using descriptive statistics and t-tests for paired differences. The level of bias between measurements before and after suctioning was evaluated using Bland-Altman limits of agreement analysis.
875
Table 2. Comparison of 50-min Pulmonary Mechanics and Indirect Calorimetry Values Measured by the E-COVX Compact Metabolic Monitor Before and 5 min (Stabilization Period) After Suctioning (n 280 5-min Means of 1-min Measurements) 95% Condence interval of the difference Lower
0.75 1.9 2.7 0.01 8.9 0.06 0.69 0.06 0.18 0.3 0.25 0.006
Before suctioning E-COVX spirometry

End-tidal CO2 (mm Hg) VO2 (mL/min) VCO2 (mL/min) Respiratory quotient Resting energy expenditure (kcal/d) Expiratory minute volume (L/min) Total respiratory rate (bpm) Peak inspiratory pressure (cm H2O) Extrinsic positive end-expiratory pressure (cm H2O) Expired tidal volume (mL) Dynamic airway resistance (cm H2O L1 s1) Dynamic compliance (mL/cm H2O)
After suctioning Mean

37.3 126.9 109.9 0.88 887.7 4.6 26.8 22.3 6.5 186.9 21.2 14.7
Mean
37.1 126.2 108.6 0.88 885.5 4.7 27 22.6 6.5 191 21.3 15.6
sd
3.5 43.3 30.5 0.1 279 1.3 8 7.8 1.6 88 11 6.7
sd
4.5 43.7 32.6 0.1 281 1.2 8.3 7.4 1.6 88 10.5 6.7
Upper
0.41 0.43 0.28 0.01 4.8 0.15 1.04 0.58 0.06 8.4 0.4 0.83
Paired differences sig (two-tailed)

0.56 0.207 0.108 0.87 0.5 0.36 0.68 0.106 0.315 0.07 0.65 0.054
Figure 1. Comparison of measured airway resistance, compliance, and minute ventilation before and after suctioning. The box-whisker plots show the median (horizontal line within the box) and the 10th and 90th percentiles (whiskers). The box length is the interquartile range. Rawa dynamic expiratory airway resistance after suctioning; Rawb dynamic expiratory airway resistance before suctioning; Compla dynamic lung compliance after suctioning; Complb dynamic lung compliance before suctioning; Mvexpa mechanical and spontaneous expired minute ventilation after suctioning; Mvexpb mechanical and spontaneous expired minute ventilation before suctioning.
repeated after the 28 ETS instances. Different ETS instances in the same patient were only performed if they could be done on different days, especially when clinical and respiratory variables had been changed, therefore producing new (different ventilator settings) study cases. The first comparisons of before/after suctioning per patient measurements in the 11 individual patients were 1) analyzed separately; and 2) pooled with all repeated discrete timepoint measurements performed subsequently. Because no differences were found between the 11 initial sets of measurements
876
in each patient versus the 28 total sets of measurements (including second or third repeated ETS events in some of the 11 patients) indicating potential biases introduced by repeated measurements, all results were pooled for the final analysis of different cases of varying clinical situation and ventilator support. Patient demographic data are summarized in Table 1. Pulmonary mechanics and indirect calorimetry did not differ between pre- and postsuction sets of measurements (Table 2). Measured dynamic airway resistance, dynamic compliance, and expiratory minute ventilation remained stable from 5 to 55 min after tracheal suctioning (Fig. 1). The series of 50-min measurements of VO2, VCO2 (Fig. 2), and REE (Fig. 3) did not change after the procedure and remained stable 55 min later. Pre- and postsuction spirometry measurements and VO2 and VCO2 did not differ among the different ventilatory modes (pressure-regulated controlled ventilation, synchronized intermittent mandatory ventilation, and pressure support ventilation) used in this series (Fig. 4). The Bland-Altman plot comparing the means of sequential pre- and postsuctioning VO2, VCO2, RQ, and REE indicated that the average paired differences were 0.6%, 1%, 0.1%, and 0.3%, respectively, showing no influence of suctioning on accuracy of measurements. This nonsignificant difference was further verified by calculating the paired REE ratio differences (paired t-tests) and the Bland-Altman biases of REE ratio means between the first 1-min sets of measurements (125 min) and the second sets (26 50 min) of measurements during both periods, before and after suctioning. These two subcohorts (halves) of the within-groups sets of measurements (125 vs 26 50 sets of 1-min measurements) were equally shown not to differ during either the presuctioning (1.2%; 95% confidence intervals 14.9% to 12.6%) or the postsuctioning (0.3%; 95% confidence intervals
Figure 2. Bland-Altman plots of average paired differences of means of oxygen consumption (VO2) (mL/min) (a) and carbon dioxide production (VCO2) (mL/min) (b) measured by E-COVX before and after suctioning in 28 individual suctioning episodes (n 280 pre- and 280 postendotracheal suctioning (ETS) 5-min means of 1400 1400 1-min paired measurements, respectively).
Figure 3. Series of 50-min means (bars) of resting energy

expenditure (REE) (kcal/d) calculated by E-COVX before and after suctioning in 28 individual suctioning episodes, showing an impressive similarity over a wide range of values. Patients age and body mass explain the different values recorded between individual suctioning episodes (age range, 6 17 yr).
Figure 4. Comparison of measured oxygen consumption

(VO2) (mL/min) and carbon dioxide production (VCO2) (mL/min) before and after suctioning in different ventilatory modes. The box-whisker plots show the median (horizontal line within the box) and the 10th and 90th percentiles (whiskers). The box length is the interquartile range. PRVC pressure-regulated controlled ventilation; SIMV synchronized intermittent mandatory ventilation; PS pressure support ventilation.
13.6% to 13%) period (Fig. 5). Hemodynamic measurements and SaO2 remained stable throughout the study.
DISCUSSION
We showed that in critically ill children without lung pathology, pulmonary mechanics and indirect calorimetry measurements are not influenced after uneventful open ETS for the early period extending from 5 min to 1 h after the procedure. We also showed that the new compact modular metabolic monitor E-COVX is able to reliably and correctly measure, record, and calculate continuous spirometry and metabolic indices 5 min after uneventful suctioning as accurately as before suctioning, without the previously described need for extra calibrations or prolonged waiting periods for stabilization.6 Provided the
patient is kept well sedated, the accuracy of measurements is ascertained at different ventilator modalities and at different timepoints before or after suctioning, during an earlier or later period. Using the brief intervention of ETS and disconnection of the ventilator, we did not demonstrate significant changes in lung mechanics in children without severe or primary lung pathology. Unavoidably, patients with secondary mild-to-moderate lung inflammation, complicating sepsis, or head injury were also included in our study. However, these patients did not have any significant pulmonary mechanics pathology and tolerated ETS uneventfully. Similarly, in ventilated children with variable lung pathology,
877
Figure 5. Bland-Altman plot of average paired differences of ratio means of resting energy expenditure (REE) between the first 1-min sets of measurements (125 min) and the second sets (26 50 min) before (a) and after (b) suctioning (n 140 5-min means of 700 1-min paired measurements for each semiperiod). there was no evidence that uneventful suctioning reduces airway resistance.11 Additionally, in a smaller sample size study of 10 mechanically ventilated adults with mild-to-moderate acute respiratory failure, lung volume returned to baseline within 10 min of suctioning with an open system. Again, in patients without severe lung disease, these changes were transient and rapidly reversible.12 Research data suggest that the duration of ETS and the size of ETs might have a significant impact on measurement results.13 In a randomized controlled trial there was no difference in dynamic compliance, expired airway resistance, or oxygen saturation between the experimental group receiving a single standardized suctioning procedure followed 5 min later by a standardized recruitment maneuver and the control group receiving only the single suctioning procedure, either immediately after the recruitment maneuver or after 25 min.14 In our study, we used brief hand ventilation for lung recruitment, and any changes that might have occurred immediately after suctioning (stabilization period, data not shown) had been reliably reversed by the start of (5 min) the monitoring period and throughout the monitoring period. In a randomized crossover study comparing outcomes after physiotherapy and suctioning in children on full ventilatory support, no significant group changes in expired tidal volume or respiratory compliance after either treatment were recorded, but only a tendency for resistance to decrease after physiotherapy but not ETS.15 In agreement with our results, other researchers, measuring airway and pulmonary resistances according to the end-inspiratory and endexpiratory occlusion methods before and after ETS, found that both resistances increased transiently only, but returned to baseline values at 1 min after ETS without exhibiting any change thereafter16 or affecting the physiological and alveolar dead space.17 The accuracy of metabolic measurements in our study may have been related to the circulatory and
878
ventilatory stability of our patients, indirectly indicated by the postsuctioning stability of the heart rate, arterial blood pressure, total respiratory rate, minute ventilation, and ETco2 values. It has been previously shown in predicting the 24-h energy expenditure that 30-min indirect calorimetry was within 20% of 24-h measurements for 89% of intervals, but its accuracy was maximized if a 30-min study was performed when minute ventilation, heart rate, systolic blood pressure, and respiratory rate were near the days average.18 In a prospective simultaneous clinical comparison study, poor agreement exceeding a possible clinical acceptability of 20% was found between the Deltatrac II and E-COVX or Evita 4 metabolic monitors, leading to the conclusion that the E-COVX provides less accurate measurements of metabolic gas exchange in stable ventilated patients.19 This study, however, was criticized for performing simultaneous measurements with sampling volumes of 150 200 mL/min, affecting the VO2 and VCO2 accuracy of measurements, which varied according to the minute volume.20 Additionally, it has been previously shown that the Deltatrac II measures VO2 with a mean error of 9.4% and the VCO2 with 1.2%, which are higher than the 0.6% and 1% respective mean errors of the early postsuctioning accuracy in our E-COVX study.4 In a study in ventilated adult patients, averaging of continuous VO2 data with the E-COVX module resulted in only small errors, mainly related to inaccurate tidal volume measurements because of the water accumulation in the pneumotachograph with water bath humidifiers.21 Although in a previous report using water bath humidifiers in lightly sedated adults, the E-COVX monitor did not provide measurements within a clinically accepted range when compared with the Deltatrac II8; in this study performed in well-sedated children only, we avoided periods of invalid data by only using heat- and moisture-exchange filters (excluding heated water bath humidification). Similarly, we avoided additional inaccuracies by carefully recording patients
tidal volumes on the ventilator, thereby selecting the most appropriate flow sensor (P-Lite 15300 mL, D-Lite 300 mL). This study has some limitations. First, it includes patients without lung injury, whose lungs are more easily rerecruited, so that we cannot extrapolate these results to patients with acute lung injury. Second, the patients were sedated or well synchronized to the ventilator. Therefore we cannot predict what differences might occur in an awake patient breathing on his own but ineffectively triggering the ventilator. Although the similarity of the results obtained in this study to results of studies including patients with pulmonary diseases and/or requiring recruitment maneuvers might suggest that these findings could be generalizable to other populations of critically ill patients,11,1316 such a hypothesis should definitely be tested in future studies. In conclusion, our results show that in well-sedated children without lung pathology, pulmonary mechanics, and gas exchange measurements are not influenced after uneventful open ETS as early as 5 min after the procedure. The E-COVX is able to reliably record spirometry and metabolic indices after uneventful suctioning as accurately as before suctioning, at different ventilator modalities, and at different timepoints before or after suctioning, and does not need prolonged waiting periods for stabilization. REFERENCES
1. Boullata J, Williams J, Cottrell F, Hudson L, Compher C. Accurate determination of energy needs in hospitalized patients. J Am Diet Assoc 2007;107:393 401 2. Briassoulis G. Nutritional assessment in the critically ill child. Curr Pediatr Rev 2006;2:233 43 3. Joosten KF, Jacobs FI, van Klaarwater E, Baartmans MG, Hop WC, Merila inen PT, Hazelzet JA. Accuracy of an indirect calorimeter for mechanically ventilated infants and children: the influence of low rates of gas exchange and varying FIO2. Crit Care Med 2000;28:3014 8 4. Melendez JA, Veronesi M, Barrera R, Ferri E, Miodownik S. Determination of metabolic monitor errors and precision under clinical conditions. Clin Nutr 2001;20:54751 5. Scheeren TW, Krossa M, Merila inen P, Arndt JO. Error in measurement of oxygen and carbon dioxide concentrations by the DeltatracII metabolic monitor in the presence of desflurane. Br J Anaesth 1998;80:521 4
6. McClave SA, Snider HL, Ireton-Jones C. Can we justify continued interest in indirect calorimetry? Nutr Clin Pract 2002; 17:133 6 7. McLellan S, Walsh T, Burdess A, Lee A. Comparison between the Datex-Ohmeda M-COVX metabolic monitor and the Deltatrac II in mechanically ventilated patients. Intensive Care Med 2002;28:870 6 8. Meyer R, Briassouli E, Briassoulis G, Habibi P. Evaluation of the M-COVX metabolic monitor in mechanically ventilated adult patients. e-SPEN 2008;3:e2329 9. Maggiore SM, Lellouche F, Pigeot J, Taille S, Deye N, Durrmeyer X, Richard JC, Mancebo J, Lemaire F, Brochard L. Prevention of endotracheal suctioning-induced alveolar derecruitment in acute lung injury. Am J Respir Crit Care Med 2003; 167:121524 10. Cereda M, Villa F, Colombo E, Greco G, Nacoti M, Pesenti A. Closed system endotracheal suctioning maintains lung volume during volume-controlled mechanical ventilation. Intensive Care Med 2001;27:648 54 11. Morrow B, Futter M, Argent A. Effect of endotracheal suction on lung dynamics in mechanically-ventilated paediatric patients. Aust J Physiother 2006;52:121 6 12. Ferna ndez MD, Piacentini E, Blanch L, Ferna ndez R. Changes in lung volume with three systems of endotracheal suctioning with and without pre-oxygenation in patients with mild-to-moderate lung failure. Intensive Care Med 2004;30:2210 5 13. Moriyama S, Utoh J, Okamoto K, Tanaka M, Kunitomo R, Hara M, Kitamura N. Direct expiratory gas analysis after hypothermic cardiopulmonary bypass. Ann Thorac Cardiovasc Surg 1999;5:150 5 14. Morrow B, Futter M, Argent A. A recruitment manoeuvre performed after endotracheal suction does not increase dynamic compliance in ventilated paediatric patients: a randomised controlled trial. Aust J Physiother 2007;53:1639 15. Main E, Castle R, Newham D, Stocks J. Respiratory physiotherapy vs. suction: the effects on respiratory function in ventilated infants and children. Intensive Care Med 2004; 30:1144 51 16. Main E, Stocks J. The influence of physiotherapy and suction on respiratory deadspace in ventilated children. Intensive Care Med 2004;30:11529 17. Guglielminotti J, Desmonts JM, Dureuil B. Effects of tracheal suctioning on respiratory resistances in mechanically ventilated patients. Chest 1998;113:1335 8 18. Smyrnios NA, Curley FJ, Shaker KG. Accuracy of 30-minute indirect calorimetry studies in predicting 24-hour energy expenditure in mechanically ventilated, critically ill patients. JPEN J Parenter Enteral Nutr 1997;21:168 74 19. Singer P, Pogrebetsky I, Attal-Singer J, Cohen J. Comparison of metabolic monitors in critically ill, ventilated patients. Nutrition 2006;22:1077 86 20. Meyer R, Habibi P. Comparing methods for measuring energy expenditure in the critically ill. Nutrition 2007;23:281 21. Donaldson L, Dodds S, Walsh TS. Clinical evaluation of a continuous oxygen consumption monitor in mechanically ventilated patients. Anaesthesia 2003;58:455 60
879
Long-Acting Local Anesthetics Attenuate FMLP-induced Acute Lung Injury in Rats

Marcus T. Schley, MD* Matthias Casutt, MD* Christoph Haberthu r, PhD* Martin Dusch, MD Roman Rukwied, PhD Martin Schmelz, MD, PhD Joachim Schmeck, MD, PhD Guido K. Schu pfer, MD, PhD* Christoph J. Konrad, MD, PhD*
BACKGROUND: Endothelin-1 (ET-1) is a mediator of lung diseases and a potent pulmonary vasoconstrictor. In addition to thromboxane A2, it participates in the formation of lung edema. Both lidocaine and mepivacaine attenuate the increase of pulmonary arterial pressure (PAP) and lung edema development. We examined the effects of procaine, bupivacaine, and ropivacaine on experimentally evoked PAP increase and ET-1 release. METHODS: PAP and lung weight were measured in isolated rat lungs during perfusion with Krebs-Henseleit hydroxyethyl starch buffer. Bupivacaine, ropivacaine, or procaine was added to the solution at concentrations of 102107 mg/kg. ET-1 levels were measured in the perfusate by enzyme-immunoassay, and thromboxane A2 levels were assayed by radioimmunoassay. N-formyl-l-leucinemethionyl-l-phenylalanine was used to activate human polymorphonuclear neutrophils. RESULTS: Bupivacaine, ropivacaine, and procaine significantly attenuated increases of PAP (P 0.05) and resulted in a reduction of lung weight in these treatment groups compared with the sham group (P 0.05). The long-acting anesthetics bupivacaine and ropivacaine (P 0.05), but not procaine, reduced ET-1 levels, produced low inflammation rates, and did not affect lung structures at doses from 103 to 106 mg/kg. CONCLUSION: Bupivacaine and ropivacaine attenuated N-formyl-l-leucine-methionyll-phenylalanine-induced PAP, reduced lung edema, and diminished ET-1 release. Lidocaine and mepivacaine are more effective in reducing PAP and edema formation, but long-acting local anesthetics also inhibit ET-1 depletion and therefore have increased anti-inflammatory properties.
(Anesth Analg 2009;109:880 5)
ajor surgical procedures are performed under regional anesthesia, either alone or as a supplement to general anesthesia. The use of regional anesthesia leads to reduced postoperative mortality, particularly by improving pulmonary function and also decreasing cardiac or ileus complications.13 In animal experiments, endothelin-1 (ET-1), a potent pulmonary vasoconstrictor, and thromboxane A2 (TXA2) have been shown to contribute to lung edema.46 The long-acting local anesthetic (LA) ropivacaine decreased the TXA2 stable analog U46619-evoked pulmonary arterial pressure (PAP). In addition, bupivacaine attenuated the
From the *Department of Anaesthesiology and Operative Intensive Care Medicine, Kantonsspital, Lucerne, Switzerland; Department of Anaesthesiology and Intensive Care Medicine, University of Heidelberg, Heidelberg, Germany; and Department of Anaesthesiology, University of Mainz, Mainz, Germany. Accepted for publication April 7, 2009. The first two authors equally contributed to this work. Supported by a grant from the Medical Faculty Mannheim, University of Heidelberg. Address correspondence and reprints requests to Dr. Christoph J. Konrad, Department of Anesthesiology and Operative Intensive Care Medicine, Kantonsspital Lucerne, CH-6000 Lucerne 16, Switzerland. Address e-mail to christoph.konrad@ksl.ch. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae5ef1
TXA2-induced contraction of aortic rings, increased the activated clotting time, and prolonged the effects of the TXA2-receptor antagonist SQ29548.79 LAs also inhibited TXA2-mediated platelet aggregation but this particular antithrombotic effect still has to be verified.10 Neutrophil function, however, was not altered by bupivacaine, whereas ropivacaine dose-dependently suppressed the Ca2 response in human neutrophils and had an inhibitory effect on the formation of oxygen radicals, hydrogen peroxides, and hydroxides.11 Conversely, ropivacaine did not impair chemotaxis or phagocytosis and failed to decrease protein kinase C activity.11 Inhibition of protein kinase C, for instance by systemic administration of LAs, has been found to excite N-methyl-daspartate receptor activation, thereby preventing pain and hyperalgesia.12 Anti-inflammatory properties of both bupivacaine and ropivacaine have been indicated by their inhibiting effect on lipopolysaccharide (LPS)-evoked release of proinflammatory cytokines (tumor necrosis factor-, interleukin [IL]-1beta, IL-6) from macrophages, their attenuation on LPS-induced mRNA upregulation for intercellular adhesion molecule 1, and their blockade of LPS-evoked increase of neutrophils in rat lungs.1315 The functional impact of long-acting LAs on lung perfusion, however, is still unclear. As shown by our
880
previous experiments, the short-acting LAs lidocaine and mepivacaine attenuated the N-formyl-l-leucinemethionyl-l-phenylalanine (FMLP)-induced increase of PAP and lung edema. Intriguingly, lidocaine had a distinct dose-dependent effect on PAP but mepivacaine did not show this effect.16 In the FMLP-induced lung injury model, we examined the dose-dependent inhibitory effects of the long-acting LAs bupivacaine and ropivacaine on PAP increase. Previous studies have demonstrated that the increase in PAP can be attenuated by inhibiting ET-1 receptor binding.5 Therefore, we additionally investigated the anti-inflammatory potency of bupivacaine and ropivacaine by quantifying ET-1 and TXA2 levels in the lung perfusate as major mediators contributing to lung edema development. Finally, the results obtained for long-acting LAs were compared with effects gained by the administration of a shortacting LAs of the ester type.
Assessment of Lung Weight and Pulmonary Arterial Pressure

PAP and weight of the isolated lungs were recorded continuously by means of pressure (R-10 Series, Rikadenki, Freiburg, Germany) and weight (KWS3073, Hottinger-Baldwin, Darmstadt, Germany) transducers. Relative changes of weight, analyzed before, during, and after lung injury, determined lung edema formation. A circuit system connected to a reservoir was used to exclude impact of perfusion pressure changes caused by the injection of the tested fluids. Because of the constant perfusion flow rate, any alteration in perfusion pressure directly reflects alterations in pulmonary vascular resistance. The final component concentrations in the buffer were starch 50 g/L; Na 138 mmol/L; K 4.5 mmol/L; Mg2 1.33 mmol/L; Ca2 2.38 mmol/L; Cl 135 mmol/L, HCO3 12 mmol/L, and glucose 12 mmol/L. The osmolality was approximately 330 mOsm/kg (Mikro-Osmometer; Roebling Messtechnik, Berlin, Germany). The pH of the perfusate was adjusted to 7.4 using 1 M NaHCO3. Intermittently, perfusate samples were taken for measurements of Po2, Pco2, and O2 saturation (ABL 330; Radiometer Copenhagen, Copenhagen, Denmark). The impact of Pco2 was investigated at both physiologic and pathophysiologic concentrations before the study, and values of Pco2 were kept constant in physiologic ranges throughout the experiment to avoid an influence of Pco2 changes on PAP.
METHODS
The study protocol and experimental procedure were approved by the animal subject protection committee of the University of Heidelberg, as well as by the responsible regional governmental committee. The Council of the American Physiology Society principles guiding the care, handling, and use of the animals were followed.
Isolated Ventilation Lung Perfusion

Sprague-Dawley rats (Charles River Laboratories, Kent, UK) of both sexes weighing 320 375 g were used. Anesthetized lung donors (ketamine/xylazine 80/10 mg/kg) were intubated through a tracheostomy with a 16-gauge cannula and the lungs were ventilated with atmospheric air containing 4% CO2 at 25 breaths/min, 7 mL/kg tidal volume, and 0.51 cm H2O positive end-expiratory pressure (Statham, PD 23, Hato Rey, Puerto Rico). After laparotomy, all donors were given heparin sodium (1000 IE/kg) for anticoagulation through a renal artery. After median sternotomy, the pleurae and pericardium were opened and the thymus excised. A ligature was passed through the transverse sinus to encircle the aorta and pulmonary artery trunk. A primed, blunt cannula (1.0/1.8 mm inner diameter/outer diameter) was passed into the pulmonary artery trunk and secured with the previously placed ligature. The inferior vena cava was clamped, the left atrial appendage excised, and the right ventricular outflow tract opened. The tracheal cannula was then clamped and the lung block excised. Isolated lungs were placed on an electronic scale (Hottinger; Baldwin Messtechnik Type U1, Darmstadt, Germany) in a temperature-controlled (37C) and humidified chamber and perfused with cell-free and plasma-free KrebsHenseleit hydroxyethyl starch buffer (KHHB) solution (Haes-steril 10%, 200/0.5; Fresenius AG, Bad Homburg, Germany) in a recirculating system (circulating volume 20 mL). Colloid osmotic pressure was maintained at 2325 mm Hg and flow rate kept constant at 15 mL/min.
Measurement of Endothelin-1
ET-1 was measured by a commercially available enzyme-linked immunoassay test kit with a detection limit of 0.05 fmol/mL (Amersham, Braunschweig, Germany). The cross-reactivity of the ET-1 antibody was 100% with ET-1, 100% with ET-2, 0.001% with ET-3, 0.07% with big ET-1, and 0.0006% with sarafotoxin 6b. Materials were purchased from Roth, Sigma and Biozym, Germany.
Radioimmunoassay of TX A2
Thromboxane B2 is the stable hydrolysis metabolite of TXA2 and was assayed serologically from 100 L of recirculating KHHB by radioimmunoassay (detection limit 10 pg/mL). Rat anti-TX was purchased from Paesel (Frankfurt, Germany), [3H]-labeled thromboxane B2 from New England Nuclear (Dreieich, Germany), and precipitating goat anti-rat antibodies from Calbiochem-Behring (Frankfurt, Germany).
Isolation, Preparation, and Stimulation of Human Granulocytes

Human polymorphonuclear neutrophils (PMN) were obtained from healthy donors. As described previously, human granulocytes were isolated from 100 mL of heparinized human blood by Percoll density gradient (55% Percoll: D 1.77 g/mL; 69% Percoll: D 1.095 g/mL).17 The cells were counted and incubated in a bouillon solution (RPMI/fetal calf
881
serum 5%) after final separation and washing for 90 min. The granulocytes were isolated after incubation and rewashed as follows: 190 106 (110 106250 106). The viability in the trypan blue exclusion test was more than 95%. Percoll/HEPES (Sigma, Munich, Germany) and Hanks balanced salt solution (GIBCO, Paisley, Scotland) were used for cell isolation. FMLP was used as chemoattractant peptide to activate human granulocytes.
Histological Preparation
After pretreatment with liquid nitrogen, lungs were fixed in formalin (10%) and embedded in paraffin. The 4-mm slices were stained with hematoxylin and eosin and analyzed with light microscopy (Leica, Wetzlar, Germany). Image processing was performed with Leica Qwin Software, Version May 1997 (Leica, Wetzlar, Germany). Examinations were performed by two independent observers, who analyzed leukocytes semiquantitatively in the lung tissue by microscopy. Figure 1. The effect of long-acting (bupivacaine and ropivacaine) and short-acting (procaine) local anesthetics on pulmonary arterial pressure (PAP) after N-formyl-l-leucinemethionyl-l-phenylalanine (FMLP)-induced activation of human granulocytes. Under sham conditions (native KrebsHenseleit hydroxyethyl starch buffer solution [KHHB], with FMLP-activated human granulocytes) in six lung preparations, PAP almost increased to 8.2 mm Hg 1.11 (gray rectangle). A significant dose-dependent attenuation of PAP increase could be observed in all local anesthetics above a concentration of 105 mg/kg (P 0.05, analysis of variance [ANOVA]), whereas procaine already showed a significant effect at 107 mg/kg (P 0.05, ANOVA).
A total of 114 isolated lung preparations were randomly assigned to three groups, i.e., a procaine, a bupivacaine, and a ropivacaine group. LAs were added at final concentrations of 107, 106, 105, 104, 103, and 102 mg/kg to the KHHB solution used for the lung perfusion, which covers the range of clinically analyzed LA plasma concentrations of about 1 g/mL.18,19 In addition to each drug concentration, a sham group was investigated in six lungs. After the preparation of the lungs, an equilibration period of 30-min KHHB perfusion was performed to ensure a consistent vascular tone and permeability. Thereafter, LAs were perfused for 10 min, which was followed by a 60-min stimulation period performed by FMLP (1 M) perfusion to activate PMN, as described previously.20 The duration of FMLP-induced lung injury, however, was limited to a maximum FMLP perfusion period of 120 min (data not shown). LAs were investigated at increasing concentrations. Before each LA concentration increase, during the 30-min equilibration period, human granulocytes (190 106 cells) were added to the KHHB perfusate. TXA2 and ET-1 concentrations were determined in the perfusate samples at 5, 15, 30, and 60 min after granulocyte activation.
Figure 2. All local anesthetics used caused a significant

attenuation of the weight gain after N-formyl-l-leucinemethionyl-l-phenylalanine (FMLP)-induced activation (P 0.05, analysis of variance [ANOVA]). Data are presented as mean sem of all dosages in each group. A dose dependency was not observed (data not shown).
RESULTS
Bupivacaine, ropivacaine, and procaine attenuated the FMLP-induced increase of PAP (P 0.05). At the lowest concentrations of 107 mg/kg, the short-acting anesthetic procaine was most effective in mitigating the FMLP-induced PAP increase (P 0.05, ANOVA). Bupivacaine and ropivacaine required higher concentrations, i.e., a minimum of 105 mg/kg (P 0.05, ANOVA), to attenuate a PAP increase (Fig. 1). Administration of bupivacaine, ropivacaine, and procaine resulted in a significant reduction of lung weight compared with the sham treatment (P 0.05, ANOVA). All administered LAs were equally effective in evoking weight loss (Fig. 2), and a dosedependent effect was not observed (data not shown).
Statistics
For statistical analysis, data are presented as mean se of the mean. Normally distributed data (Shapiro-Wilks test) were analyzed by one-way analysis of variance (ANOVA) followed by the Scheffe multiple-range test (STATISTICA Version 5.1 for Windows). For paired samples, t-tests were used to analyze differences within the groups. A value of P 0.05 was considered statistically significant.
882
Local Anesthetics and Inflammation
Figure 3. Effect of long-lasting and short-lasting local anesthetics

on endothelin-1 (ET-1) level determined after N-formyl-lleucine-methionyl-l-phenylalanine (FMLP)-induced granulocyte activation (mean values at all dosages). Bupivacaine and ropivacaine both reduced the ET-1 level (P 0.05, analysis of variance [ANOVA]).
Figure 5. A compact accumulation of polymorphonuclear

neutrophils (PMN) and extravasation of perfusate is identifiable in the sham group (A). Pretreatment with local anesthetics attenuated the invasion of PMN in the procaine (B), bupivacaine (C), and ropivacaine (D) groups. Histological samples of bupivacaine and ropivacaine were obtained after pretreatment with a concentration of 103 kg/mg, and the sample of procaine after treatment with 104 mg/kg. Magnification 25, hematoxylin/eosin staining.
Figure 4. Thromboxane A2 (TXA) levels (as stable metabolite

thromboxane B2 [TXB2]) were also determined after granulocyte activation. No significant changes could be observed in TXA release compared with the sham group. Mean values at all dosages are shown.
Bupivacaine and ropivacaine reduced the FMLPprovoked release of ET-1 (P 0.05, ANOVA) to a similar extent, whereas procaine had no effect on ET-1 levels (Fig. 3). TXA2 release was not inhibited by either short-acting or long-acting LAs (Fig. 4). FMLP activation of the human granulocytes resulted in a compact invasion in the lung tissue followed by an acute granulocytic alveolitis. We observed PMN accumulated nearly ubiquitously in the tissue and edema of all alveolar structures. Furthermore, the number and form of lung alveoli or lung ducts were altered. Also, the interstitial space was infiltrated with granulocytes and macrophages (Fig. 5A). Both long- and short-acting LAs were able to attenuate granulocyte invasion in the lung tissue. Although the granulocyte invasion was still moderate in the procaine group (Fig. 5B), the inflammation rate was very slow, and lung structures remained unaffected in the bupivacaine and ropivacaine groups (Figs. 5C and 5D). This protective effect could be observed by using concentrations of 103106 mg/kg.
such as tumor necrosis factor-, prostaglandins, or ILs.1315 The present results obtained in a rat acute lung injury model demonstrate a reduction of the metabolite ET-1. No change of the mediator TXA2, however, was observed in the presence of the longacting LAs bupivacaine and ropivacaine. Furthermore, these LAs in the rat lung attenuated both the FMLP-induced increase of PAP and the development of edema formation. Direct comparison with other LAs revealed that the short-acting procaine apparently was more effective but the long-acting LAs also significantly attenuated lung injury.
Anti-inammatory Properties of LAs

In general, these data support the previously assumed suggestion that LAs interfere with inflammatory defense systems by inhibiting granulocytic adherence and attenuating granulocytes migration into inflamed sections.21,22 Long- and short-acting LAs obviously differ in their anti-inflammatory properties in a dose-dependent manner. This was shown by the chemoattractant and priming agents FMLP and LPS in human PMNs, in which both short- and long-acting LAs inhibited PMNs interaction at clinically relevant concentrations.23,24 These findings support the hypothesis that LAs have anti-inflammatory properties.25 The present results obtained in rat lungs corroborate the anti-inflammatory role of LAs in the context of facilitated lung ventilation under pathophysiologic conditions. In particular, pathophysiologic conditions imply major pneumonia risks that are promoted and maintained by high pulmonary ventilation pressures. These may cause increased endothelial leakage and enhanced lung edema formation.1,2 Increased lung edema consequently enhances pulmonary resistance, which again would require enhanced pulmonary
DISCUSSION
LAs contribute to reduction of inflammation by reducing the release of proinflammatory mediators,
883
pressures for ventilation. Thereby, the risk of endothelial damage is increased, lung edema formation is facilitated, and thus predictors that determine pulmonary morbidity are increased.1,2 To overcome this vicious cycle, the pressure of ventilation should be kept at low levels. This can be achieved by the prevention of edema formation or by inhibiting endothelial damage. The present experimental study demonstrates in an animal ex vivo model that pretreatment with LAs reduces edema formation, endothelial damage, and pulmonary resistance. These data indicate that septic pulmonary risk factors may be diminished by LAs.
Thromboxane A2 Release
Similarly, the long-acting LAs bupivacaine and ropivacaine did not change FMLP-induced TXA2 release. However, TXA2 may be involved in the mediation of ET-1induced vasoconstriction. Pretreatment with type A endothelin receptor antagonist LU135252 significantly reduced the pressure reaction and generation of TXA2 after air embolism.5
Clinical Implications and Limitations

High plasma concentrations of LAs may be toxic for the central nervous system and heart tissue. Of these, for instance, bupivacaine was more cardiotoxic than ropivacaine, an effect that could not be alleviated with clonidine.32 In contrast, patients receiving epidural anesthesia during major surgery are less likely to suffer from thromboembolic complications than patients receiving general anesthesia.33 LAs abolish hypercoagulability without impairing normal aggregation or coagulation processes, thus demonstrating profound anti-inflammatory properties.22 Accordingly, in the present experimental study, we found that both short- and long-acting LAs lead to antiinflammatory effects in the lung. The FMLP-induced pulmonary arterial hypertension was attenuated, lung edema ameliorated, and ET-1 release reduced. Thus, LAs may provide an additional clinical tool for lung protection. However, the clinical implication of such an experimental ex vivo animal study is limited and should not be overstated. Additional clinical in vivo trials are needed to confirm the role of LAs in reducing pneumonia risks. ACKNOWLEDGMENTS We thank I. Rossbach, Experimental Pain Research, Department of Anesthesiology, for her editorial assistance, and M. Lehma and A. Hagebeuker, Department of Anesthesiology, Theresia Hospital Mannheim, and J. Christophel, Center for Medical Research Mannheim, for their technical assistance. The molecular biological analyses were performed by Prof. M. Bauer, University of Jena. REFERENCES
1. Ballantyne JC, Carr DB, deFerranti S, Suarez T, Lau J, Chalmers TC, Angelillo IF, Mosteller F. The comparative effects of postoperative analgesic therapies on pulmonary outcome: cumulative meta-analyses of randomized, controlled trials. Anesth Analg 1998;86:598 612 2. Rodgers A, Walker N, Schug S, McKee A, Kehlet H, van Zundert A, Sage D, Futter M, Saville G, Clark T, MacMahon S. Reduction of postoperative mortality and morbidity with epidural or spinal anaesthesia: results from overview of randomised trials. BMJ 2000;321:1493 3. Kehlet H, Holte K. Effect of postoperative analgesia on surgical outcome. Br J Anaesth 2001;87:6272 4. Nakano J, Takizawa H, Ohtoshi T, Shoji S, Yamaguchi M, Ishii A, Yanagisawa M, Ito K. Endotoxin and pro-inflammatory cytokines stimulate endothelin-1 expression and release by airway epithelial cells. Clin Exp Allergy 1994;24:330 6 5. Schmeck J, Koch T, Patt B, Heller A, Neuhof H, van Ackern K. The role of endothelin-1 as a mediator of the pressure response after air embolism in blood perfused lungs. Intensive Care Med 1998;24:60511
Pulmonary Hypertension and Lung Edema

The mechanisms of LAs action on pulmonary hypertension remain largely unknown; however, there is evidence that LAs specifically activate transient receptor potential V1 (TRPV1) channels.26 Even though involvement of TRPV1 channels has not been demonstrated in this study, these channels play a major role in blood flow regulation. For instance, TRPV1 activation might reduce pulmonary resistance via release of vasodilatory neuropeptides, such as calcitonin generelated peptide, from nociceptive afferent nerve fibers.27 Moreover, mRNA for TRPV1 has been detected in granulocytes; interestingly, it does not mediate calcium increase or inward currents as in nociceptors.28 In contrast to the putative mechanisms mentioned earlier, ET-1, a peptide produced primarily by vascular endothelial cells, was found to be the most potent and long-lasting endogenous vasoconstrictor substance yet discovered.5 Elevated ET-1 levels have been detected in various states of lung pathology, such as primary pulmonary hypertension, asthma, and sepsis, and ET-1s involvement in elevating pulmonary vascular resistance has been found recently by selective receptor blocking experiments.5 Moreover, inappropriate activation of the ET-1 system has been clearly shown in patients with almost all types of pulmonary arterial hypertension. Here, we demonstrated that administration of bupivacaine, ropivacaine, and procaine resulted in an attenuation of the FMLP-induced increase of PAP and reduced lung edema and granulocyte invasion in the lung tissue when compared with sham treatment. Accordingly, it can be presumed that the quantitative granulocyte invasion was mainly responsible for the lung weight increase.29 In addition, antagonism of ET-1 receptors in the treatment of pulmonary arterial hypertension has been performed successfully previously in an animal model5 and human clinical trials.30 However, the impact of continuous epidural regional anesthesia with LAs on ET-1 release has not been conclusively shown and is still the subject of controversy. For instance, continuous epidural anesthesia with bupivacaine 0.125% performed in 20 patients during abdominal aortic surgery did not have beneficial effects on ET-1 levels.31
884
Local Anesthetics and Inflammation
6. Schmeck J, Heller A, Groschler A, Recker A, Neuhof H, Urbaschek R, Koch T. Impact of endothelin-1 in endotoxininduced pulmonary vascular reactions. Crit Care Med 2000;28:28517 7. Hahnenkamp K, Nollet J, Strumper D, Halene T, Rathman P, Mortier E, Van Aken H, Knapp J, Durieux ME, Hoenemann CW. Bupivacaine inhibits thromboxane A2-induced vasoconstriction in rat thoracic aorta. Anesth Analg 2004;99:97102 8. Fischer LG, Honemann CW, Patrie JT, Durieux ME, Rich GF. Ropivacaine attenuates pulmonary vasoconstriction induced by thromboxane A2 analogue in the isolated perfused rat lung. Reg Anesth Pain Med 2000;25:18794 9. Kohrs R, Hoenemann CW, Feirer N, Durieux ME. Bupivacaine inhibits whole blood coagulation in vitro. Reg Anesth Pain Med 1999;24:326 30 10. Lo B, Honemann CW, Kohrs R, Hollmann MW, PolanowskaGrabowska RK, Gear AR, Durieux ME. Local anesthetic actions on thromboxane-induced platelet aggregation. Anesth Analg 2001;93:1240 5 11. Mikawa K, Akamarsu H, Nishina K, Shiga M, Obara H, Niwa Y. Effects of ropivacaine on human neutrophil function: comparison with bupivacaine and lidocaine. Eur J Anaesthesiol 2003; 20:104 10 12. Hahnenkamp K, Durieux ME, Hahnenkamp A, Schauerte SK, Hoenemann CW, Vegh V, Theilmeier G, Hollmann MW. Local anaesthetics inhibit signalling of human NMDA receptors recombinantly expressed in Xenopus laevis oocytes: role of protein kinase C. Br J Anaesth 2006;96:77 87 13. Zhang XW, Thorlacius H. Inhibitory actions of ropivacaine on tumor necrosis factor-alpha-induced leukocyte adhesion and tissue accumulation in vivo. Eur J Pharmacol 2000;392:R13 14. Huang YH, Tsai PS, Huang CJ. Bupivacaine inhibits COX-2 expression, PGE2, and cytokine production in endotoxinactivated macrophages. Acta Anaesthesiol Scand 2008;52:530 5 15. Blumenthal S, Borgeat A, Pasch T, Reyes L, Booy C, Lambert M, Schimmer RC, Beck-Schimmer B. Ropivacaine decreases inflammation in experimental endotoxin-induced lung injury. Anesthesiology 2006;104:9619 16. Konrad CJ, Schuepfer GK, Neuburger M, Schley M, Schmelz M, Schmeck J. Reduction of pulmonary edema by short-acting local anesthetics. Reg Anesth Pain Med 2006;31:254 9 17. Hjorth R, Jonsson AK, Vretblad P. A rapid method for purification of human granulocytes using percoll. A comparison with dextran sedimentation. J Immunol Methods 1981;43:95101 18. Groeben H, Schwalen A, Irsfeld S, Stieglitz S, Lipfert P, Hopf HB. Intravenous lidocaine and bupivacaine dose-dependently attenuate bronchial hyperreactivity in awake volunteers. Anesthesiology 1996;84:5339 19. Tuominen M, Pitkanen M, Rosenberg PH. Postoperative pain relief and bupivacaine plasma levels during continuous interscalene brachial plexus block. Acta Anaesthesiol Scand 1987; 31:276 8
20. Hammerschmidt S, Wahn H. Comparable effects of HOCl and of FMLP-stimulated PMN on the circulation in an isolated lung model. Am J Respir Crit Care Med 1997;156:924 31 21. Eriksson AS, Sinclair R, Cassuto J, Thomsen P. Influence of lidocaine on leukocyte function in the surgical wound. Anesthesiology 1992;77:74 8 22. Hollmann MW, Durieux ME. Local anesthetics and the inflammatory response: a new therapeutic indication? Anesthesiology 2000;93:858 75 23. Tsai PS, Buerkle H, Huang LT, Lee TC, Yang LC, Lee JH. Lidocaine concentrations in plasma and cerebrospinal fluid after systemic bolus administration in humans. Anesth Analg 1998;87:601 4 24. Takao Y, Mikawa K, Nishina K, Maekawa N, Obara H. Lidocaine attenuates hyperoxic lung injury in rabbits. Acta Anaesthesiol Scand 1996;40:318 25 25. Fischer LG, Bremer M, Coleman EJ, Conrad B, Krumm B, Gross A, Hollmann MW, Mandell G, Durieux ME. Local anesthetics attenuate lysophosphatidic acid-induced priming in human neutrophils. Anesth Analg 2001;92:10417 26. Leffler A, Fischer MJ, Rehner D, Kienel S, Kistner K, Sauer SK, Gavva NR, Reeh PW, Nau C. The vanilloid receptor TRPV1 is activated and sensitized by local anesthetics in rodent sensory neurons. J Clin Invest 2008;118:76376 27. Inoue R, Jensen LJ, Shi J, Morita H, Nishida M, Honda A, Ito Y. Transient receptor potential channels in cardiovascular function and disease. Circ Res 2006;99:119 31 28. Heiner I, Eisfeld J, Luckhoff A. Role and regulation of TRP channels in neutrophil granulocytes. Cell Calcium 2003;33: 533 40 29. Clavijo LC, Carter MB, Matheson PJ, Wills-Frank LA, Wilson MA, Wead WB, Garrison RN. Platelet-activating factor and bacteremia-induced pulmonary hypertension. J Surg Res 2000;88:173 80 30. Leuchte HH, Meis T, El Nounou M, Michalek J, Behr J. Inhalation of endothelin receptor blockers in pulmonary hypertension. Am J Physiol Lung Cell Mol Physiol 2008;294:L7727 31. Piper SN, Boldt J, Schmidt CC, Maleck WH, Brosch C, Kumle B. Hemodynamics, intramucosal pH and regulators of circulation during perioperative epidural analgesia. Can J Anaesth 2000;47:6317 32. Gulec S, Aydin Y, Uzuner K, Yelken B, Senturk Y. Effects of clonidine pre-treatment on bupivacaine and ropivacaine cardiotoxicity in rats. Eur J Anaesthesiol 2004;21:2059 33. Hollmann MW, Wieczorek KS, Smart M, Durieux ME. Epidural anesthesia prevents hypercoagulation in patients undergoing major orthopedic surgery. Reg Anesth Pain Med 2001;26:21522
885
Obstetric Anesthesiology
Section Editor: Cynthia A. Wong
Focused Reviews
Intraoperative Awareness During General Anesthesia for Cesarean Delivery

Kay Robins, FRCA* Gordon Lyons, FRCA, MD
Intraoperative awareness is defined as the spontaneous recall of an event occurring during general anesthesia. A move away from rigid anesthetic protocols, which were designed to limit drug transmission across the placenta, has reduced the incidence of awareness during cesarean delivery to approximately 0.26%. Nevertheless, it remains an undesirable complication with potential for the development of posttraumatic stress disorder. Assessing depth of anesthesia remains a challenge for the anesthesia provider as clinical signs are unreliable and there is no sensitive and specific monitor. Bispectral Index monitoring with the goal of scores 60 has been recommended to prevent awareness. Induction drugs vary in their ability to produce amnesia and the period of hypnotic effect is affected by the rate at which they are redistributed. After initiation of anesthesia, volatile anesthetics should be administered to a target of 0.7 minimum alveolar anesthetic concentration, which has been shown to consistently achieve mean Bispectral Index scores 60. Because of its rapid uptake, nitrous oxide remains an important adjunct to reduce the risk of awareness during emergency cesarean delivery. In the absence of fetal compromise, there is no rationale for an inspired oxygen concentration above 0.33. Deeper levels of anesthesia reduce the incidence of awareness; current evidence does not suggest an increased risk of tocolysis or fetal morbidity.
(Anesth Analg 2009;109:886 90)
THE DILEMMA OF OBSTETRIC ANESTHESIA

The objectives of general anesthesia for cesarean delivery are to keep mother and fetus adequately oxygenated, while limiting fetal drug transmission and maintaining maternal comfort. Crawford1 called this conflict the dilemma of obstetric anesthesia and analgesia and said it epitomized the challenge and the attraction of the specialty. The balance of this conflict has changed over the years. Intraoperative recall during general anesthesia was unreported with the spontaneous breathing and ether of Mendelsons day, but this changed with the introduction of succinylcholine in the late 1950s when endotracheal intubation and muscle relaxation were popularized. Initially, anesthesia was provided largely by thiopental and nitrous oxide2 and was associated with an incidence of awareness up to 26%.3 A reluctance to load with a volatile anesthetic, and concern about lack of care for an anesthetized newborn
From the *Department of Anaesthesia, York Hospital, York; and Department of Obstetric Anaesthesia, St. James University Hospital, Leeds, UK. Accepted for publication April 24, 2009. Address correspondence and reprint requests to Kay Robins, FRCA, Department of Anaesthesia, York Hospital, York YO61 1PS, UK. Address e-mail to kayrobins14@hotmail.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af83c1
from an undeveloped neonatal service, might have helped make this frequent incidence of recall seem an acceptable side effect. The addition of halothane 0.5% (0.66 minimum alveolar anesthetic concentration [MAC]) to the anesthetic moved the balance further in the maternal direction, reducing awareness to around 1%,4 and throughout the 1970s, this was widely regarded as an acceptable incidence. The balance shifted further toward maternal comfort when it was demonstrated that awareness at cesarean delivery could be reduced by more generous doses of thiopental and more liberal use of a volatile anesthetic.5 This practice became more widely disseminated in the 1990s, a time in which access to neonatal resuscitation support became more widely available. Additionally, anesthesia providers were taking advantage of the electronic monitoring revolution (including measurement of end-tidal gas concentrations), which offered a dynamic alternative to the traditional recipe approach. Today, the incidence of awareness during anesthesia in the United States is believed to be between 0.1% and 0.2% of all patients undergoing general anesthesia, representing 20,000 40,000 cases per year.6 The risk appears to be higher when muscle relaxants are used and during cesarean delivery.7 In Australia and New Zealand, the Australian and New Zealand College of Anesthesia (ANZCA) Trial studied 1095 cesarean
886
Table 1. Terminology of Awareness37,44

Consciousness Recall Amnesia Wakefulness/ responsiveness Explicit memory Priming Learning State in which information from patients surroundings can be processed Ability to retrieve stored memories Absence of recall. Event not retained in long-term memory Unequivocal communication with an anesthetized patient without subsequent recall Recall of specific intraoperative clinical events Presentation of material to an anesthetized patient Evidence of communication or detection of priming through postoperative tests but without recall Postoperative evidence of priming but without recall
Implicit memory
deliveries and interviewed 763 women postoperatively; two women had recall giving an incidence of 0.26%.8
DEFINITIONS AND SCOPE

Awareness is defined as the spontaneous postoperative recall of an event that occurred during general anesthesia (Table 1).9 One difficulty with explicit memory of perioperative events is distinguishing between recall of genuine intraoperative events and emergence phenomena, because voices, the baby crying, and wound pain are part of the postoperative experience. A wider definition of recall takes in a spectrum that ranges from dreams, through recall of specific events, to full consciousness with paralysis and pain. Dreaming is often thought to be indicative of light anesthesia but more likely occurs during emergence from anesthesia and recovery.10 Crawford1 took the view that unpleasant dreams did reflect awareness and that the two should always be linked. Anecdotal reports have shown that, even when the content can be linked to intraoperative events, dreams are not necessarily unpleasant.5
INDUCTION OF ANESTHESIA
Monitoring cerebral function to detect awareness has advanced considerably in recent years, but the perfect monitor has yet to be developed. The most widely studied brain function monitor, Bispectral Index (BIS) monitoring, is easy to initiate but even rapid application may delay delivery of the fetus in an emergency cesarean delivery. A BIS monitor was used in 32% of 1095 general anesthetics studied as part of the ANZCA trial. Of note, 30% of Category 1* and 37% of Category 4* cesarean deliveries were monitored.8 Clearly, the limiting factor in the use of monitoring
*Category 1: emergency cesarean delivery indicated because of presence of condition which is of immediate threat to the life of the woman or fetus; Category 4: elective procedure, cesarean delivery can be scheduled to suit the woman and staff. Vol. 109, No. 3, September 2009
was not the urgency of the procedure. Another consideration is whether, within the context of an emergency cesarean delivery, a BIS score target 60 is attainable predelivery. When the target anesthetic concentration is 0.8 MAC or above, it seems that mean BIS scores 60 can be achieved,11,12 but without a commitment to this level of volatile anesthetic delivery at the outset, the rationale for BIS monitoring is lost. Intraoperative brain function monitoring during cesarean delivery has yet to become a mandatory requirement by any governing or regulatory agency. The risk of recall is increased with a rapid sequence induction of anesthesia as tracheal intubation and surgical incision follow in rapid sequence. There may be insufficient time to allow adequate uptake and distribution of volatile anesthetic to prevent awareness12 before redistribution causes brain levels of the induction drug to decrease. The choice and dose of induction drug then becomes critical. Many regard thiopental as the drug of choice but a single induction dose is soon redistributed with rapid recovery of consciousness. Recommended doses range from 3 to 7 mg/kg; in the ANZCA Trial the mean dose was 4.9 mg/kg.8,1317 Not surprisingly, larger doses of hypnotic drug result in a lower incidence of recall.5 The wide variation in recommended dose may reflect how anesthesiologists view their role in the conflict between fetal drug transmission and maternal comfort. There is general agreement that doses 4 mg/kg are unlikely to lead to fetal depression, and that doses in excess of 7 mg/kg are liable to do so.1 The degree of concern for maternal awareness might be expected to decide where, between those limits, the choice lies. At low doses, thiopental is mildly amnesic18 but it does not produce retrograde amnesia.19 Although thiopental remains the drug of choice, a new generation of anesthesiologists is largely untrained in its use. Propofol is now the most widely used IV drug in anesthesia but there are concerns over its capacity to produce neonatal depression and adequate depth of anesthesia. Celleno et al. examined the maternal electroencephalogram with either thiopental 5 mg/kg or propofol 2.4 mg/kg. Half of the propofol group had rapid low voltage (8 9 Hz) waves on their electroencephalogram suggestive of a light plane of anesthesia compared with 10% of the thiopental group.20 Another disadvantage of propofol is its long effect-site equilibration time, which slightly prolongs the period from injection to hypnosis. Other studies have provided no evidence for the superiority of thiopental compared with propofol.21 However, severe maternal bradycardia has been reported with propofol combined with succinylcholine.22 Propofol has a greater amnesic effect than thiopental23 through interference with long-term memory.24 Although there are no data on which dose is best to avoid awareness, 2.5 mg/kg is commonly used.25 Worldwide, there is little doubt that propofol is used for
887
cesarean delivery despite these concerns and, to date, without the accumulation of adverse reports.8,23 Ketamine is used as an induction drug 2% of the time.8 It is associated with less responsiveness and recall than thiopental when used at a dose of 1 mg/kg,26 but the sympathomimetic effects limit its use in preeclampsia, and when there are concerns about hypertension. The associated hallucinations and emergence phenomena are another problem, although both are dose related and possibly occur less frequently in obstetric patients.27 Ketamine may be of use to reduce hypotension at induction in the setting of hemodynamic instability. Benzodiazepines are used infrequently as sole induction drugs, although they may be used occasionally to supplement an induction sequence.8 Midazolam produces more profound amnesia than propofol,28 impairing both explicit and implicit memory,29 but the onset of hypnosis is slow and neonatal depression is slow to resolve. It does not produce retrograde amnesia.28
MAINTENANCE OF ANESTHESIA
The rapid redistribution of induction drugs underlines the importance of introducing an adequate volatile anesthetic as soon after induction as is practical. In some centers, the skin is prepared and the drapes applied before induction of anesthesia. Although this might be in the best interests of a fetus in need of immediate delivery, emergency surgery and inadequate uptake of the volatile anesthetic are known risk factors for awareness.12 Depth of anesthesia can be considered in terms of the MAC that is required to achieve anesthesia in 50% of the patients. MAC may be reduced in pregnancy by 25% 40%, possibly because of increased pain thresholds or analgesia administered in labor. Lower BIS scores were observed for similar anesthetic concentrations in pregnant compared with nonpregnant patients. A comparison between parturients with and without prior labor undergoing cesarean delivery found that prior labor was associated with lower intraoperative BIS values during sevoflurane/nitrous oxide general anesthesia.30 The rapid uptake of nitrous oxide makes it a useful adjunct despite being a weak anesthetic. The choice of concentration is secondary to that of the inspired oxygen requirement. The administration of 100% oxygen may improve 1-min Apgar scores31 but oxygenfree radicals have been detected in newborns after maternal administration of high oxygen concentrations,32 and resuscitation of neonates with oxygen was associated with poorer Apgar scores than air.33 A common recommendation is that 50% oxygen should be given,13,1517 but 33% has been shown to result in similar outcome provided there is no fetal compromise.34 Nitrous oxide has little influence on monitors
Gin T, Chan MTV. Pregnancy reduces the bispectral index during isoflurane anesthesia (abstract). Anesthesiology 1997; 87:A305.
of cerebral function35,36 but a concentration of 70% contributes around 0.5 MAC and, if less is given, a corresponding increase in volatile anesthetic is needed to compensate. Several textbooks recommend that the MAC of volatile drug administered predelivery should be approximately 0.51,15 despite evidence that this policy is associated with an incidence of awareness close to 1%.5 This finding is consistent with predelivery BIS scores at 0.5 MAC in 50% nitrous oxide that range between 57 and 64.12 In a small sample, anesthesia with 0.2% end-tidal isoflurane in 50% nitrous oxide gave BIS scores between 70 and 80 with evidence of learning but not spontaneous recall.37 When MAC was increased to 0.8 in nonobstetric patients, BIS scores between 40 and 60 were achieved but the incidence of awareness was still 0.21%.38 One point to consider is whether the target MAC should represent the MAC for the volatile drug alone or include the contribution of nitrous oxide. Because the effect of nitrous oxide on memory is uncertain, prudent advice would be to regard the target MAC as that of the volatile drug alone.39 However, increasing the concentration of the volatile drug introduces a new conflict as all volatile drugs are tocolytic and uterine contractility and tone decrease in a dose-dependent manner. The uterus will contract in response to oxytocin, however, provided MAC is 0.8 1.0.40 These operational limits should provide sufficient scope for adequate anesthesia without penalty. In pregnancy, reduced functional residual capacity and increased minute ventilation increase the rate of equilibration of blood and inspired concentration of the volatile anesthetic, although the pregnancyinduced increase in cardiac output counteracts this to some degree. Equilibration between inspired and brain concentrations may take 4 12 min depending on the volatile anesthetic. Uptake of volatile anesthetic therefore needs to be accelerated. McCrirrick et al.41 described an overpressure technique with initial vaporizer settings in excess of MAC to speed equilibration, but measurement of end-tidal vapor concentrations has offered a more dynamic approach. Indeed, investigators found no difference in the incidence of recall when noncesarean patients were randomized to adjustment of the vaporizer setting to deliver a target end-tidal concentration or BIS monitoring with a target BIS score 60.38 Unfortunately, because this study was underpowered, equivalence between the two techniques cannot be assumed. Isoflurane and sevoflurane are favored because of rapid uptake; for the former, the target end-tidal concentration should be in excess of 0.7%,11 and for the latter, an end-tidal concentration of 1.5% achieved mean predelivery BIS scores of 60.12 After delivery, the concentration of nitrous oxide may be increased and opioids may be administered. This will result in a reduction in reflex activity but not necessarily a reduction in the incidence of recall. The
888
Awareness During Cesarean Delivery
Table 2. Techniques to Avoid Awareness

Beware drug and equipment errors Brain function monitoring (e.g., Bispectral Index monitoring to achieve scores 60) Thiopental dose 57 mg/kg Target end-tidal volatile anesthetic monitoring to achieve concentration 0.8 MACa Highest concentration of nitrous oxide compatible with maternal and fetal oxygen requirements Opioid analgesia after delivery Consider benzodiazepines after delivery
a
There is no evidence of fetal morbidity with increased depth of anesthesia.
volatile anesthetic should be continued until completion of the operation, but in the event of uterine atony, it can be reduced and a small dose of midazolam or ketamine substituted. Suggested techniques to reduce the risk of awareness during cesarean delivery are summarized in Table 2.
BRAIN FUNCTION MONITORING

Routine brain function monitoring of patients undergoing general anesthesia is controversial, although in one study, it was shown to result in an 82% reduction in the incidence of awareness in patients undergoing procedures considered at high risk for awareness, including cesarean delivery.9 However, a low BIS score does not guarantee unconsciousness.42 Whether routine monitoring of brain function in the specific setting of general anesthesia for cesarean delivery can reduce the incidence of awareness has not been studied.
may help align expectation with experience and reduce the risk of litigation.45 When awareness occurs, a full account with precise details should be recorded in the medical record for future reference. An apology costs nothing and might avert legal proceedings; denial is unhelpful. Symptoms consistent with posttraumatic stress disorder, sleep disturbance, nightmares, irritability, and lack of concentration that can interfere with work, may follow. Counseling is recommended, but its efficacy is unknown. The anesthesiologist may also be distressed by the incident.46 Hull and Thorburn47 believe that awareness cannot occur without negligence and equated light anesthesia with inadequate anesthesia. An analysis of 81 incidents of awareness found that 32 occurred because of avoidable drug and equipment errors.46 The alternative view is that a low incidence of awareness during general anesthesia for cesarean delivery is unavoidable, but the difficulty of mounting a successful defense is acknowledged.47 REFERENCES
1. Crawford JS. Principles and practice of obstetric anaesthesia. 5th ed. Oxford: Blackwell Science, 1984 2. Hamer Hodges RJ, Bennet JR, Tunstall ME, Knight RF. General anaesthesia for operative obstetrics. Br J Anaesth 1959;31:152 63 3. Crawford JS. Awareness during operative obstetrics under general anaesthesia. Br J Anaesth 1971;43:179 82 4. Moir DD. Anesthesia for Caesarean section: an evaluation of a method using low concentration of halothane and 50 percent of oxygen. Br J Anaesth 1970;42:136 42 5. Lyons G, Macdonald R. Awareness during Caesarean section. Anaesthesia 1991;46:62 4 6. Sebel PS. The incidence of awareness during anesthesia: a multicenter United States study. Anesth Analg 2004;99:8339 7. Sandin R, Enlaund G, Samuelson P, Lennmarken C. Awareness during anaesthesia: a prospective case study. Lancet 2000; 355:70711 8. Paech MJ, Scott KL, Clavisi O, Chua S, McDonnell N. The ANZCA Trials Group. A prospective study of awareness and recall associated with general anaesthesia for caesarean section. Int J Obstet Anesth 2008;17:298 303 9. Myles PS, Leslie K, McNeil J, Forbes A, Chan MTV. Bispectral index monitoring to prevent awareness during anaesthesia: the B-Aware randomised controlled trial. Lancet 2004;363:1757 63 10. Leslie K, Skrzypek H, Paech MJ, Kurowski I, Whybrow T. Dreaming during anesthesia and anesthetic depth in elective surgical patients: a prospective cohort study. Anesthesiology 2007;106:33 42 11. Yeo SN, Lo WK. Bispectral index in assessment of adequacy of general anaesthesia for lower segment caesarean section. Anaesth Intensive Care 2002;30:36 40 12. Chin K, Yeo S. Bispectral index values at sevoflurane concentrations of 1% and 1.5% in lower segment cesarean delivery. Anesth Analg 2004;98:1140 4 13. Malinow AM. General anesthesia for cesarean delivery. In: Norris MC, ed. Obstetric anesthesia. 2nd ed. Philadelphia: Lippincott Williams & Wilkins 1998;37598 14. Kuczkowski KM, Reisner LS, Liu D. Anesthesia for cesarean section. In: Chestnut DH, ed. Obstetric anesthesia, principles and practice. 3rd ed. Philadelphia: Mosby Elsevier 2004;421 46 15. Biribo MA. Anesthesia for cesarean section. In: Birnbach D, Gatt S, Datta S, eds. Textbook of obstetric anesthesia. Philadelphia: Churchill Livingstone 2000;239 44 16. Paech MJ General anesthesia for cesarean section. In: Palmer CM, DAngelo R, Paech MJ, eds. Handbook of obstetric anesthesia. Oxford: Bios 2002;10513 17. Yentis S, May A, Malhotra S. Analgesia, anaesthesia and pregnancy. 2nd ed. Cambridge: Cambridge University Press, 2007
THE FETUS
Catecholamine secretion during light anesthesia promotes uterine vasoconstriction and tocolysis. Depressant effects from transplacental drug transmission are usually responsible for a lower 1-min Apgar score in neonates after general anesthesia compared with neuraxial techniques, but by 5 min, differences have largely disappeared. Provided neonatal resuscitative support is available, the effects of general anesthesia are wholly reversible and the uterine incision to delivery time is more important than the induction to delivery time for good neonatal outcome. Evidence is lacking that an awareness avoidance approach to general anesthesia has untoward neonatal effects beyond the first few minutes of life.
AVOIDING LITIGATION
Intraoperative awareness is one of several major patient concerns when undergoing general anesthesia; Klafta and Roizen43 showed that up to 54% of patients worry about the possibility of pain, paralysis, and mental distress during surgery. Current advice is that patients considered to be high risk should be informed of the possibility of awareness, when circumstances permit.44 It may not be appropriate to do this before an emergency cesarean delivery when anxiety can be extreme. When possible, a preoperative discussion
889
18. Veselis RA, Reinsel R, Feschenko VA, Wronski M. The comparative amnesic effects of midazolam, propofol, thiopental and fentanyl at equisedative concentrations. Anesthesiology 1997; 87:749 64 19. Dundee JW, Pandit SK. Studies on drug induced amnesia with intravenous anaesthetic agents in man. Br J Clin Pract 1972; 26:164 6 20. Celleno D, Capogna G, Tomassetti M, Costantino P, Di Feo G, Nisini R. Neurobehavioural effects of propofol on the neonate following elective caesarean section. Br J Anaesth 1989; 62:649 54 21. Gin T, OMeara ME, Kan AF, Leung RKW, Tan P, Yau G. Plasma catecholamines and neonatal condition after induction of anaesthesia with propofol or thiopentone at Caesarean section. Br J Anaesth 1993;70:311 6 22. Baraka A. Severe bradycardia following propofol-suxamethonium sequence. Br J Anaesth 1988;61:4823 23. Duggal K. Propofol should be the induction agent of choice for caesarean section under general anaesthesia. Int J Obstet Anesth 2003;12:2759 24. Polster MR, Gray PA, OSullivan G, McCarthy RA, Park GR. Comparison of the amnesic effects of midazolam and propofol. Br J Anaesth 1993;70:612 6 25. Dailland P, Cockshott ID, Lirzin JD, Jacquinot P, Jorrot JC, Devery J, Harmey JL, Conseiller C. Intravenous propofol during cesarean section: placental transfer, concentrations in breast milk and neonatal effects. A preliminary study. Anesthesiology 1989;71:82734 26. Baraka A, Louis F, Noueihid R, Diab M, Dabbous A, Sibai A. Awareness following different techniques of general anesthesia for Caesarean section. Br J Anaesth 1989;62:645 8 27. Shulterus R, Hill C, Dharamraj C, Banner T, Berman L. Wakefulness during cesarean section after anesthetic induction with ketamine, thiopental, or ketamine and thiopentone combined. Anesth Analg 1986;65:723 8 28. Ghoneim MM, Block RI, Sum Ping ST, El-Zahaby HM, Hinrichs JV. The interactions of midazolam and flumazenil on human memory and cognition. Anesthesiology 1993;79:118392 29. Bulach R, Myles PS, Russnak M. Double-blinded randomized controlled trial to determine extent of amnesia with midazolam given immediately before general anaesthesia. Br J Anaesth 2005;94:300 5 30. Yoo KY, Jeong CW, Kang MW, Kim SJ, Chung ST, Shin MH, Lee J. Bispectral index values during sevoflurane-nitrous oxide general anesthesia in women undergoing cesarean delivery: a comparison between women with and without prior labor. Anesth Analg 2008;106:182732 31. Piggott SE, Bogod DG, Rosen M, Rees GAD. Isoflurane with either 100% oxygen or 50% nitrous oxide in oxygen for caesarean section. Br J Anaesth 1990;61:255 62
32. Khaw KS, Wang CC, Ngan Kee W, Pang CP, Rogers MS. The effects of high inspired oxygen fraction during elective caesarean section under spinal anaesthesia on maternal and fetal oxygenation and lipid peroxidation. Br J Anaesth 2002;88:18 23 33. Saugstad OM, Rootwelt T, Aalen O. Resuscitation of asphyxiated newborn infants with room air or oxygen: an international controlled trial: the Resair 2 Study. Pediatrics 1998;102:e1 34. Lawes EG, Newman B, Campbell MJ, Irwin M, Dolenska S, Thomas TA. Maternal inspired oxygen concentration and neonatal status for caesarean section under general anaesthesia. Comparison of effects of 33% or 50% oxygen in nitrous oxide. Br J Anaesth 1988;61:250 4 35. Barr G, Jakobsson JG, Owall A, Anderson RE. Nitrous oxide does not alter bispectral index: a study with nitrous oxide as sole agent and as adjunct to intravenous anaesthesia. Br J Anaesth 1999;82:82730 36. Anderson RE, Jakobsson J. Entropy of EEG during anaesthetic induction: a comparative study with propofol or nitrous oxide as sole agent. Br J Anaesth 2004;92:16770 37. Lubke G, Kerssens C, Gershon R, Sebel P. Memory formation during general anesthesia for emergency cesarean sections. Anesthesiology 2000;92:1029 34 38. Avidan MS, Zhang L, Burnside BA, Finkel J, Searleman AC, Selvidge JA, Saager L, Turner MS, Rao S, Bottros M, Hantier C, Jacobsohn E, Evers AS. Anesthesia awareness and the bispectral index. N Engl J Med 2008;358:1097108 39. Sneyd JR, Mathews DM. Memory and awareness during anaesthesia. Br J Anaesth 2008;100:7423 40. Yildiz K, Dogru K, Dalgic H, Serin IS, Sezer Z, Madenoglu H, Boyad A. Inhibitory effects of desflurane and sevoflurane on oxytocin-induced contractions of isolated pregnant human myometrium. Acta Anaesthesiol Scand 2005;49:13559 41. McCrirrick A, Evans GH, Thomas TA. Overpressure isoflurane at caesarean section: a study of arterial isoflurane concentrations. Br J Anaesth 1994;72:122 4 42. Mychaskiw G, Horowitz M, Sachdev V, Heath BJ. Explicit intraoperative recall at a bispectral index of 47. Anesth Analg 2001;92:808 9 43. Klafta JM, Roizen M. Current understanding of patients attitudes toward and preparation for anaesthesia: a review. Anesth Analg 1996;83:1314 21 44. American Society of Anesthesiologists Task Force on Intraoperative Awareness. Practice advisory for intraoperative awareness and brain function monitoring. Anesthesiology 2006;104:847 64 45. Guerra F. Awareness during anaesthesia. Can Anaesth Soc J 1980;27:178 46. Bergman IJ, Kluger MT, Short TG. Awareness during general anaesthesia: a review of 81 cases from the Anaesthetic Incident Monitoring Study. Anaesthesia 2002;57:549 56 47. Hull C, Thorburn J. Controversies: awareness is due to negligence during general anaesthesia for Caesarean section. Int J Obstet Anesth 1997;6:178 80
890
Awareness During Cesarean Delivery
Economics, Education, and Policy

Section Editor: Franklin Dexter
Anesthesiologists with Substance Use Disorders: A 5-Year Outcome Study from 16 State Physician Health Programs
Gregory E. Skipper, MD* Michael D. Campbell, PhD Robert L. DuPont, MD
BACKGROUND: Anesthesiologists have a higher rate of substance use disorders than other physicians, and their prognoses and advisability to return to anesthesiology practice after treatment remain controversial. Over the past 25 yr, physician health programs (PHPs), created under authority of state medical regulatory boards, have become primary resources for management and monitoring of physicians with substance abuse and other mental health disorders. METHODS: We conducted a 5-yr, longitudinal, cohort study involving 904 physicians consecutively admitted to 1 of 16 state PHPs between 1995 and 2001. This report analyzed a subset of the data involving the 102 anesthesiologists among the subjects and compared them with other physicians. The main outcome measures included relapse (defined as any unauthorized addictive substance use, including alcohol), return to anesthesiology practice, disciplinary actions, physician death, and patient harm. RESULTS: Anesthesiologists were significantly less likely to enroll in a PHP because of alcohol abuse (odds ratio [OR] 0.4 [confidence interval {CI}: 0.20.6], P 0.001) and much more likely to enroll because of opioid abuse (OR 2.8 [CI: 1.74.4], P 0.001). Anesthesiologists had a higher rate of IV drug use, 41% vs 10% (OR 6.3 [CI: 3.810.7], P 0.001). During similar periods of monitoring, anesthesiologists received more drug tests, 101 vs 82 (mean difference 19 [CI: 335], P 0.02); however, anesthesiologists were less likely to fail at least one drug test during monitoring, 11% vs 23% (OR 0.4 [CI: 0.20.9], P 0.02). There was no statistical difference among rates of program completion, disciplinary actions, return to practice, or deaths, and there was no report of significant patient harm from relapse in any record. CONCLUSIONS: Anesthesiologists in our sample treated and monitored for substance disorders under supervision of PHPs had excellent outcomes similar to other physicians, with no higher mortality, relapse rate, or disciplinary rate and no evidence in their records of patient harm. It is postulated that differences of study design account for contradictory conclusions from other reports.
(Anesth Analg 2009;109:8916)
mong physicians, anesthesiologists have an unusually high incidence of substance use disorders. For example, a survey of 260 anesthesiologists from the Medical College of Wisconsin graduating between 1958 and 1988 reported that 32% used drugs to get
From the *Departments of Medicine and Psychiatry, University of Alabama School of Medicine, Montgomery, Alabama; and Institute for Behavior and Health, Rockville, Maryland. Supported by the Robert Woods Johnson Foundation. All authors participated in study design, implementation, and writing and editing the article. Skipper had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Skipper and DuPont were paid as Co-Principal Investigators and Campbell was paid by the Institute for Behavior and Health to assist in study design, statistical analysis, and review. Skipper is the medical director of the Alabama Physician Health Program. Neither DuPont nor Campbell have any affiliation with Physician treatment or Physician Health Programs. A. Thomas McLellon, PhD, was the recipient of the grant and contributed significantly to development and implementation of the larger study. A steering committee of the Federation of State Physician Health Programs oversaw the larger study design, implementation, and interpretation and reviewed this article for comment. Address correspondence and reprint requests to Gregory E. Skipper, MD, 19 S Jackson St., Montgomery, AL 36117. Address e-mail to gregskipper@usa.net. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc39d
high and 15.8% had been drug dependent.1 Physician health programs (PHPs) are specialized programs granted authority in most states by regulatory boards to manage and monitor physicians after treatment for substance use disorders and other problems.2 Anesthesiologists are consistently overrepresented (approximately 2.5 times the rate of the average physician) in reports from these programs3,4 and similarly over-represented in substance abuse treatment centers that specialize in treatment of physicians.5 Underwriters have identified such high rates of substance abuse among anesthesiologists that some disability insurance companies no longer insure anesthesiologists.6 Most physicians managed and monitored by PHPs have reported 75%90% success rates 5 or more years after treatment for substance use disorders;7,8 however, there is controversy regarding anesthesiologists prognoses, especially concerning risk of returning to anesthesiology practice.9 A recent editorial suggested that substance-abusing anesthesiologists should not be permitted to return to anesthesiology practice after treatment for substance use disorders even with strict monitoring.10 This attitude springs in part from a survey by Menk et al.,9 which reported poor outcomes
891
for substance-abusing anesthesiology residents: only a 34% successful reentry for those using parenteral opioids and 26 deaths (14% of the 180 reported cases), half attributed to drug relapse. A similarly designed but more recent survey of anesthesiology training program directors regarding substance-abusing residents between 1991 and 2001 reported comparably poor findings, noting a lower but still significant death rate (9%).11 In contrast are studies from PHPs reporting good outcomes for anesthesiologists, comparable with other physicians, with low risk of suicide and low risk of patient harm.2,3,12 This study is the first long-term outcome report based on actual data from records of anesthesiologists from a cross-section of 16 state monitoring programs reviewed 5 or more years after treatment for substance use disorders.
The overrepresentation of anesthesiologists in the participant sample (odds ratio [OR] 2.9 [confidence interval {CI}: 2.4 3.6], P 0.001) is consistent with findings from previous studies of physician enrollment in substance abuse treatment programs.5
Lost to Follow-Up
During the study period, 82 of the 862 participants (9.5%) moved out of their state programs jurisdiction. We had no access to any continuing records for those participants and so they were not included in the analyses for this study. Those lost to follow-up included 13 anesthesiologists and 69 nonanesthesiologists.* We therefore performed analyses comparing 83 anesthesiologists with 697 nonanesthesiologists for whom 5 yr of follow-up data were available.
SPSS for Windows version 15 was used for the analyses. Demographic and outcome variables for anesthesiologists and nonanesthesiologists were analyzed using 2 and t-test statistics for comparisons of proportions and means, respectively. Univariate (unadjusted) ORs with 95% CIs were computed to compare the two physician groups on selected binomial characteristics and outcomes. In addition, binary logistic regression analysis was used to produce adjusted multivariate ORs for the two groups on the same variables, controlling for the effects of year of enrollment and program location. Because the CIs for the adjusted and unadjusted ORs were essentially unchanged, it was concluded that there was effective homogeneity by time and location. Therefore, the ORs presented in the Results section are the unadjusted univariate ORs, and the P values are by Fishers exact test. The adjusted multivariate ORs and P values are provided in Table 3.
METHODS
Design
The study used the dataset from a 5-yr, longitudinal, cohort study reported previously, involving 904 physicians with diagnoses of substance abuse or dependence consecutively admitted to 1 of 16 state PHPs between 1995 and 2001.13 The characteristics and outcomes of a subset of 83 anesthesiologists were compared with those of nonanesthesiologists. We restricted the comparisons with objective data from official records (for example, treatment services, attendance, sanctions by the program, reports to licensing boards) and from laboratory records (urine tests and other specimens). To protect the confidentiality of the physicians, members of each programs medical records department collected the data. Data were collected between November 2006 and January 2007 under training, supervision, and monitoring by the authors. All components of this study were reviewed and approved by the IRB of the Treatment Research Institute.
RESULTS
The study was based on treatment records from 16 programs that had previously participated in a survey of 42 PHPs conducted by the authors. That original study described the structure, function, funding, and overall characteristics of the PHPs as well as the intervention, evaluation, referral for treatment, and monitoring activities after the treatment was provided. The 26 PHPs that did not participate in record review were contacted, and all claimed lack of resources and/or regulatory impediments as the reason for declining to participate. The programs that did and did not participate in the follow-up study were not
*Comparisons between those lost to follow-up and those retained in the study revealed no significant differences between groups on gender, age, primary substance of abuse at admission, history of prior treatment, or treatment participation status (mandatory versus voluntary). Among those lost to follow-up, there were no significant differences between anesthesiologists and nonanesthesiologists on these same variables. Over half of the anesthesiologists (54%) and the nonanesthesiologists (68%) who could not be followed for 5 yr had transferred in good standing to PHPs in other states.
Participant Sample
Of the 904 participants in the original study, 42 (4.6%) were residents, all of whom were excluded from this study because they constituted a population of physicians who were both younger than the average practicing physician and therefore at higher risk of substance abuse; although there were no significant differences between residents and practicing physicians (including anesthesiology residents) on any outcome variables measured, their numbers were deemed too small to be conclusive. Residents excluded from the study included 6 in anesthesiology training programs and 36 in other specialties. Of the remaining 862 participants, 96 (11.1%) were anesthesiologists. At the time these participants enrolled in PHPs, anesthesiologists comprised 4.1% of the approximately 749,000 physicians (excluding residents) providing patient care in the United States.14
892
Anesthesiologists with Substance Use Disorders
Table 1. Characteristics of Anesthesiologists and Other Physicians Participating in State Physician Health Programs for Substance Use Disordersa Characteristic
Age at enrollment Mean sd Range Gender Male Female Enrollment status Mandatory Voluntary History of treatment Yes No Type of agreement Dependence (5 yr) Diagnosis/abuse Primary drug of abuse Alcohol Opioids Stimulants Sedatives Other IV drug use history Yes No Number of substances Single Multiple Months in testing period Mean sd Range Number of tests Mean sd Range
Anesthesiologists (n 83)
42 6 2660 71 (86) 12 (14) 49 (59) 34 (41) 25 (30) 58 (70) 76 (92) 7 (8) 23 (28) 46 (55) 7 (8) 2 (2) 5 (6) 32 (41) 46 (59) 43 (52) 40 (48) 49 22 282 101 72 2384
Other physicians (n 697)

45 9 2775 599 (86) 95 (14) 393 (57) 303 (43) 273 (39) 422 (61) 611 (88) 86 (12) 361 (52) 217 (32) 50 (7) 25 (4) 36 (5) 64 (10) 584 (90) 339 (49) 358 (51) 47 26 0155 82 68 1435
P*
0.01 0.87 0.73 0.12 0.37
0.01
0.001 0.63 0.50 0.02
a Values are number (percentage) unless otherwise indicated. * From t-test for independent means or 2 test for comparison of proportions (two-tailed) as appropriate.
statistically or clinically significantly different for evaluation, referral, treatment, supervision, support, and monitoring practices. The 16 participating programs tended to be large: 31% were in the largest quarter of programs. The mean number of physicians in each program was 56 (range, 11119). Although these 16 programs may not be considered nationally representative, they showed no obvious clinical, administrative, or organizational differences from those not participating. The 780 participants in this study (83 anesthesiologists and 697 other physicians) were distributed among the 16 programs so that, on average, there were 5 anesthesiologists (range, 112) and 44 nonanesthesiologists (range, 6 95) per PHP. Anesthesiologists did not constitute more than 17% of the participants in any of the 16 programs. Descriptive characteristics of anesthesiologists and nonanesthesiologists are presented in Table 1. On average, program enrollees were in their forties with males comprising 86% of each group. The majority of physicians in both groups, approximately 58%, were mandated to participate in the program. According to intake records, 30% of anesthesiologists and 39% of
the other physicians had a history of treatment for substance use when they enrolled in the program. In each group, about 90% of enrollees signed a 5-yr dependence agreement, indicating that a diagnosis of substance dependence had been made and the physician agreed to be monitored for at least 5 yr. The others signed a diagnostic monitoring agreement, which is a more limited and shorter-duration agreement used when a diagnosis of substance dependence was not made. The two groups differed regarding the primary substance of abuse as recorded in their intake records: the majority of nonanesthesiologists (52%) were enrolled because of alcohol-related problems, whereas for most anesthesiologists (55%) the primary drug of abuse was an opioid. Thus, anesthesiologists were significantly less likely than their peers to enroll in a PHP because of alcohol abuse (OR 0.4 [CI: 0.2 0.6], P 0.001) and much more likely to enroll because of abuse of opioids (OR 2.8 [CI: 1.7 4.4], P 0.001). Another significant difference between the groups was that 41% of the anesthesiologists had a history of IV drug use compared with 10% of the nonanesthesiologists (OR 6.3 [CI: 3.8 10.7], P 0.001).
893
Table 2. Drug Testing Outcomes and Program and Occupational Status of Anesthesiologists and Other Physicians at 5-yr or More Follow-Up from Signing a Monitoring Contract with a State Physician Health Program for Substance Use Disordersa Outcome
Positive drug test Yes No Reported to board Yes No Program status Completed contract Contract extended Failed to complete Occupational status Licensed or practicing medicine Licensed or working (not clinical) Retired or left practice voluntarily License revoked Died Unknown
a Values are number (percentage). * From 2 test for comparison of proportions (two-tailed).
9 (11) 74 (89) 15 (18) 68 (82) 59 (71) 15 (18) 9 (11) 63 (76) 1 (1) 4 (5) 6 (7) 5 (6) 4 (5)

156 (23) 534 (77) 140 (20) 556 (80) 445 (64) 112 (16) 140 (20) 508 (73) 38 (6) 27 (4) 78 (11) 24 (3) 22 (3)
P*
0.02 0.77 0.09
0.21
Random drug testing was required of the physicians participating in the programs. Data presented in Table 1 show that both the anesthesiologists and the other physicians were subject to testing for an average period of about 48 mo. During this time, the mean number of tests (101) administered to anesthesiologists was higher than the number (82) administered to nonanesthesiologists; however, because the CI for the mean difference between the groups was large (mean difference 19 [CI: 335], P 0.02), it cannot be concluded that anesthesiologists were routinely tested more frequently than other physicians. Table 2 compares anesthesiologists and nonanesthesiologists on primary outcome measures examined in this study: positive drug tests during monitoring, physicians reported to the licensing board, program status at 5-yr follow-up, occupational status at followup, and deaths. The PHP records, which chronicled each instance in which a program participant tested positive for drugs, revealed that 11% of anesthesiologists had at least one positive test compared with 23% of nonanesthesiologists. Although this difference was statistically significant, examination of the OR indicated a wide CI with the upper bound approaching 1 (OR 0.4 [CI: 0.2 0.9], P 0.02). Therefore, we cannot report with confidence that anesthesiologists were less likely than other physicians to test positive for drugs. Approximately 20% of the participants in both groups were reported to their state licensing agencies because of noncompliance with the terms of the PHP agreement or relapse. At the end of the 5-yr follow-up period, 71% of anesthesiologists and 64% of nonanesthesiologists had completed their contracts and were no longer required to be monitored (OR 1.4 [CI: 0.9 2.3], P 0.23). Another 18% of anesthesiologists and 16% of nonanesthesiologists had their contracts extended beyond the initial monitoring period (OR 1.2 [CI: 0.6 2.1], P 0.64). The
894
reasons for continued monitoring included relapse, failure to comply with requirements, such as group attendance or therapy, or, in some cases, voluntary continuance to help prevent relapse and/or demonstrate continued recovery to others. Although a larger proportion of nonanesthesiologists (20%) failed to complete the program than anesthesiologists (9%), the odds of failing to complete were not significantly smaller for anesthesiologists (OR 0.5 [CI: 0.21.0], P 0.05). These results indicate that anesthesiologists were no more likely than other physicians to complete the program, to fail to complete, or to extend the monitoring period beyond the original 5 yr specified in their agreements. The final outcome examined was participants occupational status at follow-up. As shown in Table 2, there were no overall differences between the two groups in the distribution of participants among the various occupational status categories used in the study. A primary category of interest was the extent to which physicians who had participated in the programs were licensed and practicing medicine at the 5-yr follow-up. The study found that the proportion of anesthesiologists (76%) continuing their medical practice was not significantly different than that for nonanesthesiologists (73%) (OR 1.2 [CI: 0.72.0], P 0.60). Additionally, there were no statistically significant differences between anesthesiologists and nonanesthesiologists in regard to the percentage who had their licenses revoked or the percentage reported to have died (Table 3). The record review sought evidence of any patient harm associated with relapse. None was detected in this cohort of anesthesiologists.
DISCUSSION
As in other studies, anesthesiologists were significantly overrepresented, further documenting a reported higher rate of substance abuse of 22.7 times
Table 3. Selected Characteristics and Outcomes of Anesthesiologists and Other Physicians in State Physician Health Programs for Substance Use Disorders, with Adjusted and Unadjusted Odds Ratiosa Unadjusted univariate odds ratios Characteristic/outcome
Gender Male Primary drug of abuse Alcohol Opioids IV drug use history Yes Number of substances Multiple Prior treatment Yes Enrollment status Mandatory Positive drug test Yes Reported to board Yes Program status Completed contract Contract extended Failed to complete Occupational status Licensed or practicing medicine License revoked Died
Adjusted multivariate odds ratios* OR (95% CI)

1.1 (0.52.0) 0.4 (0.20.6) 2.9 (1.84.6) 5.7 (3.49.8) 0.9 (0.61.4) 0.7 (0.41.1) 1.2 (0.82.0) 0.4 (0.20.8) 0.8 (0.51.5) 1.5 (0.92.5) 1.1 (0.62.1) 0.5 (0.20.9) 1.2 (0.72.1) 0.6 (0.21.4) 1.7 (0.64.7)
71 (86) 23 (28) 46 (55) 32 (41) 40 (48) 25 (30) 49 (59) 9 (11) 15 (18) 59 (71) 15 (18) 9 (11) 63 (76) 6 (7) 5 (6)

599 (86) 361 (52) 217 (32) 64 (10) 358 (51) 273 (39) 393 (57) 156 (23) 140 (20) 445 (64) 112 (16) 140 (20) 508 (73) 78 (11) 24 (3)
OR (95% CI)
1.1 (0.62.0) 0.4 (0.20.6) 2.8 (1.74.4) 6.3 (3.810.7) 0.9 (0.61.4) 0.7 (0.41.1) 1.1 (0.71.8) 0.4 (0.20.9) 0.9 (0.51.6) 1.4 (0.92.3) 1.2 (0.62.1) 0.5 (0.21.0) 1.2 (0.72.0) 0.6 (0.31.5) 1.8 (0.74.8)
P
0.87 0.001 0.001 0.001 0.63 0.12 0.73 0.02 0.77 0.23 0.64 0.05 0.60 0.35 0.22
P
0.88 0.001 0.001 0.001 0.61 0.15 0.76 0.01 0.72 0.17 0.69 0.05 0.57 0.28 0.30
OR odds ratio; CI condence interval. a Values are number (percentage); all odds ratios are anesthesiologists/other physicians. * Adjusted for year of enrollment and program location; Wald test P values for year and location were not signicant at P 0.05.
compared with other physicians. This finding has been consistent throughout all reports. Exploring the contradictory reports regarding prognosis for substance-abusing anesthesiologists, earlier studies were essentially of two designs: survey studies of training program directors versus longitudinal studies of anesthesiologists in monitoring by individual state PHPs. The survey studies,9,11 which reported much poorer outcomes, surveyed training program directors regarding the number of residents encountered with substance disorders over the preceding 10 yr and their outcomes. Considering the tendency for confidential handling of this type of information, the data from these reports were likely skewed toward poor outcomes, such as relapse or death, which would less likely remain confidential and more likely to be remembered. It was also noted in one of the survey studies that many of the training program directors tenure was less than the 10-yr study period for which they were being asked to recall cases. The reported numbers themselves from these survey studies suggest that reports were not complete. For example, 31 of 230 physicians in the Collins et al. survey were currently in active treatment leading to their finding of a 0.89% point prevalence of active addiction. Because drug treatment seldom lasts more
than 4 mo, it is reasonable to expect, from the cohort of more than 100 training programs, that approximately 93 residents per year, or 930 per 10 yr, would be treated. Because only 230 cases were identified, this represents only approximately 25% of the total expected cases. In the Menk et al. survey of the 159 anesthesia training programs in the United States, 113 responded reporting 180 cases, an average of 1.6 case reports per responding program for a 10-yr period, clearly a smaller number than would be expected, again suggesting under-reporting. Our data demonstrating similar or better outcomes (survival, total abstinence, completion of monitoring, return to work in profession, and retention of medical license) for anesthesiologists compared with other specialties are consistent with other outcome studies from single PHP.3,4,12 Anesthesiologists received similar treatment as other physicians under PHP care; however, their monitoring often had the following added features. 1) Witnessed naltrexone administration. 2) Regular periodic hair testing, which is more effective than urine testing because fentanyl and similar compounds are extremely short lived and are very difficult to detect (a fact well known to anesthesiologists). It is also more difficult to cheat on a hair test, adding to its value testing in a high-risk population.
895
Periodic hair testing may more effectively discourage drug use because those being monitored with hair tests know that any drug use will more likely be detected. 3) Enhanced security measures in and around the operating room to prevent diversion. (Being careful with drug access and disposal by using witnesses, automated distribution devices, and monitoring cameras, and spectrometric scanning of discard wastage.) The earlier more pessimistic studies regarding the ability of anesthesiologists to remain drug-free did not note which, if any, subjects were in active PHP monitoring. Our study, in contrast, was limited to the experiences of physicians in active PHP care management, which included active monitoring using more sophisticated means of detecting any return to alcohol or other drug use and high quality addiction treatment, factors that possibly account for better outcomes. Additionally, the articles of Menk and Collins involved anesthesiology residents exclusively, whereas our study excluded residents. We found no evidence indicating patient harm had occurred associated with any relapse. Although the value of this finding may be limited because data were restricted to a review of anesthesiologists records, it is consistent with the Domino et al. study, which found no evidence of patient harm among 33 anesthesiologists over 10 yr in Washington state.12 Sivarajan et al.15 examined data from the American Society of Anesthesiology malpractice database seeking evidence of patient harm from substance abuse. Of the 2715 closed anesthesia claims, in only 7 was substance abuse or chemical dependence noted by the anesthesiology reviewer in the claim summary. Two of the seven cases involved nurse anesthetists who were abusing substances under the supervision of anesthesiologists. Three of the five claims in which a substance-abusing anesthesiologist delivered anesthesia care involved serious patient harm (brain damage or death) because of lack of vigilance or judgment during anesthesia. Two of these three claims involved anesthesiologists who were alcoholics and the third involved an anesthesiologist who left the care of the patient to smoke a cigarette. The two alcoholic anesthesiologists had been unavailable to provide care: one because of alcohol intoxication and the other who left to attend rehabilitation without providing backup care for a chronic pain patient. In summary, of 2715 malpractice claims against anesthesiologists 5 involved substance-abusing anesthesiologists, 4 of whom were alcoholics and the other a smoker. No closed claims involving drugaddicted anesthesiologists were noted. This indicates a remarkably low rate of patient harm from substance-abusing anesthesiologists. The special
stigma directed toward opiate-addicted anesthesiologists, especially those using IV opiates, does not appear to be warranted.
CONCLUSION
This study supports the finding that anesthesiologists have a significantly higher rate of substance abuse by a factor of 2.71 when compared with other physicians. Programs to prevent and/or detect substance use in this relatively high-risk group would therefore seem especially justified but have been almost nonexistent.16 Although any incidence of overdose death or suicide is unacceptable, the rates of these phenomena were small and not higher among anesthesiologists compared with other physicians. There is now considerable evidence, corroborated by this study, that anesthesiologists managed by PHPs have good prognoses. REFERENCES
1. Lutsky I, Hopwood M, Abram SE, Jacobson GR, Haddox JD, Kampine JP. Psychoactive substance use among American anesthesiologists: a 30-year retrospective study. Can J Anaesth 1993;40:91521 2. Available at: http://www.fsmb.org/pdf/1995_grpol_Physician_ Impairment.pdf. Accessed June 1, 2009 3. Pelton C, Ikeda RM. The California Physicians Diversion Programs experience with recovering anesthesiologists. J Psychoactive Drugs 1991;23:42731 4. Paris RT, Canavan DI. Physician substance abuse impairment: anesthesiologists vs. other specialties. J Addict Dis 1999;18:17 5. Talbott GD, Gallegos KV, Wilson PO, Porter TL. The Medical Association of Georgias Impaired Physicians Program. Review of the first 1000 physicians: analysis of specialty. JAMA 1987;257:292730 6. Guadagnino C. MDs challenged on disability insurance. Physician News Digest. Jan 2002. Available at: http://physiciansnews. com/cover/102.html. Accessed June 1, 2009 7. Pelton C. Physician diversion program: Californias experience with successful graduates. J Psychoactive Drugs 1993;25:159 64 8. Shore JH. The Oregon experience with impaired physicians on probation. JAMA 1987;257:2931 4 9. Menk EJ, Baumgarten RK, Kingsley CP, Culling RD, Middaugh R. Success of reentry into anesthesiology training programs by residents with a history of substance abuse. JAMA 1990;263:3060 2 10. Berge KH, Seppala MD, Lanier WL. The Anesthesiology Communitys approach to opioid- and anesthetic-abusing personnel: time to change course. Anesthesiology 2008;109:762 4 11. Collins GB, McAllister MS, Jensen M, Gooden TA. Chemical dependency treatment outcomes of residents in anesthesiology: results of a survey. Anesth Analg 2005;101:1457 62 12. Domino KB, Hornbein TF, Polissar NL, Renner C, Johnson J, Alberti S, Hankes L. Risk factors for relapse in health care professionals with substance use disorders. JAMA 2005;293:1453 60 13. McLellan T, Skipper GE, Campbell M, DuPont R. Five year outcomes in a cohort study of physicians treated for substance use disorders in the United States. BMJ 2008;337:a2038 14. American Medical Association. Physician Characteristics and Distribution in the US, 20022003 Edition. Chicago, IL: AMA Press, 2002:15 15. Sivarajan M, Posner KI, Caplan RA, Gild WM, Cheney FW. Substance abuse among anesthesiologists. Anesthesiology 1994; 80:704 16. Fitzsimmons MG, Baker KH, Lowenstein E, Zapol WM. Random drug testing to reduce the incidence of addiction in anesthesia residents: preliminary results from one program. Anesth Analg 2008;107:630 5
896
Brief Report
Seventh and Eighth Year Follow-Up on Workforce and Finances of the United States Anesthesiology Training Programs: 2007 and 2008
Sachin Kheterpal, MD, MBA Kevin Tremper, PhD, MD Amy Shanks, MS Michelle Morris, MS
We sent follow-up financial and workforce surveys to 121 United States anesthesiology training programs in 2007 and 2008. Seventy-four respondents (61%) demonstrated a continued increase in the institutional support for faculty and stabilization in the number of open positions. Institutional support per faculty full time equivalent with certified nurse anesthetist support removed averages $109,000. A 7% open faculty position rate is characterized by a preponderance of generalists (31%) and pediatric (21%) anesthesiologists.
(Anesth Analg 2009;109:8979)
ecause of a decrease in interest of medical students in anesthesiology in the mid 1990s, there were fewer resident graduates starting in the year 2000.13 When the resulting decrease in the number of anesthesiologists available for the workforce was combined with increasing faculty salaries, the financial status of academic anesthesia programs worsened.4 6 For the past 8 yr, surveys have been sent to the United States anesthesiology training programs. These surveys have demonstrated a progressive increase in institutional financial support for anesthesiology training departments: from a mean of $34,000/faculty full time equivalent (FTE)/yr in the year 2000 to a mean of $120,000/FTE/yr in the year 2006.4 At the same time, the percent of open faculty positions in academic departments has been progressively decreasing from a high of approximately 10% in the year 2000 to a low of 5% in 2006.4 We sought to update survey data for the academic years 2007 and 2008.
percentile values were calculated as descriptive data using SPSS Version 15 (SPSS, Chicago, IL) and MedCalc Version 10 (Mariakerke, Belgium). The 95% confidence intervals for the 25th percentile and 75th percentile value (without transformation) were compared with the median value for each data point to identify any overlaps. Confidence intervals were derived using the nonparametric order statistics technique first described by Wilks.7
RESULTS
The surveys were distributed to 121 chairs who are members of the Society of Academic Anesthesiology Associations. The response rates for 2007 and 2008 were 60% and 61%, respectively. The descriptive statistics are presented in Tables 17. The 95% confidence intervals (data not shown) for the 25th and 75th percentiles did not overlap with the means or medians reported in these tables. The average academic anesthesiology department has 53 faculty, 8 fellows, 44 residents, 7 interns (66% of the departments offer internships), and 32 certified registered nurse anesthetists (CRNA) (95% of the departments have CRNAs). In the departments which had open faculty positions in 2008 (56 of 74 responders 76%), there was an average of five open positions (95% confidence interval of 46). This translates to an overall 7% national open position rate in the sampled departments, defined as: (average number of reported open positions at institutions reporting open positions/average number of positions at all institutions) (percentage of institutions reporting open positions). The most sought-after specialty was a generalist followed by pediatric, cardiac, and critical care, respectively (Table 3). In 2008, there were four open CRNA positions per department, translating to a national average open percentage of 13%
897
METHODS
This survey has been presented in detail in previous publications.4 For the year 2008, the survey was modified to also request which type of faculty anesthesiologist positions were unfilled: generalist, pediatric, cardiac, critical care, pain, regional, ambulatory, obstetric, or neuro. The electronic mail surveys were sent starting in the Fall of each year, and reminder surveys were sent to nonresponders every 2 wk for the next 8 wk. Mean, median, 25th percentile, and 75th
From the Department of Anesthesiology, University of Michigan Health System, Ann Arbor, Michigan. Accepted for publication May 1, 2009. Address correspondence and reprint requests to Sachin Kheterpal, MD, MBA, Department of Anesthesiology, University of Michigan Health System, 1500 E. Medical Center Dr., Ann Arbor, MI 48109. Address e-mail to sachinkh@med.umich.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b0fef6
Table 1. 2008 Faculty, Fellows, Residents, and Certied Registered Nurse Anesthetists (CRNAs) Stafng Mean
Faculty Fellows Residents (CA-1,2,3) Internsa CRNAsa 53 8 44 11 34
Table 4. Revenue: Sources and Total

2007 2008 n 72 $13,000,000 $23,000,000 $20,000,000 $31,900,000 n 72 $370,000 $430,000 $427,000 $490,000 n 72 $71,500 $2,000,000 $520,000 $2,650,000 n 72 $1700 $29,000 $11,000 $43,600 n 72 $3,000,000 $6,400,000 $5,800,000 $8,800,000 n 72 $56,000 $136,000 $109,000 $202,500 n 72 $17,800,000 $32,000,000 $26,000,000 $40,000,000 n 72 $491,000 $595,000 $570,000 $660,000 n 72 $17,900,000 $32,300,000 $26,100,000 $41,000,000 n 72 $516,000 $605,000 $588,000 $682,000
Median
45 5 39 10 21
25%
30 2 26 7 12
75%
68 11 59 12 39
a 66% of programs have interns and 95% of programs have CRNAs. These data reect programs that have interns and CRNAs, respectively.
Table 2. Open Faculty Position Data and Open Certied Registered Nurse Anesthetist (CRNA) Positions 2007
Number of faculty Open faculty positions (N 56)a Departments with open positions (%) Open faculty positionsb (%) Number of CRNA Open CRNA positions Departments with open positions (%) Open CRNA positionsb (%) 45 4 76 7 23 4 75 14
2008
53 5 76 7 32 4 68 13
Data are presented as means. a For those that have open positions. b These percentages are overall and include the departments with no open positions.
Table 3. Open Faculty Position by Subspecialty (2008) Type

Generalist Pediatric Cardiac Critical care Regional Pain Ambulatory Obstetric Neuro
Because of rounding, the percentages do not add up to 100%.
Open positions (%)

31 21 12 11 6 6 4 4 4
(Table 2). For the fiscal year ending June 2008, the departments averaged total revenue of $595,000/FTE (Table 4). This total revenue was composed of clinical revenue ($430,000/FTE), research revenue ($29,000/FTE), and institutional support ($136,000/FTE) (Table 4). When the portion of institutional support used for CRNA salaries is removed, the overall support is $109,000/FTE. This is a progressive increase for 2007 and 2008 (Table 5). The departments billed an average of 11,400 units/faculty FTE in 2008, and the average unit charge increased to $96.00 in 2008. The average dollars collected per unit billed was $35 (Table 6).
Clinical revenue 25th percentile Mean Median 75th percentile Clinical revenue/FTE 25th percentile Mean Median 75th percentile Research revenue 25th percentile Mean Median 75th percentile Research revenue/FTE 25th percentile Mean Median 75th percentile Institutional support 25th percentile Mean Median 75th percentile Institutional support/FTE 25th percentile Mean Median 75th percentile Total revenue 25th percentile Mean Median 75th percentile Total revenue/FTE 25th percentile Mean Median 75th percentile Total expenses 25th percentile Mean Median 75th percentile Total expenses/FTE 25th percentile Mean Median 75th percentile
n 71 $11,700,000 $19,000,000 $18,000,000 $24,300,000 n 71 $317,000 $413,000 $412,000 $464,000 n 71 $34,000 $1,800,000 $237,000 $2,100,000 n 71 $1000 $28,000 $9500 $34,000 n 70 $2,300,000 $5,200,000 $4,900,000 $7,500,000 n 70 $56,000 $126,000 $103,000 $183,000 n 70 $15,500,000 $26,000,000 $21,000,000 $34,000,000 n 70 $464,000 $566,000 $524,000 $627,000 n 69 $15,200,000 $25,800,000 $22,000,000 $33,700,000 n 69 $460,000 $555,000 $520,000 $620,000
Data are in actual dollars. Medical ination rates for years 2000 2006 were 4.1%, 4.6%, 4.7%, 4.0%, 4.4%, 4.2%, 4.0%, and 4.4% based upon the Medical Care subset of the Consumer Price Index-Urban published by the United States Bureau of Labor and Statistics, data series CUUR0000SAM. $10,000 in 2000 medical care dollars would be $13,414 in 2007 and $14,000 in 2008 medical care dollars based upon these ination rates. FTE faculty full time equivalent.
DISCUSSION
The 2007 and 2008 surveys seem to follow the same trend as the previous surveys. The observed open faculty position percentage compares well with the data received from the yearly Society of Academic Anesthesiology Chairman Salary Survey distributed by the University of Florida, which demonstrated 6%
898
Brief Report
open positions in 2008 (personal communication with Rebecca Lovely, University of Florida, Gainesville, Florida). The number of faculty open positions, although decreasing from the year 2000, has stabilized at approximately 7% (Table 2). Faculty support per FTE (with CRNA support removed) is still increasing, reaching a new high of $109,000/FTE in 2008. A concurrent decrease in reimbursement rate and anesthesia units billed/FTE is observed and confirms recent Medical Group Management Association data (Table 6).8 With the recent increase in Medicare funding for anesthesiologists and the recent revision of the Medicare Teaching
Table 5. Itemized Institutional Support Year

Total support 25th percentile Mean Median 75th percentile Support/FTE 25th percentile Mean Median 75th percentile Total support less CRNA $ 25th percentile Mean Median 75th percentile Total support less CRNA $/FTE 25th percentile Mean Median 75th percentile Hospital support 25th percentile Mean Median 75th percentile Medical school support 25th percentile Mean Median 75th percentile Other support 25th percentile Mean Median 75th percentile Total support
Table 6. Anesthesia Units, Charges, and Collections 2008 (n 74)

n 72 $3,000,000 $6,400,000 $5,700,000 $8,800,000 n 72 $56,000 $136,000 $109,000 $202,000 n 70 $2,700,000 $4,900,000 $4,000,000 $7,100,000 n 70 $48,500 $108,600 $101,000 $150,000 n 72 $2,000,000 $5,000,000 $4,500,000 $7,100,000 n 72 $0 $720,000 $470,000 $1,100,000 n 72 $0 $500,000 $0 $600,000 $6,400,000 (n 72) Anesthesia units/FTE 25th percentile Mean Median 75th percentile Anesthesia value charge 25th percentile Mean Median 75th percentile Dollars collected per anesthesia unit 25th percentile Mean Median 75th Percentile Medicaid payment per unit 25th percentile Mean Median 75th percentile
2007 (n 72)
n 70 $2,300,000 $5,200,000 $4,800,000 $7,500,000 n 70 $56,000 $126,000 $103,000 $183,000 n 70 $2,000,000 $4,200,000 $3,700,000 $6,100,000 n 70 $49,000 $100,000 $88,000 $127,000 n 70 $1,800,000 $4,000,000 $3,500,000 $5,700,000 n 70 $88,000 $800,000 $600,000 $1,400,000 n 70 $0 $400,000 $0 $330,000 $5,200,000 (n 70)
2007 (n 71)
n 66 9900 12,100 11,300 14,000 $75 $89 $85 $100 $27 $32 $32 $37 $12 $16 $15 $18
2008 (n 70)
n 63 9000 11,400 10,800 13,300 $80 $96 $92 $106 $27 $35 $34 $40 $14 $16 $16 $18
Table 7. Clinical Anesthetizing Locations 2007

Operating rooms Nonoperating room sites Faculty assigned/day Labor and delivery Intensive care unit Acute pain service Pain clinic Preoperative clinic Total Faculty/clinical site Research revenue/faculty 34 6 1.3 1.4 1.1 1.7 0.9 46.4 1.0 $28,000
2008
39 7 1.4 1.7 1.1 1.9 0.9 53 1.0 $29,000
Data are presented as means. Totalling the number of operating room sites and the number of faculty assigned to nonoperating room sites is presented only as a crude way of comparing faculty full time equivalent numbers among departments and clinical locations covered. It does not account for how operating rooms are covered, or any faculty requirements associated with call, or other clinical, academic, or administrative commitments.
Data are in actual dollars. FTE faculty full time equivalent; CRNA certied registered nurse anesthetist.
REFERENCES
1. Tremper KK, Barker SJ, Gelman S, Reves JG, Saubermann AJ, Shanks AM, Greenfield MLVH, Anderson ST. A demographic, service, and financial survey of anesthesia training programs in the United States. Anesth Analg 2003;96:1432 46 2. Tremper KK, Reves JG, Barker SJ, Saubermann AJ, Gelman S. Financial environment of acadamic anesthesia. In: Lake, CL Johnson JO, eds. Advances in anesthesia. Carlsbad, CA: Mosby, Inc., 2001:135 3. Tremper KK, Shanks A, Sliwinski M, Barker SJ, Hines R, Tait AR. Faculty and finances of United States anesthesiology training programs: 20022003. Anesth Analg 2004;99:118592 4. Kheterpal S, Tremper KK, Shanks A, Morris M. Six-year follow-up on work force and finances of the United States anesthesiology training programs: 2000 to 2006. Anesth Analg 2009;108:26372 5. Tremper KK, Shanks A, Morris M. Trends in the financial status of United States anesthesiology training programs: 2000 to 2004. Anesth Analg 2006;102:51723 6. Tremper KK, Shanks A, Morris M. Five-year follow-up on the work force and finances of United States anesthesiology training programs: 2000 to 2005. Anesth Analg 2007;104:863 8 7. Wilks S. Order statistics. Bull Am Math Soc 1948;54:6 50 8. Abouleish A. The fallacy of the field of dreams business plan: a downward trend in anesthesiology productivity. ASA Newsl 2007; 71:301 9. Medicare Teaching Rule. Text of H.R. 6331: Medicare improvements for patients and providers act of 2008. Available at: http://www.govtrack.us/congress/billtext.xpd?billh110 6331. Accessed 28 Dec 2008
Rule, the charge-reimbursement gap may shrink in upcoming years as those new laws go in effect.9 These two revisions of Medicare payment may have a positive effect on academic training departments. However, Medicare is not the majority payor at most facilities.2 In addition, this increase in departmental support appears to parallel the increasing salaries provided to academic anesthesiology faculty. The most recent salary data demonstrate that the 50th percentile for an assistant professor has increased to $288,000/yr (2008 Society of Academic Anesthesiology Chairman Salary Survey, personal communication with Rebecca Lovely, University of Florida, Gainesville, Florida). Our data suffer from the potential errors associated with the survey methodology itself and have been described earlier.1,3 6 A skewed response population or errors in the respondents understanding the survey questions are the two most common errors that may affect accuracy.
899
Numbers of Simultaneous Turnovers Calculated from Anesthesia or Operating Room Information Management System Data
Franklin Dexter, MD, PhD* Eric Marcon, PhD John Aker, MS, CRNA Richard H. Epstein, MD
BACKGROUND: More personnel are needed to turn over operating rooms (ORs) promptly when there are more simultaneous turnovers. Anesthesia and/or OR information management system data can be analyzed statistically to quantify simultaneous turnovers to evaluate whether to add an additional turnover team. METHODS: Data collected for each case at a six OR facility were room, date of surgery, time of patient entry into the OR, and time of patient exit from the OR. The number of simultaneous turnovers was calculated for each 1 min of 122 4-wk periods. Our end point was the reduction in the daily minutes of simultaneous turnovers exceeding the number of teams caused by the addition of a team. RESULTS: Increasing from two turnover teams to three teams reduced the mean daily minutes of simultaneous turnovers exceeding the numbers of teams by 19 min. The ratio of 19 min to 8 h valued the time of extra personnel as 4.0% of the time of OR staff, surgeons, and anesthesia providers. Validity was suggested by other methods of analyses also suggesting staffing for three simultaneous turnovers. Discrete-event simulation showed that the reduction in daily minutes of turnover times from the addition of a team would likely match or exceed the reduction in the daily minutes of simultaneous turnovers exceeding the numbers of teams. Confidence intervals for daily minutes of turnover times achieved by increasing from two to three teams were calculated using successive 4-wk periods. The distribution was sufficiently close to normal that accurate confidence intervals could be calculated using Students t distribution (Lilliefors test P 0.58). Analysis generally should use 13 4-wk periods as increasing the number of periods from 6 to 13 significantly reduced the coefficient of variation of the averages but not increasing the number of periods from 6 to 9 or from 9 to 13. CONCLUSION: The number of simultaneous turnovers can be calculated for each 1 min over 1 yr. The reduction in the daily minutes of simultaneous turnovers exceeding the number of teams achieved by the addition of a turnover team can be averaged over the years 13 4-wk periods to provide insight as to the value (or not) of adding an additional team.
(Anesth Analg 2009;109:900 5)
either anesthesiologists nor surgeons desire prolonged operating room (OR) turnover times.1,2 One effective way to reduce turnover times is to use additional personnel (e.g., extra anesthesia technicians and/or postanesthesia care unit nurses).35 In the United States, such personnel must usually be paid for by the facility due to regulatory issues.6 Anesthesiologists have an incentive to persuade administrators to increase the OR nursing budget to include more turnover personnel,
From the *Division of Management Consulting, Departments of Anesthesia and Health Management and Policy, University of Iowa, Iowa; Department of Manufacturing System Management and Maintenance, Jean Monnet University of Saint Etienne, France; Department of Anesthesia, University of Iowa, Iowa; and Department of Anesthesiology, Jefferson Medical College, Philadelphia, Pennsylvania. Accepted for publication May 19, 2009. Franklin Dexter is Section Editor of Economics, Education, and Policy for the Journal. This manuscript was handled by Steve Shafer, Editor-in-Chief, and Dr. Dexter was not involved in any way with the editorial process or decision.
as reducing excessive idle time can increase anesthesia group productivity.7 Because quantitative analysis enhances the persuasiveness of arguments,8 we studied how OR and/or anesthesia information management system data can be analyzed statistically to quantify simultaneous turnovers to evaluate whether to add an additional turnover team (e.g., housekeeper and anesthesia technician). We studied simultaneous turnovers being as the number of personnel needed to turn over
FD receives no funds personally other than his salary from the State of Iowa, including no travel expenses or honoraria; has tenure with no incentive program; and owns no health care stocks other than indirectly through mutual funds. RHE is President of Medical Data Applications, Ltd., whose CalculatOR software includes the analyses considered in this article. Address correspondence and reprint requests to Franklin Dexter, MD, PhD, Division of Management Consulting, Department of Anesthesia, University of Iowa, Iowa City, IA 52242. Address e-mail to Franklin-Dexter@UIowa.edu or web site www.FranklinDexter.net. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b08855
900
Figure 1. Gantt charts showing that the effect of adding a turnover team on daily minutes of surgeon experienced turnover times can be predicted by minutes of simultaneous turnovers exceeding the number of teams. In the schematic, time is plotted along the horizontal axis. Each row represents an operating room (OR). The panes show turnover times for three turnover teams, two teams, and one team. The long light gray bars represent times that patients are in ORs. The dark gray bars are the cleanup and setup times. The white bars represent delays contributing to turnover times (i.e., nothing is happening in the OR because the turnover team is elsewhere). With three teams, the minutes of simultaneous turnovers exceeding the threshold of three teams 0 min. With two teams, the minutes of simultaneous turnovers exceeding the threshold of two teams 10 min. Thus, an increase from two teams to three teams results in a 10 min reduction in the daily minutes for which the number of simultaneous turnovers exceeds the numbers of teams, where 10 min 10 0 min. In comparison, an increase from two turnover teams to three teams reduces the total surgeon experienced turnover time by 10 min. The two end points match. With one team, the minutes of simultaneous turnovers exceeding the threshold of one team 40 min. An increase from one team to two teams results in a 30 min reduction in the daily minutes of simultaneous turnovers exceeding the numbers of teams, where 30 min 40 10 min. In comparison, an increase from one turnover team to two teams reduces the total minutes of surgeon experienced turnover time by 30 min. For this example, the two end points match.
ORs promptly is positively correlated with the number of simultaneous turnovers (Figs. 1 and 2). We developed the statistical method using 9 yr of data from an outpatient facility with six staffed ORs, validated the method, and then applied the method to a large tertiary suite with several months of data.
METHODS
There were 53,716 cases performed at the outpatient facility during its 2345 workdays from December 20, 1998, to April 26, 2008. All procedures were elective. No cases were performed on holidays or weekends. Data used for each case were, room, date of surgery, time of patient entry into the OR, and time of patient exit from the OR, entered into an Excel spreadsheet (Microsoft, Redmond, WA). Using Visual Basic
for Applications, turnover times were calculated for each case and set equal to 90 min when longer than 90 min.9,10 The resulting 37,282 turnovers totaled 799,787 min. Turnovers were considered to be simultaneous if they overlapped in time by at least 1 min. The number of simultaneous turnovers was calculated for each 1 min over the 122 studied 4-wk periods. In routine practice, we use 13 4-wk periods (see Results). An array stored in 1 min increments was created that stored the number of simultaneous turnovers, as described previously for studies of staffing in postanesthesia care units.11 The daily minutes with the number of simultaneous turnovers exceeding a threshold number of teams was calculated. By this we mean, precisely, the daily minutes of simultaneous turnovers during which there were more ORs to be turned over than there were turnover teams available. For example, if the threshold were three teams, and if for 1 min there were five simultaneous turnovers, then that minute contributed 2 min to the measured value, because there were two ORs without assistance for that 1 min interval. Depending on a facilitys organization, a turnover team may not be needed for the entire turnover. Results are proportional if a turnover team is needed for only a fraction of the turnover time. Two-sided 95% confidence intervals (CIs) were calculated for the reduction in the daily minutes for which the number of simultaneous turnovers exceeded a specified number of teams, with reductions calculated for additions of one turnover team. This approach of studying incremental differences was described previously for studying anesthesia staffing in afternoons.12,13 The averages of 4-wk periods of daily differences were treated as following a normal distribution such that CI were calculated using Students t-distribution. Lilliefors test was used to confirm this assumption.14 As a potential alternative end point, we calculated the percentage of turnover time attributable to turnovers occurring when the number of simultaneous turnovers exceeded the number of teams. To calculate these percentages, we used larger bin sizes of 8-wk periods to have no bins with zero in the numerator. The sum over each 8-wk period of the minutes of simultaneous turnovers exceeding the number of turnover times was divided by the total minutes of turnover time during the period. Two-sided 95% CI were calculated by taking the Freeman-Tukey transformation of the counts from each 8-wk period, applying the Students t-distribution to the transformed values, and then taking the inverse.15,16 The percentages were pooled because personnel responsible for turnovers generally work 8 h shifts, 5 days a week, corresponding to when most turnovers occur.10 The percentages were pooled also because successive turnovers times were correlated.10 As another potential alternative end point, the daily peak number of simultaneous turnovers was calculated by using bin sizes of 1 day. The Clopper-Pearson
901
Figure 2. Simulated reductions in daily minutes of surgeon experienced turnover times from the addition of a turnover team
match or exceed the reduction in daily minutes of simultaneous turnovers exceeding threshold of the number of teams. The methodology is described in the last two paragraphs of the Methods. The figure is presented in the same format as Figure 3 for comparison, with dark boxes in both representing our recommended study end point. The comparisons between adjacent bars are analogous to the comparisons made in Figure 1. The light boxes in this figure show the extra time experienced by the physicians, called makespan in the operations research/industrial engineering fields. Unlike Figures 35, there is no box around three turnover teams, because with simulation, any number of turnover teams can be studied. The figure shows that the reduction in total turnover time of the suite (e.g., as experienced by surgeons) from the addition of one team will be at least the reduction in minutes of simultaneous turnovers achieved by an increase in the threshold number of teams by one team. As explained in the second to last paragraph of the Results, the reduction in total turnover time will be no more than that achieved by increasing the number of teams to one team per operating room. Consequently, as the numbers of teams is increased, the absolute accuracy of the method of Figure 3 is improved.
method was used to calculate one-sided 95% CI for the cumulative distribution of the percentage of days with the peak number of simultaneous turnovers exceeding a threshold number of simultaneous turnovers.17,18 We will show, below, that our recommended method of analysis is two-sided 95% CI for the reduction in the daily minutes for which the number of simultaneous turnovers exceeds the number of turnover teams, as achieved by the addition of one extra team. We considered the appropriate number of 4-wk periods for use in routine monitoring. Deciding to increase staffing, hiring the additional personnel, and training can typically take around half a year. Budgeting usually is reevaluated annually. Thus, 6 and 13 4-wk periods (i.e., 24 and 52 wk, respectively) serve as the minimum and maximum intervals. We calculated the coefficient of variation (CV) of the moving average over the n 122 4-wk periods using 6 periods (n 117), using 9 periods (n 114), and using 13 periods (n 110). Comparison of coefficients of variation between different numbers of 4-wk periods was performed asymptotically using two-group analyses.19 Finally, discrete event simulation was used to study the influence of the numbers of turnover teams on both our recommended end point and the reduction in the total daily minutes of turnovers at the suite, matching the comparison showed in Figure 1. The 53,716 cases studied were performed in 14,070 combinations of six ORs and 2345 workdays. The mean standard deviations of OR times 1.41 0.95 h and of turnover times 0.32 0.28 h. ARENA (v8.0, Rockwell Software, Sewickley, PA) was used to simulate 14,070 identical OR workdays. For each OR on
902
Staffing for Simultaneous Turnovers
each workday, an OR time was simulated using a log-normal distribution with the observed mean and standard deviation.20 If 9 h, a turnover time and another OR time were simulated. If the sum of the three was 9 h, then all three are performed. Another turnover time and OR time were simulated. If adding those two to the preceding three exceeded 9 h, then only the first three were used. If the sum of the five events did not exceed 9 h, then all five events were included. The process was continued until 9 h was exceeded. The resulting duration of the simulated workday was 7.71 0.53 h. The simulation model was used for testing the five scenarios in which the maximum number of turnover teams working in the surgical suite was specified. For the first scenario, six turnover teams were assumed always to be available, equal to the number of ORs. Thus, no OR ever waited for cleanup or setup, and these times were considered to be the turnover time. Simulations were repeated with five teams reduced stepwise to one team. With fewer teams, ORs sometimes waited for cleaning to start, and the turnover time was increased, as shown in Figure 1. For example, without an anesthesia technician, cleaning and setup of anesthesia equipment may wait until the anesthesia provider returns from the postanesthesia care unit. The result was that the OR finished later in the workday, increasing the work hours of the surgeons, anesthesia providers, and OR nurses. This increase in makespan (a term used in the field of operations research/industrial engineering), measured in minutes, quantified the increased waiting experienced by surgeons. The number of simulated OR-day combinations
Figure 3. Average reduction in minutes of simultaneous turnovers per day exceeding the threshold number of turnover teams.
For example, (A) calculate the total minutes of simultaneous turnovers exceeding the threshold of two turnover teams. A 5-min period with four simultaneous turnovers would contribute 10 min, where 10 min (5 min) (4 simultaneous turnovers 2 teams). (B) Repeat using three turnover teams. The value of 19 min in the figure equals the average of AB. The error bars are 95% confidence intervals. The figure shows that each increase in the threshold by one team is associated with large decreases in the incremental reduction in turnover time resulting from a further 1 increase in the number of teams. However, the data were measured with 23 teams, analogous to the situation described in Figures 1 and 2. This is indicated by the value of 19 min labeled with a bold box. The facility had 23 anesthesia technicians during the study period.
was sufficient for widths of 95% CIs for all reported end points to be 0.6% of the value of the end point. Being so narrow, these CIs are not displayed.
RESULTS
Figure 3 shows the average reduction in minutes of turnovers per day from each unit increase in the number of teams. Increasing from 1 to 2 teams reduced the daily minutes for which the number of simultaneous turnovers exceeded the numbers of teams by 82 min/day. Increasing from 2 to 3 teams reduced the daily minutes by 19 min. Increasing from 3 to 4 teams reduced the daily minutes by 2.8 min/day. Figure 4 provides further insight into the large changes resulting from each incremental increase in the numbers of teams in Figure 3. The number of simultaneous turnovers exceeded two teams for 6.2% of turnover times. In contrast, one team was exceeded for much (30%) of the turnover time, whereas three teams were exceeded for little (0.8%) of the turnover time. Figure 5 shows that if 6.2% were acceptable, the consequence would be that on most days (81%) there would be at least one event when the number of simultaneous turnovers exceeds the number of teams. If a manager wanted to reduce this rate of exceedance to below 1 event/week (20% of days) or even 1 event/3 weeks (6.7%), four teams would be required. We examined further how to analyze statistically the incremental reduction in minutes for which the number of simultaneous turnovers exceeds the number of teams by the addition of another team. First, Figure 6 shows a histogram of the n 122 averages over 4-wk periods of the reduction in the
Figure 4. Percentages of overall minutes of turnover times attributable to minutes of turnovers occurring when the number of simultaneous turnovers exceeded the threshold number of teams on the horizontal axis. For example, when the number of simultaneous turnovers exceeded two teams, there was a total of 6.2% of all turnover time. The datum for three turnover teams is listed with a square because the facility had 23 anesthesia teams and results of the analyses (e.g., Fig. 3) suggest that three turnover teams is appropriate. Two-sided 95% confidence intervals are present but are sufficiently narrow to be obscured by the circles. daily minutes of turnover times achieved by increasing the threshold of exceeded turnovers from 2 to 3 teams. The distribution was sufficiently close to normal for Students t-distribution to give accurate CIs (Lilliefors test P 0.58). Second, the number of 4-wk periods for use in routine monitoring was considered. The CV of the moving average using six 4-wk periods was 27%, using 9 periods was 25%, and using 13 periods was 22%. The improvement in the CV was significant between 6 and 13 periods (P 0.039, Z 2.06) but not between 6 and 9 periods (P 0.36, Z 0.92) or between 9 and 13 periods (P
903
Table 1. Application of Method to 24 Operating Room Tertiary Surgical Suite 3 vs 4 4 vs 5 5 vs 6

Reduction in minutes per day of simultaneous turnovers exceeding number of turnover teams Mean 66 27 9 Lower 95% CI 61 23 7 Upper 95% CI 71 31 12 Maximum potential reduction in minutes of turnover time per day by increasing number of turnover teams Mean 102 36 9
Figure 5. Percentage of workdays with at least 1 min for which

the number of simultaneous turnovers exceeded the listed number of teams. The datum for three turnover teams is listed with a square because the facility had two to three anesthesia teams, and results of the analyses (e.g., Fig. 3) suggest that three turnover teams is appropriate. One-sided 95% upper confidence bounds are given for the cumulative distribution of the percentage of days with the peak number of simultaneous turnovers exceeding the number of turnover teams.
The Lower 95% CI and Upper 95% CI values refer, respectively, to lower and upper 95% condence bounds for the mean calculated using 13 4-wk periods of data. The maximum potential reduction in minutes of turnover times achieved by a unit increase in the number of teams was calculated by summing the successive decreases in the mean reduction to the right of the baseline turnover team level. For example, an increase in the number of teams from 3 to 4 can directly reduce total turnover time by at most 102 min, where 102 min 66 27 9. An increase in the number of teams from 4 to 5 can directly reduce total turnover time by at most 36 min, where 36 min 27 9. CI condence interval.
Figure 6. Histogram of the averages over 4-wk periods of daily

differences of turnover times achieved by increasing the number of turnover teams from 2 to 3. The difference corresponds to the middle bar labeled 19 min in Figure 3. The dotted vertical line shows 19 min. The best fit normal probability density curve is overlaid (P 0.58, n 122 4-wk periods).
0.25, Z 1.15). Therefore, we recommend the use of 1 yr of data. Third, discrete-event simulation was performed using the raw datas parameter values. Figure 2 shows that the reduction in daily minutes of turnover times from the addition of a turnover team matches or exceeds the daily minutes for which the number of simultaneous turnovers exceeds the number of teams. In the example in Figure 1, the two match. The maximum potential reduction in total turnover time equals the reduction in daily minutes of simultaneous turnovers achieved by increasing the number of teams
904
Staffing for Simultaneous Turnovers
to the number of ORs. We describe the relevance by referring to Figure 3. First, when the number of teams is relatively low, the reduction in the minutes for which the number of simultaneous turnovers exceeds the number of turnover teams as achieved by increasing the number of teams by 1 underestimates the reduction in the total daily minutes of turnover time. However, the underestimation is not sufficiently large to alter the managerial decision as to whether to add a team. When the number of turnover teams is relatively high, compared with the numbers of ORs, the reduction in minutes during which the number of simultaneous turnovers exceeds the numbers of teams achieved by increasing by one team is an accurate estimate. Second, the maximum potential reduction in total turnover time of the suite can be calculated from the observed data without simulation because there is an upper and lower limit to the potential benefit of increasing the number of teams. For example, suppose the anesthesia group were sharing analysis results with hospital administration showing the potential value in hiring another anesthesia technician and housekeeper to increase from 2 turnover teams to 3 teams. There was a 19 min reduction in the minutes for which the number of simultaneous turnovers exceeded the number of teams (Figure 3). The corresponding increases from three turnover teams to four teams was 2.8 min, from four turnover teams to five teams was 0.23 min, and from five turnover teams to six teams was 0.01 min (Fig. 3). Further increases in the number of teams would result in savings of 0 min, because there are six ORs. Summing these values (using the raw data to avoid rounding) gives 24 min. The corresponding saving of extra minutes of turnover time from the addition of one team is at least the observed 19 min but no more than 24 min. Table 1 gives an example of application of the statistical method to a larger suite using 1 yr of data, as displayed for regular use. We expect a 27 min reduction per day in times when the number of simultaneous turnovers exceeds the number of turnover
teams by an increase from four to five teams. The maximum potential direct reduction in turnover time experienced by the surgeons, anesthesia providers, and nurses would be 36 min.
DISCUSSION
The methodology we propose to decide whether to add one turnover team is straightforward and can easily be implemented using the steps listed in the first three paragraphs of the Methods. Any vendors spreadsheet supporting addition of programming code can be used for this purpose. Figure 3s mean 19 min reduction in minutes of simultaneous turnovers by adding one turnover team should be balanced against increases in working times of staff who participate in the team. An 8 h day has 480 min. The ratio of 19 to 480 min values the time of extra personnel as 4.0% of the time of OR staff, surgeons, and anesthesia providers, since 4.0% 19 min/480 min. In other words, the facility would be paying for 480 min of work by the extra turnover team member to save 19 min of idle OR time. The value of 4.0% can be compared with thresholds for ORs and physicians waiting for patients. Applying national compensation data, the value of patients time averages 5.3% that of OR and physicians.21 At a tertiary suite, patients entered the holding area after the OR was ready 5.0% of the time.22 Exceeding 3 teams at a 4.0% rate is close to these referenced values of 5.3% and 5.0%. Our methodology does not differentiate among times of the day. Most turnovers occur in the middle of the workday (e.g., 10:00 am to 2:00 pm) not at the end. If part-time people were to be hired to improve turnover times, their work hours can best be determined by calculating which hours of the day have the largest numbers of prolonged turnovers. The methodology is described in Ref. 10 and reviewed in Ref. 23. An alternative approach to our analysis of Figure 3 would be to use the discrete-event simulation of Figure 2 under routine circumstances to evaluate the impact of adding turnover teams.24 However, developing such models tends to be expensive and needs special software and expertise. Detailed data on workflow and processes would be needed to evaluate the impact of adding individual members of a team (e.g., one anesthesia technician). Thus, we recommend discrete-event simulation for research as used in the current article but not for routine use.10,11,16,24 REFERENCES
1. Vitez TS, Macario A. Setting performance standards for an anesthesia department. J Clin Anesth 1998;10:166 75
2. Eappen S, Flanagan H, Lithman R, Bhattacharyya N. The addition of a regional block team to the orthopedic operating rooms does not improve anesthesia-controlled times and turnover time in the setting of long turnover times. J Clin Anesth 2007;19:8591 3. Cendan JC, Good M. Interdisciplinary work flow assessment and redesign decreases operating room turnover time and allows for additional caseload. Arch Surg 2006;141:659 4. Smith MP, Sandberg WS, Foss J, Massoli K, Kanda M, Barsoum W, Schubert A. High-throughput operating room system for joint arthroplasties durably outperforms routine processes. Anesthesiology 2008;109:2535 5. Heslin MJ, Doster BE, Daily SL, Waldrum MR, Boudreaux AM, Smith AB, Peters G, Ragan DB, Buchalter S, Bland KI, Rue LW. Durable improvements in efficiency, safety, and satisfaction in the operating room. J Am Coll Surg 2008;206:10839 6. Semo JJ. Ambulatory surgical centers: a manual for anesthesiologists. Washington, DC: American Society of Anesthesiologists, 2006:44, 81 7. Masursky D, Dexter F, Nussmeier NA. Operating room nursing directors influence on anesthesia group operating room productivity. Anesth Analg 2009;107:1989 96 8. Kadous K, Koonce L, Towry KL. Quantification and persuasion in managerial judgment. Contemp Account Res 2005;22:643 86 9. Dexter F, Abouleish AE, Epstein RH, Whitten CW, Lubarsky DA. Use of operating room information system data to predict the impact of reducing turnover times on staffing costs. Anesth Analg 2003;97:1119 26 10. Dexter F, Epstein RH, Marcon E, Ledolter J. Estimating the incidence of prolonged turnover times and delays by time of day. Anesthesiology 2005;102:1242 8 11. Marcon E, Dexter F. Observational study of surgeons sequencing of cases and its impact on post-anesthesia care unit and holding area staffing requirements at hospitals. Anesth Analg 2007;105:119 26 12. Dexter F, Traub RD. Determining staffing requirements for a second shift of anesthetists by graphical analysis of data from operating room information systems. AANA J 2000;68:31 6 13. Dexter F, Epstein RH. Optimizing second shift OR staffing. AORN J 2003;77:82530 14. Sprent P. Applied nonparametric statistical methods. New York: Chapman and Hall, 1989:49 58 15. Mosteller F, Youtz C. Tables of the Freeman-Tukey transformations for the binomial and Poisson distributions. Biometrika 1961;48:433 40 16. Dexter F, Marcon E, Epstein RH, Ledolter J. Validation of statistical methods to compare cancellation rates on the day of surgery. Anesth Analg 2005;101:46573 17. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 1934;26:404 13 18. Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17: 85772 19. Miller GE. Asymptotic test statistics for coefficients of variation. Commun Statist Theory Meth 1991;20:3351 63 20. Dexter F, Macario A, Manberg PJ, Lubarsky DA. Computer simulation to determine how rapid anesthetic recovery protocols to decrease the time for emergence or increase the phase I post anesthesia care unit bypass rate affect staffing of an ambulatory surgery center. Anesth Analg 1999;88:1053 63 21. Dexter F, Traub RD. Statistical method for predicting when patients should be ready on the day of surgery. Anesthesiology 2000;93:110714 22. Wachtel RE, Dexter F. Simple method for deciding what time patients should be ready on the day of surgery without procedure-specific data. Anesth Analg 2007;105:127 40 23. McIntosh C, Dexter F, Epstein RH. Impact of service-specific staffing, case scheduling, turnovers, and first-case starts on anesthesia group and operating room productivity: tutorial using data from an Australian hospital. Anesth Analg 2006;103:1499 516 24. Albert F, Marcon E. How to combine transversal staff skills in operating rooms? Information control problems in manufacturing, Proceedings of the 12th International Federation of Automatic Control Symposium on Information Control Problems in Manufacturing 2006;3:677 82
905
Neurosurgical Anesthesiology and Neuroscience

Section Editor: Adrian W. Gelb
The Effect on Cerebral Tissue Oxygenation Index of Changes in the Concentrations of Inspired Oxygen and End-Tidal Carbon Dioxide in Healthy Adult Volunteers
Martin M. Tisdall, MD* Christopher Taylor, FRCA* Ilias Tachtsidis, PhD Terence S. Leung, PhD Clare E. Elwell, PhD Martin Smith, FRCA*
BACKGROUND: A variety of near-infrared spectroscopy devices can be used to make noninvasive measurements of cerebral tissue oxygen saturation (ScO2). The ScO2 measured by the NIRO 300 spectrometer (Hamamatsu Photonics, Japan) is called the cerebral tissue oxygenation index (TOI) and is an assessment of the balance between cerebral oxygen delivery and utilization. We designed this study to investigate the effect of systemic and intracranial physiological changes on TOI. METHODS: Fifteen healthy volunteers were studied during isocapneic hyperoxia and hypoxemia, and normoxic hypercapnea and hypocapnea. Absolute cerebral TOI and changes in oxy- and deoxyhemoglobin concentrations were measured using a NIRO 300 spectrometer. Changes in arterial oxygen saturation (Sao2), ETco2, heart rate, mean arterial blood pressure (MBP), and middle cerebral artery blood flow velocity (Vmca) were also measured during these physiological challenges. Changes in cerebral blood volume (CBV) were subsequently calculated from changes in total cerebral hemoglobin concentration. RESULTS: Baseline TOI was 67.3% with an interquartile range (IQR) of 65.2%71.9%. Hypoxemia was associated with a median decrease in TOI of 7.1% (IQR 9.1% to 5.4%) from baseline (P 0.0001) and hyperoxia with a median increase of 2.3% (IQR 2.0%2.5%) (P 0.0001). Hypocapnea caused a reduction in TOI of 2.1% (IQR 3.3% to 1.3%) from baseline (P 0.0001) and hypercapnea an increase of 2.6% (IQR 1.4%3.7%) (P 0.0001). Changes in Sao2 (P 0.0001), ETco2 (P 0.0001), CBV (P 0.0003), and MBP (P 0.03) were significant variables affecting TOI. Changes in Vmca (P 0.7) and heart rate (P 0.2) were not significant factors. CONCLUSION: TOI is an easy-to-monitor variable that provides real-time, multisite, and noninvasive assessment of the balance between cerebral oxygen delivery and utilization. However, TOI is a complex variable that is affected by Sao2 and ETco2, and, to a lesser extent, by MBP and CBV. Clinicians need to be aware of the systemic and cerebral physiological changes that can affect TOI to interpret changes in this variable during clinical monitoring.
(Anesth Analg 2009;109:906 13)
erebral oxygenation monitoring is widely used to assess the balance between cerebral metabolic supply and demand but standard bedside methods of measuring cerebral oxygenation have significant limitations. Jugular venous oxygen saturation is a global, flowweighted measure that may miss regional ischemia,1 whereas intraparenchymal brain tissue oxygen tension is a hyperfocal measure and its ability to identify ischemia is dependent on the location of the probe.2 In addition to being invasive, these techniques are also associated with
From the *Department of Neuroanaesthesia and Neurocritical Care, The National Hospital for Neurology and Neurosurgery, University College London Hospitals; and Department of Medical Physics and Bioengineering, University College London, London, UK. Accepted for publication April 20, 2009. Supported by the Wellcome Trust (Grant no. 075608) (to MT), the Engineering Physical Science Research Council (Grant no. GR/ N14248/01) (to IT), and Hamamatsu Photonics, KK (to TL). This work was undertaken at University College London Hospitals and partially funded by the Department of Healths NIHR Biomedical Research Centres funding scheme. The work was also made possible in part by a donation in memory of Karolyn Margaret Jones.
a degree of technical difficulty and are not widely available outside specialist centers.3 There is therefore a need for a noninvasive, bedside measure of cerebral oxygenation that can provide real-time data from several regions of the brain simultaneously. Near-infrared spectroscopy (NIRS) is a noninvasive technique based on the transmission and absorption of near-infrared light (7001000 nm) at multiple wavelengths as it passes through tissue. NIRS allows interrogation of the cerebral cortex using reflectance spectroscopy via optodes, light transmitting and detecting devices, placed
Professor Clare E. Elwell has received lecture honoraria and Dr. Terence S. Leung salary funding from Hamamatsu Photonics, KK, Japan. Address correspondence and reprint requests to Martin Smith, FRCA, Department of Neuroanaesthesia and Neurocritical Care, Box 30, The National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK. Address e-mail to martin.smith@uclh.nhs.uk. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aedcdc
906
on the scalp.4 Oxygenated hemoglobin (O2Hb) and deoxygenated hemoglobin (HHb) have different absorption spectra, and cerebral oxygenation and hemodynamic status can be determined by their relative absorption of near-infrared light. Biological tissue is a highly scattering medium but if the average path length of light through tissue is known, the modified Beer-Lambert law (MBL), which assumes constant scattering losses, allows calculation of absolute changes in chromophore concentration.5 Earlier NIRS methodology was predominantly limited to differential spectroscopy methods that provide trend monitoring of the changes in tissue chromophore concentration (e.g., O2Hb and HHb).5 These variables are generally unfamiliar to clinicians, even if the changes are quantified in micromolar units. Technical developments, for example, the use of spatially resolved spectroscopy (SRS), have allowed the introduction of clinical monitors that incorporate an absolute measure of cerebral tissue hemoglobin oxygen saturation (ScO2), an easily accessible and continuous measure of the balance between cerebral tissue oxygen delivery and utilization.6 There are a variety of NIRS instruments available that measure ScO2 in some form.7 The NIRO 300 spectrometer (Hamamatsu Photonics, Japan) uses four wavelengths (778, 813, 850, and 913 nm) and the MBL to measure changes in O2Hb and HHb concentrations, and the SRS technique to measure absolute ScO2, which is expressed as the cerebral tissue oxygenation index (TOI) and is displayed as a simple percentage value.8 The application of SRS and the validity of TOI have been described in normal adult volunteers9,10 and in clinical scenarios.1113 The depth sensitivity of TOI has also been evaluated by selective internal and external carotid artery clamping during carotid surgery and shows high sensitivity and specificity to intracerebral changes in adults.11 Furthermore, because NIRS interrogates arterial, venous, and capillary blood within the field of view, the derived saturation represents a tissue oxygen saturation measured from these three compartments6 and can be used to identify tissue hypoxia/ischemia.14 The development of indices such as TOI has been motivated in part by the desire to provide clinicians with an easily accessible measure of cerebral tissue oxygenation. The optical measurement of TOI is derived from the proportion of O2Hb relative to total Hb (HbT) concentration in the field of view.8 However, exactly what TOI represents in physiological terms is complex and likely to be influenced by a number of inputs. These have been summarized as15:
oxygen carrying ability of hemoglobin, CBF cerebral blood flow, and [Hb] blood Hb concentration. Most of these variables affect either cerebral oxygen delivery or utilization, and TOI should therefore be a reasonable measure of the balance between the two. However, when interpreting TOI, or other NIRS measures of Sco2, it is important to understand how physiological variables might affect the measured saturation. From Eq. 1 it is clear that changes in Sao2 will affect TOI and, although of fundamental importance, this relationship has not previously been studied. Because NIRS interrogates arterial, venous, and capillary blood, TOI will also be affected by variation in the cerebral arterial:venous volume ratio (AVR). While the AVR is typically 1:3 (25% arterial and 75% venous),16 the actual ratio depends on individual anatomy, local physiology, and pathological states.17 Changes in arterial Paco2 also induce changes in cerebral AVR18 and Paco2 is therefore similarly likely to affect TOI. This study was designed to investigate these relationships by observing the effects on cerebral TOI of changes in cerebral oxygen delivery during isocapneic hyperoxia and hypoxemia, and normoxic hypercapnea and hypocapnea, in healthy volunteers.
METHODS
The study was approved by the Joint Research Ethics Committee of the National Hospital for Neurology and Neurosurgery and the Institute of Neurology, University College London.
NIRS Measurements
After obtaining informed written consent, a NIRO 300 monitor was used to measure absolute TOI using SRS and changes in HbO2 and HHb concentrations using the MBL during a variety of physiological challenges in 15 healthy adult volunteers. The sourcedetector optode pair was fixed in a black rubber holder with a source-detector separation of 5 cm over the right side of the forehead in the midpupillary line, avoiding the sinuses. The optode holder was secured to the head using an elasticated crepe bandage to prevent optode movement and covered with a lightabsorbing cloth to eliminate stray light. NIRS data were collected at 6 Hz.
Other Measurements
Sao2 was measured using a pulse oximeter, modified to provide beat-to-beat recording (Novametrix Medical Systems, Wallingford, CT), with the probe attached to the subjects left ear. Mean arterial blood pressure (MBP) and heart rate (HR) were measured noninvasively using a Portapress and finger probe (Biomedical Instrumentation, TNO Institute of Applied Physics, Belgium). MBP was recorded from the analog output of the Portapress continuously at 100 Hz and the signal was later resampled to 6 Hz. Blood flow velocity in the right middle cerebral artery (Vmca) was measured by the same experienced operator (MT)
TOI Sao2
Vv Va Vv
CMRO2
k CBF [Hb]
100 (1)
where Sao2 arterial oxyhemoglobin saturation, Vv and Va venous and arterial blood volume, respectively, CMRO2 cerebral metabolic rate for oxygen, k
907
using 2 MHz transcranial Doppler ultrasonography (TCD) (Nicolet, UK), as a surrogate of CBF.19 Mean Vmca was calculated from the CBF velocity envelope using a trapezoidal integration function (MatLab, Mathworks, USA). Inspired oxygen fraction (Fio2) and ETco2 were measured using an in-line gas analyzer (Hewlett Packard, UK) and a CO2SMO optical sensor (Novametrix Medical Systems), respectively.
Study Protocol
A modified anesthetic machine delivered inspired gas to the subjects via a Mapelson E (Ayres T-piece) breathing system incorporating a mouthpiece and 50 cm expiratory limb. The study was divided into four challenge periods with a rest period between each. Each challenge period was preceded by 5 min of data collection at normoxia and normocapnea. Three cycles of the following physiological challenges were performed in each volunteer. Hypoxemia Nitrogen was added to the inspired gas to induce a gradual decrease in Sao2 to 80% and, immediately after this was achieved, Fio2 was returned to baseline (normoxia) for 5 min. ETco2 was continuously fed back to the subjects so that they could adjust their minute ventilation and maintain normocapnea throughout this part of the study. Hyperoxia Fio2 was increased to 100% for 5 min and then returned to normoxia for 5 min. The subjects again adjusted their minute ventilation using the ETco2 feedback system to maintain normocapnea. Hyperventilation The subjects hyperventilated to reduce ETco2 by 1.5 kPa below baseline. This was maintained for 5 min and then a normal breathing rate was resumed, allowing ETco2 to return to baseline over approximately 5 min. Hypercapnea Approximately 6% CO2 was added to the inspired gas and titrated to induce an increase in ETco2 of 1.5 kPa. This was maintained for 5 min and the inspired CO2 fraction was then returned to zero for another 5 min. At the end of the study, a venous blood sample was obtained and the Hb concentration measured using a coximeter (ABL 700, Radiometer Copenhagen, Denmark).
Figure 1. Schematic of data analysis for hypoxemia paradigm using Sao2 data to define data windows for summary analysis. Table 1. Median and Interquartile Range (IQR) for Baseline Values of Measured Variables Before the Start of Hypoxemias (n 15) Median
Fio2 (%) SaO2 (%) ETco2 (kPa) HR (min1) MBP (mm Hg) Vmca (cm/s) Hb (g/dL) TOI (%) 21.0 99.2 5.4 62.5 77.6 43.2 14.7 68.3
IQR
21.021.0 98.299.2 5.25.7 60.071.5 70.388.8 37.951.1 13.7514.95 65.271.9
FiO2 inspired oxygen fraction; SaO2 arterial oxygen saturation; ETCO2 end-tidal carbon dioxide tension; HR heart rate; MBP mean arterial blood pressure; Vmca middle cerebral artery blood ow velocity; Hb hemoglobin; TOI tissue oxygenation index.
CBV
[HbT] MWHb 100 [Hb] CSLVH brain
(2)
Data Analysis
Absolute change in O2Hb and HHb concentrations was calculated from changes in light attenuation using the MBL and the UCL4 algorithm, assuming a differential pathlength factor of 6.26.20 Changes in HbT concentration were subsequently calculated ([HbT] [HbO2] [HHb]) and converted to changes in cerebral blood volume (CBV) using the formula:
908
Cerebral Tissue Oxygen Saturation
where CBV change in CBV (mL/100g of brain), [HbT] change in total hemoglobin concentration (mol/L), MWHb molecular weight of hemoglobin (64 500 g/mole), [Hb] the large vessel hemoglobin concentration (g/L), CSLVH the cerebral small vessel to large systemic vessel hematocrit, and the brain density (1.05 g/mL). The start and end of each challenge period was identified from the Sao2, Fio2, or ETco2 data according to the phase of the study. To enable comparison between subjects and across physiological challenges, 8 points were selected within each individual period of alteration of Fio2 or ETco2 (the challenge period) so that the time between adjacent points represented an eighth of the total time course of the challenge period. This produced 9 timepoints with Point 1 representing the point just before the start and Point 9 the end of the challenge period (Fig. 1). The recovery period was similarly divided, producing Points 9 (just before start of recovery) to 17 (end of recovery period). At each timepoint, the mean of the preceding 10 s of data was calculated and used for analysis. Data from the three experimental cycles of each physiological challenge
Figure 2. Median and interquartile range (n 15) for variable values during hypoxemia (*P 0.05, P 0.01, P 0.001, P 0.0001 for change from baseline).
were averaged to give a single course for each subject. For each of the challenges, a mean of the variables from the two NIRS channels was calculated. Group median changes from baseline at each timepoint were produced. Statistical analysis was performed using SAS software (v9.1, SAS Institute, USA). Percentage changes from baseline for Vmca and absolute changes from baseline for other measured variables were compared using nonparametric analysis of variance with post hoc pairwise comparisons.21 P values 0.05 were considered significant. Multiple regression analysis was performed using change in TOI as the dependent variable and changes in other variables as regression variables. Regression variables which were not significant were then removed and the regression analysis repeated.
(P 0.0001) and hyperoxia with a median increase of 2.3% (IQR 2.0%2.5%) (P 0.0001). Hypocapnea caused a reduction in TOI of 2.1% (IQR 3.3% to 1.3%) from baseline (P 0.0001) and hypercapnea an increase of 2.6% (IQR 1.4%3.7%) (P 0.0001). Figures 25 show the group data for the measured variable values during the four physiological challenges. Multiple regression analysis confirmed that changes in Sao2 (P 0.0001), ETco2 (P 0.0001), CBV (P 0.0003), and MBP (P 0.03) were significant variables affecting TOI. Percentage changes in Vmca (P 0.7) and HR (P 0.2) were not significant factors. The regression analysis was repeated using only the significant regression variables to determine the regression values:
TOI 0.53 Sao2 1.13 EtCo2 2.35 CBV 0.01 MBP (3)
The overall adjusted r value was 0.82 (P 0.0001). The standardized coefficients and P values for each variable are shown in Table 2.
RESULTS
Fifteen adult volunteers (10 male and 5 female) with median age 31 (range, 2739) yr were recruited into the study. Baseline values for the measured variables are shown in Table 1. The Sao2 of all subjects reached 80% at the nadir of the hypoxemic challenge, with a median reduction in SaO2 of 15.8% (interquartile range [IQR] 18.4% to 14.1%). The median increase in Fio2 at the mouthpiece during hyperoxia was 72% and all subjects responded with a significant increase in Sao2 from baseline (P 0.0001) with a median increase of 0.7% (IQR 0.5% 0.9%). During hyperventilation, the median reduction in ETco2 was 1.5 kPa (IQR 1.7 to 1.4 kPa), and during hypercapnea, the median increase in ETco2 was 1.7 kPa (IQR 1.51.9 kPa). Hypoxemia was associated with a median decrease in TOI of 7.1% (IQR 9.1% to 5.4%) from baseline
DISCUSSION
Hyperoxia and hypercapnea resulted in an increase in cerebral TOI, whereas TOI was reduced during hypoxemia and hyperventilation. TOI is predominantly affected by Sao2 and ETco2 and, to a lesser extent, by CBV and MBP. There was a large variation in TOI among individuals in this study. The median baseline TOI was 68.3% with an IQR of 65.2%71.9%. This variability has been reported previously.22,23 The normal range for TOI varies between 60% and 75%, with a coefficient of
909
Figure 3. Median and interquartile range (n 15) for variable values during hyperoxia (*P 0.05, P 0.01, P 0.001, P 0.0001 for change from baseline).
Figure 4. Median and interquartile range (n 15) for variable values during hyperventilation (*P 0.05, P 0.01, P 0.001, P 0.0001 for change from baseline).
variation for absolute baseline values of almost 10%, and this might limit the usefulness of isolated measurements of TOI.23 Rasmussen et al.24 estimated capillary Hb oxygen saturation as the mean of arterial and jugular bulb oxygen saturation and compared this derived value with TOI during changes in inspired oxygen and carbon dioxide fractions. Although these variables were correlated, there was a wide variation between TOI and the modeled capillary oxygen saturation, also suggesting that the potential for using absolute TOI values to define ischemic thresholds or
910
guide targeted therapy in the clinical environment might be limited. Al-Rawi and Kirkpatrick12 attempted to determine the reduction in TOI that is associated with cerebral ischemia by studying the effect of carotid artery clamping during carotid surgery. Electroencephalography was used to define the presence of cerebral ischemia, and no patient with a percentage reduction in TOI 13% showed electroencephalography evidence of ischemia. The potential application of TOI and other NIRS measures of ScO2 as trend monitors of incipient cerebral hypoxia/ischemia is
Figure 5. Median and interquartile range (n 15) for variable values during hypercapnea (*P 0.05, P 0.01, P 0.001, P 0.0001 for change from baseline).
Table 2. Standardized Regression Estimates (standard ) and P Values for Multiple Regression Analysis Variables Shown in Eq. 1 Variable
Standard P
Sao2
0.71 0.0001
ETco2
0.37 0.0001
CBV
0.09 0.0005
MBP
0.04 0.03
CBV cerebral blood volume; MBP mean arterial blood pressure.
therefore attractive but, for their interpretation to be clinically valid, clinicians must understand which systemic and cerebral physiological variables affect the measured cerebral saturation. In our study, hypoxemia (Sao2 approximately 80%), was associated with a median decrease in TOI of 7.1% from a median baseline of 68.3%. Although it is likely that the reduction in TOI was related to the reduction in Sao2, it is important to exclude other causes. There was a small (0.2 kPa) but statistically significant reduction in median ETco2 at the nadir of hypoxemia. This was likely because of the hypoxemic stimulus to hyperventilate despite the application of an ETco2 biofeedback loop. However, multiple regression analysis demonstrates that this magnitude of change in ETco2 in isolation would induce a reduction in TOI of only 0.2% and it is therefore unlikely to be contributing to the large reduction in TOI that we observed. In agreement with other investigators,25 we observed a significant increase in HR and small change in MBP during hypoxemia, but these are also unlikely to have affected TOI. There was an increase in Vmca and CBV during hypoxemia and this finding is in keeping with previous studies that identified the threshold for hypoxic vasodilatation in healthy volunteers occurring at Sao2 of around 90%.25 The small increase in Sao2 (median 0.7%) during hyperoxia was associated with a median increase in
TOI of 2.3%. Although this degree of change in TOI is unlikely to be of clinical significance, it warrants an explanation from a physiological perspective. Despite subjects attempting to maintain isocapnea, there was a small but statistically significant decrease in ETco2 (median 0.3 kPa) that was likely related to two linked effects. Increasing oxyhemoglobin saturation decreases the affinity of Hb for carbon dioxide (the Haldane effect),26 thereby reducing carbon dioxide uptake from tissue. This is likely to translate to a reduction in Paco2 and therefore in ETco2. It has also been suggested that the Haldane effect-mediated CO2 retention in the respiratory centers of the brain might induce a hyperventilatory response that would in turn result in decreased Paco2 and ETco2.27 Hyperoxia caused a reduction in Vmca and CBV and this might in part be related to the small reduction in Paco2. However, arterial-spin-labeled magnetic resonance imaging studies indicate that normobaric hyperoxia has a direct cerebral vasoconstrictive effect and an indirect effect mediated via the reduced Paco2.27 Reductions in CBV and Vmca would tend to reduce TOI, but, during hyperoxia, we observed the opposite, i.e., an increase in TOI that was of greater magnitude (median increase 2.3%) than the associated increase in Sao2 (median increase 0.7%). This might, in part, be related to an increase in dissolved blood oxygen. However, this is unlikely to be the only explanation because of the modest increase in dissolved oxygen that occurs during hyperoxia. Assuming a constant CMRO2, our data suggest that the combined effect of the small increase in arterial oxygen content and reduced CBV and Vmca (and therefore presumably of CBF) is an overall increase in cerebral oxygen delivery. The increase in TOI that we observed therefore seems to indicate that the reduction in CBF during hyperoxia was small and
911
compensated for by the increase in dissolved blood oxygen. Alternatively, our results could be explained by a decrease in CMRO2 during hyperoxia, although we believe that this is unlikely during the short time course of the study. Hyperventilation caused a reduction in TOI in association with a small increase in Sao2 and reduction in CBV. The latter is presumably related to the known cerebral arteriolar vasoconstrictive effects of reduced Paco228 because we also observed a simultaneous reduction in Vmca. Despite the small increase in Sao2 during hyperventilation, TOI was reduced and this is likely to be explained by a reduction in CBV and CBF. It is of note that the CBV and Vmca responses had different time courses during this phase of the study. Vmca returned to baseline during the recovery period, whereas CBV returned toward, but did not reach, baseline by the end of the study. This suggests that the autoregulatory processes which attempt to maintain a stable CBF might entail mechanisms beyond changes in arteriolar calibre. Hypercapnea resulted in an increase in TOI in association with a small increase in SaO2. The latter is likely to be related to the tendency of the subjects to hyperventilate in the presence of a high Paco2 despite the application of the biofeedback mechanism. During the early part of the hypercapneic challenge, there was an increase in Vmca that returned toward baseline before the end of the challenge period, again suggesting the presence of autoregulatory mechanisms in addition to carbon dioxide effects on arteriolar caliber. CBV increased during hypercapnea and returned toward, but did not reach, baseline values during the recovery period. The time course of the hypercapneainduced increase in TOI suggests that the TOI changes were more likely to be related to the increase in CBV than the increase in Vmca because TOI also returned toward, but did not reach, baseline by the end of the recovery period. A similar increase in cerebral tissue oxygen tension in response to hypercapnea which outlasts the CO2 changes has also been demonstrated in a rat model.29 The high temporal resolution of the noninvasive techniques used in this study offers a unique opportunity to investigate the relative time courses of changes in cerebral oxygenation and hemodynamic variables. Such an analysis is beyond the scope of this study, and to fully understand how the complex interactions between changes in Pao2, Paco2, and CBF interact to produce changes in TOI, we are undertaking further analysis within the context of a recently published mathematical model that was specifically developed to aid the interpretation of cerebral NIRS data.30 Assuming constant CMRO2 and Hb concentration during the challenge periods, our data indicate that several variables can affect the value of TOI. Change in Sao2 is the most important and ETco2 the second most important. Although changes in CBV and MBP
912
were statistically significant, their standardized values were an order of magnitude lower than that for Sao2 (0.09 and 0.04 respectively, vs 0.71) and their effects are unlikely to have major clinical relevance. It is of note that CBV was a significant variable affecting TOI, whereas percentage change in Vmca was not. Regression analysis confirmed that the changes in ETco2 alone can account for the observed changes in Vmca and this is likely to explain why Vmca was not a significant independent variable affecting TOI. When we performed multiple regression analysis with changes in ETCO2 omitted, changes in Vmca became a significant factor affecting TOI (analysis not shown) and this tends to confirm this hypothesis. There are several limitations to our study. Because we wished to avoid the placement of arterial lines for blood gas analysis in volunteers, we used Sao2 as a measure of arterial oxygenation and ETco2 as a surrogate for Paco2. Measurement of Sao2 does not account for dissolved oxygen, but the consequence of this is likely to be negligible during normoxia and hypoxia. It might, however, become relevant during hyperoxia and impact on the results of that part of our study. Changes in ETco2 are accurate surrogates of changes in Paco2 in healthy subjects and the use of ETco2 is therefore unlikely to have affected our results.31 We used TCD-derived Vmca as a surrogate for CBF and this relationship relies on there being no change in the diameter of the insonated vessel or in the angle of insonation during the measurements. Magnetic resonance imaging studies have confirmed that basal middle cerebral arterial diameter does not change substantially during the type of physiological challenges that we used in this study.19 Furthermore, the TCD studies were performed by an experienced operator (MT) using a proprietary probe head fixation system to minimize artifact from probe movement. Because TOI is the balance between cerebral oxygen delivery and utilization, changes in CMRO2 are likely to affect TOI. Although it is unlikely that CMRO2 changed during the physiological challenges in our healthy subjects, we did not measure CMRO2 and cannot therefore exclude changes in CMRO2 as a confounding factor. Finally, our calculation of CBV, measured from MBL-derived changes in O2Hb and HHb concentrations, is likely to be prone to some error because these variables do not exclusively measure the intracerebral compartment.7,11 However, they do not apply to the SRS measurement of TOI which has high sensitivity and specificity to intracerebral changes.11 In conclusion, TOI is an easy-to-monitor variable that provides a real-time, noninvasive assessment of regional tissue cerebral oxygenation. The predominant factors determining TOI are Sao2 and ETco2, with changes in MBP and CBV having limited effects. The variability of TOI in healthy volunteers is likely to limit its clinical usefulness as a one off measure of cerebral oxygenation, but changes in TOI have great
potential as a trend monitor for the identification of hypoxia/ischemia. However, further studies in braininjured patients are required to determine the magnitude of the reduction in TOI that is associated with incipient cerebral hypoxia/ischemia. Importantly though, clinicians need to be aware of the systemic and cerebral physiological changes that can affect TOI, in order to interpret changes in this variable during clinical monitoring. REFERENCES
1. Schell RM, Cole DJ. Cerebral monitoring: jugular venous oximetry. Anesth Analg 2000;90:559 66 2. Gupta AK, Hutchinson PJ, Al-Rawi P, Gupta S, Swart M, Kirkpatrick PJ, Menon DK, Datta AK. Measuring brain tissue oxygenation compared with jugular venous oxygen saturation for monitoring cerebral oxygenation after traumatic brain injury. Anesth Analg 1999;88:549 53 3. Tisdall MM, Smith M. Multimodal monitoring in traumatic brain injury: current status and future directions. Br J Anaesth 2007;99:617 4. Okada E, Delpy DT. Near-infrared light propagation in an adult head model. II. Effect of superficial tissue thickness on the sensitivity of the near-infrared spectroscopy signal. Appl Opt 2003;42:291522 5. Delpy DT, Cope M, van der Zee P, Arridge S, Wray S, Wyatt J. Estimation of optical pathlength through tissue from direct time of flight measurement. Phys Med Biol 1988;33:1433 42 6. Smith M. Perioperative uses of transcranial perfusion monitoring. Neurosurg Clin N Am 2008;19:489 502 7. Owen-Reece H, Smith M, Elwell CE, Goldstone JC. Near infrared spectroscopy. Br J Anaesth 1999;82:418 26 8. Suzuki S, Takasaki S, Ozaki T, Kobayashi Y. A tissue oxygenation monitor using NIR spatially resolved spectroscopy. Proc SPIE 1999;3597:58292 9. Tisdall MM, Tachtsidis I, Leung TS, Elwell CE, Smith M. Near-infrared spectroscopic quantification of changes in the concentration of oxidized cytochrome c oxidase in the healthy human brain during hypoxemia. J Biomed Opt 2007;12:024002 10. Quaresima V, Sacco S, Totaro R, Ferrari M. Noninvasive measurement of cerebral hemoglobin oxygen saturation using two near infrared spectroscopy approaches. J Biomed Opt 2000;5: 2015 11. Al-Rawi PG, Smielewski P, Kirkpatrick PJ. Evaluation of a near-infrared spectrometer (NIRO 300) for the detection of intracranial oxygenation changes in the adult head. Stroke 2001;32:2492500 12. Al-Rawi P, Kirkpatrick P. Tissue oxygen index: thresholds for cerebral ischemia using near-infrared spectroscopy. Stroke 2006;37:2720 5 13. McLeod AD, Igielman F, Elwell C, Cope M, Smith M. Measuring cerebral oxygenation during normobaric hyperoxia: a comparison of tissue microprobes, near-infrared spectroscopy, and jugular venous oximetry in head injury. Anesth Analg 2003;97:851 6 14. Kurth CD, Levy WJ, McCann J. Near-infrared spectroscopy cerebral oxygen saturation thresholds for hypoxia-ischemia in piglets. J Cereb Blood Flow Metab 2002;22:335 41
15. Tachtsidis I, Tisdall M, Delpy DT, Smith M, Elwell CE. Measurement of cerebral tissue oxygenation in young healthy volunteers during acetazolamide provocation: a transcranial Doppler and near-infrared spectroscopy investigation. Adv Exp Med Biol 2008;614:389 96 16. An H, Lin W. Cerebral venous and arterial volumes can be estimated separately in humans using magnetic resonance imaging. Magn Reson Med 2002;48:583 8 17. Watzman HM, Kurth CD, Montenegro LM, Rome J, Steven JM, Nicolson SC. Arterial and venous contributions to near-infrared cerebral oximetry. Anesthesiology 2000;93:94753 18. Ito H, Ibaraki M, Kanno I, Fukuda H, Miura S. Changes in the arterial fraction of human cerebral blood volume during hypercapnia and hyocapnia measured by positron emission tomography. J Cereb Blood Flow Metab 2005;25:8527 19. Valdueza JM, Balzer JO, Villringer A, Vogl TJ, Kutter R, Einhaupl KM. Changes in blood flow velocity and diameter of the middle cerebral artery during hyperventilation: assessment with MR and transcranial Doppler sonography. Am J Neuroradiol 1997;18:1929 34 20. Duncan A, Meek JH, Clemence M, Elwell CE, Tyszczuk L, Cope M, Delpy DT. Optical pathlength measurements on adult head, calf and forearm and the head of the newborn infant using phase resolved optical spectroscopy. Phys Med Biol 1995;40: 295304 21. Siegel S, Castellan NJ. Nonparametric statistics for the behavioural sciences. 2nd ed. Singapore: McGraw-Hill, 1988 22. Sorensen LC, Greisen G. Precision of measurement of cerebral tissue oxygenation index using near-infrared spectroscopy in preterm neonates. J Biomed Opt 2006;11:054005 23. Thavasothy M, Broadhead M, Elwell C, Peters M, Smith M. A comparison of cerebral oxygenation as measured by the NIRO 300 and the INVOS 5100 Near-Infrared Spectrophotometers. Anaesthesia 2002;57:999 1006 24. Rasmussen P, Dawson EA, Nybo L, van Lieshout JJ, Secher NH, Gjedde A. Capillary-oxygenation-level-dependent near-infrared spectrometry in frontal lobe of humans. J Cereb Blood Flow Metab 2007;27:108293 25. Gupta AK, Menon DK, Czosnyka M, Smielewski P, Jones JG. Thresholds for hypoxic cerebral vasodilation in volunteers. Anesth Analg 1997;85:81720 26. Dash R, Bassingthwaighte J. Blood HbO2 and HbCO2 dissociation curves at varied O2, CO2, pH, 2,3-DPG and temperature levels. Ann Biomed Eng 2004;32:1676 93 27. Floyd TF, Clark JM, Gelfand R, Detre JA, Ratcliffe S, Guvakov D, Lambertson CJ, Eckenhoff RG. Independent cerebral vasoconstrictive effects of hyperoxia and accompanying arterial hypocapnia at 1 ATA. J Appl Physiol 2003;95:2453 61 28. Greenberg J, Alavi A, Reivich M, Kuhl D, Uzzell B. Local cerebral blood volume response to carbon dioxide in man. Circ Res 1978;43:324 31 29. Hare G, Kavanagh B, Mazer C, Hum K, Kim S, Coackley C, Barr A, Baker AJ. Hypercapnia increases cerebral tissue oxygen tension in anesthetized rats. Can J Anaesth 2007;50:1061 8 30. Banaji M, Mallet A, Elwell CE, Nicholls P, Cooper CE. A model of brain circulation and metabolism: NIRS signal changes during physiological challenges. PLoS Comput Biol 2008; 4:e10000212 31. Whitesell R, Asiddao C, Gollman D, Jablonski J. Relationship between arterial and peak expired carbon dioxide pressure during anesthesia and factors influencing the difference. Anesth Analg 1981;60:508 12
913
Deep Hypothermia Attenuates Microglial Proliferation Independent of Neuronal Death After Prolonged Cardiac Arrest in Rats
Tomas Drabek, MD* Samuel A. Tisherman, MD, FACS, FCCM* Lauren Beuke* Jason Stezoski* Keri Janesko-Feldman* Manuella Lahoud-Rahme, MD* Patrick M. Kochanek, MD, FCCM*
INTRODUCTION: Conventional resuscitation of exsanguination cardiac arrest (CA) victims is generally unsuccessful. Emergency preservation and resuscitation is a novel approach that uses an aortic flush to induce deep hypothermia during CA, followed by delayed resuscitation with cardiopulmonary bypass. Minocycline has been shown to be neuroprotective across a number of brain injury models via attenuating microglial activation. We hypothesized that deep hypothermia and minocycline would attenuate neuronal death and microglial activation and improve outcome after exsanguination CA in rats. METHODS: Using isoflurane anesthesia, rats were subjected to a lethal hemorrhagic shock. After 5 min of no flow, hypothermia was induced with an aortic flush. Three groups were studied: ice-cold (IC) flush, room-temperature (RT) flush, and RT flush followed by minocycline treatment (RT-M). After 20 min of CA, resuscitation was achieved via cardiopulmonary bypass. Survival, Overall Performance Category (1 normal, 5 death), Neurologic Deficit Score (0%10% normal, 100% max deficit), neuronal death (Fluoro-Jade C), and microglial proliferation (Iba1 immunostaining) in hippocampus were assessed at 72 h. RESULTS: Rats in the IC group had lower tympanic temperature during CA versus other groups (IC, 20.9C 1.3C; RT, 28.4C 0.6C; RT-M, 28.3C 0.7C; P 0.001). Although survival was similar in all groups (RT, 6/9; IC, 6/7; RT-M, 6/11), neurological outcome was better in the IC group versus other groups (Overall Performance Category: IC, 1 1; RT, 3 1; RT-M, 2 1; P 0.05; Neurologic Deficit Score: IC, 8% 9%; RT, 55% 19%; RT-M, 27% 16%; P 0.05). Histological damage assessed in survivors showed selective neuronal death in CA1 and dentate gyrus, similar in all groups (P 0.15). In contrast, microglial proliferation was attenuated in the IC group versus all other groups (P 0.01). CONCLUSIONS: Deeper levels of hypothermia induced by the IC versus RT flush resulted in better neurological outcome in survivors. Surprisingly, deep hypothermia attenuated microglial activation but not hippocampal neuronal death. Minocycline had modest benefit on neurologic outcome in survivors but did not attenuate microglial activation in brain. Our findings suggest a novel effect of deep hypothermia on microglial proliferation during exsanguination CA.
(Anesth Analg 2009;109:914 23)
urrently, the outcomes from traumatic exsanguination cardiac arrest (CA) show that more than 50% of deaths caused by trauma occur at the scene13 where
From the *Safar Center for Resuscitation Research, University of Pittsburgh; Departments of Anesthesiology, Critical Care Medicine, Surgery, University of Pittsburgh School of Medicine; and Childrens Hospital of Pittsburgh, Pittsburgh, Pennsylvania. Accepted for publication April 16, 2009. Supported by a Starter Grant from the Society of Cardiovascular Anesthesiologists (to TD), the Laerdal Foundation for Acute Medicine (to TD), and a Seed Grant from the Department of Anesthesiology, University of Pittsburgh (to TD). Drs. Kochanek and Tisherman are co-patent holders with the University of Pittsburgh on Emergency Preservation and Resuscitation. Address correspondence and reprint requests to Tomas Drabek, MD, Safar Center for Resuscitation Research, 3434 Fifth Ave., Pittsburgh, PA 15260. Address e-mail to drabekt@anes.upmc.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b0511e
medical care is limited. Less than 10% of the patients who become pulseless from trauma survive.2 However, in an appropriate setting, some of those traumatic injuries could be surgically repairable.4 Emergency preservation and resuscitation (EPR) is a novel approach for resuscitation of exsanguination CA victims.5 EPR uses cold aortic flush to induce deep hypothermic preservation for prolonged CA to buy time for transport, damage control surgery, and delayed resuscitation with cardiopulmonary bypass (CPB). In prior studies of EPR, we used a dog model to maximize clinical relevance. Because of the lack of molecular tools available for use in dogs, we recently developed a rat EPR model to study the cellular and molecular mechanisms underlying deep hypothermic neuroprotection. Understanding cellular and molecular mechanisms of secondary damage in ischemiareperfusion injury after CA and the impact of deep
914
hypothermia on these cascades would allow us to define specific targets for future interventions, assess markers of reversibility, and screen novel therapies. Although the early brain injury in CA is initiated by energy failure and resultant neuronal death cascades, microglial activation has been suggested to be an additional mechanism of delayed neuronal death, most likely through releasing neurotoxic substances.6 Pharmacological modulation of microglial proliferation may help to improve outcome after CA. Recently, studies in several central nervous system insults have shown benefit from treatment with minocyline, a drug that attenuated microglial activation and proliferation.7 We hypothesized that deeper levels of intraarrest hypothermia would improve functional outcome, attenuate neuronal death, and attenuate microglial proliferation compared with more moderate hypothermia. We also hypothesized that minocycline would further augment the hypothermic protection via attenuating microglial activation, neuronal death, and improve outcome.
METHODS
We used the rat EPR model described in detail previously.8 All rats received humane care in compliance with the Guide for the Care and Use of Laboratory Animals (www.nap.edu/catalog/5140.html). The study protocol has been approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh. Adult male Sprague-Dawley rats (350 375 g) were obtained from Hilltop Lab Animals (Scottdale, PA) and housed for at least 3 days before the experiment under 12-h light/dark cycle with unrestricted access to food and water. On the day of the experiment, rats were anesthetized with 4% isoflurane in a transparent acrylic jar. After tracheal intubation with a 14-gauge IV catheter (Becton Dickinson, Sandy, UT), rats were mechanically ventilated using a piston ventilator (Harvard Ventilator Model 683; Harvard Rodent Apparatus, South Natick, MA) with a tidal volume of 0.8 mL/100 g and a frequency of 20 24/min to maintain normocapnia, and a positive end-expiratory pressure of 4 cm H2O. Anesthesia was maintained with 1.5%2% isoflurane in Fio2 0.5. After shaving and prepping with povidone iodine, bilateral femoral and right jugular cutdowns were performed. The left femoral artery and vein were cannulated for arterial blood pressure monitoring and blood sampling. Electrocardiogram, respiration, and arterial and central venous pressure were continuously monitored and recorded (Polygraph; Grass Instruments, Quincy, MA). The right femoral artery was cannulated with a 20-gauge catheter (Becton Dickinson, Sandy, UT) that served as an arterial CPB cannula. The right jugular vein was cannulated with a modified 5-hole 14-gauge IV cannula advanced to the right atrium to be used for venous drainage during the hemorrhage phase and later as a venous CPB cannula. Rectal and tympanic
probes were used to monitor the temperature. Baseline blood samples were obtained, and hemodynamic values were recorded. Removed blood volume was replaced with an electrolyte-balanced crystalloid Plasma-Lyte A (Baxter, Deerfield, IL) in a ratio 1:3. Heparin sodium was administered to achieve activated clotting time 400 s (Hemochron Jr. Signature, ITC, Edison, NJ). Three groups were studied: 1) ice-cold (IC) flush group (n 7), 2) room-temperature (RT) flush group (n 9), and 3) room-temperature flush group followed by minocycline treatment, 20 mg/kg 1 h after resuscitation and 90 mg/kg IP at 24 and 48 h (RT-M, n 11). Rats in the RT and IC groups received the same volume of vehicle (phosphate buffered saline [PBS]). After instrumentation, intubated rats were weaned to spontaneous ventilation of isoflurane 2% at Fio2 0.25 via a nose cone mask. After 5-min equilibration period, rapid exsanguination (12.5 mL of blood over 5 min) was performed via the internal jugular catheter. The shed blood was collected. After the rapid exsanguination phase, CA was ensured with IV administration of 9 mg of esmolol (0.9 mL) and 0.2 mEq of potassium chloride (0.1 mL). After 5 min of CA, 270 mL of either an RT or an IC flush solution (PlasmaLyte A) was instilled via the right femoral artery catheter at 50 mL/min. The flush was drained from the jugular vein catheter. After 20 min of CA, resuscitation was started with CPB. Heating and cooling were achieved with a circulating water bath around the oxygenator and a forced-air blower blowing air over the rat covered by a transparent semi-closed lid. Blood samples for biochemistry and hematology were obtained at 5, 15, 30, 45, and 60 min CPB time and processed immediately using a point-of-care blood analyzer (Stat Profile, Nova Biomedical, Waltham, MA). Arterial blood gas management followed -stat principles. pH and electrolyte values outside of the normal range were corrected during CPB and intensive care unit phases by adjustments in ventilation and/or administration of sodium bicarbonate, calcium chloride, and potassium chloride. Additional blood obtained from an isoflurane-anesthetized donor rat was used to maintain hematocrit 25%. CPB support was gradually weaned after 60 min. Mechanical ventilation with a Fio2 of 1.0 was continued while maintaining normocapnia for additional 2 h. Using a midline laparotomy incision, a Mini-mitter probe (Mini-Mitter, Sunriver, OR) was introduced into the peritoneal cavity to allow postoperative temperature control and continuous monitoring of heart rate and movement. Surviving rats were tracheally extubated 2 h later after removal of catheters and placed separately in a temperature-controlled cage (34.5C for 4 h) with supplemental oxygen for 18 h and free access to food and water. Weight and neurologic status were assessed daily, using Overall Performance Category (OPC; 1 normal, 2 mild disability, 3 moderate
915
disability, 4 severe disability, and 5 death or brain death) and a modified Neurologic Deficit Score9 (NDS; 0%10% normal, 100% maximum deficit). Postoperatively, the rats that did not resume normal eating and drinking habits received supplemental subcutaneous injections of 0.45NS/D5W (10 mL twice daily). At 72 h after resuscitation, blood samples were obtained, and the rats were euthanized with an isoflurane overdose and perfused via the left ventricle with normal saline followed by 10% neutral buffered formalin.
RESULTS
Baseline body weight before the experiment was similar in all groups (RT, 387 3 g; IC, 385 1 g, RT-M, 386 10 g; P 0.05). The surgical time did not differ among groups (RT, 98 7 min; IC, 99 19 min; RT-M 99 12 min, P 0.05). After cooling, rats in the IC group had significantly lower temperature during CA versus other groups (tympanic, 21C vs 28C; rectal, 20C-25C vs 27C30C, P 0.001) (Fig. 1). Rats in the IC group had higher mean arterial blood pressure during flush (35 5 vs 27 6 mm Hg, P 0.05). Retained volume, i.e., the amount of flush that was not drained from the jugular catheter, was also lower in the IC group versus RT and RT-M groups combined (29 7 vs 42 4 mL, P 0.001). Heart rate increased more slowly during resuscitation in the IC versus RT group (P 0.01) (Fig. 2, Panel A). Mean arterial pressure was higher in the IC group versus RT group over time (P 0.05; Fig. 2, Panel B). After discontinuation of temperature control, body core temperature increased more rapidly in the IC group. This steady increase reached statistical significance versus the RT group but not the RT-M group (P 0.05 IC versus RT group, Fig. 3). However, all groups achieved normothermia at 24 h after resuscitation. While pH and base excess were similar at 5 min after the start of resuscitation, lactate was lower in the IC group versus other groups (P 0.05) (Table 1). The survival rate was not different among groups. Neurological outcome was significantly better in the IC versus other groups (P 0.05) (Table 2, Fig. 4). The markers of microglial activation but not neuronal death were attenuated in the IC versus other groups (Figs. 57). Selective vulnerability of CA1 neurons and hilar neurons in the dentate gyrus was observed (Figs. 6 and 7), with proliferated microglial cells with ameboid-shaped cell bodies and shortened, retracted processes. Microglial activation was attenuated in the IC group (Figs. 6 and 7, Panel E). There was a positive correlation between OPC and neuronal death (r 0.566, P 0.018). Similarly, there was a correlation between F-JC and Iba-1 (r 0.513, P 0.035). In contrast, there was only a trend for a correlation between Iba-1 and neurologic outcome (Iba-1 versus OPC, r 0.345, P 0.176; Iba-1 versus NDS, r 0.393, P 0.119). We performed a formal necropsy on all rats that died before completion of the study. However, we were not able to determine the cause of death in all rats. The common findings were pulmonary edema and/or pulmonary hemorrhage. The time of death was between 15 and 45 h of resuscitation in all the groups (RT, 29.5 14.1 h; IC, 13.1 h; RT-M, 24.0 7.5 h).
Histology
The tissue samples were processed for embedding in paraffin. The resulting paraffin blocks were sequentially sectioned at 5 m. All sections were stained with Fluoro-Jade C (F-JC).10 For the Iba-1 staining, sections were washed in PBS, incubated in 0.3% H2O2 in methanol for 30 min to inhibit endogenous peroxidase activity, washed in PBS, and blocked in PBS containing 1.5% normal goat serum and 1% bovine serum albumin for 2 h at RT. The sections were then incubated with a rabbit anti-Iba1 polyclonal antibody (1:500, Serotec) overnight at 4C, washed in PBS, and incubated with a fluorescein isothiocyanateconjugated goat antirabbit IgG antibody (Invitrogen) for 2 h at RT. For control staining, normal rabbit IgG was used as the primary antibody. After the reaction, the sections were counterstained with 4,6-diamidino-2phenylindole, dehydrated in ethanol steps, and mounted. Adjacent sections obtained at approximately 4.3 mm from bregma were used for assessing neuronal death and microglial activation within the selective brain regions. A photograph of representative sections of dentate gyrus and CA1 region was taken under 10 magnification. F-JC positive neurons and Iba-1positive activated microglia (characterized by ameboid cell body and retracted processes without thin ramifications)11 were then counted using National Institutes of Health Image-J software by an observer masked to the treatment group.
Repeated measures analysis of variance was performed, followed by StudentNewmanKeuls post hoc tests, to identify differences in hemodynamic and arterial blood gas parameters and temperature among groups. One-way analysis of variance was used to compare histologic damage among groups. The 2 test was used to test the differences in proportions of OPC among groups. KruskalWallis H test was used to compare NDS among groups. MannWhitney U-test was used to compare two groups if KruskalWallis H test indicated there were differences between groups. Pearson and Spearman correlations between variables were determined as appropriate. A P value 0.05 was considered statistically significant.
916
Hypothermia Attenuates Microglia
DISCUSSION
Exsanguination CA is a relatively unexplored form of CA. Resuscitation of exsanguination CA victims with conventional cardiopulmonary resuscitation technique has a poor prognosis because of a volume-depleted and
Figure 1. Tympanic and rectal temperatures during cardiac arrest (CA). P 0.001 ice-cold (IC) flush group versus other groups.
trauma-disrupted circulatory system. In the civilian setting, 50% of deaths caused by trauma occur at the scene and another 30% within hours from injury.12 More aggressive treatments with thoracotomy and aortic cross-clamping have also not improved the poor outcome in these patients.2 The surgeon cannot obtain hemostasis and resuscitation before vital organs (particularly brain and heart) have suffered irreversible ischemic damage. However, in an appropriate setting, many of those injuries would be technically repairable. Traditionally, deep hypothermic circulatory arrest (DHCA) has been used in cardiac surgery to provide a bloodless field and enable repair of the congenital cardiac malformations or acquired pathologies with considerable success. Hypothermia for deep hypothermic circulatory arrest (DHCA) is used in a protective rather than a therapeutic fashion; the use of hypothermia in EPR is a much more challenging
situation. In resuscitation of exsanguination CA, including EPR, the rapid onset of cooling can only be initiated after a period of normothermic CA. Adjuncts to hypothermia would thus be of great potential benefit. Recently, we modified the rat EPR model to produce a screening tool to study mechanisms of neuronal death and evaluate novel therapeutic adjuncts to hypothermia. Also it should be recognized that the use of a normothermic control group is not feasible because the rats would not survive the insult if maintained normothermic throughout the period of emergency preservation. In our paradigm, we use 5 min of hemorrhagic shock followed by ice-cold or room-temperature flush initiated 5 min after CA. This is a clinically relevant delay that would allow cannulation of a large vessel.13 Flushing with either IC or RT saline resulted in a brain temperature of 21C or 28C, respectively. Better protection achieved after IC flush was likely reflected by
917
Figure 2. Heart rate (top panel) and

mean arterial blood pressure (bottom panel). a P 0.05 room-temperature (RT) flush versus room-temperature flush followed by minocycline (RT-M) group; b P 0.05 ice-cold (IC) versus RT and RT-M groups; c P 0.05 RT versus IC group. BL baseline; HS end of hemorrhagic shock; CA cardiac arrest; CPB cardiopulmonary bypass; ICU intensive care.
lower lactate levels at 5 min after reperfusion and better neurologic function in survivors at 72 h. The latter was not affected by minocycline treatment. Despite the functional benefit with IC flush, neuronal death seen in traditionally selectively vulnerable brain regions did not differ among groups. We also noted robust microglial activation surrounding the dying neurons. Surprisingly, deep hypothermia (21C) was able to attenuate microglial activation but not neuronal damage. It is possible that hypothermia-induced attenuation of microglial activation contributed to the improved neurologic outcome in the IC group. After discontinuation of the postoperative hypothermia, rats in the IC group spontaneously rewarmed more quickly than rats in other groups. The attenuation of microglial activation in the IC group thus could not be explained by unintentional prolonged postoperative hypothermia. While the early brain injury in CA is believed to result from release of excitatory mediators, energy
918
failure, oxidative stress, damage to mitochondria and endoplasmic reticulum, and cell signaling pathway disturbances in neurons, secondary damage could also be triggered by microglia, that transform into phagocytes. Microglial activation starts immediately after ischemia and thus precedes the morphologically detectable neuronal damage. Microglial activation has been suggested to contribute to delayed neuronal death, most likely through releasing neurotoxic substances, including reactive oxygen radicals, nitric oxide, and proinflammatory cytokines.6 Microglial activation could contribute to neuronal death or microglial-mediated synaptic injury and/or neuronal dysfunction, which could mediate cognitive deficits even in the absence of overt neuronal death. Additional studies focused on these secondary injury mechanisms in our model are warranted. Microglia could also have a protective role,14 18 possibly in delayed repair after injury via elaboration of growth factors. Thus, there may be a specific time window for benefit from inhibition of
Figure 3. Postoperative body core

temperature within the first 24 h. P 0.05 room-temperature (RT) versus ice-cold (IC) group.
Table 1. Biochemical and Hematological Values After 20 min Cardiac Arrest Treated by Emergency Preservation and Resuscitation BL1
pHa RT IC RT-M Pao2 RT IC RT-M Paco2 RT IC RT-M BE RT IC RT-M Lactate RT IC RT-M Hct RT IC RT-M Glucose RT IC RT-M 7.37 0.04 7.40 0.04 7.39 0.04 246 29 328 111 262 83 47 6 42 6 44 3 12 1.0 2.8 1.1 2.2 1.1 0.9 1.1 0.6 1.2 1.1 38 3 37 3 37 3 227 51 208 42 228 18
BL2
7.45 0.03* 7.40 0.05 7.36 0.06 119 17 271 52 244 85 30 4 37 6 43 6 1.9 2.4 0.7 2.1 0.6 1.8 1.6 0.6 1.7 0.8 1.8 1.0 32 3 34 2 34 1 224 62 232 49 241 54
CPB5
6.99 0.05 7.00 0.08 6.95 0.07 454 32 536 49 474 33 34 3 37 9 38 4 21 1.4 20.7 2.0 21.8 1.5 7.0 0.9 5.3 1.0 6.9 0.8 25 2 26 4 26 1 210 20 213 27 223 35
CPB60
7.41 0.05 7.36 0.12 7.42 0.03 386 31 382 51 374 33 42 7 40 9 44 7 2.7 3.2 0.9 2.4 2.2 2.6 6.7 1.2 5.7 2.0 6.5 1.6 29 3 29 3 30 2 225 52 255 75 248 84
ICU120
7.40 0.04 7.45 0.04 7.40 0.04 273 97 351 126 349 154 48 7 44 1 48 7 4.9 2.7 6.9 2.2 6.3 3.3 2.5 0.7 4.2 2.6 3.1 2.1 30 3 32 4 32 2 152 29 177 51 175 64
FINAL
7.46 0.06 7.49 0.06 7.48 0.03 497 96 379 138 529 51 33 6 33 8 27 3 0.6 3.0 0.6 1.6 1.1 3.0 2.5 1.4 3.6 1.4 4.3 1.5 28 3 28 5 26 3 149 33 207 38 132 33
IC ice-cold ush group; RT room-temperature ush group; RT-M RT ush followed by minocycline treatment group; BL baseline; CPB cardiopulmonary bypass; CPB5 5 min after start of CPB; CPB60 at the end of CPB; ICU intensive care unit; ICU120 2 h after CPB; FINAL at 72 h; pHa arterial pH; BE base excess; Hct hematocrit. * P 0.05 RT versus RT-M group. P 0.05 IC versus RT and RT-M groups.
the early microglial contribution to damage. Recent studies also suggested that the severity of neuronal injury determines microglial release of toxic versus protective effectors.15 To visualize microglia, we chose to use anti-Iba-1 staining. Iba-1 is a calcium-binding protein expressed specifically in activated microglia,19 with its peak occurring at 4 7 days after injury.20 While resident microglia exist in a ramified state, after brain
injury they migrate toward the lesion, their cell body becomes ameboid-shaped, the processes shorten, and become virtually indistinguishable from macrophages. Minocycline is a widely used antibiotic with antiinflammatory and antiapoptotic properties which has been tested in several models of neurologic injury, including global2123 and focal brain ischemia,24 27 traumatic brain injury,28,29 spinal cord injury,30,31 and
919
intracerebral hemorrhage.32 Most recently, minocycline showed favorable results in a clinical trial in acute stroke patients.33 It penetrates the blood-brain barrier,34 reduces tissue injury, and improves functional recovery.21,35,36 The primary effect of minocycline is probably inhibition of activation of microglia.21,22,24,30,37 Surprisingly, minocycline was also reported to be more protective than brief hypothermia after focal cerebral ischemia.26,27 Specifically, inhibition of p38 mitogen-activated protein kinase activation in microglia has been suggested as a key mechanism underlying minocycline antiinflammatory effects, although other mechanisms may also be involved. In preliminary studies, we did not observe a beneficial effect with a lower dose of minocycline (3 mg/kg IV followed by 45 mg/kg IP; data not shown). Thus, we chose to use the high-dose minocycline (20 mg/kg IV followed by 90 mg/kg IP), which was previously used by others in similar settings. Table 2. Overall Performance Categories (OPC) After 20-min Cardiac Arrest Treated by Emergency Preservation and Resuscitation No. of ratsa OPC
5 4 3 2 1 Death Severe disability Moderate disability Mild disability Normal
RT

IC

RT-M

a Each dot represents one rat. No differences among groups in survival rate (P 0.05). Favorable neurological outcome (assessed by OPC) was signicantly better in the IC group versus other groups (P 0.05). IC ice-cold ush group; RT room-temperature ush group; RT-M RT ush followed by minocycline treatment group.
In our study, hypothermia attenuated microglial activation. Temperatures used in our study (21C28C) were generally lower than those used in other studies of mild-to-moderate hypothermia. Postischemic hypothermia (32C for 24 h) suppressed microglial activation after hypoxic-ischemic injury in the developing brain.38 Even a brief period of hypothermia (33C for 2 h) attenuated neuroinflammation after experimental stroke and brain inflammation induced by IV injection of lipopolysaccharide.39 A similar effect of hypothermia was observed in microglial cell cultures stimulated by lipopolysaccharide.40 Although many studies used minocycline as a drug suppressing microglial activation, we did not see any effect of minocycline on microglia activation or neuronal death. This striking lack of effect could be potentially explained by the fact that minocycline was added to augment the protective effects of preexisting moderate hypothermia (28C). It is possible that minocycline could not add further benefit to hypothermia. Moderate hypothermia in the 28C group was limited to the intraischemic time, followed by mild hypothermia for 6 h. Previous studies suggested that the onset of microglial activation starts at 24 h and peaks at 4 7 days.20 In our study, we administered minocycline up to 72 h. However, we cannot rule out that hypothermia delayed or modified the course of microglial activation, and therefore the dosing regimen or assessment time were not optimal. The lack of effect of minocycline in our EPR paradigm is not entirely surprising. Previously, we tested 14 pharmacological adjuncts to hypothermia. Using our similar moderate hypothermia canine model with 20-min CA, only the antioxidant tempol showed some benefit.41
Figure 4. Neurologic Deficit Score after 20-min cardiac arrest treated by emergency preservation and resuscitation. Boxes represent interquartile ranges. The line across each box indicates the median, and the whiskers are the highest and lowest values. *P 0.05, roomtemperature (RT) versus ice-cold (IC) group.
920
Recently, there has been increasing evidence suggesting the neuroprotective role of microglia in central nervous system pathologies. Selective ablation of microglial cells before cerebral ischemia in vivo revealed a marked neuroprotective potential of proliferating
Figure 5. Neuronal death and microglial activation after 20-min cardiac arrest (CA) treated by emergency preservation and resuscitation (EPR) with either room-temperature (RT) flush, ice-cold (IC) flush, or room-temperature flush followed by minocycline treatment (RT-M) in the dentate gyrus region of hippocampus. The line across each box indicates the median, and the whiskers are the highest and lowest values. The round marker and the asterisk represent outliers of the respective groups.
microglia, serving as an endogenous pool of neurotrophic molecules such as IGF-1.16 Microglia cells were also shown to protect neurons by direct engulfment of invading neutrophil granulocytes that infiltrate ischemic lesions in an in vitro model.18 Despite robust microglial proliferation, we have not, however, observed neutrophil accumulation in our model. In our study, NDS assessments were not very tightly coupled to hippocampal cell loss. Rats in the IC group that achieved favorable OPC and NDS scores still had substantial neuronal injury. Advanced neurobehavioral testing will be needed in a future study to define the association between hippocampal neuronal loss, microglial activation, and neurocognitive outcome. Our exploratory study was focused on the histological markers of injury. Previously we have shown that motor deficits observed in this complex model persist up to Day 7.42 This would require delaying the period of water maze tests until after Day 7. The time of completion of water maze tasks would then occur outside of the peak microglial activity. We have observed a significant correlation between neuronal death and neurological outcome. A larger number of animals would be necessary to appropriately test the hypothesis that microglial activation contributes to neurologic deficits. We cannot exclude that injuries in other brain regions or extracerebral injuries played a role and influenced the neurologic outcome. In conclusion, deeper levels of hypothermia compared with moderate hypothermia (21C vs 28C) induced by aortic flush resulted in better neurologic outcome in survivors. Surprisingly, hypothermia attenuated microglial activation but not hippocampal neuronal death. Minocycline did not improve either neurologic outcome or attenuate microglial activation in brain. Our preliminary findings suggest a potentially novel effect of
Figure 6. Neuronal death and microglial proliferation after exsanguination cardiac arrest and emergency preservation and resuscitation with either room-temperature (RT) or ice-cold (IC) flush in the CA1 region of hippocampus. Blue staining is 4,6-diamidino-2-phenylindole, identifying neurons, and green staining is Fluoro-Jade C, Panels AC, identifying dying neurons, or anti-Iba-1 staining visualizing microglia (Panels DF). Microglial activation is attenuated in the IC group. Representative samples from each group are shown. A, 10: Hippocampal neuronal loss in a rat from the RT group. Full CA1 loss. B, 10: CA1 region in a rat from the IC group. Intensive neuropil staining between CA1 and dentate gyrus (DG). C, 10: Hippocampal neuronal loss in a rat from RT-M group. D, 10: Microglial activation in CA1-CA2 regions of hippocampus in a rat from the RT group. E, 10: Microglial activation was attenuated in a rat from the IC group. F, 10: Microglial activation was marked in a rat from RT-M group despite high dose minocycline treatment. Scale bar in Panel A 80 m.
921
Figure 7. Neuronal death and microglial activation after exsanguination cardiac arrest and emergency preservation and
resuscitation with either room-temperature (RT) or ice-cold (IC) flush in the dentate gyrus (DG) region of hippocampus. Blue staining is 4,6-diamidino-2-phenylindole, identifying neurons, and green staining is Fluoro-Jade C, identifying dying neurons (Panels AC), or anti-Iba-1 staining visualizing microglia (Panels DF). Microglial activation (Panels DF) is attenuated in the IC group (Panel E). Representative samples from each group are shown. Hippocampal neuronal loss in DG. Hilar neurons are selectively injured in all groups (A, 10, RT group; B, 10, IC group; C, 10, RT-M group, respectively). D, 10: Microglial activation in hippocampal DG in a rat from the RT group. E, 10: Microglial activation was attenuated in a rat from the IC group. F, 10: Microglial activation was marked in a rat from RT-M group despite high dose minocycline treatment. Scale bar in Panel A 80 m.
hypothermia on microglial activation during deep hypothermia. Further studies with comprehensive neurobehavioral testing will be needed to further elucidate the role of microglia on functional outcome. REFERENCES
1. Bellamy R, Safar P, Tisherman SA, Basford R, Bruttig SP, Capone A, Dubick MA, Ernster L, Hattler BG Jr, Hochachka P, Klain M, Kochanek PM, Kofke WA, Lancaster JR, McGowan FX Jr, Oeltgen PR, Severinghaus JW, Taylor MJ, Zar H. Suspended animation for delayed resuscitation. Crit Care Med 1996;24: S24 47 2. Rhee PM, Acosta J, Bridgeman A, Wang D, Jordan M, Rich N. Survival after emergency department thoracotomy: review of published data from the past 25 years. J Am Coll Surg 2000; 190:288 98 3. Acosta JA, Yang JC, Winchell RJ, Simons RK, Fortlage DA, Hollingsworth-Fridlund P, Hoyt DB. Lethal injuries and time to death in a level I trauma center. J Am Coll Surg 1998;186:528 33 4. Shoemaker WC, Peitzman AB, Bellamy R, Bellomo R, Bruttig SP, Capone A, Dubick M, Kramer GC, McKenzie JE, Pepe PE, Safar P, Schlichtig R, Severinghaus JW, Tisherman SA, Wiklund L. Resuscitation from severe hemorrhage. Crit Care Med 1996;24:S1223 5. Safar P, Tisherman SA, Behringer W, Capone A, Prueckner S, Radovsky A, Stezoski WS, Woods RJ. Suspended animation for delayed resuscitation from prolonged cardiac arrest that is unresuscitable by standard cardiopulmonary-cerebral resuscitation. Crit Care Med 2000;28:N214 8 6. Gehrmann J, Banati RB, Wiessner C, Hossmann KA, Kreutzberg GW. Reactive microglia in cerebral ischaemia: an early mediator of tissue damage? Neuropathol Appl Neurobiol 1995;21:277 89 7. Stirling DP, Koochesfahani KM, Steeves JD, Tetzlaff W. Minocycline as a neuroprotective agent. Neuroscientist 2005;11: 308 22 8. Drabek T, Stezoski J, Garman RH, Han F, Henchir J, Tisherman SA, Stezoski SW, Kochanek PM. Exsanguination cardiac arrest in rats treated by 60 min, but not 75 min, emergency preservation and delayed resuscitation is associated with intact outcome. Resuscitation 2007;75:114 23 9. Neumar RW, Bircher NG, Sim KM, Xiao F, Zadach KS, Radovsky A, Katz L, Ebmeyer E, Safar P. Epinephrine and sodium bicarbonate during CPR following asphyxial cardiac arrest in rats. Resuscitation 1995;29:249 63
10. Schmued LC, Hopkins KJ. Fluoro-Jade B: a high affinity fluorescent marker for the localization of neuronal degeneration. Brain Res 2000;874:12330 11. Koshinaga M, Suma T, Fukushima M, Tsuboi I, Aizawa S, Katayama Y. Rapid microglial activation induced by traumatic brain injury is independent of blood brain barrier disruption. Histol Histopathol 2007;22:129 35 12. Trunkey D. Initial treatment of patients with extensive trauma. N Engl J Med 1991;324:1259 63 13. Bregman D, Nichols AB, Weiss MB, Powers ER, Martin EC, Casarella WJ. Percutaneous intraaortic balloon insertion. Am J Cardiol 1980;46:261 4 14. Neumann J, Gunzer M, Gutzeit HO, Ullrich O, Reymann KG, Dinkel K. Microglia provide neuroprotection after ischemia. FASEB J 2006;20:714 6 15. Lai AY, Todd KG. Differential regulation of trophic and proinflammatory microglial effectors is dependent on severity of neuronal injury. Glia 2008;56:259 70 16. Lalancette-Hebert M, Gowing G, Simard A, Weng YC, Kriz J. Selective ablation of proliferating microglial cells exacerbates ischemic injury in the brain. J Neurosci 2007;27:2596 605 17. Hayashi Y, Tomimatsu Y, Suzuki H, Yamada J, Wu Z, Yao H, Kagamiishi Y, Tateishi N, Sawada M, Nakanishi H. The intraarterial injection of microglia protects hippocampal CA1 neurons against global ischemia-induced functional deficits in rats. Neuroscience 2006;142:8796 18. Neumann J, Sauerzweig S, Ronicke R, Gunzer F, Dinkel K, Ullrich O, Gunzer M, Reymann KG. Microglia cells protect neurons by direct engulfment of invading neutrophil granulocytes: a new mechanism of CNS immune privilege. J Neurosci 2008;28:596575 19. Ito D, Imai Y, Ohsawa K, Nakajima K, Fukuuchi Y, Kohsaka S. Microglia-specific localisation of a novel calcium binding protein, Iba1. Brain Res Mol Brain Res 1998;57:19 20. Ito D, Tanaka K, Suzuki S, Dembo T, Fukuuchi Y. Enhanced expression of Iba1, ionized calcium-binding adapter molecule 1, after transient focal cerebral ischemia in rat brain. Stroke 2001; 32:1208 15 21. Fan LW, Lin S, Pang Y, Rhodes PG, Cai Z. Minocycline attenuates hypoxia-ischemia-induced neurological dysfunction and brain injury in the juvenile rat. Eur J Neurosci 2006;24:34150 22. Yrjanheikki J, Keinanen R, Pellikka M, Hokfelt T, Koistinaho J. Tetracyclines inhibit microglial activation and are neuroprotective in global brain ischemia. Proc Natl Acad Sci USA 1998;95:15769 74
922
23. Arvin KL, Han BH, Du Y, Lin SZ, Paul SM, Holtzman DM. Minocycline markedly protects the neonatal brain against hypoxic-ischemic injury. Ann Neurol 2002;52:54 61 24. Yrjanheikki J, Tikka T, Keinanen R, Goldsteins G, Chan PH, Koistinaho J. A tetracycline derivative, minocycline, reduces inflammation and protects against focal cerebral ischemia with a wide therapeutic window. Proc Natl Acad Sci USA 1999; 96:13496 500 25. Liu Z, Fan Y, Won SJ, Neumann M, Hu D, Zhou L, Weinstein PR, Liu J. Chronic treatment with minocycline preserves adult new neurons and reduces functional impairment after focal cerebral ischemia. Stroke 2007;38:146 52 26. Wang CX, Yang T, Shuaib A. Effects of minocycline alone and in combination with mild hypothermia in embolic stroke. Brain Res 2003;963:3279 27. Wang CX, Yang T, Noor R, Shuaib A. Delayed minocycline but not delayed mild hypothermia protects against embolic stroke. BMC Neurol 2002;2:2 28. Sanchez Mejia RO, Ona VO, Li M, Friedlander RM. Minocycline reduces traumatic brain injury-mediated caspase-1 activation, tissue damage, and neurological dysfunction. Neurosurgery 2001;48:13939; discussion 1399 401 29. Bye N, Habgood MD, Callaway JK, Malakooti N, Potter A, Kossmann T, Morganti-Kossmann MC. Transient neuroprotection by minocycline following traumatic brain injury is associated with attenuated microglial activation but no changes in cell apoptosis or neutrophil infiltration. Exp Neurol 2007,204:220 33 30. Stirling DP, Khodarahmi K, Liu J, McPhail LT, McBride CB, Steeves JD, Ramer MS, Tetzlaff W. Minocycline treatment reduces delayed oligodendrocyte death, attenuates axonal dieback, and improves functional outcome after spinal cord injury. J Neurosci 2004;24:218290 31. Festoff BW, Ameenuddin S, Arnold PM, Wong A, Santacruz KS, Citron BA. Minocycline neuroprotects, reduces microgliosis, and inhibits caspase protease expression early after spinal cord injury. J Neurochem 2006;97:1314 26 32. Power C, Henry S, Del Bigio MR, Larsen PH, Corbett D, Imai Y, Yong VW, Peeling J. Intracerebral hemorrhage induces macrophage activation and matrix metalloproteinases. Ann Neurol 2003;53:731 42
33. Lampl Y, Boaz M, Gilad R, Lorberboym M, Dabby R, Rapoport A, Anca-Hershkowitz M, Sadeh M. Minocycline treatment in acute stroke: an open-label, evaluator-blinded study. Neurology 2007;69:1404 10 34. Saivin S, Houin G. Clinical pharmacokinetics of doxycycline and minocycline. Clin Pharmacokinet 1988;15:355 66 35. Hewlett KA, Corbett D. Delayed minocycline treatment reduces long-term functional deficits and histological injury in a rodent model of focal ischemia. Neuroscience 2006;141:2733 36. Yenari MA, Xu L, Tang XN, Qiao Y, Giffard RG. Microglia potentiate damage to blood-brain barrier constituents: improvement by minocycline in vivo and in vitro. Stroke 2006;37: 108793 37. Fan LW, Pang Y, Lin S, Rhodes PG, Cai Z. Minocycline attenuates lipopolysaccharide-induced white matter injury in the neonatal rat brain. Neuroscience 2005;133:159 68 38. Fukui O, Kinugasa Y, Fukuda A, Fukuda H, Tskitishvili E, Hayashi S, Song M, Kanagawa T, Hosono T, Shimoya K, Murata Y. Post-ischemic hypothermia reduced IL-18 expression and suppressed microglial activation in the immature brain. Brain Res 2006;1121:35 45 39. Deng H, Han HS, Cheng D, Sun GH, Yenari MA. Mild hypothermia inhibits inflammation after experimental stroke and brain inflammation. Stroke 2003;34:2495501 40. Maekawa S, Aibiki M, Si QS, Nakamura Y, Shirakawa Y, Kataoka K. Differential effects of lowering culture temperature on mediator release from lipopolysaccharide-stimulated neonatal rat microglia. Crit Care Med 2002;30:2700 4 41. Behringer W, Safar P, Kentner R, Wu X, Kagan VE, Radovsky A, Clark RS, Kochanek PM, Subramanian M, Tyurin VA, Tyurina YY, Tisherman SA. Antioxidant Tempol enhances hypothermic cerebral preservation during prolonged cardiac arrest in dogs. J Cereb Blood Flow Metab 2002;22:10517 42. Drabek T, Fisk JA, Dixon CE, Garman RH, Stezoski J, Wisnewski SR, Wu X, Tisherman SA, Kochanek PM. Prolonged deep hypothermic circulatory arrest in rats can be achieved without cognitive deficits. Life Sci 2007;81:54352
923
General Article
The Safety of Modern Hydroxyethyl Starch in Living Donor Liver Transplantation: A Comparison with Human Albumin
Ahmed Mukhtar, MD* Fawzia Aboulfetouh, MD* Gihan Obayah, MD* Maged Salah, MD* Mohamed Emam, MD* Yehia Khater, MD* Ramzia Akram, MD Aly Hoballah, MD Mohamed Bahaa, MD Mahmoud Elmeteini, MD Alaa Hamza, MD
BACKGROUND: Intravascular volume replacement therapy is an important issue in the perioperative management of liver transplantation. There is paucity of data on the safety of hydroxyethyl starch (HES) in patients undergoing liver transplantation. We evaluated the safety of a new HES 130/0.4 in the perioperative management of liver transplantation, with a special emphasis on renal function. METHODS: Forty patients undergoing living donor liver transplantation were prospectively randomized into two groups. Patients in the ALB group (n 20) received 5% human albumin. Patients in the HES group (n 20) received third generation HES (6% HES 130/0.4). Total colloid administration was limited to 50 mL kg1 d1. The volume was given to maintain pulmonary artery occlusion pressure or central venous pressure between 5 and 7 mm Hg. If additional fluids were required, balanced crystalloid solution was used. Anesthetic and surgical techniques were standardized. Serum creatinine and cystatin C plasma levels were measured from arterial blood samples after induction of anesthesia, at the end of surgery, and on the first 4 postoperative days. RESULTS: All 40 enrolled patients completed the study. Demographic and intraoperative variables were comparable in both groups. Postoperatively, the mean sd volume was 6229 1140 mL and 4636 1153 mL in HES and ALB groups, respectively (P 0.003). There was significantly larger net cumulative fluid balance in the ALB group 1100 900 mL compared with the HES group 3047 2000 mL, P 0.029. Serum creatinine, creatinine clearance, and cystatin C plasma levels showed no significant differences between the two groups. One patient in each group developed acute renal failure requiring renal replacement therapy. CONCLUSION: The use of HES 130/0.4 as an alternative to human albumin resulted in equivalent renal outcome after liver transplantation.
(Anesth Analg 2009;109:924 30)
ypovolemia is frequently encountered during liver transplantation.1 Fluid management remains a controversial subject in perioperative medicine and organ transplantation. Colloids may be preferred to crystalloids to maintain effective cardiac output and tissue oxygenation.2 The ideal colloid therapy has not been studied in patients undergoing liver transplant. Albumin is a naturally occurring colloid and the mainstay of therapy in many centers3 because its oncotic, antiinflammatory, and antioxidant effects have been demonstrated in animal and experimental studies.4 Hydroxyethyl starch (HES) is a widely used, inexpensive alternative to human albumin for correcting hypovolemia. HES solutions are polydisperse, comprising a distribution of molecular sizes. The polyglucose chains
From the *Department of Anesthesia and Intensive Care, Cairo University, Giza; Wadi-Alneel Liver Transplant Center, Cairo, Egypt; and Department of Surgery, Ain Shams University, Cairo, Egypt. Accepted for publication April 13, 2009. Supported by Wadi-Alneel Liver Transplant Center. Address correspondence and reprint requests to Ahmed M. Mukhtar, 2 Zaafran St., Cairo, Egypt. Address e-mail to ahmed3m2003@ yahoo.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aed54f
closely resemble glycogen with predominant 1 4 bindings. The merits of HES types reside in its mean molecular weight, molar substitution, and C2/C6 ratio. Its rate of metabolism and the in vivo molecular weight are primarily determined by molar substitution and C2/C6 ratio.5 Renal dysfunction is one of the most common complications in the postoperative course after liver transplantation, with an incidence of 12%70%.6 The need for renal replacement therapy is associated with 40%90% mortality.7 Potential risk factors for renal impairment in the perioperative phase of liver transplantation include preexisting comorbidities, ischemic, or toxic insults to the kidneys during surgery as a result of hemodynamic instability and hypovolemia or repeated use of nephrotoxic drugs.8 Multiple authors have expressed concern about HES-associated renal dysfunction.9,10 Although some studies have found HES solutions to have little effect on renal function,11,12 others have shown cause for concern.13,14 However, neither patients with end-stage liver disease nor those with hypoalbuminemia were included in these studies. Based on these data, renal function was the primary end point in this study. Our hypothesis was that
924
human albumin would preserve perioperative renal function better than HES.
METHODS
After approval of the local Ethics Committee and obtaining written informed consent, we studied 40 patients with end-stage liver disease, scheduled for living donor liver transplantation. We excluded patients undergoing retransplantation, patients with a history of previous upper abdominal surgery, patients with portal vein thrombosis (diagnosed with preoperative duplex ultrasound), patients younger than 18-yr-old, and patients with primary renal dysfunction diagnosed by examination of urinary sediment and urinary chemistry, as well as appropriate ultrasonographic and radiological investigations.15 Patients with hepatorenal disease were not excluded. Patients were randomly allocated to receive either 6% HES 130/0.4 (Voluven, Fresenius Kabi, Bad Hombourg, Germany) (HES group: n 20) or albumin 5% (ALB group: n 20) as their colloid during the intraoperative period and first 4 postoperative days, with a maximum dosage of 50 mL kg1 d1. Randomization was done by the attending physician opening a sealed envelope. Anesthesia was induced with IV propofol, fentanyl, and atracurium. Anesthesia was maintained with sevoflurane adjusted between 1% and 2% in an oxygen/air mixture, a fentanyl infusion at 12 g kg1 h1, and an atracurium infusion at 0.5 mg kg1 h1. Mechanical ventilation was provided by a Primus anesthesia machine (Dra ger, Germany) using a tidal volume of 8 10 mL/kg with the respiratory rate adjusted to maintain Paco2 between 30 and 35 mm Hg. All patients were monitored for temperature, noninvasive and invasive arterial blood pressure, 5-lead electrocardiogram, peripheral oxygen saturation, end-tidal carbon dioxide tension, hourly urinary output, central venous pressure (CVP), and pulmonary artery occlusion pressure (PAOP). A pulmonary artery catheter (OPTIQ SVO2/CCO Abbott Laboratories, North Chicago, IL) was inserted into the right internal jugular vein. In both groups, Ringer acetate solution was administered routinely at 10 mL kg1 h1. The patients received a 250 mL bolus of either Voluven (HES group) or albumin 5% (ALB group) with a maximum administration of 50 mL kg1 d1 to maintain CVP between 5 and 7 mm Hg. Ringer acetate solution was given if additional fluid replacement was required. Blood transfusion was given based on a hemoglobin level (7 g/dL). Norepinephrine was administered if the mean arterial blood pressure was 70 mm Hg and if the systemic vascular resistance was 600 dyne s1 cm5. Epinephrine was administered if the mean arterial blood pressure was 70 mm Hg and the cardiac index was 2.5 L min1 m2, despite sufficient volume infusion to maintain a target cardiac index of 2.53.0 L min1 m2.
We typically tracheally extubate transplant recipients 6 8 h after admission to the intensive care unit (ICU).16 We administered propofol by continuous infusion at 0.51 mg kg1 h1 until patients were normothermic and hemodynamically stable. The criteria for extubation included normal neurologic status, stable hemodynamics requiring minimal vasopressors support (e.g., 0.1 g kg1 min1 norepinephrine), and adequate pulmonary function (e.g., spontaneous ventilation for a minimum of 30 min with respiratory rate between 10 and 20/min, Paco2 50 mm Hg, pH 7.30, and Pao2 75 mm Hg on a Fio2 40%). Postoperatively, intravascular volume replacement (albumin 5% or HES 130/0.4) was given to maintain the CVP and/or PAOP between 5 and 7 mm Hg. Packed red blood cells were given if the hemoglobin was 8 g/dL. Fresh frozen plasma was given only to maintain hemostasis (e.g., if the international normalized ratio [INR] was 1.5 with evidence of increased bleeding from the surgical drains). Platelets were transfused if the platelet count was 20,000/cm3. Urine output was measured every hour. Furosemide 20 mg was given IV if the urine output decreased to 0.5 mL kg1 h1. Enteral feeding was initiated when there was an evidence of bowel movement. Patients were discharged from the ICU once they were alert and cooperative, did not require inotropic or ventilator support, did not require IV intravascluar volume replacement, and were able to maintain oral intake.
Measured Variables
Hemodynamic Variables Heart rate, mean arterial blood pressure, PAOP, CVP, and cardiac output (using a pulmonary artery catheter) were monitored. Hemodynamic data were recorded after induction of anesthesia and before volume administration, at the end of the preanhepatic phase, at the end of the anhepatic phase, at the end of the surgery, and on the first 4 days after surgery. Assessment of Renal Function Renal function of all enrolled patients was monitored by measuring serum creatinine, serum cystatin C, and by calculating creatinine clearance (CrCl) using a complete 24-h urine collection. Blood samples were drawn between 8 and 9 am. Serum creatinine was determined by means of Jaffe reaction.17 The serum cystatin C concentration was measured by means of latex-amplified nephelometry using the N Latex Cystatin C diagnostic kit (Dade Behring Diagnostic, Manheim, Germany) and the BN-II system (Dade Behring Diagnostic).18 Serum creatinine and cystatin C were measured after induction of anesthesia, at the end of surgery, and on the first 4 days postoperatively. CrCl was measured on postoperative Days 1, 3, and 5. Both
925
creatinine and CrCl were measured at the time of discharge from the hospital. Laboratory Variables Liver function tests (aspartate aminotransferase, alanine aminotransferase, INR, and bilirubin), serum albumin level, and serum amylase were obtained for the first 4 days in the ICU. Other Variables Intravascular volume replacement therapy, including crystalloid and colloid infusion, and blood and plasma transfusion were recorded daily during the first 4 days in the ICU. Urine output, the dose of diuretics, and fluid balance (calculated as fluid input minus fluid output) were collected. The duration of mechanical ventilation (the time between ICU admission and tracheal extubation), start of enteral feeding, ICU stay, and hospital stay were documented. All complications including rejection episodes, renal replacement therapy, pulmonary complications, nosocomial infection, and reoperation were recorded.
Table 1. Demographic and Perioperative Data ALB (n 20)

Age (yr) Gender (M/F) Body mass index Child classification Child B/C Diagnosis HCV HCV HCC MELD Diabetes CrCl (mL/min) CrCl 50 mL/min CrCl 3050 mL/min CrCl 30 mL/min GWR (%) Ischemia time (min) Operative time (min) Mechanical ventilation (h) Reoperation Start of enteral feeding (d) ICU stay (d) Hospital stay (d) 51 6 16/4 29.9 5.3 4/16 17 3 15 (1219) 7 77 (36194) 15 5 0 1 0.13 104 14 645 99 12.5 2.2 3 2.1 0.4 4.5 0.5 23 4
HES (n 20)
55 5.8 19/1 26.2 4 3/17 18 2 15 (820) 8 100 (24129) 11 6 3 1.1 0.2 94 45 640 104 12 3 1 2.2 0.5 5 0.2 27 3
Statistics
Power analysis was performed using Students t-test for independent samples on CrCl 24 h after surgery because it was the main outcome variable in this study. Our previous study showed that the standard deviation of CrCl in posttransplant patients was nearly 20 mL/min.19 Based on the assumption that a mean difference of 20 mL/min was considered a clinically significant difference between groups and taking power 0.8 and error 0.05, a minimum sample size of 17 patients was calculated for each group. Twenty patients in each group were included to compensate for possible dropouts. Categorical variables were assessed using 2 or Fischers exact test where appropriate. Normally distributed data are presented as mean (sd) and were analyzed using Students t-test and two-way analyses of variance with repeated measures and post hoc Dunnett test as appropriate. Data not normally distributed (tested by Kolmogorov-Smirnov test) are presented as median (range) and were analyzed with MannWhitney U-test or the KruskalWallis test as appropriate. The software SPSS v15.0 for Windows (SPSS, Chicago, IL) was used for statistical analysis.
Data are mean SD, median (range), or number of patients. ALB albumin; HES hydroxyethyl starch; CrCl creatinine clearance; MELD model for end-stage liver disease; GWR graft weight ratio; ICU intensive care unit; HCV hepatitis C virus; HCC hepatocellular cardinoma.
RESULTS
All 40 enrolled patients completed the study. Demographic data were comparable in the two groups (Table 1). All three patients with severe renal impairment (baseline CrCl 30 mL/min) were randomized to the HES treatment group. Both study groups were comparable regarding ischemia time, operative time, duration of postoperative mechanical ventilation, and the start of enteral feeding. Comparable results were also found in ICU stay and hospital stay (Table 1). One patient in each group needed renal replacement therapy during the ICU stay on the sixth and seventh postoperative days. No postoperative pulmonary complications were recorded between the study groups. Mild to moderate rejection was documented in both groups by biopsy (three in HES group and four in the ALB group). One patient in each group died during the hospital stay. The cause of death was sepsis and multiple organ failure 2 wk after surgery. The mean sd volume of the study colloid administered introperatively was 3500 1000 mL in the ALB group comparable with that administered in the HES group, which was 3080 417. The mean volume was 6229 1140 mL and 4636 1153 mL in the HES and ALB groups, respectively, P 0.003. This resulted in significantly greater net cumulative fluid balance in the HES group 3047 2000 mL compared with the ALB group 1100 900 mL, P 0.029. The use of crystalloids, packed red blood cells, fresh frozen plasma, and urine output did not differ between the two groups. However, use of diuretics was
Analysis of Effect Size

Post hoc power analysis was based on univariate repeated measures analysis of variance to determine the effects of IV fluid type on CrCl. The univariate model was used after establishing that the assumptions of sphericity and compound symmetry were met (P 0.05). Greenhouse-Geisser correction was used for violation of the sphericity assumption.20,21 Post hoc power calculations were performed using the computer program G*Power 3.0 for Windows (Franz Faul, Universita t Kiel, Germany).
926
HES and Liver Transplantation
Table 2. Infused Volumes, Diuretics, and Urine Output Data Are Mean SD or Median (Range) Intraoperative
Colloids (mL) ALB HES Crystalloids (mL) ALB HES Urine output (mL) ALB HES Furosemide (mg) ALB HES PRBCs (units) ALB HES FFP ALB HES 3500 1000 3080 417 4944 1261 4416 874 1360 681 1390 659 80 (30180) 40 (20160) 4 (06) 4 (010) 2 (08) 0 (06)
DISCUSSION
The main result of this study was that the use of modern HES 130/0.4 exerted no adverse effect on kidney function in living donor liver transplantation when compared with albumin 5%. The incidence of postoperative renal impairment in orthotopic liver transplantation is common, as high as 70%, which is still associated with considerable morbidity and mortality.22 Considering the effect of colloid therapy on renal function in patients undergoing liver transplantation, it is extremely important to be able to make use of highly sensitive indicators of glomerular filtration rate to identify renal dysfunction early and assess its severity. In addition to assessing serum creatinine and 24-h CrCl, we measured serum cystatin C, a protein of low molecular weight (13 kDa) that is constantly produced by all nucleated cells and is not affected by age, gender, or muscle mass.23 Several studies have found cystatin C superior to creatinine as a marker of kidney function immediately after liver transplantation.24,25 This study found no deleterious effects on renal function with the use of HES 130/0.4 as a primary colloid therapy during surgery and in the early postoperative period, as measured by serum creatinine, CrCl, and plasma cystatin C. Serum cystatin C tended to be higher in the HES group, although it did not reach a statistically significant difference, which might be explained by the fact that three patients with severe renal impairment were randomly assigned to the HES group. In kidney transplant patients, osmotic nephrosis was found after receiving the less-metabolizable HES specification (HES 200/0.62).26,27 Different HES molecules have different physico-chemical properties according to molecular weight, molar substitution (mole hydroxyethyl residue per mole glucose subunit), and the C2/C6 ratio. The higher the molar substitution, the slower is the breakdown and elimination of the molecule. Plasma clearance of HES 130/0.4 is at least 20 times faster than that of hetastarch and pentastarch.28 Hemodynamics in both groups in our study were similar over the entire study period. However, we needed significantly more HES than albumin to maintain the intravascular volume in our target population. In line with our result, Persson and Grande29 demonstrated that the plasma-expanding effects after hemorrhage were significantly greater with 20 mL/kg of 5% human albumin than with the same amount of 6% HES 130/0.4. In our center, we did not routinely use albumin 20% to correct the serum albumin level after liver transplantation.19 Accordingly, patients in the HES group experienced a more severe degree of hypoalbuminemia compared with the other group, with an absolute minimum serum albumin concentration reaching 0.7 g/dL in the HES group at the end of surgery. The plasma albumin concentration is the intravascular
Postoperative
4636 1153 6229 1140* 12579 2054 12221 2073 8859 1730 8607 1306 20 (080) 65 (0180)* 4 (08) 2 (08) 0 (06) 0 (08)
ALB albumin; HES hydroxyethyl starch; PRBCS packed red blood cells; FFP fresh frozen plasma. * Signicance compared with the other group (P 0.05).
significantly greater in the HES group compared with the ALB group (Table 2). Hemodynamic variables did not differ between the two groups except for CVP and PAOP, which were greater in the ALB group compared with the HES group on Days 1 and 2 (Tables 3 and 4). Serum creatinine and CrCl were similar in both groups throughout the entire study period (Figs. 1 and 2). The cystatin C plasma level tended to be higher in the HES group on Days 2 and 3, but this difference did not reach statistical significance (P 0.08) (Fig. 3). There were no significant differences in other laboratory data (bilirubin, alanine aminotransferase, aspertate aminotransferase, and INR) between the two study groups. The serum albumin concentration was higher in the ALB group. Serum amylase was higher in the HES group (Table 5).
Effect Size Analysis

Based on the observed pooled standard deviation in CrCl of 24.47 mL/min and the mean correlation among the five measurements of CrCl was 0.601, our study with a sample size of 40 subjects who were randomized into a balanced within-between (mixed repeated measures) design with an assumed error probability of 0.05 and error probability of 0.2 allowed detection of a between-factor effect size of 0. 38. This translates to a between-group mean CrCl difference of 18.36 mL/min, and it represents the minimum detectable mean difference, which could be found with this study by a power of 80%. Based on these data, our study has a power of 86.3% to detect a presumed minimum clinically significant difference of 20 mL/min in CrCl.
927
Table 3. Intraoperative Hemodynamic Data Baseline

HR (bpm) ALB HES MAP (mm Hg) ALB HES CVP(mm Hg) ALB HES PAOP(mm Hg) ALB HES CO (L/min) ALB HES 83 12.5 75.3 10.4 87.5 6 94.6 19.6 83 72 10 4 11 3 8.2 2.9 93
End of preanhepatic
94 13 91 17 82 11 83 13 63 61 10 3 82 10.2 2.4 9.4 3.9
End of anhepatic
97 9 97 18 84 13 75 13 3.8 2 4.5 2.3 63 7.6 2.8 8.5 2 8 2.5
End of surgery
95 7 98 18 85.5 19 92 7 6 1.9 6 2.7 10 3 93 10 3.4 12 2.1
Values are mean SD. ALB albumin; HES hydroxyethyl starch; HR heart rate; MAP mean arterial blood pressure; CVP central venous pressure; PAOP pulmonary artery occlusion pressure; CO cardiac output.
Table 4. Postoperative Hemodynamic Data Day 1

HR (bpm) ALB HES MAP (mm Hg) ALB HES CVP (mm Hg) ALB HES PAOP (mm Hg) ALB HES CO (L/min) ALB HES 115.8 5.5 117 4 91 6 93 5.6 4.8 0.7 2.3 0.8* 8.8 0.7 6.5 1* 9.5 1 10 1.6
Day 2
112 4 115 4 102 2.5 98 7 5.6 1.8 2.8 0.7* 9.6 1.8 6.8 0.7* 9 1.3 8.8 1.1
Day 3
106 4 106 5 91 9 94 8 51 4 0.8 8.8 1 92 7.7 0.4 7.8 0.8
Day 4
97 5 102 9 100 10 103 4 5.6 1 6.6 2.8 91 8 0.8 6.6 0.8 7.5 0.5
Values are mean SD. ALB albumin; HES hydroxyethyl starch; HR heart rate; MAP mean arterial blood pressure; CVP central venous pressure; PAOP pulmonary artery occlusion pressure; CO cardiac output. * Signicance compared with the other group (P 0.05).
protocol. Values are mean sd. HES hydroxyethyl starch.
Figure 1. Change in serum creatinine throughout the study
Figure 2. Change in creatinine clearance throughout the study protocol. Values are mean sd. HES hydroxyethyl starch. Postoperative fluid overload is a risk factor for postoperative pulmonary complications after liver transplantation.30 In our study, the duration of mechanical ventilation and postoperative pulmonary complications were the same in both study arms. One major concern about the use of HES preparations is the potential for inducing disorders in coagulation.31 In this study, routine coagulation tests did not
albumin mass divided by the plasma volume. Dilution secondary to HES infusion and redistribution secondary to altered vascular permeability may be the cause of the rapid reduction in serum albumin concentration that was observed in patients assigned to the HES group.
928
Figure 3. Change in serum cystatin C throughout the study protocol. Values are mean sd. HES hydroxyethyl starch. Table 5. Postoperative Laboratory Investigation ALB (n 20)
Bilirubin (mg/dL) Day 1 Day 2 Day 3 Day 4 AST (U/L) Day 1 Day 2 Day 3 Day 4 ALT (U/L) Day 1 Day 2 Day 3 Day 4 INR Day 1 Day 2 Day 3 Day 4 Amylase Day 1 Day 2 Day 3 Day 4 7.8 2.5 3.3 1.8 3.8 2.0 4.3 2.7 609 424 316 232 183 190 74 36 611 472 548 400 425 300 288 200 2.6 1.4 2.2 0.3 1.5 0.2 1.3 0.1 51 27 67 41 85 70 53 25
There were several limitations in this study. We did not study the influence of volume therapy on the integrity of renal tubular function by using new sensitive markers, such as N-acetyl--d-glucosaminidase and glutathione transferase-. Most of our patients had nearly normal kidney function before transplant. We cannot conclude that HES 130/0.4 is as safe as albumin 5% in liver-transplanted patients with more severe renal impairment before transplant. We also did not gather long-term follow-up data in this study. In conclusion, the perioperative use of HES 130/0.4 up to 50 mL kg1 d1 had no impact on renal function or patient outcome during early hospitalization in patients undergoing orthotropic liver transplantation. Further, larger studies are warranted to address the nonrenal safety of using HES in livertransplanted patients. REFERENCES
1. Schrier RW. Decreased effective blood volume in edematous disorders: what does this mean? J Am Soc Nephrol 2007;18: 2028 31 2. Funk W, Baldinger V. Microcirculatory perfusion during volume therapy. A comparative study using crystalloid or colloid in awake animals. Anesthesiology 1995;82:975 82 3. Merritt WT. Practice patterns and anesthesia-related costs for liver transplantation. Liver Transpl Surg 1997;3:449 50 4. Margarson MP, Soni N. Serum albumin: touchstone or totem? Anaesthesia 1998;53:789 803 5. Treib J, Baron JF, Grauer MT, Strauss RG. An international view of hydroxyethyl starches. Intensive Care Med 1999;25:258 68 6. Chuang FR, Lin CC, Wang PH, Cheng YF, Hsu KT, Chen YS, Lee CH, Chen CL. Acute renal failure after cadaveric related liver transplantation. Transplant Proc 2004;36:2328 30 7. Faenza S, Santoro A, Mancini E, Pareschi S, Siniscalchi A, Zanzani C, Pinna AD. Acute renal failure requiring renal replacement therapy after orthotopic liver transplantation. Transplant Proc 2006;38:11412 8. Biancofiore G, Davis CL. Renal dysfunction in the perioperative liver transplant period. Curr Opin Organ Transplant 2008;13:2917 9. Dickenamm MJ, Filipovic M, Schneider MC, Brunner FP. Hydroxyethylstarch-associated transient acute renal failure after epidural anaesthesia for labour analgesia and Caesarean section. Nephrol Dial Transplant 1998;13:2706 10. Wiedermann CJ. Renal impairment in cardiac surgery patients receiving hydroxyethyl starch. Intensive Care Med 2004;30:519 20; author reply 21 11. Boldt J, Brosch C, Ducke M, Papsdorf M, Lehmann A. Influence of volume therapy with a modern hydroxyethylstarch preparation on kidney function in cardiac surgery patients with compromised renal function: a comparison with human albumin. Crit Care Med 2007;35:2740 6 12. Sakr Y, Payen D, Reinhart K, Sipmann FS, Zavala E, Bewley J, Marx G, Vincent JL. Effects of hydroxyethyl starch administration on renal function in critically ill patients. Br J Anaesth 2007;98:216 24 13. Winkelmayer WC, Glynn RJ, Levin R, Avorn J. Hydroxyethyl starch and change in renal function in patients undergoing coronary artery bypass graft surgery. Kidney Int 2003;64:1046 9 14. Mahmood A, Gosling P, Vohra RK. Randomized clinical trial comparing the effects on renal function of hydroxyethyl starch or gelatine during aortic aneurysm surgery. Br J Surg 2007;94:42733 15. Arroyo V, Cardenas A, Campistol JM, Gines P. Acute renal failure in liver disease. In: Davison A, Stewart CJ, Grunfeld JP, Kerr DNS, Ritz E, Winearls CG, eds. Oxford textbook of clinical hepatology. London: Oxford Press, 2005:1564 79 16. Abofetouh F, Khater Y, Mukhtar A, Salah M, Khedr H, Hamed H, Badawy S. Perioperative management in adult and pediatric living related liver transplantation: an Egyptian experience. Int Anesthesiol Clin 2006;44:12736
HES (n 20)
6.2 2.7 4 2.1 3.1 1.6 3.4 2 590 381 494 370 250 200 130 78 600 414 635 530 568 441 345 300 2.1 0.2 2.0 0.3 1.4 0.27 1.3 0.2 207 151* 147 105* 188 82* 152 93*
Values are mean SD. ALB albumin; HES hydroxyethyl starch; AST aspartate aminotransferase; ALT alanine aminotransferase; INR international normalized ratio. *Signicance compared with the other group (P 0.05).
differ between the two groups. Our study sample size was not sufficient to address whether HES 130/0.4 can alter the coagulation profile. Moreover, in vitro coagulation tests do not represent the balance of coagulation as it occurs in vivo, which may explain why the many in vitro tests are not good predictors of bleeding in patients with liver disease.32 The serum amylase level was significantly higher in the HES group throughout our study protocol. Although it was judged to be clinically irrelevant, an increase in amylase may confound the diagnosis of acute pancreatitis, which occurs in 3% of patients after liver transplantation.33
929
17. Habazin-Novak V, Plestina R. A microcolorimetric determination of creatinine in serum or plasma by the Jaffes reaction. Arh Hig Rada Toksikol 1979;30:2532 18. Finney H, Newman DJ, Gruber W, Merle P, Price CP. Initial evaluation of cystatin C measurement by particle-enhanced immunonephelometry on the Behring nephelometer systems (BNA, BN II). Clin Chem 1997;43:1016 22 19. Mukhtar A, EL Masry A, Moniem AA, Metini M, Fayez A, Khater YH. The impact of maintaining normal serum albumin level following living related liver transplantation: does serum albumin level affect the course? A pilot study. Transplant Proc 2007;39:32148 20. Thomas L. Retrospective power analysis. Conserv Biol 1997;11: 276 80 21. Maxwell SE, Kelley K, Rausch JR. Sample size planning for statistical power and accuracy in parameter estimation. Annu Rev Psychol 2008;59:537 63 22. Gainza FJ, Valdivieso A, Quintanilla N, Errazti G, Gastaca M, Campo M, Lampreabe I, Ortiz-de-Urbina J. Evaluation of acute renal failure in the liver transplantation perioperative period: incidence and impact. Transplant Proc 2002;34:250 1 23. Page MK, Bukki J, Luppa P, Neumeier D. Clinical value of cystatin C determination. Clin Chim Acta 2000;297:6772 24. Samyn M, Cheeseman P, Bevis L, Taylor R, Samaroo B, BuxtonThomas M, Heaton N, Rela M, Mieli-Vergani G, Dhawan A. Cystatin C, an easy and reliable marker for assessment of renal dysfunction in children with liver disease and after liver transplantation. Liver Transpl 2005;11:344 9 25. Biancofiore G, Pucci L, Cerutti E, Penno G, Pardini E, Esposito M, Bindi L, Pelati E, Romanelli A, Triscornia S, Salvadorini MP, Stratta C, Lanfranco G, Pellegrini G, Del Prato S, Salizzoni M, Mosca F, Filipponi F. Cystatin C as a marker of renal function immediately after liver transplantation. Liver Transpl 2006;12:28591
26. Cittanova ML, Leblanc I, Legendre C, Mouquet C, Riou B, Coriat P. Effect of hydroxyethylstarch in brain-dead kidney donors on renal function in kidney-transplant recipients. Lancet 1996;348:1620 2 27. Pillebout E, Nochy D, Hill G, Conti F, Antoine C, Calmus Y, Glotz D. Renal histopathological lesions after orthotopic liver transplantation (OLT). Am J Transplant 2005;5:1120 9 28. Jungheinrich C, Sauermann W, Bepperling F, Vogt NH. Volume efficacy and reduced influence on measures of coagulation using hydroxyethyl starch 130/0.4 (6%) with an optimised in vivo molecular weight in orthopaedic surgery: a randomised, double-blind study. Drugs R D 2004;5:19 29. Persson J, Grande PO. Volume expansion of albumin, gelatin, hydroxyethyl starch, saline and erythrocytes after haemorrhage in the rat. Intensive Care Med 2005;31:296 301 30. Golfieri R, Giampalma E, Morselli Labate AM, dArienzo P, Jovine E, Grazi GL, Mazziotti A, Maffei M, Muzzi C, Tancioni S, Sama C, Cavallari A, Gavelli G. Pulmonary complications of liver transplantation: radiological appearance and statistical evaluation of risk factors in 300 cases. Eur Radiol 2000;10: 1169 83 31. Treib J, Haass A, Pindur G. Coagulation disorders caused by hydroxyethyl starch. Thromb Haemost 1997;78:974 83 32. Segal JB, Dzik WH. Paucity of studies to support that abnormal coagulation test results predict bleeding in the setting of invasive procedures: an evidence-based review. Transfusion 2005; 45:141325 33. Krokos NV, Karavias D, Tzakis A, Tepetes K, Ramos E, Todo S, Fung JJ, Starzl TE. Acute pancreatitis after liver transplantation: incidence and contributing factors. Transpl Int 1995; 8:17
930
Pain Medicine
Section Editor: Spencer S. Liu
Brief Report
The Impact of Asynchronous Electroacupuncture Stimulation Duration on Cold Thermal Pain Threshold
Shu-Ming Wang, MD* Eric C. Lin, BS Inna Maranets, MD* Zeev N. Kain, MD, MBA
The durations of asynchronous electroacupuncture can affect the resultant hypoalgesia. Healthy volunteers were randomized to receive different durations (0 min, 20 min, 30 min, or 40 min) of asynchronous electroacupuncture stimulations (alternating low/high [2/100 Hz] frequency at 5 mA). Using a human experimental cold thermal pain threshold model, we found that 30 min of asynchronous 2/100 Hz stimulation resulted in the most significant hypoalgesic effect that was sustained for at least 60 min after stimulation compared with 0-, 20-, or 40-min stimulations (P 0.05). We conclude that the most optimal duration for asynchronous electroacupuncture stimulation is 30 min.
(Anesth Analg 2009;109:9325)
custimulations have been widely incorporated into a comprehensive clinical pain management program.19 Electroacupuncture stimulation (EAS) is a technique that applies a small electrical current to needles inserted into the acupuncture points.10 This technique is intended to achieve synergistic or additive analgesic benefits of traditional acupuncture and electrical stimulation.6,10 The advantages of EAS over traditional acupuncture manipulation include better quantification of the stimulation delivered and reproducibility.6 Asynchronous EAS was developed based on the findings that by alternating low and high frequency electrical stimuli (e.g., 2 Hz of electrical stimulation alternating with 100 Hz of electrical stimulation) through needles into acupuncture points enhances hypoalgesia more than either frequency alone.11,12 The duration of a single frequency of electroacupuncture affects the resultant hypoalgesia.3,4 We therefore conducted the following study to determine whether
From the *Department of Anesthesiology, Yale School of Medicine; Department of Anesthesiology, The Yale Center for Advancement of Perioperative Health, Yale School of Medicine, New Haven, Connecticut; and Departments of Anesthesiology, Pediatrics, and Psychiatry and Human Behavior, University of California, Irvine School of Medicine, Irvine, California. Accepted for publication March 27, 2009. None of the authors have any relationships with any company or organization that has a potential or vested interest in the outcome of the study. Partial data from this study were presented at the annual 2007 ASA meeting in San Francisco, California. Address correspondence and reprint requests to Shu-Ming Wang, MD, Department of Anesthesiology, Yale School of Medicine, 333 Cedar St., New Haven, CT 06520. Address e-mail to shu-ming.wang@yale.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ad9292
the duration of asynchronous EAS can also affect the resultant hypoalgesia using a human experimental pain model.
METHODS
After approval of the Human Investigation Committee of Yale School of Medicine, healthy volunteers (aged 18 yr and older) were recruited. Exclusion criteria included a history of any systemic medical or psychological illness, daily medication usage, illicit substance usage, prior experience in acupuncture, and pregnancy. All volunteers were informed that the purpose of the study was to test whether different durations of EAS can affect their ability to sense cold pain. After obtaining informed consent and baseline demographic information, including State-Trait Anxiety Inventory, education and belief in acupuncture treatment, all participants were positioned on a stretcher comfortably in an experimental room with the temperature set at 70F. To eliminate the possibility of bias by the subjects, no watch or clock was allowed in the study area. Using a Datex-Ohmeda portable patient monitor, the changes of hemodynamic variables, which included continuous heart rate monitoring and noninvasive arterial blood pressure measurements set at every 5 min, were recorded during the entire study period. The preselected sites were marked on the medial side of the right lower leg before the initiation of cold thermal pain threshold (CTPT). CTPT was delivered by PATHWAY (Medoc Medical, Israel) through a 3 3 cm Peltier thermal stimulation probe using advanced thermal stimulator version 2.4. The first research assistant assessed the baseline CTPT at these sites (Fig. 1) in a random order for a total of nine times (three times per site) as described in previous
932
Figure 2. Locations of acupuncture points and dermatomal distributions. Figure 1. Locations of premarked testing site and dermatomal distributions.
studies.3,13 All participants were instructed to press the stop button when they sensed the cold sensation turning into pain, and the corresponding temperatures were automatically recorded into the PATHWAY. The participants were randomized into one of the four durations based on a computer-generated randomization table. Only the acupuncturist knew that the number represented a selected duration of EAS. The acupuncturist then inserted the acupuncture needles (Siren, L type No 5 [0.25 40 mm], Shizuoka, Japan) into the selected acupuncture points on the left leg (Fig. 2). All participants were informed that they might or might not feel a de qi* sensation and/or a vibrating sensation during the intervention period. For participants assigned to the sham control group, i.e., 0 min group, the acupuncture needles were inserted superficially without eliciting any de qi sensation. For participants receiving 20-, 30-, or 40-min of EAS, the acupuncture needles were placed until the acupuncturist experienced a de qi sensation. The depth
*A subjective sensation, described by subjects who received acupuncture treatment, as numbness, aching, soreness, or distension. From the acupuncturists perspective, this subjective sensation coincides with the sensation of the needle getting caught during application of acupuncture needle. Vol. 109, No. 3, September 2009
of needle insertion at that point was about 0.751.5 cm. Once the acupuncture needles were placed, the acupuncturist connected a pair of electrodes from a Han stimulator (Beijing, China) to the needles. Asynchronous EAS (2/100 Hz) was delivered via Han stimulator and the intensity gradually adjusted to 5 mA over a 3-min period. For participants assigned to receive 0 min of stimulation, the electrodes were connected to an inactive Han stimulator for 20 min. Only the acupuncturist had a timer, so the EAS would be terminated when the proper assigned duration of EAS was reached. None of the research assistants was present during the EAS period. Once the assigned EAS was completed, the State-Trait Anxiety Inventory-S was reassessed and the CTPT was immediately reassessed post-EAS, then subsequently every 5 min for a total of 60 min by the second research assistant at the same premarked sites. This procedure assured blindness of the research assistants to prevent bias.
The sample size was calculated based on a previous study14 in which the difference in mean visual analog scale scores for pain between volunteers who received real EAS versus those who received sham EAS was 0.65 with a standard deviation of 0.5. Based on our calculation, 14 subjects per group were needed to achieve a power of 90% in detecting 10% hypoalgesic
933
Table 1. The Baseline Demographic Data 0 min (control) (n 14)

Age Gender (M/F) Educationa Beliefb STAI-S STAI-T STAI-Sc 35 12 5/9 17 3 74 13 29 6 33 6 25 2
20 min (n 14)
32 10 5/9 17 3 76 10 28 5 28 6 24 3
30 min (n 14)
37 12 5/9 17 2 75 14 30 7 32 6 24 3
40 min (n 14)
37 17 5/9 16 2 72 25 30 8 32 9 24 5
P
0.8 1 0.4 0.1 0.6 0.2 0.7
Data are presented as mean SD. STAI-S Baseline State Anxiety Score; STAI-T Baseline Trait Anxiety Score. a The highest education level subject received. b The level of belief regarding acupuncture treatment; 0 represents none at all and 100 represents total belief in acupuncture treatment. c Postintervention State Anxiety Score.
Figure 3. Cold thermal pain thresholds (% se) before and after asynchronous electroacupuncture stimulations. differences between the treatment groups with an value of 0.05. All hemodynamic and CTPT data were directly recorded into PATHWAY and exported into SPSS 16 for Mac (SPSS, Chicago, IL) and analyzed using repeated measure analysis with post hoc least significant difference and P 0.05 considered to be statistically significant. Baseline demographic data and an anxiety questionnaire were analyzed using one-way analysis of variance. Because the durations of EAS were different, we only compared the differences in temperature at which volunteers sensed cold pain among groups between 40 and 80 min after the onset of intervention. As illustrated in Figure 3, participants receiving 30 min of EAS had significant hypoalgesia as compared with those receiving 0 min (P 0.008), 20 min (P 0.005), and 40 min (P 0.024). The asynchronous EAS-induced hypoalgesia did not differ significantly between participants receiving 0 min vs 20 min (P 0.59), 0 min vs 40 min (P 0.57), and 20 min EAS vs 40 min of EAS (P 0.69). At 0 min post-EAS, participants in the 30-min group sensed cold pain at a significantly lower temperature than those who received 0 min (P 0.01) and 20 min (P 0.013) but not lower than those who received 40 min of EAS (P 0.93). However, at 5 min post-EAS, the participants who received 30 minutes of
RESULTS
Fifty-six subjects participated in this randomized, controlled trial, and no subject withdrew from the trial. There was no difference in age, education, gender, anxiety, and level of belief in acupuncture among the groups (Table 1). There were no differences in the arterial blood pressure or heart rate among the groups before, during, or after EAS.
934
Brief Report
asynchronous EAS reported sensing cold pain at a much lower temperature than those who received 0 min (P 0.006), 20 min (P 0.005), and 40 min (P 0.03). The above phenomenon persisted for 60 min post-EAS (Fig. 3).
DISCUSSION
Under the conditions of this study, we found that 30 min of asynchronous 2/100 Hz (5 mA) EAS at ST36 and SP6 resulted in significant hypoalgesia compared with 0 min, 20 min, or 40 min. The hypoalgesic effect was sustained for 60 min post-EAS. Similar to previous electrical stimulation studies,3,14 the asynchronous EAS-induced hypoalgesia is not restricted to segmental pain inhibition. As illustrated in Figures 1 and 2, EAS at left ST36 and SP6 (L4, L5, and S2 dermatomal distribution) and the hypoalgesia was tested at the right premarked sites (L3, L4, and S2 dermatomal distribution). A previous manual acupuncture study demonstrated that the peak of hypoalgesia was between 20 and 40 min sustained up to 60 min,15 and the peak hypoalgesia of single frequency EAS also occurred at about 30 min and was sustained up to 75 min during stimulation.4,16 We were not able to demonstrate the similar phenomena as described in these previous studies.4,15,16 In our study, the temperatures at which subjects perceived cold pain after receiving 40 min of asynchronous EAS were significantly higher than the temperatures in those who received 30 min of the same EAS. Future studies should explore the potential mechanism(s) contributing to the development of this phenomenon. Interestingly, the optimal duration for alternating frequencies of EAS as demonstrated in our study is the same as the time setting used in other electrical nerve stimulators that are commercially available to deliver alternating electrical frequency stimulations to the target nerves.5,9 Therefore, we speculate that the analgesia resulting from alternating frequencies of EAS, transcutaneous nerve stimulation, or percutaneous nerve stimulation may share a similar underlying mechanism. The limitations of this study are as follows: 1) we did not assess the level of hypoalgesia while the stimulation was ongoing, thus we could not establish how the hypoalgesia developed over time, and 2) we did not have an adequate duration of CTPT assessment after stimulation. Based on a previous single frequency EAS study, the resulting hypoalgesic effects gradually return to control values within 35 min after termination of stimulation.16 As a result, we decided
a priori to assess the CTPTs up to 60 min after EAS was terminated. In addition, the temperatures at which subjects sensed a CTPT were directly recorded into PATHWAY. Thus we did not discover that there was any fading even at 60 min post-EAS until after the study was completed and at the time of data analysis. In conclusion, the duration of asynchronous 2/100 Hz EAS indeed affects the resultant hypoalgesia, and 30 min of stimulation seems to be the optimal duration. REFERENCES
1. Ghoname EA, Craig WF, White PF, Ahmed EA, Hamza MA, Gajraj NM, Vakharia AS, Noe CE. The effect of stimulus frequency on the analgesic response to percutaneous electrical nerve stimulation in patients with chronic low back pain. Anesth Analg 1999;88:8412 2. Ghoname EA, Craig WF, White PF, Ahmed HE, Hamza MA, Henderson, Gajraj BN, Huber PJ, Gatchel RJ. Percutaneous electrical nerve stimulation for low back pain: a randomized crossover study. J Am Med Assoc 1999;281:818 23 3. Leung AY, Kim SJ, Schulteis G, Yaksh T. The effect of acupuncture duration on analgesia and peripheral sensory thresholds. BMC Complemt Altern Med 2008;8:18 29 4. Andersson SA, Ericson T, Holmgren E, Lindqvist G. Electroaccupuncture and pain threshold. Lancet 1973;2:564 9 5. Hamza MA, White PF, Ahmed HE, Ghoname EA. Effect of the frequency of transcutaneous electrical nerve stimulation on the postoperative opioid analgesic requirement and recovery profile. Anesthesiology 1999;91:1232 8 6. Ulett GA, Han SP, Han JS. Electroacupuncture: mechanisms and clinical application. Biol Psychiat 1998;44:129 38 7. Chen XH, Guo SF, Chang CG, Han JS. Optimal conditional for elicit maximal electroacupuncture analgesia with dense and disperse mode stimulation. Am J Acupunct 1994;22:4753 8. Chen L, Tang J, White PF, Sloninsky A, Wender RH, Naruse R, Kariger R. The effect of location of transcutaneous electrical nerve stimulation on postoperative opioid analgesic requirement: acupoint versus non-acupoint stimulation. Anesth Analg 1998;87:1129 34 9. Hamza MA, White PF, Craig WF, Ghoname ES, Ahmed HE, Proctor TJ, Noe CE, Vakharia AS, Gajraj J. Percutaneous electrical nerve stimulation: a novel analgesic therapy for diabetic neuropathic pain. Diabetes Care 2000;23:36570 10. Wang SM, Kain ZN, White PF. Acupuncture analgesia: II. Clinical consideration. Anesth Analg 2008;106:61121 11. Han JS. Neurochemical basis of acupuncture analgesia. Annu Rev Pharmacol Toxicol 1982;22:193220 12. Han JS. Acupuncture and endorphins. Neurosci Lett 2004;361: 258 61 13. Leung A, Khadivi B, Duann JR, Cho ZH, Yaksh T. The effect of ting point (tendinomuscular meridians) electroacupuncture on thermal pain: a model for studying the neuronal mechanism of acupuncture analgesia. J Altern Complement Med 2006;12:74350 14. Zhang WT, Jin Z, Huang J, Zhang L, Zeng YW, Luo F, Chen CAN, Han JS. Modulation of cold pain in human brain by electric acupoint stimulation: evidence from fMRI. Neuroreport 2003;14:1951 6 15. Research Group of Acupuncture Anesthesia, Peking Medical College. Effect of acupuncture on the pain threshold of human skin. Natl Med J China 1973;3:1517 16. Andersson SA, Ericson T, Hoimgren E, Lindqvist G. Electroacupuncture. Effect on pain threshold measured with electrical stimulation of teeth. Brain Res 1973;63:393 6
935
Pain Mechanisms
Section Editors: Tony L. Yaksh/Quinn H. Hogan
Central Administration of Minocycline and Riluzole Prevents Morphine-Induced Tolerance in Rats

Bohlool Habibi-Asl, PharmD, PhD Kambiz Hassanzadeh, PharmD Mohammad Charkhpour, PharmD, PhD
BACKGROUND: Long-term exposure to opiates induces tolerance to the analgesic effect. The neurobiological mechanism of this phenomenon is not completely clear. In this study, we evaluated the effects of central administration of minocycline (a tetracycline derivative) and riluzole (an antiglutamatergic drug) on morphineinduced tolerance in rats. METHODS: Groups of rats received daily morphine (10 mg/kg, IP) in combination with saline (10 L/rat, intracerebroventricular [ICV]) or 1% Tween 80 (10 L/rat, ICV) or minocycline (60, 120, and 240 g/10 L per rat, ICV) or riluzole (20, 40, 80 g/10 L per rat, ICV). Nociception was assessed using hotplate apparatus (55C 0.5C). Hotplate latency was recorded when the rat licked its hindpaw. Baseline latencies were determined once per day for each rat, then morphine (10 mg/kg) was injected. After 20 min, the above-mentioned drugs were administered and postdrug latency was measured 10 min after the injection of drugs or vehicles. RESULTS: Results showed that ICV administration of minocycline and riluzole delayed morphine-induced tolerance. Morphine tolerance was complete after 8 days in the control groups but was complete in the groups treated with minocycline (120 g/10 L per rat) and riluzole (80 g/10 L per rat) on the 13th day. In addition, our results showed that minocycline and riluzole increased the total analgesic effect of morphine (area under the curve of the percentage of maximal possible effect values). CONCLUSION: The effects of minocycline on nitric oxide and the glutamatergic system and the effect of riluzole on the glutamate system are potentially important mechanisms in delaying morphine-induced tolerance.
(Anesth Analg 2009;109:936 42)
pioids such as morphine are the most widely used drugs for the alleviation of moderate to severe chronic pain. Systemically administered morphine produces antinociception via actions at both spinal and supraspinal sites.1 Repeated use of opioids induces tolerance that results in a loss of drug effect or the requirement for escalating doses to produce pain relief. The neurobiological mechanisms of the development of opioid tolerance are multifaceted and only partially understood. There are several lines of evidence that suggest that N-methyl-d-aspartate (NMDA) glutamate receptors (NMDARs) are involved in the plasticity that arises from long-term administration of morphine.25 This idea was suggested by Trujillo and Akil2 who reported that the NMDA receptor antagonist, MK-801 (dizocilpine), inhibited the development of tolerance to the
From the Department of Pharmacology and Toxicology, Faculty of Pharmacy, Tabriz University of Medical Sciences, Tabriz, Iran. Accepted for publication April 13, 2009. Address correspondence and reprint requests to Kambiz Hassanzadeh, Department of Pharmacology and Toxicology, Faculty of Pharmacy, Tabriz University (Medical Sciences), Tabriz 5166414766, Iran. Address e-mail to hassanzadehk@tbzmed.ac.ir. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae5f13
antinociceptive effect of morphine and morphine physical dependence. Using behavioral studies, we and others have shown that a variety of NMDA receptor antagonists have the ability to inhibit the development of opiate tolerance and dependence.2 8 Other studies have shown that activation of NMDARs can lead to neurotoxicity under many circumstances.9 For instance, peripheral nerve injury has been shown to activate spinal cord NMDARs, which results in not only intractable neuropathic pain but also neuronal cell death because of apoptosis.10 12 There are also several lines of evidence which suggest that activation of NMDARs leads to toxic calcium influx, which activates numerous enzymes, including neuronal nitric oxide (NO) synthase (NOS). NO is able to further increase excitotoxicity by enhancing glutamate release from presynaptic neurons and inhibiting glial glutamate transporters.1315 Minocycline, a semisynthetic tetracycline derivative, is able to provide neuroprotection against global ischemia in gerbils and focal brain ischemia in rats.16,17 Another study strongly suggested that minocycline acts at an earlier plasmalemmal step by limiting glutamate release and the ensuing [Ca2] elevation in target neurons. Minocycline may prevent the activation of this [Ca2]-dependent intracellular pathway,
936
thus inhibiting neuronal death. A decrease in neuronal excitability, together with a marked decrease in glutamate release, may explain the cytoprotective properties of minocycline.18 20 Many studies have indicated that the protective effect of minocycline was associated with reduced activation of inducible NOS and interleukin-1b-converting enzyme, which are mainly expressed by microglia.19 Riluzole is the only proven effective medicine for amyotrophic lateral sclerosis (ALS) because it has been demonstrated to delay the time of death in ALS patients.20 Riluzole is an antiglutamatergic drug, which interferes with responses mediated by excitatory amino acids, even though it does not interact with any known binding site on NMDA, kainate, or -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptors.20 In addition, it has been shown that coadministration of riluzole with morphine decreased the intensity of the withdrawal syndrome, reflecting a reduction in physical dependence.21 In view of these data, both minocycline and riluzole have a common mechanism of action on the glutamatergic system; therefore, in this study, we evaluated the effect of intracerebroventricular (ICV) administration of minocycline and riluzole on morphine tolerance in rats.
Verication of Cannula Placement

At the end of all experiments, methylene blue solution (5 L/rat, ICV) was injected into the cannula and the animals were killed by an overdose of ether followed by decapitation. The brain of each animal was dissected out and cut in the coronal plane to verify the placement of the guide cannula and distribution of methylene blue in the ventricles. Only data from animals that showed uniform distribution of methylene blue in the ventricles were considered for statistical analysis. In all, six animals were discarded because the placement of the guide cannula was incorrect.
Drug Treatment
Morphine sulfate (Sigma-Aldrich; Sigma-Aldrich, Germany) (10 mg/kg, daily) was dissolved in double distilled water and was injected IP using 1-mL insulin syringes. Minocycline (Sigma-Aldrich) (60, 120, and 240 g/10 L per rat) was dissolved in double distilled water and infused ICV using a Hamilton microsyringe. Riluzole (Sigma-Aldrich) (20, 40, and 80 g/10 L per rat) was dissolved (1% Tween 80 in sterile 0.9% normal saline) and infused ICV using a Hamilton syringe. Dizocilpine (Sigma-Aldrich) (1 g/10 L per rat) was dissolved in double distilled water and infused ICV using a Hamilton microsyringe. Two control groups were included, which received either morphine, IP 1% Tween 80 in saline normal 0.9%, ICV or morphine, IP distilled water 0.9%, ICV. Volume of infusion was 10 L at a rate of 10 L/min in each rat.
METHODS
Animals
Male Wistar rats (Razi Institute, Tehran, Iran) weighing 250 300 g (eight animals in each group) were used in this study. They were kept in a temperature-controlled room (24C 0.5C) and maintained on a 12-h light/ dark cycle (light on 08:00 am) with free access to food and water. All experiments were executed in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 85-23, revised 1985) and were approved by the research and ethics committee of Tabriz University of Medical Sciences.
Assessment of Nociception
Nociception was assessed using the hotplate apparatus (55C 0.5C).23 Hotplate latency was recorded when the rat licked its hindpaw. A cut-off time (40 s) was imposed to prevent tissue damage.24 Hotplate response latencies (s) are expressed as the percentage of maximal possible effect (%MPE) using the equation below:
%MPE
Intracerebroventricular Cannula Implantation

Rats were anesthetized with sodium pentobarbital (50 mg/kg, IP) (Merck, Germany) and a stainless steel guide cannula (23 gauge) was implanted stereotaxically into the lateral cerebral ventricle (coordinates: 0.8 mm posterior, 1.3 mm midline to lateral, and 3.5 mm ventral) with respect to bregma.22 A stainless steel guide (30 gauge) was placed into the guide cannula as a dummy cannula to maintain patency. After surgery, a recovery period of 7 days was allowed before experiments. During the recovery period, animals were habituated to the testing environment including transfer to the experimental room and twice daily handling, weighing, restraining on the platform for 1 min, and gently removing and replacing the dummy cannula. Animals were also habituated to the hotplate apparatus and testing started after the recovery period of 7 days in all groups.
Post-drug latency (s) Baseline latency (s) Cut-off value (s) Baseline latency (s)
100
Baseline latency was determined once per day for each rat, before daily injection of morphine (10 mg/kg). After 20 min, the drugs were administered and the postdrug latency was measured 10 min after the injection of drugs or vehicles (30 min after the injection of morphine). The %MPE was then calculated for that day. The experiments continued until there was no significant difference in %MPE between the vehicle- or drug-treated groups and the saline group.
Evaluation of the Global Analgesic Effect

To evaluate the global analgesic effect and to allow a comparison of the effects from different behavioral tests, the area under the curve (AUC) of the %MPE was calculated. To calculate the AUC, the trapezoidal
937
Figure 1. Analgesic effect of daily administration of morphine (10 mg/kg, IP) in combination with distilled water (10 L/rat) or 1% Tween 80 in normal saline (10 L/rat). Developed tolerance to the analgesic effect of morphine was complete on the 8th day when there were no significant differences in percentage of maximal possible effect (%MPE) between the control groups and the saline group. Each bar represents the mean of %MPE sem for eight rats. M morphine; DW distilled water; Mino minocycline; Rilu riluzole.
Figure 2. Effect of daily ICV injections of minocycline (60, 120, and 240 g/10 L per rat) on tolerance to the analgesic effect of morphine. Each bar represents the mean of percentage of maximal possible effect (%MPE) sem for eight rats. One-way analysis of variance (ANOVA) followed by Tukeys test were used to analyze the statistical significances. P values 0.05 were considered to be significant in all analyses. *P 0.05; **P 0.01; ***P 0.001 when compared with the control group. Arrow represents the day of morphine tolerance. M morphine; DW distilled water; Mino minocycline. 0.9% or distilled water. The analgesic effect of morphine decreased on the 8th day compared with the 1st day and, because there were no significant differences between the control groups and the saline-treated animals on Day 8, this was considered the day of morphine tolerance initiation.
rule was used. The AUC (113 days) was calculated using the trapezoidal rule from the observed values.
Evaluation of Tolerance Induction

To evaluate the induction of tolerance, groups of rats received either saline or morphine (10 mg/kg, IP) saline or morphine (10 mg/kg, IP) minocycline (the most effective dose) or morphine (10 mg/kg, IP) riluzole (the most effective dose) for 8 days, then on the 9th day (1 day after morphine tolerance in the control group) several doses of morphine (10, 25, 50, and 100 mg/kg, IP) were administered to generate analgesic dose-response curves. Morphine antinociceptive 50% effective dose (ED50) values for each of the drug groups were derived using linear regression of %MPE of the morphine dose. Differences in the ED50 estimations were determined using the confidence interval method at P 0.05.25
Evaluation of the Effect of Minocycline on MorphineInduced Tolerance to the Analgesic Effect

Minocycline delayed the onset of morphineinduced tolerance. Minocycline (60, 240, and 120 g/10 L per rat) delayed morphine tolerance for 4, 4 and 5 days, respectively (Fig. 2). Analysis of the AUC of hotplate latency (Table 1), which allowed evaluation of the global effect, showed that repeated treatment with minocycline before morphine enhanced the effectiveness of morphine. On the other hand, minocycline (120 g/10 L per rat) had the greatest AUC of %MPE (289.3) compared with the control group (177.5) or minocycline (60 g/10 L per rat) (264.7) or minocycline (240 g/10 L per rat) (234.7) (Table 1). In addition, the results in Figure 3 show a significant shift to the right in the dose-response curve for animals who received morphine (10 mg/kg) saline (10 L/rat) compared with those receiving saline (10 L/rat) or morphine (10 mg/kg) minocycline (120 g/10 L per rat). A significant shift to the right in ED50 in the control group (88.64) compared with saline (29.5) or morphine (10 mg/kg) minocycline (120 g/10 L per rat) (61.1) treated animals was also seen.
Data Analysis
Data are expressed as the mean of %MPE sem of eight rats per group. One-way analysis of variance followed by Tukeys test were used to analyze statistical significance in multiple comparisons. P values 0.05 were considered to be significant in all analyses. *P 0.05, **P 0.01, and ***P 0.001 indicate a significant difference as compared with the saline group for that day. The AUC113 data were analyzed by one-way analysis of variance.
RESULTS
Induction of Tolerance to the Antinociceptive Effect of Morphine
As shown in Figure 1, daily administration of morphine (10 mg/kg, IP) induced tolerance to the antinociceptive effect of this drug in both the control groups which received 1% Tween 80 in saline normal
938
Minocycline and Riluzole Decreased Morphine Tolerance
Evaluation of the Effect of Riluzole on Morphine-Induced Tolerance to the Analgesic Effect

Riluzole (20, 40, and 80 g/10 L per rat) also delayed morphine tolerance. Riluzole (20 and 40 g/10 L per rat) delayed morphine tolerance for 4 days; however, the results indicated that riluzole (80 g/10 L per rat) decreased the development of
Table 1. The Global Analgesic Effect of Morphine in the Control and Treatment Groups During 13 Days of Experimentation Minocycline
Control mino 60 g/10 L per rat 120 g/10 L per rat 240 g/10 L per rat
AUC
177.5 264.7* 289.3* 234.7*
sem
4.2 3.7 2.6 2.9
Riluzole
Control rilu 20 g/10 L per rat 40 g/10 L per rat 80 g/10 L per rat Dizo 1 g/10 L per rat
AUC
156 216 197 261* 363.8*
sem
3.4 3.1 2.4 3.4 4.1
Area under the curve (AUC) of percentage of maximal possible effect (%MPE) was calculated for each group for 13 days. To calculate the AUC, the trapezoidal rule was used. One-way analysis of variance followed by Tukeys test was used to analyze the differences between the control and treatment groups. Mino Minocycline; Rilu Riluzol; Dizo Dizocilpine. P values less than 0.05 were considered to be signicant in all analyses. *P 0.001; P 0.01; P 0.05 when compared with the control group.
Figure 3. Hotplate responses and percent of maximal possible effect (%MPE) to various morphine doses (10, 25, 50, and100 mg/kg, IP) administered on Day 9 after 8 days of continuous ICV drug infusion. Each point represents the mean sem of eight rats. Different morphine doses administered on Day 9 demonstrated a significant shift to the right in the dose-response curve and antinociceptive ED50 values in animals treated with morphine saline compared with those receiving saline or morphine minocycline (120 g/10 L per rat) or morphine riluzole (80 g/10 L per rat).
Figure 4. Effect of daily ICV injections of riluzole (20, 40, and 80 g/10 L per rat) on tolerance to the analgesic effect of morphine. Each bar represents the mean of percentage of maximal possible effect (%MPE) sem for eight rats. One-way analysis of variance (ANOVA) followed by Tukeys test were used to analyze the statistical significances. P values 0.05 were considered to be significant in all analyses. *P 0.05; **P 0.01; ***P 0.001 when compared with the control group. Arrow represents the day of morphine tolerance. M morphine; Rilu riluzole.
Comparison of the Effect of Minocycline or Riluzole with Dizocilpine on Attenuation of Morphine Tolerance Development
To examine the possible mechanism involved in morphine-induced tolerance, we tested the effect of ICV administration of dizocilpine (1 g/10 L per rat) (a noncompetitive NMDA receptor antagonist), because this agent has well-established effects on opioid tolerance and was used as a positive control in this study. As shown in Figure 6, dizocilpine attenuated morphine tolerance and delayed tolerance initiation for 6 days at this dose. Dizocilpine (1 g/10 L per rat) did not have a significant analgesic effect, thus the effects of this drug are not related to analgesia.
morphine tolerance and could delay morphine tolerance for 5 days (Fig. 4). Riluzole (80 g/10 L per rat) had the greatest AUC of %MPE (261) compared with the control group (156) or riluzole (20 g/10 L per rat) (216) or riluzole (40 g/10 L per rat) (197) (Table 1). Figure 3 shows a significant shift to the right in the dose-response curve for animals that received morphine (10 mg/kg) saline (10 L/rat) compared with those receiving saline (10 L/rat) or morphine (10 mg/kg) riluzole (80 g/10 L per rat). Furthermore, a significant shift to the right in ED50 in the control group (88.64) was observed when compared with saline (29.5) and morphine (10 mg/kg) riluzole (80 g/10 L per rat) (66.8) treated animals.
DISCUSSION
Tolerance is a behavioral adaptation to the prolonged use of opioid drugs, such as morphine. The cellular mechanism underlying the development of morphine tolerance remains controversial. In the current study, we investigated the effect of minocycline and riluzole on morphine-induced tolerance to the analgesic effect. The main findings of this study indicate that ICV administration of minocycline and riluzole can prevent the development of morphine tolerance.
Evaluation of the Analgesic Effects of Minocycline and Riluzole

Administration of the most effective doses of minocycline (120 g/10 L per rat) or riluzole (80 g/10 L per rat) on morphine-induced tolerance did not have a significant analgesic effect and there were significant differences between the administration of these drugs and saline (Fig. 5).
939
Figure 5. Analgesic effects of daily ICV injections of dizocilpine (1 g/10 L per rat), minocycline (120 g/10 L per rat), riluzole (80 g/10 L per rat), and saline (10 L/rat). Each bar represents the mean of percentage of maximal possible effect (%MPE) sem for eight rats. One-way analysis of variance (ANOVA) followed by Tukeys test were used to analyze the differences between the saline and treatment groups. P values 0.05 were considered to be significant in all analyses. Mino minocycline; Rilu riluzole; Dizo dizocilpine.
doses of minocycline (120 g/10 L per rat) and riluzole (80 g/10 L per rat) were compared with dizocilpine (1 g/10 L per rat) used as a positive control. Each bar represents the mean of percentage of maximal possible effect (%MPE) sem for eight rats. M morphine; DW distilled water; Mino minocycline; Rilu riluzole; Dizo dizocilpine.
Figure 6. Effects of daily ICV injections of the most effective
Several studies have indicated that repeated administration of opiates can activate the NMDA receptor through the G-protein associated with the opioid receptor and/or intracellular mechanisms.26,27 This opiate-related activation of NMDARs may initiate subsequent intracellular changes, such as the production of NO and/or the activation of protein kinase C (PKC). Both NO and PKC have been shown to be critical in the development of morphine tolerance.28,29 Previous studies have shown that minocycline, a semisynthetic tetracycline derivative, exhibits neuroprotective effects against neuronal damage in animal disease models.18,30 33 Our results showed that minocycline (120 g/10 L per rat) postponed morphine tolerance for 5 days and minocycline (60 and 240 g/10 L per rat) shifted morphine tolerance from the 8th to the 12th day (Fig. 2). On the other hand, the total analgesic effect of morphine (AUC of %MPE) significantly increased in all treatment groups (Table 1). Although minocycline (120 g/10 L per rat) delayed morphine tolerance more than other doses, there was no significant difference among the three doses in their effects on morphine tolerance. Furthermore, the results shown in Figure 5 demonstrate that the doses of minocycline administered in this study did not have an analgesic effect. Therefore, the action of minocyline in preventing morphine tolerance is not related to its analgesic effect. Several reports attribute the neuroprotective effects of minocycline to various intracellular signaling pathways, including antioxidant systems,34 inhibition of NOS19 blockade of inflammatory responses,35 prevention of the activation of Ca2-dependent intracellular pathways, and a marked decrease in glutamate release.20 In addition, another study showed that NMDAinduced neuronal death involved proliferation and activation of microglial cells and that minocycline
940
completely prevented NMDA toxicity and the preceding activation and proliferation of microglial cells. These results support the notion that microglial activation contributes to excitotoxic neuronal death, which can be inhibited by antiinflammatory compounds, such as minocycline.18 The mechanism underlying the role of glial cells in the effects of morphine on naive mice is unclear. It is possible that morphine acts directly on microglia, triggering alterations in their morphology, metabolism, and function.36 Furthermore, a recent study indicated that systemic administration of minocycline attenuated morphine antinociceptive tolerance.37 Mika et al.37 concluded that the effect of minocycline on morphine tolerance is related to microglia. Their results provide evidence that systemic administration of minocycline in mice influences morphines effectiveness and delays the development of morphine tolerance by attenuating microglial activation and its markers. According to the above-mentioned studies and our findings, it is concluded that the effect of minocycline in this study is possibly related to its neuroprotective property, its effect on preventing glutamate release, and its inhibition of microglial cells and NOS. Figure 4 showed that riluzole decreased tolerance to the analgesic effect of morphine. Riluzole is the only proven effective medicine for ALS and was demonstrated to delay the time of death in these patients.38,39 This is thought to result from the neuroprotective effect of riluzole, which has a complex mechanism of action involving several effects: inhibition of voltagedependent sodium channels,40,41 high-voltage-activated calcium and potassium channels,41 and inhibition of PKC, suggesting involvement in antioxidative processes.42 Our results showed that riluzole (20, 40, and 80 g/10 L per rat) delayed morphine tolerance. As shown in Figure 4, low doses of riluzole (20 and 40 g/10 L per rat) only delayed morphine tolerance for 2 days and 1 day, respectively; however, riluzole (80 g/10 L per rat) delayed morphine tolerance for
5 days. On the other hand, the total analgesic effect of morphine (AUC of %MPE) significantly increased in animals treated with morphine riluzole (80 g/10 L per rat). The results shown in Figure 5 indicate that the most effective dose of riluzole in morphine tolerance (80 g/10 L per rat) did not have a significant analgesic effect. Previous studies have indicated that the most important mechanism of action for riluzole is its effect on glutamatergic transmission: riluzole inhibited glutamate release from presynaptic terminals through a mechanism linked to G-protein signaling.43 Riluzole also affects neurotransmission mediated by AMPA/kainate receptors and reduces NMDA-evoked responses.44,45 Finally, riluzole enhances high-affinity glutamate uptake in rat spinal cord synaptosomes in vitro and after treatment in vivo.46,47 A recent study demonstrated that riluzole significantly increased glutamate uptake mediated by transporters, such as GLAST, GLT1, and EAAC1.48 As previously mentioned, it has been shown that the neuroprotectant riluzole has a direct inhibitory action on PKC, which has a critical role in the development of morphine tolerance,49 and it has been confirmed that riluzole at a concentration 30 mM protects against excitotoxic neuronal injury induced by NMDA or kainate in mouse cortical cultures.42 Furthermore, there is evidence that riluzole inhibits the electrophysiological responses mediated by rat kainate and NMDA receptors expressed in Xenopus oocytes.50 In this study, we examined the analgesic effect of three other doses of morphine in combination with saline or minocycline (120 g/10 L per rat) or riluzole (80 g/10 L per rat) on the 9th day (the day after morphine tolerance in the control group). The significant shift to the right in the dose-response curve and ED50 for animals receiving morphine (10 mg/kg) saline (10 L/rat) showed that these animals were tolerant to morphine analgesia. The ED50 in those rats that received morphine minocycline or riluzole indicated that these drugs could prevent tolerance to the analgesic effect of morphine and that they showed significant differences in ED50 compared with the control group. Furthermore, our results showed that dizocilpine (1 g/10 L per rat), a noncompetitive NMDA receptor antagonist, which has a well-established effect on opioid tolerance, could attenuate morphine tolerance and the effects of minocycline and riluzole were similar to this drug, thus dizocilpine could help to elucidate the mechanisms involved in morphine tolerance. Dizocilpine alone did not have a significant analgesic effect; however, the global analgesic effect of this drug when combined with morphine was greater than morphine minocycline or morphine riluzole. Similarly, dizocilpine delayed morphine tolerance from the 8th day to the 14th day. In conclusion, our results showed that ICV administration of minocycline and riluzole could delay morphine tolerance; however, further studies are needed to clarify the exact mechanism involved.
REFERENCES
1. Barton C, Basbaum AI, Fields HL. Dissociation of supraspinal and spinal actions of morphine: a quantitative evaluation. Brain Res 1980;188:48798 2. Trujillo KA, Akil H. Inhibition of morphine tolerance and dependence by the NMDA receptor antagonist MK-801. Science 1991;251:857 3. Trujillo KA. Effects of non-competitive N-methyl-d-aspartate receptor antagonists on opiate tolerance and physical dependence. Neuropsychopharmacology 1995;13:3017 4. Trujillo KA. The neurobiology of opiate tolerance, dependence and sensitization: mechanisms of NMDA receptor-dependent synaptic plasticity. Neurotox Res 2002;4:37391 5. Mao J. NMDA and opioid receptors: their interactions in antinociception, tolerance and neuroplasticity. Brain Res Rev 1999;30:289 304 6. Habibi-Asl B, Hassanzadeh K. Effects of ketamine and midazolam on morphine induced dependence and tolerance in mice. DARU 2004;12:1015 7. Habibi-Asl B, Hassanzadeh K, Moosazadeh S. Effects of ketamine and magnesium on morphine induced tolerance and dependence in mice. DARU 2005;13:110 5 8. Habibi-Asl B, Hassanzadeh K, Khezri E, Mohammadi S. Evaluation the effects of dextromethorphan and midazolam on morphine induced tolerance and dependence in mice. Pak J Biol Sci 2008;11:1690 5 9. Rothman SM, Olney JW. Glutamate and the pathophysiology of hypoxic-ischemic brain damage. Ann Neurol 1986;19:10511 10. Mao J, Mayer DJ, Hayes RL, Lu J, Price DD. Differential roles of NMDA and non-NMDA receptor activation in induction and maintenance of thermal hyperalgesia in rats with painful peripheral mononeuropathy. Brain Res 1992;598:271 8 11. Mao J, Price DD, Mayer DJ. Mechanisms of hyperalgesia and opiate tolerance: a current view of their possible interactions. Pain 1995;62:259 74 12. Mao J, Price DD, Zhu J, Lu J, Mayer DJ. The inhibition of nitric oxide-activated poly(ADP-ribose) synthetase attenuates transsynaptic alteration of spinal cord dorsal horn neurons and neuropathic pain in the rat. Pain 1997;72:355 66 13. Montague PR, Gancayco CD, Winn MJ, Marchase RB, Friedlander MJ. Role of NO production in NMDA receptor mediated neurotransmitter release in cerebral cortex. Science 1994;263: 9737 14. Pogun S, Dawson V, Kuhar MJ. Nitric oxide inhibits 3Hglutamate transport in synaptosomes. Synapse 1994;18:21 6 15. Trotti D, Rossi D, Gjelsdal O, Levy LM, Racagni G, Danbolt NC. Peroxynitrite inhibits glutamate transporter subtypes. J Biol Chem 1996;271:5976 9 16. Yrja nheikki J, Keina nen R, Pellikka M, Kfelt TH, Koistinaho J. Tetracyclines inhibit microglial activation and are neuroprotective in global brain ischemia. Proc Natl Acad Sci USA 1998;95:15769 74 17. Yrja nheikki J, Tikka T, Keinanen R, Goldsteins G, Chan PH, Koistinaho JA. Tetracycline derivative, minocycline, reduces inflammation and protects against focal cerebral ischemia with a wide therapeutic window. Proc Natl Acad Sci USA 1999;96:13496 500 18. Tikka TM, Koistinaho JE. Minocycline provides neuroprotection against N-methyl-d-aspartate neurotoxicity by inhibiting microglia. J Immunol 2001;166:752733 19. Sadowski T, Steinmeyer J. Minocycline inhibits the production of inducible nitric oxide synthase in articular chondrocytes. J Rheumatol 2001;28:336 40 20. Jose CG, Javier E, Mara CG, Francisco J, Fernandez G, Jose SP. Neuroprotectant minocycline depresses glutamatergic neurotransmission and Ca2 signalling in hippocampal neurons. Eur J Neurosci 2007;26:248195 21. Debono MW, LeGuern L, Canton T. Inhibition by riluzole of electrophysiological responses mediated by rat kainate and NMDA receptors expressed in Xenopus oocytes. Eur J Pharmacol 1993;235:2839 22. Sepulveda J, Astorga JG, Contreras E. Riluzole decreases the abstinence syndrome and physical dependence in morphinedependent mice. Eur J Pharmacol 1999;379:59 62 23. Paxinos G, Watson C. The rat brain in stereotaxic coordinates. 4th ed. London: Academic Press, 1998
941
24. Eddy NB, Leimback D. Synthetic analgesics. II. Dithienylbutenyl and dithienylbutylamines. J Pharmacol Exp Ther 1953;107: 38593 25. McCarthy RJ, Kroin JS, Tuman KJ, Penn RD, Ivankovich AD. Antinociceptive potentiation and attenuation of tolerance by intrathecal co-infusion of magnesium sulfate and morphine in rats. Anesth Analg 1998;86:830 6 26. Elliott K, Kest B, Man A, Inturrisi, CE. N-methyl-d-aspartate (NMDA) receptors, mu and kappa opioid tolerance and perspectives on new analgesic drug development. Neuropharmacology 1995;13:34756 27. Eric JN, Aghajanian GK. Molecular and cellular basis of addiction. Science 1997;278:58 63 28. George FK. Neuroadaptive mechanisms of addiction: study on the extended amigdala. Neuropsychopharmacology 2003;13:44252 29. Liu JG, Anand KJ. Protein kinases modulate the cellular adaptations associated with opioid tolerance and dependence. Brain Res Brain Res Rev 2000;38:119 30. Wang X, Zhu S, Drozda M, Zhang W, Stavrovskaya IG, Cattaneo E. Minocycline inhibits caspase-independent and -dependent mitochondrial cell death pathways in models of Huntingtons disease. Proc Natl Acad Sci USA 2003;100:104837 31. Zhu S, Stasvrovskaya IG, Drozda M, Kim BYS, Ona V, Li M, Sarang S. Minocycline inhibits cytochrome C release and delays progression of amyotrophic lateral sclerosis in mice. Nature 2002;417:74 8 32. Hunter CL, Quintero EM, Gilstrap L, Bhat NR. Granholm A. Minocycline protects basal forebrain cholinergic neurons from mu p75-saporin immunotoxic lesioning. Eur J Neurosci 2004;19:330516 33. He Y, Appel S, Le W. Minocycline inhibits microglial activation and protects nigral cell after 6-hydroxydopamine injection into mouse striatum. Brain Res 2001;909:18793 34. Kraus RL, Pasiweczny R, Lariosa WK, Turner MS, Jiang A, Trauger JW. Antioxidant properties of minocycline. Neuroprotection in an oxidative stress assay and direct radicalscavenging activity. J Neurochem 2005;94:819 27 35. Stirling DP, Khodarahmi K, Steeves JD, Tetzlaff W. Minocycline as neuroprotective agent. Neuroscientist 2005;11:308 22 36. Watkins LR, Hutchinson MR, Johnston IN, Maier SF. Glia: novel counterregulators of opioidanalgesia. Trends Neurosci 2005;28:6619 37. Mika J, Wawrzczak-Bargiela A, Osikowicz M, Makuch W, Przewlocka B. Attenuation of morphine tolerance by minocycline and pentoxifylline in naive and neuropathic mice. Brain Behav Immun 2009;23:75 84
38. Bryson HM, Fulton B, Benfield P. Riluzole: a review of its pharmacodynamic and pharmacokinetic properties and therapeutic potential in amyotrophic lateral sclerosis. Drugs 1996;52:549 63 39. Mitsumoto H. Riluzolewhat is its impact in our treatment and understanding of amyotrophic lateral sclerosis? Ann Pharmacother 1997;31:779 81 40. Urbani A, Belluzzi O. Riluzole inhibits the persistent sodium current in mammalian CNS neurons. Eur J Neurosci 2000;12:356774 41. Zona C, Siniscalchi A, Mercuri NB, Bernardi G. Riluzole interacts with voltage-activated sodium and potassium currents in cultured rat cortical neurons. Neuroscience 1998;85:931 8 42. Noh KM, Hwang JY, Shin HC, Koh JY. A novel neuroprotective mechanism of riluzole: direct inhibition of protein kinase C. Neurobiol Dis 2000;7:375 83 43. Wang SJ, Wang KY, Wang WC. Mechanisms underlying the riluzole inhibition of glutamate release from rat cerebral cortex nerve terminals (synaptosomes). Neuroscience 2004;125:191201 44. Albo F, Pieri M, Zona C. Modulation of AMPA receptors in spinal motor neurons by the neuroprotective agent riluzole. J Neurosci Res 2004;78:200 7 45. De Sarro G, Siniscalchi A, Ferreri G, Gallelli L, De Sarro A. NMDA and AMPA/kainate receptors are involved in the anticonvulsant activity of riluzole in DBA/2 mice. Eur J Pharmacol 2000;408:2534 46. Azbill RD, Mu X, Springer JE. Riluzole increases highaffinity glutamate uptake in rat spinal cord synaptosomes. Brain Res 2000;871:175 80 47. Dunlop J, Beal McIlvain H, She Y, Howland DS. Impaired spinal cord glutamate transport capacity and reduced sensitivity to riluzole in a transgenic superoxide dismutase mutant rat model of amyotrophic lateral sclerosis. J Neurosci 2003;23:1688 96 48. Fumagalli E, Funicello M, Rauen T, Gobbi M, Mennini T. Riluzole enhances the activity of glutamate transporters GLAST, GLT1 and EAAC1. Eur J Pharmacol 2008;578:171 6 49. Rang HP, Dale MM, Ritter JM. Pharmacology. 5th ed. London: Churchil Livingstone, 2005 50. Koh JY, Kim DK, Hwang JY, Kim YH, Seo JH. Antioxidative and proapoptotic effects of riluzole on cultured cortical neurons. J Neurochem 1999;72:716 23
942
The Effect of a Peripheral Block on Inammation-Induced Prostaglandin E2 and Cyclooxygenase Expression in Rats
He le ` ne Beloeil, MD, PhD Marc Gentili, MD, PhD Dan Benhamou, MD Jean-Xavier Mazoit, MD, PhD
BACKGROUND: Peripheral inflammatory pain is associated with an upregulation of spinal cord COX-2 (cyclooxygenase-2), with a subsequent increase in central prostaglandin E2 (PGE2) levels associated with the development of hyperalgesia. In this study, we evaluated the effect of bupivacaine administered via a nerve block or via a systemic route on the spinal expression of PGE2 and COX in a model of peripheral inflammation in rats. METHODS: All rats randomly received three injections: 1) a left subcutaneous hindpaw injection (0.2 mL with either carrageenan 2% w/v or saline), 2) a left sciatic block (0.2 mL with either bupivacaine 0.5% or saline), and 3) a systemic injection (subcutaneous interscapular with 0.2 mL with either bupivacaine 0.5% or saline). Local edema, thermal, and mechanical hyperalgesia as well as cerebrospinal fluid PGE2 concentration and COX-1 and COX-2 expression in the spinal cord in dorsal root ganglions were measured. RESULTS: We confirmed that a bupivacaine block attenuates hyperalgesia and local inflammation in a model of inflammatory pain. This effect was associated with an inhibition of the increase in COX-2 expression induced by peripheral inflammation in dorsal root ganglions and cord. The subsequent production of PGE2 in cerebrospinal fluid was also impaired. Systemic bupivacaine did not modify either the hyperalgesia and local inflammation or COX expression. CONCLUSION: These results constitute a key element strongly suggesting that local anesthetics act at a different level when administered systematically or via a nerve block.
(Anesth Analg 2009;109:94350)
rostaglandin E2 (PGE2) is the predominant neurotransmitter released after surgical trauma and has been associated with inflammation and pain.1,2 Human and animal data have demonstrated an increase in spinal PGE2 after peripheral inflammation positively correlated with pain.3,4 Animal models of peripheral inflammation have demonstrated an upregulation of spinal cord COX-2 (cyclooxygenase-2) with a subsequent increase in central PGE2 levels associated with the development of allodynia and hyperalgesia.5,6 The prevention of hyperalgesia by analgesic drugs is associated with the prevention of the changes in PGE2 concentration in cerebrospinal fluid (CSF) in animals.7 Local anesthetics (LAs) can inhibit components of the inflammatory response. We previously reported that LAs via a nerve block could attenuate the inflammatory hindpaw edema and hyperalgesia induced by hindpaw injection of carrageenan (CARR) in rats.8 In the same study, the systemic administration of bupivacaine was ineffective in
From the Univ Paris-Sud, Laboratoire dAnesthe sie, UPRES EA 3540, F-94276 Le Kremlin Bice tre, France. Accepted for publication April 27, 2009. Address correspondence and reprint requests to He le ` ne Beloeil, MD, PhD, Laboratoire dAnesthe sie, UPRES EA 3540, Faculte de Me decine, Universite de Paris-Sud, 94276 Le Kremlin Bice tre, France. Address e-mail to helene.beloeil@bct.aphp.fr. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181aff25e
preventing hindpaw edema and hyperalgesia, but it regulated the systemic inflammatory response elicited by the peripheral inflammation in rats. Bupivacaine was indeed effective in suppressing systemic tumor necrosis factor- (TNF) and interleukin 1-. The mechanisms by which LAs affect inflammation and hyperalgesia are still not clear. Although the involvement of both spinal PGE2 and COX in pain transmission has been demonstrated, the effect of LAs on the release of these pain mediators into the CSF of animals with peripheral inflammatory hyperalgesia has been poorly investigated. The purpose of this study, therefore, was to evaluate the effect of bupivacaine administered via a nerve block or via a systemic route on the spinal expression of PGE2 and COX in a model of peripheral inflammation in rats.
METHODS
Young adult male Sprague-Dawley rats weighing 200 250 g were used. Guidelines of the International Association for the Study of Pain were followed.9 Our institutional committee on research in animals approved this study. The animals were kept on a 12-h light/dark cycle with free access to food and water. The rats were handled repeatedly for at least 3 days before the experiments to habituate them to the investigators and the testing paradigm.
943
Solutions
CARR was prepared fresh before each experiment (0.2 mL of 2% w/v solution of lambda CARR in saline (Sigma Chemical Co.). Bupivacaine 0.5% w/v (5 g/L) was used as the LA.
Table 1. Summary of Treatment Groups Left hindpaw (SC)

Saline Saline Saline CARR CARR CARR
Groups
Control Systemic bupi Bupi block CARR CARR bupi block CARR systemic bupi
Left sciatic area

Saline Saline Bupi Saline Bupi Saline
Systemic (back)
Saline Bupi Saline Saline Saline Bupi
Sciatic Blockade Technique

Before nerve block injections, rats were anesthetized briefly with isoflurane (2% 4% inspired concentration in 100% oxygen) by face mask. The block was initiated by introducing a 23-gauge needle posteromedially to the greater trochanter pointed in an anteromedial direction. Once bone was touched, the needle was withdrawn 1 mm, and the drug was injected. The final volume of injectate was 0.2-mL test solution. The left leg was always used for blocks. The efficacy of the block was tested by measuring paw withdrawal latency in response to a radiant thermal stimulus applied using a Hargreaves-type testing device (cf. Behavioral Measurements section) 30 min after the injection. Failure to remove the hindpaw after 22 s was regarded as a dense thermal nocifensive blockade. All animals receiving a sciatic blockade with bupivacaine had a dense blockade 30 min after the injection of bupivacaine. In a separated group of eight animals receiving a bupivacaine sciatic block, bupivacaine blood levels were determined 6 h later using gas chromatography. The blood levels achieved were below the level of detection or very low (0.015 mg/L), excluding a systemic effect of the bupivacaine administered via a nerve block.
Bupi bupivacaine; CARR carrageenan; SC subcutaneous.
blockade. This test was repeated three times on each hindpaw for each rat. The development of mechanical hypersensitivity after hindpaw inflammation was assessed by the application of calibrated von Frey filaments. Animals were placed on a mesh floor in individual plastic boxes and allowed to acclimate to their environment. von Frey filaments were then applied vertically to the plantar surface of both hindpaws. Filaments were applied three times over 2 s. If no response was elicited, a larger diameter filament was applied in the same manner. The filaments were applied in increasing order until brisk withdrawal or paw flinching was elicited, which was considered a positive response. This withdrawal threshold was determined three times, and the mean withdrawal threshold was used for data analysis.
Paw Circumference
To evaluate the edema, we used a technique previously described.10 The paw circumference (PC) was measured by a thread, to the nearest millimeter, at the metatarsal level.
Experimental Groups
Animals were randomly assigned to one of six experimental groups (n 5 8/group) as described in Table 1. Each animal received three injections at time 0: 1) a left subcutaneous hindpaw injection (0.2 mL with either CARR 2% w/v or saline), 2) a left sciatic block (0.2 mL with either bupivacaine 0.5% or saline), and 3) a systemic injection (subcutaneous interscapular with 0.2 mL with either bupivacaine 0.5% or saline). Bupivacaine was injected at only one site (sciatic or back) in an animal.
PGE2 in CSF
CSF was collected 6 h after drug injections. The amounts of PGE2 were measured with a commercial enzyme-linked immunosorbent assay kit (R&D systems) according to the manufacturers instructions. The assay detection limits were 10 pg/mL. At the end of each experiment, rats were killed with an overdose of pentobarbital (100 mg/kg administered IV).
COX-1 and COX-2 Behavioral Measurements

All behavioral tests were performed by a single investigator. Nociceptive responses to a thermal stimulus were evaluated by measuring paw withdrawal latency in response to a radiant thermal stimulus applied using a Hargreaves-type testing device (Ugo Basile, Milan, Italy). It consists of a movable infrared source, which the operator glides below a glass pane, upon which a three-compartment enclosure for the rat is positioned. When the rat withdraws its paw, the instrument automatically detects the withdrawal latency to the nearest 0.1 s. The paw withdrawal latency was evaluated 6 h after initial injections. Failure to remove the hindpaw after 22 s was regarded as dense thermal nocifensive
944
Nerve Block, COX-2, and PGE2
COX-1 and COX-2 expression was measured by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR). The animals were anesthetized using isoflurane and decapitated for exsanguination. The spinal cord (L3L5) separated in right and left portions and the corresponding dorsal root ganglions (DRG) were rapidly immersed in RNAlater (Quiagen, Courtaboeuf France) and kept frozen at 80C. After tissue homogenization in ice-cold tubes, total RNA was extracted using Trizol (Invitrogen, Cergy Pontoise, France). Reverse transcription was performed using 0.5-g total RNA using the Superscript II reverse transcriptase (Invitrogen, France). TaqMan real-time PCR assays for COX-1, COX-2 microsomial ribonucleic acid (mRNA), and 18S
Figure 1. Paw circumference (mm) assessing edema 6 h after injection in the hindpaw injection side. Results are expressed as median with 25th and 75th percentiles. *P 0.01 versus control group. P 0.01 versus carr group. Bupi bupivacaine; carr carrageenan; syst systemic.
ribosomial ribonucleic acid (rRNA) were performed on an ABI Prism 7300 sequence detector (Applied Biosystems). After 2 min at 50C and 10 min at 95C, 40 cycles of denaturation/annealing-extending (95C for 15 s, 60C for 1 min) were performed. Primers and probes were purchased from Applied Biosystems. The oligonucleotide primer pairs were purchased from Applied Biosystems, Courtaboeuf France: Rn 00566881_m1 and Rn00568225_1 for COX-1 and COX-2, respectively. COX-1 and COX-2 mRNA are expressed relative to 18S rRNA.
side, no significant differences were observed among the groups (data not shown).
Thermal Nociceptive Withdrawal Latencies and von Frey Filament Mechanical Withdrawal Thresholds
Six hours after injection of CARR, a significant decrease in heat withdrawal latency and in the mechanical withdrawal threshold were observed in the CARR group compared with that in the control group, indicating thermal hyperalgesia in the CARR group (Figs. 2 and 3). In the groups receiving hindpaw CARR and a sciatic block with bupivacaine, heat withdrawal latency and mechanical threshold were significantly greater than in the CARR group, indicating partial prevention of CARR-induced hyperalgesia. The motor blockade observed after the injection of the nerve block was not present 6 h later. Systemic bupivacaine was ineffective in prevention of CARR-induced mechanical hyperalgesia (Figs. 2 and 3). In the nonhindpaw injection side, no significant differences were observed among the groups (data not shown).
As the behavioral and the PC data were not normally distributed, differences among groups were assessed using nonparametric tests (KruskalWallis and MannWhitney U-test with the Bonferroni correction). The results are expressed as the median with 25th and 75th percentiles. PGE2 production in CSF, COX-1, and COX-2 expression was compared among groups using the KruskalWallis test followed by a MannWhitney U-test with the Bonferroni correction. The results are expressed as median with 25th and 75th percentiles. A P value below 0.05 was considered the minimum level of statistical significance.
Production of PGE2 in CSF

The basal PGE2 concentration in CSF was 112 (96 246) pg/mL (median and interquartile range) (Fig. 4). In the group receiving CARR, PGE2 production was increased and was significantly different from the control group. In the groups receiving hindpaw saline and bupivacaine (systemic or block), PGE2 production was not significantly different from the production observed in the control group. In the group receiving hindpaw CARR and a bupivacaine block, PGE2 production was not significantly different from the production observed in the control group, showing that a bupivacaine block inhibited the increased production of PGE2 induced by a hindpaw CARR injection. In the group receiving hindpaw
RESULTS
Evaluation of Hindpaw Edema by PC
Six hours after the injection of CARR, a significantly increased PC was observed in the CARR group compared with that in the control group (Fig. 1). In the groups receiving CARR and an ipsilateral sciatic block with bupivacaine, the mean PC was significantly less than that in the CARR group. The group receiving CARR plus systemic bupivacaine had mean PCs that were not significantly different from those of the CARR group (Fig. 1). In the nonhindpaw injection
945
Figure 2. Thermal nociceptive withdrawal latency (s) assessing thermal hyperalgesia in the hindpaw injection side 6 h after injection. Results are expressed as median with 25th and 75th percentiles. *P 0.01 versus control group. P 0.01 versus carr group. Bupi bupivacaine; carr carrageenan; syst systemic.
Figure 3. Withdrawal threshold (g) to von Frey filaments assessing mechanical hyperalgesia in the hindpaw injection side, 6 h after injection. Results are expressed as median with interval range. *P 0.01 versus control group. P 0.01 versus carr group. Bupi bupivacaine; carr carrageenan; syst systemic.
CARR and systemic bupivacaine, PGE2 production was not significantly different from the production observed in the CARR group.
COX-1 and COX-2 Expression in Cord and DRGs

In control animals, COX-1 was expressed in both cord and DRG at low levels, whereas COX-2 was almost not expressed (Tables 2 and 3). The COX mRNA/18S rRNA level was unchanged in these control animals. The expression of COX-1 was not modified by any treatment at any time. Contrary to COX-1, COX-2 mRNA expression was increased in the animals treated with CARR. Six hours after CARR injection, COX-2 expression significantly increased in both left (side of CARR injection) DRGs and cord (Fig. 5). Importantly, COX-2 mRNA expression markedly increased in the opposite side of
946
the spinal cord (Fig. 6) but not in the opposite DRG (Fig. 6). In animals treated with a bupivacaine block, the increase of COX-2 mRNA was impaired in the ipsilateral cord and DRG. This was significantly less compared with the CARR group. Moreover, a bupivacaine block was able to impair the increase in COX-2 expression in cord on the opposite side. The increase of expression in COX-2 after CARR was not modified by bupivacaine via a systemic route.
DISCUSSION
In this study, we confirmed that a bupivacaine block attenuates thermal hyperalgesia, mechanical allodynia, and local inflammation in a model of inflammatory pain. We and others had already shown such
Figure 4. Spinal prostaglandin E2

(PGE2) production at 6 h in cerebrospinal fluid (CSF). Results are expressed as median with interval range. *P 0.01 versus control group. Bupi bupivacaine; carr carrageenan; syst systemic.
Table 2. Expression of Cyclooxygenase (COX)-1 Relative to 18S in the Spinal Cord and Dorsal Root Ganglions (DRGs) Measured by Quantitative Reverse-Transcriptase Polymerase Chain Reaction (qRT-PCR) Cord left
Control Systemic bupi Bupi block CARR CARR bupi block CARR systemic bupi 0.69 (0.420.85) 0.55 (0.360.91) 0.32 (0.160.4) 0.51 (0.260.58) 0.5 (0.021.1) 0.55 (0.290.68)
Cord right
0.56 (0.370.86) 0.79 (0.641.05) 0.56 (0.430.85) 0.59 (0.561.63) 0.3 (0.020.9) 0.34 (0.230.56)
DRG left
0.18 (0.10.31) 0.25 (0.080.46) 0.36 (0.130.56) 0.2 (0.10.58) 0.25 (0.010.79) 0.2 (0.030.2)
DRG right
0.21 (0.150.28) 0.36 (0.030.76) 0.2 (0.160.54) 0.28 (0.171.7) 0.4 (0.20.8) 0.2 (0.120.9)
Results are presented as median (interquartile range). Bupi bupivacaine; CARR carrageenan.
Table 3. Expression of Cyclooxygenase (COX)-2 Relative to 18S in the Spinal Cord and Dorsal Root Ganglions (DRGs) Measured by Quantitative Reverse-Transcriptase Polymerase Chain Reaction (qRT-PCR) Cord left
Control Systemic bupi Bupi block CARR CARR bupi block CARR systemic bupi 0.29 (0.10.33) 0.68 (0.111.27) 0.69 (0.250.96) 37.76 (27.545.24)* 6.69 (5.898.96)* 42.56 (40.2345.63)*
Cord right
0.1 (0.080.36) 0.8 (0.561.77) 0.30 (0.020.3) 7.80 (4.2510.15)* 2.03 (0.62.36)* 8.96 (7.899.34)*
DRG left
0.15 (0.070.4) 0.06 (0.050.08) 0.55 (0.450.85) 3.8 (3.225)* 0.36 (00.56) 5.69 (5.636.56)*
DRG right
0 (00.01) 0.09 (0.080.1) 0.01 (0.010.59) 0.56 (0.50.6) 0.1 (0.050.5) 0.5 (0.470.56)
Results are presented as median (interquartile range). Bupi bupivacaine; CARR carrageenan. * Versus control, P 0.01. Versus CARR, P 0.01.
a property,8,11 and we wanted to better explain the mechanisms involved. This effect was associated with an inhibition of the increase in COX-2 expression induced by peripheral inflammation in DRG and cord. The subsequent production of PGE2 in CSF was also impaired. Systemic bupivacaine did not modify either the hyperalgesia and local inflammation or the COX expression. As already published,5,12,13 the inflammatory process induced by CARR did not increase the expression of COX-1. COX-1 mRNA was observed at a low concentration, but the concentration did not increase
at any time in any group with any treatment. In contrast, COX-2 expression was significantly increased both in DRG and in cord 6 h after CARR injection.6,12,13 Interestingly, COX-2 expression was also increased in the opposite side of the spinal cord. Activation of TNF-, interleukin-1, and COX-2 in contralateral DRGs has been shown, and this is a remarkable observation because somatic afferent peripheral fibers project to ipsilateral DRGs and are not thought to have contralateral connections at the level of DRGs.12,13 We previously observed that hindpaw CARR-induced activation of cytokines and p-p38
947
Figure 6. Expression of cyclooxygenase (COX)-2 microsomial Figure 5. Expression of cyclooxygenase (COX)-2 microsomial
ribonucleic acid (mRNA) (relative to 18S ribosomial ribonucleic acid [rRNA]) in the left dorsal root ganglion (DRG) and spinal cord measured by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR). *P 0.01 versus control group. P 0.01 versus carr group. ribonucleic acid (mRNA) (relative to 18S ribosomial ribonucleic acid [rRNA]) in the right dorsal root ganglion (DRG) and spinal cord measured by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR). *P 0.01 versus control group. P 0.01 versus carr group.
mitogen-activated protein kinase in bilateral lumbar DRGs involves a regional or segmental, rather than systemic, mechanism.14 Potential mechanisms of this bilateralization are not clearly understood and need further study. In this study, there was minor but no significant hyperalgesia on the contralateral side of the local inflammation created by CARR. A bilateralization of hyperalgesia has, however, been observed 2 h after CARR injection in the rat hindpaw.15 Subsequent PGE2 production in CSF was also increased after inflammation as has been described in animals16 and humans after surgery.3 As previously reported, LAs, via a nerve block, attenuated the inflammatory hindpaw edema and hyperalgesia induced by hindpaw injection of CARR in rats. This was associated with an inhibition of the inflammatory-induced increase of COX-2 expression in DRG and spinal cord and subsequent PGE2 production in CSF. Many other markers of central and peripheral nerve sensitization are blocked by a nerve block, such as cytokines. We previously observed that a bupivacaine nerve block regulated the systemic cytokine response elicited by peripheral inflammation in rats.8 A prolonged nerve block itself (70% ethanol) can decrease not only the local inflammatory reaction
948
observed after CARR injection but also its systemic consequences.17 On the other hand, in vitro studies have reported a direct specific effect of LAs on the eicosanoid system. Lidocaine was shown in vitro to inhibit PGE2 release18 and PLA2 activity.19,20 Products of COX (PG1 and thromboxane A2) were inhibited by topical lidocaine/prilocaine applied on burned skin in ex vivo experiments.21 An effect of bupivacaine on PGE2 EP1 receptor signaling has also been observed in vitro.22 In an animal model of neuropathic pain, an interaction between intrathecal lidocaine and the eicosanoid system in the spinal cord was reported by Ma et al.23 Although these studies suggested that LAs may inhibit prostaglandin production, Kroin et al.24 did not observe any decrease in CSF PGE2 in sham-operated rats using spinal bupivacaine. They also observed that spinal bupivacaine produced a large postsurgical CSF PGE2 upregulation, isoflurane a moderate increase, and propofol did not allow any increase in CSF PGE2. Samad et al.2 were the first to report that a sciatic block administered before complete Freunds adjuvant-induced inflammation could reduce but not eliminate COX2-mRNA induction in the spinal cord or PGE2 levels in CSF. To prolong the effect of the sciatic block, however, they used bupivacaine microspheres, which have been reported to
produce inflammation.25 A recent study reported that bupivacaine administered locally in patients undergoing dental surgery had no effect on local PGE2 concentration and COX-2 expression when compared with a COX-2 inhibitor and less effect when compared with local lidocaine.26 However, the lack of a control group in this particular study did not allow us to conclude that bupivacaine had an effect on the increased PGE2 and COX-2 levels usually observed after surgery without analgesia in experimental studies. Although a bupivacaine nerve block decreased spinal PGE2 and COX-2 after peripheral inflammation, systemic bupivacaine modified neither hyperalgesia and local inflammation nor COX expression in our study. We previously reported that systemic administration of bupivacaine was ineffective in preventing hindpaw edema and hyperalgesia, but it regulated the systemic inflammatory response elicited by peripheral inflammation in rats.8 Many arguments strongly suggest that LAs act at a different level when administered systematically or via a nerve block. Studies in humans have demonstrated little effect of IV lidocaine on normal pain thresholds but profound effects on hyperalgesia-related phenomena.27 Intravenous lidocaine can specifically inhibit visceral nociceptive reflexes and spinal neurons in the rat.28 In patients, continuous IV administration of lidocaine during and after abdominal surgery improves rehabilitation and shortens hospital stay.29,30 Meanwhile, a nerve block can inhibit pain, edema, and hyperalgesia, which are local and systemic consequences of a peripheral inflammation. This effect is thought to be related to the conduction blockade initiated by the nerve block. Inhibition of axonal transport by LAs was only reported in vitro.31 Recently, we showed that the proinflammatory cytokine TNF- was transported along the axon after peripheral inflammation in rats, and this slow retrograde (from the periphery to the DRG) transport was inhibited by bupivacaine in a dose-dependent manner (Deruddre et al. submitted). This phenomenon could explain the effects of a nerve block on inflammation, and these effects could be of interest in the treatment of inflammatory disease. For example, patients with complex regional pain syndrome frequently develop bilateral pain and neurovascular signs and symptoms after a unilateral injury, which could be treated by a prolonged nerve block.32 We found that the antihyperalgesic and antiinflammatory effects of a bupivacaine block in a peripheral model of inflammation was associated with an inhibition of the increase in COX-2 expression induced by peripheral inflammation in DRG and cord. The subsequent production of PGE2 in CSF was also impaired. Systemic bupivacaine did not modify either the hyperalgesia and local inflammation or COX expression. These results constitute one more key element strongly suggesting that LAs act at a different level when administered systematically or via a nerve block.
ACKNOWLEDGMENTS The authors thank Mrs. Re gine Le Guen for her technical assistance. REFERENCES
1. Baba H, Kohno T, Moore KA, Woolf CJ. Direct activation of rat spinal dorsal horn neurons by prostaglandin E2. J Neurosci 2001;21:1750 6 2. Samad TA, Moore KA, Sapirstein A, Billet S, Allchorne A, Poole S, Bonventre JV, Woolf CJ. Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature 2001;410:4715 3. Buvanendran A, Kroin JS, Berger RA, Hallab NJ, Saha C, Negrescu C, Moric M, Caicedo MS, Tuman KJ. Upregulation of prostaglandin E2 and interleukins in the central nervous system and peripheral tissue during and after surgery in humans. Anesthesiology 2006;104:40310 4. Ebersberger A, Grubb BD, Willingale HL, Gardiner NJ, Nebe J, Schaible HG. The intraspinal release of prostaglandin E2 in a model of acute arthritis is accompanied by an up-regulation of cyclo-oxygenase-2 in the spinal cord. Neuroscience 1999; 93:775 81 5. Guay J, Bateman K, Gordon R, Mancini J, Riendeau D. Carrageenan-induced paw edema in rat elicits a predominant prostaglandin E2 (PGE2) response in the central nervous system associated with the induction of microsomal PGE2 synthase-1. J Biol Chem 2004;279:24866 72 6. Hay CH, Trevethick MA, Wheeldon A, Bowers JS, de Belleroche JS. The potential role of spinal cord cyclooxygenase-2 in the development of Freunds complete adjuvant-induced changes in hyperalgesia and allodynia. Neuroscience 1997;78:84350 7. Bianchi M, Martucci C, Ferrario P, Franchi S, Sacerdote P. Increased tumor necrosis factor-alpha and prostaglandin E2 concentrations in the cerebrospinal fluid of rats with inflammatory hyperalgesia: the effects of analgesic drugs. Anesth Analg 2007;104:949 54 8. Beloeil H, Ababneh Z, Chung R, Zurakowski D, Mulkern RV, Berde CB. Effects of bupivacaine and tetrodotoxin on carrageenan-induced hind paw inflammation in rats (Part 1): hyperalgesia, edema, and systemic cytokines. Anesthesiology 2006;105:128 38 9. Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 1983;16:109 10 10. Fletcher D, Kayser V, Guilbaud G. Influence of timing of administration on the analgesic effect of bupivacaine infiltration in carrageenin-injected rats. Anesthesiology 1996;84:1129 37 11. Gentili ME, Mazoit JX, Samii KK, Fletcher D. The effect of a sciatic nerve block on the development of inflammation in carrageenan injected rats. Anesth Analg 1999;89:979 84 12. Ichitani Y, Shi T, Haeggstrom JZ, Samuelsson B, Hokfelt T. Increased levels of cyclooxygenase-2 mRNA in the rat spinal cord after peripheral inflammation: an in situ hybridization study. Neuroreport 1997;8:2949 52 13. Pham-Marcou TA, Beloeil H, Sun X, Gentili M, Yaici D, Benoit G, Benhamou D, Mazoit JX. Antinociceptive effect of resveratrol in carrageenan-evoked hyperalgesia in rats: prolonged effect related to COX-2 expression impairment. Pain 2008;140:274 83 14. Beloeil H, Ji RR, Berde CB. Effects of bupivacaine and tetrodotoxin on carrageenan-induced hind paw inflammation in rats (Part 2): cytokines and p38 mitogen-activated protein kinases in dorsal root ganglia and spinal cord. Anesthesiology 2006; 105:139 45 15. Estebe JP, Gentili ME, Le Corre P, Leduc C, Moulinoux JP, Ecoffey C. Contralateral effect of amitriptyline and bupivacaine for sciatic nerve block in an animal model of inflammation. Br J Anaesth 2004;93:7059 16. Ibuki T, Matsumura K, Yamazaki Y, Nozaki T, Tanaka Y, Kobayashi S. Cyclooxygenase-2 is induced in the endothelial cells throughout the central nervous system during carrageenaninduced hind paw inflammation; its possible role in hyperalgesia. J Neurochem 2003;86:318 28 17. Pham-Marcou TA, Gentili M, Asehnoune K, Fletcher D, Mazoit JX. Effect of neurolytic nerve block on systemic carrageenaninduced inflammatory response in mice. Br J Anaesth 2005; 95:243 6
949
18. Goel RK, Tavares IA, Nellgard P, Jonsson A, Cassuto J, Bennett A. Effect of lignocaine on eicosanoid synthesis by pieces of human gastric mucosa. J Pharm Pharmacol 1994;46:319 20 19. Kunze H, Nahas N, Traynor JR, Wurl M. Effects of local anaesthetics on phospholipases. Biochim Biophys Acta 1976;441:93102 20. Vadas P, Stefanski E, Pruzanski W. Potential therapeutic efficacy of inhibitors of human phospholipase A2 in septic shock. Agents Actions 1986;19:194 202 21. Jonsson A, Cassuto J, Tarnow P, Sinclair R, Bennett A, Tavares IA. Effects of amide local anaesthetics on eicosanoid formation in burned skin. Acta Anaesthesiol Scand 1999;43:618 22 22. Honemann CW, Heyse TJ, Mollhoff T, Hahnenkamp K, Berning S, Hinder F, Linck B, Schmitz W, van Aken H. The inhibitory effect of bupivacaine on prostaglandin E(2) (EP(1)) receptor functioning: mechanism of action. Anesth Analg 2001;93:628 34 23. Ma W, Du W, Eisenach JC. Intrathecal lidocaine reverses tactile allodynia caused by nerve injuries and potentiates the antiallodynic effect of the COX inhibitor ketorolac. Anesthesiology 2003;98:203 8 24. Kroin JS, Buvanendran A, Watts DE, Saha C, Tuman KJ. Upregulation of cerebrospinal fluid and peripheral prostaglandin E2 in a rat postoperative pain model. Anesth Analg 2006;103:334 43 25. Drager C, Benziger D, Gao F, Berde CB. Prolonged intercostal nerve blockade in sheep using controlled-release of bupivacaine and dexamethasone from polymer microspheres. Anesthesiology 1998;89:969 79
26. Gordon SM, Chuang BP, Wang XM, Hamza MA, Rowan JS, Brahim JS, Dionne RA. The differential effects of bupivacaine and lidocaine on prostaglandin E2 release, cyclooxygenase gene expression and pain in a clinical pain model. Anesth Analg 2008;106:3217 27. Koppert W, Zeck S, Sittl R, Likar R, Knoll R, Schmelz M. Low-dose lidocaine suppresses experimentally induced hyperalgesia in humans. Anesthesiology 1998;89:134553 28. Ness TJ. Intravenous lidocaine inhibits visceral nociceptive reflexes and spinal neurons in the rat. Anesthesiology 2000;92:168591 29. Marret E, Rolin M, Beaussier M, Bonnet F. Meta-analysis of intravenous lidocaine and postoperative recovery after abdominal surgery. Br J Surg 2008;95:1331 8 30. Herroeder S, Pecher S, Schonherr ME, Kaulitz G, Hahnenkamp K, Friess H, Bottiger BW, Bauer H, Dijkgraaf OG, Durieux ME, Hollmann MW. Systemic lidocaine shortens length of hospital stay after colorectal surgery: a double-blinded, randomized, placebo-controlled trial. Ann Surg 2007;246:192200 31. Kanai A, Hiruma H, Katakura T, Sase S, Kawakami T, Hoka S. Low-concentration lidocaine rapidly inhibits axonal transport in cultured mouse dorsal root ganglion neurons. Anesthesiology 2001;95:675 80 32. Dadure C, Motais F, Ricard C, Raux O, Troncin R, Capdevila X. Continuous peripheral nerve blocks at home for treatment of recurrent complex regional pain syndrome I in children. Anesthesiology 2005;102:38791
950
A Peripherally Acting Nav1.7 Sodium Channel Blocker Reverses Hyperalgesia and Allodynia on Rat Models of Inammatory and Neuropathic Pain
Erin McGowan, BS Scott B. Hoyt, PhD Xiaohua Li, BS Kathryn A. Lyons, BS Catherine Abbadie, PhD
BACKGROUND: Voltage-gated sodium channels (Nav1) are expressed in primary sensory neurons where they influence excitability via their role in the generation and propagation of action potentials. Recently, human genetic data have shown that one sodium channel subtype, Nav1.7, plays a major role in pain. We performed these studies to characterize the antinociceptive effects of N-[(R)-1-((R)-7-chloro-1isopropyl-2-oxo-2,3,4,5-tetrahydro-1 H -benzo[ b ]azepin-3-ylcarbamoyl)-2-(2fluorophenyl)-ethyl]-4-fluoro-2-trifluoromethyl-benzamide (BZP), a non-central nervous system (CNS) penetrant small molecule with high affinity and preferential selectivity for Nav1.7 over Nav1.8 and Nav1.5. METHODS: BZP was evaluated in rat preclinical models of inflammatory and neuropathic pain and compared with standard analgesics. Two models were used: the complete Freunds adjuvant model of inflammatory pain and the spinal nerve ligation model of neuropathic pain. BZP was also evaluated in a motor coordination assay to assess its propensity for CNS side effects. RESULTS: In preclinical models of chronic pain, BZP displayed efficacy comparable with that of leading analgesics. In the complete Freunds adjuvant model, BZP produced reversal of hyperalgesia comparable with nonsteroidal antiinflammatory drugs, and in the spinal nerve ligation model, BZP produced reversal of allodynia comparable with gabapentin and mexiletine. Unlike the CNS penetrant compounds gabapentin and mexiletine, BZP did not induce any impairment of motor coordination. CONCLUSIONS: These data suggest that a peripherally acting sodium channel blocker, preferentially acting through Nav1.7, could provide clinical relief of chronic pain without the CNS side effects typical of many existing pain treatments.
(Anesth Analg 2009;109:9518)
oltage-gated sodium channels (Nav1) are multisubunit protein complexes composed of a poreforming, voltage-sensing -subunit and two smaller -subunits.1,2 They are essential for the generation and propagation of action potentials and have been shown to play a central role in primary afferent ectopic discharges that originate from the injury site or the dorsal root ganglia (DRG).3,4 Nine different mammalian -subunits (Nav1.11.9) have been cloned (for nomenclature see Ref. 5). Over the last decade, there has been considerable debate regarding which -subunit would provide the best target for the development of novel analgesics. Compelling genetic evidence linking Nav1.7 to pain in humans has been provided by the identification of gain-of-function and loss-of-function mutations in
From the Merck Research Laboratories, Department of Pharmacology and Medicinal Chemistry Rahway, New Jersey. Accepted for publication April 13, 2009. All authors are Merck employees. Address correspondence and reprint requests to Catherine Abbadie, PhD, Alcon Labs, 6201 South Freeway, Fort Worth, TX 76134-2099. Address e-mail to catherine.abbadie@alconlabs.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181b01b02
SCN9A, the gene that encodes Nav1.7.6,7 Gain-offunction mutations that result in increased Nav1.7 activity are responsible for inherited erythromelalgia, a burning pain syndrome of the skin of the extremities in response to warmth or moderate exercise.8 10 Nine mutations of SCN9A have been identified in patients with erythromelalgia (see references in Ref. 7). These mutations cause a hyperpolarizing shift in the voltage dependence of channel activation, which in turn allows the channel to be activated by smaller depolarizations, resulting in enhanced Nav1.7 activity. In addition, recovery from inactivation (repriming) has been shown to be faster for some mutations. Alterations in the activation or deactivation of Nav1.7 might contribute to the hyperexcitability of DRG neurons that underlies some forms of hyperalgesia. Gain-of-function mutations of Nav1.7 can also cause paroxysmal extreme pain disorder, which is characterized by ocular, mandibular, and/or rectal pain.11,12 In contrast, mutations leading to loss of Nav1.7 function have been linked to a congenital insensitivity to pain.13 Individuals with these mutations have the ability to detect tactile thresholds and differences in temperature but are unable to perceive any form of pain. Functional studies show that channelopathy-associated insensitivity to pain is
951
Figure 1. A, Structure of BZP, N-[(R)-1-((R)-7-chloro-1-isopropyl-2-oxo-2,3,4,5-tetrahydro-1H-benzo[b]azepin-3-ylcarbamoyl)2-(2-fluorophenyl)-ethyl]-4-fluoro-2-trifluoromethyl-benzamide. B, Plasma levels versus time of BZP when administered at 1 mg/kg IV (n 2) or 3 mg/kg PO (n 3).
caused by loss-of-function of Nav1.7,13,14 in contrast with the genetic basis of inherited erythermalgia and paroxysmal extreme pain disorder associated with gain-of-function.9,10 Collectively, these studies provide strong genetic evidence for the importance of Nav1.7 in pain. The present studies were undertaken to characterize the antinociceptive effects of N-[(R)-1-((R)7-chloro-1-isopropyl-2-oxo-2,3,4,5-tetrahydro-1 H-benzo[b]azepin-3-ylcarbamoyl)-2-(2-fluorophenyl)ethyl]-4-fluoro-2-trifluoromethyl-benzamide (BZP), a small molecule that displayed high affinity and preferential selectivity for Nav1.7 over Nav1.8.15 BZP was screened against a broad panel of ion channels, receptors, and enzymes and was found to have no significant in vitro activity for known pain targets other than Nav1 channel blockade.16 BZP was therefore further evaluated in chronic models of inflammatory and neuropathic pain. Because sodium channel blockers currently approved for pain relief are often brainpenetrant, they frequently elicit central nervous system (CNS) side effects, such as sedation and impairment. In light of this, BZP was also evaluated in a motor coordination assay to assess the probability of CNS side effects and to determine a preclinical therapeutic index between side effects and antinociceptive effects.
METHODS
Animals
Male Sprague-Dawley rats (220 280 g) were purchased from Charles River Laboratories (Wilmington, MA). The procedures used in these studies were approved by the Institutional Animal Care and Use Committee at Merck, Rahway, NJ and adhered to the guidelines of the Committee for Research and Ethical Issues.17 Rats were habituated to the testing environment by being placed in the experiment room 3 days before sensory testing. For von Frey assessment, rats were habituated 2 h/day for 2 days before evaluation. Rats were used for only one drug and one dose.
952
Effects of a Nav1.7 Blocker in Rat Chronic Pain Models
For testing in the neuropathic pain model (spinal nerve ligation [SNL]), rats were anesthetized with isoflurane and placed on a heating pad. Using aseptic techniques, the L5 spinal nerve was exposed, ligated, and transected.18 Muscle and skin were closed with 4-0 polydiaxone and wound clips, respectively. Allodynia was assessed 4 wk after SNL surgery, and only rats that developed allodynia as defined by a significant decrease in their mechanical threshold using von Frey filaments were used. Tactile allodynia was assessed with calibrated von Frey filaments (Stoelting, Wood Dale, IL) using an up-down paradigm.19 Mechanical sensitivity was determined by applying a series of eight calibrated von Frey filaments (0.36, 0.6, 1, 2, 4, 6, 8, and 15 g) to the plantar aspect of the left hindpaw. Cutoff value corresponded to 15 g. Rats that displayed preinjury baseline measurements 15 g were not included in the study. A response was indicated by brisk withdrawal of the hindpaw. For testing in the inflammatory pain model, rats were injected with complete Freunds adjuvant (CFA) (200 L, 1:1 in saline; Sigma, St. Louis, MO) intraplantar into their left paws. Animals were tested for hyperalgesia 3 days after CFA administration, using withdrawal threshold to paw pressure (Randal-Sellito, Stoelting). In the absence of inflammation, 95% of the animals reached cutoff values of 25 g. For motor coordination assessment, rats were trained on the rota-rod (Stoelting) for 3 min at a speed of 10 rpm. For testing, the speed was set at 10 rpm for 60 s and subsequently accelerated from 6 to 60 rpm. The time required for rats to fall after the beginning of acceleration was recorded.
Compound Administration
BZP (Fig. 1) was synthesized according to established procedures.15 As described by Williams et al.,16 in stably transfected human embryonic kidney 293 cells, BZP displayed selective block of Nav1.7 over Nav1.5 and Nav1.8. In a voltage-dependent fluorescence resonance energy transfer imaging assay, BZP blocked hNav1.7, Nav1.5, and Nav1.8 with IC50s of
0.03 0.02, 0.18 0.05, and 0.3 016 (mean sd) M, respectively. Similarly, in whole cell electrophysiology experiments, BZP blocked Nav1.5 and Nav1.8 more weakly than Nav1.7.16 For PO administration, rats were manually restrained for oral gavage. BZP was diluted in DMSO:PEG300:water (15:60:25). Naproxen, indomethacin (Sigma), and rofecoxib (synthesized in house) were diluted in 0.5% methylcellulose (Sigma) in water. Mexiletine and gabapentin (Sigma) were diluted in water. Compounds were administered at the doses indicated in a formulation of 2 mL/kg of body weight. Investigators were blinded to rat treatment.
as (postdose predose)/(preinjury predose) for each rat. One hundred percent corresponds to complete reversal of hyperalgesia or allodynia, equivalent to noninjured values, and 0% corresponds to values not different from baseline postinjury (Figs. 2 and 3). Results were analyzed using a two-way analysis of variance (ANOVA) (for dose and time postdose) test followed by a Bonferroni post hoc test for multiple comparisons (Prism, Graph Pad, San Diego, CA). ED50 are calculated as doses corresponding to a 50% effect (100% effect corresponding to recovery to baseline values in the absence of injury).
Pharmacokinetic Analysis
For pharmacokinetic experiments, rats received BZP at 1 mg/kg IV (n 2) or 3 mg/kg PO (n 3). For brain penetration analysis, rats received 30 or 100 mg/kg PO BZP (n 6/group). BZP plasma and brain concentrations were determined by liquid chromatography/mass spectrometry using an ABI Sciex API 3000 mass spectrometer operated in positive ion atmospheric pressure chemical ionization mode with multiple-reaction monitoring. Brain homogenate (1:3, tissue:water) and plasma were prepared for analysis by protein precipitation with acetonitrile. Extracts were chromatographed using a DuPont Zorbax SB-C8 column (50 2 mm, 5 m) and eluted at 0.2 mL/min under isocratic conditions with acetonitrile:water (77:23) containing 5 mM ammonium formate/0.1% formic acid. Under these conditions, BZP eluted at 1.4 min. Pharmacokinetic parameters were calculated with a noncompartmental model using Watson software (Watson Software System version 7.3, Thermo Fisher, Waltham, MA). The area under the plasma concentration versus time curve from 0 to 24 h (AUC0 24) was determined using linear trapezoidal interpolation in the ascending slope and logarithmic trapezoidal interpolation in the descending slope. The portion of the AUC from the last measurable concentration to infinity (AUC0-inf) was estimated by Ct/kel, where Ct represents the last measurable concentration and kel is the elimination rate constant. The latter was determined from the concentration versus time curve at the terminal phase by linear regression of the semilogarithmic plot. Oral bioavailability (F) was estimated as the AUC0-inf ratio after oral and IV administration normalized for differences in dose. Peak plasma concentration (Cmax [M]) and its time of occurrence (Tmax [h]) were obtained by inspection of the plasma concentration-time profile. The apparent half-life (t1/2) was estimated from the slope of the terminal phase of the log plasma concentration-time curve.
RESULTS
Pharmacokinetic Prole of BZP
BZP exhibited good bioavailability in rats (F 90%) after oral administration of a 3 mg/kg dose, with a Cmax of 0.33 M obtained 2 h postdose (Tmax). BZP half-life was 2.8 h after a 1 mg/kg IV dose. Plasma concentrations of BZP were dose-proportional between 1 and 100 mg/kg PO. BZP was poorly brainpenetrant, with a brain-to-plasma ratio of 0.077. Plasma samples harvested after administration of BZP in rat models of pain or motor coordination are illustrated in Figure 3D. As an example, a 10 mg/kg PO dose gave plasma concentrations of 0.4 and 0.8 M at 2 and 4 h postdose, respectively.
Rat CFA Model of Inammatory Pain

Inflammatory pain was induced by intraplantar injection of CFA into the hindpaws of rats. Three days after CFA injection, a single PO dose (1, 3, or 10 mg/kg) of BZP or vehicle was administered, and mechanical hyperalgesia was assessed at 2, 4, 8, and 24 h postdose. Two-way ANOVA revealed a significant effect of dose (F 24.80, P 0.0001) and of time (F 51.01, P 0.0001) with a significant interaction (F 16.31, P 0.0001). Subsequent Bonferroni post hoc tests comparing each dose with the vehicle group determined that hyperalgesia was significantly reversed by 1 mg/kg PO at 4 and 8 h postdose (Fig. 2A). At the 3 and 10 mg/kg doses, significant reversal of hyperalgesia was observed at 2, 4, and 8 h postdose. Maximal reversal of hyperalgesia was 69% at the 10 mg/kg dose. At 24 h postdose, hyperalgesia was not different from baseline before administering BZP (Figs. 2A and B). The pharmacokinetic versus pharmacodynamic effect of BZP is illustrated in Figure 2C. Regardless of the dose, BZP reversal of hyperalgesia was similar at 2, 4, or 8 h postdose in agreement with the concentration of BZP in plasma (Fig. 2C). The comparison of BZP antihyperalgesic effects to standard analgesics showed that BZP was as potent as rofecoxib in reversing CFA-induced hyperalgesia. The BZP ED50 value was equivalent to rofecoxib and lower than that of indomethacin or naproxen (Fig. 2D).
Data Analysis
Rats were randomly assigned to each treatment group, and the investigator evaluating animals was blinded to the animals treatment. Results are presented as mean sem. Percent reversals are calculated
953
Figure 2. Effects of N-[(R)-1-((R)-7-chloro-1-isopropyl-2-oxo-2,3,4,5-tetrahydro-1H-benzo[b]azepin-3-ylcarbamoyl)-2-(2fluorophenyl)-ethyl]-4-fluoro-2-trifluoromethyl-benzamide (BZP) in the rat complete Freunds adjuvant (CFA)-induced hyperalgesia model. A, Time course and dose-response of the effects of BZP using a Randal-Sellito apparatus to assess mechanical hyperalgesia. BZP significantly reversed CFA-induced hyperalgesia at 110 mg/kg PO with a duration of action more than 8 h postdose. % reversal of hyperalgesia calculated as indicated in Methods are shown above graph bars. Points and bars represent the mean sem. n 6/group. Significance is expressed using a two-way analysis of variance followed by a Bonferroni posttests for multiple comparisons with vehicle values (*P 0.05, ***P 0.001). B, Percent reversal of hyperalgesia versus time. BZP exerted a significant antihyperalgesic effect up to 8 h postdose. At 24 h postdose, hyperalgesia values were not different from predose values. C, Plasma levels of BZP versus percent reversal of hyperalgesia in the CFA model. All timepoints (224 h postdose) are represented in this graph. D, Comparison of BZP antihyperalgesic effects to standard analgesics. Maximal effect, regardless of time postdose is represented. BZP ED50 value was equivalent to rofecoxib and more than indomethacin or naproxen. All data are expressed as mean sem, except for plasma levels that are expressed as mean sd.
Rat SNL Model of Neuropathic Pain

Four weeks after SNL, a single PO dose (3, 10, or 30 mg/kg) of BZP or vehicle was administered, and mechanical allodynia was assessed at 2 and 4 h postdose. Two-way ANOVA revealed a significant effect of dose (F 20.75, P 0.0001) and of time (F 37.57, P 0.0001) with a significant interaction (F 15.50, P 0.0001). Subsequent Bonferroni posttests determined that a 3 mg/kg PO dose significantly reversed mechanical allodynia (Fig. 3A). Maximal reversal of allodynia was 61% at the 30 mg/kg dose (Fig. 3A). Comparison of BZPs antiallodynic effect with neuropathic pain drugs used in the clinic showed that BZP was three times more potent than gabapentin or mexiletine (Fig. 3C).
doses larger than 30 mg/kg induced significant motor impairment (Fig. 3C).
CNS Therapeutic Window

In the SNL model of neuropathic pain, mexiletine and gabapentin each exhibited an ED50 of 45 and 50 mg/kg PO, respectively (Fig. 3C). In the rota-rod model, where the ED20 indicated significant motor coordination impairment, mexiletine and gabapentin exhibited ED20s of 50 and 10 mg/kg PO, respectively. Therefore, the CNS therapeutic index was determined to be 1 for mexiletine and 1 for gabapentin (Fig. 3C). BZP exhibited an ED50 in the SNL model of 7 mg/kg PO (Fig. 3C), with no significant effects on motor coordination at 100 mg/kg PO in the rota-rod model. Additionally, BZP EC50 in rat chronic pain models was 0.4 M (Fig. 2C), with no rota-rod effects observed at concentrations up to 5 M. Thus, CNS therapeutic index of BZP was at least 12, greater than that of mexiletine or gabapentin.
Rat Rota-Rod Model of Motor Coordination

In naive rats, BZP activity was evaluated in a motor coordination model at 30 and 100 mg/kg PO. At these doses, BZP induced no significant change in running latency in the rota-rod (Fig. 3B). Gabapentin at doses larger than 10 mg/kg PO and mexiletine at
954
Figure 3. Effects of N-[(R)-1-((R)-7-chloro-1-isopropyl-2-oxo-2,3,4,5-tetrahydro-1H-benzo[b]azepin-3-ylcarbamoyl)-2-(2fluorophenyl)-ethyl]-4-fluoro-2-trifluoromethyl-benzamide (BZP) in the rat spinal nerve ligation (SNL)-induced allodynia model of neuropathic pain and in the rota-rod (RR) model of motor coordination. A, Dose-response of the effects of BZP using von Frey filaments to assess mechanical allodynia. BZP significantly reversed SNL-induced allodynia at 330 mg/kg PO. % reversal calculated as indicated in Methods are shown above graph bars. Significance is expressed comparing postdose values with vehicle values using a two-way analysis of variance followed by a Bonferroni posttests for multiple comparisons (**P 0.01, ***P 0.001). n 8/group. B, BZP effects in the RR model of motor coordination. n 6/group. BZP at 30 or 100 mg/kg PO exerted no significant effects on motor coordination 1 h postdose. C, Comparison of the antiallodynic activity versus side effect profile for BZP versus standard analgesics used for neuropathic pain. BZP ED50 value in SNL model is more than mexiletine and gabapentin. In the RR model, mexiletine and gabapentin significantly (20%) impair motor coordination.
DISCUSSION
In this study, we showed that a peripherally acting sodium channel blocker dose-dependently reversed hyperalgesia in the CFA model of inflammatory pain and allodynia in the SNL model of neuropathic pain. In the CFA-induced hyperalgesia model, BZP produced reversal of hyperalgesia comparable with nonsteroidal antiinflammatory drugs, and in the SNL model of neuropathic pain, BZP produced reversal of allodynia comparable with the clinical drugs gabapentin and mexiletine. In contrast to gabapentin and mexiletine, BZP did not induce any sedative effects. The small molecule sodium channel blocker BZP displays selectivity for Nav1.7 versus other Nav1.8 and Nav1.5 subtypes. In a voltage/ion probe reader assay in stably transfected cells, BZP displayed 10-fold selectivity for block of Nav1.7 over Nav1.5 and Nav1.8 (IC50 0.03 vs 0.3 M). Similarly, in whole cell electrophysiology experiments, BZP blocked Nav1.5 and Nav1.8 more weakly than Nav1.7.16 BZP was also screened for activity against a broad panel of other ion channels, receptors, and enzymes and was found to have no significant in vitro activity other than sodium
channel blockade.16 To our knowledge, BZP acts solely through sodium channels and no other currents that contribute to neuronal excitability. Because the EC50 in rodent models of pain was 0.4 M, we cannot exclude that BZP acts solely through Nav1.7 and not any other sodium channel subtypes. Although we calculated EC50s based on plasma concentrations, local concentration at the tissue level (i.e., nerve or DRGs) would be more relevant, albeit more difficult to determine. Over the last decade, there has been intense debate regarding the relative role of sodium channel subtypes in nociceptive transmission.20 The -subunit is necessary and sufficient to generate a functional sodium channel, and each -subunit confers unique biophysical properties to the voltage-gated sodium channel. The distribution of sodium channels is also important in determining channel function. Five sodium channels, Nav1.1, Nav1.6, Nav1.7, Nav1.8, and Nav1.9, are expressed at substantial levels in DRG of adult rats, and most, if not all, DRG neurons express multiple sodium channel isoforms.21 The expression of sodium channels in DRG neurons is dynamic. For
955
example, Nav1.3 is expressed at low levels in adult DRG neurons but upregulated in several models of neuropathic pain,22 whereas Nav1.1, Nav1.6, and Nav1.7 are downregulated after tight SNL.23,24 In addition, inflammatory or neurotrophic factors (nerve growth factor and/or glial cell neurotrophic factor) as well as prostaglandin E2 can affect the expression of specific sodium channels (see references in Ref. 20). Regarding the relative role and contribution of Nav1 subtypes to nociceptive mechanisms, data from mice deficient in one channel subtype versus data from antisense oligonucleotides knock down are contradictory. For example, Nav1.8 antisense studies have shown that Nav1.8 contributes to both neuropathic25,26 and inflammatory27 pain. However, Nav1.8-disrupted mice showed reduction in inflammatory hyperalgesia compared with wild-type control mice but no attenuation of mechanical allodynia after partial sciatic nerve ligation.28,29 Consistent with many of the Nav1.8 antisense data, A-803467, a small molecule-selective blocker of Nav1.8, is effective in various rat models of inflammatory and neuropathic pain.30 In agreement with our data showing that BZP potently inhibits inflammation-induced hyperalgesia in rats, and in addition to human genetic data supporting a role for Nav1.7 in acute nociception and inflammatory pain (see references in Ref. 7), preclinical studies demonstrate a prominent role for Nav1.7 in inflammatory pain. Both Nav1.7 mRNA and protein are upregulated in DRG in rodent models of inflammation.31,32 Because a global Nav1.7 null mutant was found to die shortly after birth, Nassar et al.33 used a Cre-loxP system to generate nociceptor-specific knockouts. Nav1.7 nociceptor-specific knockout mice are viable and apparently normal. These animals show increased mechanical and thermal pain thresholds. Remarkably, all inflammatory pain responses evoked by a range of stimuli, such as formalin, carrageenan, CFA, or nerve growth factor, are reduced or abolished in Nav1.7 null mutant mice.33 In contrast to the highly significant role for Nav1.7 in determining inflammatory pain thresholds, the development of neuropathic pain is not affected in Nav1.7 nociceptor-specific knockout mice or in double knockouts of both Nav1.7 and Nav1.8 mice.34 These data in mice deficient for Nav1.7 and/or Nav1.8 are not in agreement with our present data showing that pharmacological blockade of Nav1.7 with BZP potently reverses SNL-induced allodynia. Nav1.7 sodium channels produce a rapidly activating and inactivating current that is sensitive to tetrodotoxin. Nav1.7 appears to be important in the early phases of depolarization. It is characterized by slow transition of the channel into an inactive state when it is depolarized, allowing these channels to remain available for activation with small or slowly developing depolarizations.35 Nav1.7 amplifies small depolarizations, such as generator potentials, and therefore plays a prominent role in spontaneous firing associated with nerve injury. There are also a number
956
of lines of evidence that support a role for Nav1.8 in hyperexcitability and neuropathic pain. These include antisense knockdown of the channel,25,26 as well as the use of A-803467, a Nav1.8-selective small molecule sodium channel blocker, to reduce pain in models of nerve injury.30 Subtype selectivity does not seem to be a prerequisite for alleviating neuropathic pain because nonsubtype selective sodium channel blockers are effective in preclinical models of neuropathic pain,36,37 and approved drugs, such as lidocaine or amitriptyline, are nonsubtype selective. Several hypotheses regarding BZP mechanism of action can be proposed. Sodium channels expressed in primary afferent neurons have been implicated in the aberrant firing that follows nerve injury from peripheral inflammation,20 suggesting that BZP could inhibit the hyperexcitability of damaged or inflamed primary afferent fibers.38,39 Alternatively, because Nav1.7 channels are expressed, not only in DRG neurons but also in sympathetic ganglion neurons,40 BZP could exert its antihyperalgesic activity through sympathetic blockade. Because it is poorly brain-penetrant, BZP likely does not act via a central mechanism, such as mexiletine or gabapentin. The exact mechanism by which BZP blocks Nav1.7 channels remains to be investigated. Local anesthetic, antiarrhythmic, and antiepileptic sodium channel blockers bind to overlapping receptor sites located in the inner cavity of the channel pore.41 Because porelining segments are highly conserved among sodium channel subtypes, most blockers interact with the same region of the channel. At present it is unknown whether BZP shares the same affinity site as local anesthetic, antiarrhythmic, and antiepileptic sodium channel blockers; however, these agents can compete with BZP for similar binding sites.16 The fact that BZP blocks Nav1.7 more potently than clinically used sodium channel blockers16 suggests the existence of specific and well-defined interactions of BZP with selected domains of the channel. Sodium channel blockers currently approved for pain treatment are highly brain penetrant and elicit dose-limiting side effects, such as somnolence and sedation. In contrast, BZP is poorly brain-penetrant (brain to plasma ratio 0.08) and, in a model of motor coordination, did not induce impairment at plasma levels higher than those required for efficacy in chronic pain models. In principle, sodium channel blockade could affect motor coordination either centrally or by a peripheral effect at the neuromuscular junction. It is not yet clear whether brain penetration is required for clinical efficacy, because pain relief is obtained when lidocaine is delivered via a transdermal patch.42 Although the exact mechanism of action of the lidocaine patch is unknown,43 a local drug effect has been demonstrated in postherpetic neuralgia patients.44 Nevertheless, as a peripherally acting compound, BZP would likely be devoid of CNS side effects in the clinic.
In summary, we have shown that a peripheral sodium channel blocker acting preferentially through Nav1.7 dose-dependently reversed hyperalgesia in the CFA model of inflammatory pain and allodynia in the SNL model of neuropathic pain. In the CFA-induced hyperalgesia model, BZP produced reversal of hyperalgesia comparable with that observed with nonsteroidal antiinflammatory drugs. In the SNL model of neuropathic pain, BZP produced reversal of allodynia comparable with that observed with the clinical drugs gabapentin and mexiletine. Unlike gabapentin and mexiletine, BZP did not show any motor coordination impairment. These results suggest that a peripherally acting Nav1.7 blocker could provide clinical relief of chronic pain without the CNS side effects typical of many existing pain treatments. ACKNOWLEDGMENTS The authors thank Nina Jochnowitz for outstanding technical expertise. REFERENCES
1. Marban E, Yamagishi T, Tomaselli GF. Structure and function of voltage-gated sodium channels. J Physiol 1998;508:64757 2. Catterall WA. From ionic currents to molecular mechanisms: the structure and function of voltage-gated sodium channels. Neuron 2000;26:1325 3. Wall PD, Gutnick M. Properties of afferent nerve impulses originating from a neuroma. Nature 1974;248:740 3 4. Devor M, Janig W, Michaelis M. Modulation of activity in dorsal root ganglion neurons by sympathetic activation in nerveinjured rats. J Neurophysiol 1994;71:38 47 5. Goldin AL, Barchi RL, Caldwell JH, Hofmann F, Howe JR, Hunter JC, Kallen RG, Mandel G, Meisler MH, Netter YB, Noda M, Tamkun MM, Waxman SG, Wood JN, Catterall WA. Nomenclature of voltage-gated sodium channels. Neuron 2000;28: 365 8 6. Dib-Hajj SD, Cummins TR, Black JA, Waxman SG. From genes to pain: Nav1.7 and human pain disorders. Trends Neurosci 2007;30:555 63 7. Drenth JP, Waxman SG. Mutations in sodium-channel gene SCN9A cause a spectrum of human genetic pain disorders. J Clin Invest 2007;117:36039 8. Drenth JP, te Morsche RH, Guillet G, Taieb A, Kirby RL, Jansen JB. SCN9A mutations define primary erythermalgia as a neuropathic disorder of voltage gated sodium channels. J Invest Dermatol 2005;124:1333 8 9. Dib-Hajj SD, Rush AM, Cummins TR, Hisama FM, Novella S, Tyrrell L, Marshall L, Waxman SG. Gain-of-function mutation in Nav1.7 in familial erythromelalgia induces bursting of sensory neurons. Brain 2005;128:184754 10. Waxman SG, Dib-Hajj SD. Erythromelalgia: a hereditary pain syndrome enters the molecular era. Ann Neurol 2005;57:785 8 11. Fertleman CR, Ferrie CD. Whats in a namefamilial rectal pain syndrome becomes paroxysmal extreme pain disorder. J Neurol Neurosurg Psychiatry 2006;77:1294 5 12. Fertleman CR, Baker MD, Parker KA, Moffatt S, Elmslie FV, Abrahamsen B, Ostman J, Klugbauer N, Wood JN, Gardiner RM, Rees M. SCN9A mutations in paroxysmal extreme pain disorder: allelic variants underlie distinct channel defects and phenotypes. Neuron 2006;52:76774 13. Cox JJ, Reimann F, Nicholas AK, Thornton G, Roberts E, Springell K, Karbani G, Jafri H, Mannan J, Raashid Y, Al-Gazali L, Hamamy H, Valente EM, Gorman S, Williams R, McHale DP, Wood JN, Gribble FM, Woods CG. An SCN9A channelopathy causes congenital inability to experience pain. Nature 2006;444: 894 8 14. Ahmad S, Dahllund L, Eriksson AB, Hellgren D, Karlsson U, Lund PE, Meijer IA, Meury L, Mills T, Moody A, Morinville A, Morten J, ODonnell D, Raynoschek C, Salter H, Rouleau GA, Krupp JJ. A stop codon mutation in SCN9A causes lack of pain sensation. Hum Mol Genet 2007;16:2114 21 Vol. 109, No. 3, September 2009
15. Hoyt SB, London C, Ok H, Gonzalez E, Duffy JL, Abbadie C, Dean B, Felix JP, Garcia ML, Jochnowitz N, Karanam BV, Li X, Lyons KA, McGowan E, Macintyre DE, Martin WJ, Priest BT, Smith MM, Tschirret-Guth R, Warren VA, Williams BS, Kaczorowski GJ, Parsons WH. Benzazepinone Nav1.7 blockers: potential treatments for neuropathic pain. Bioorg Med Chem Lett 2007;17:61727 16. Williams BS, Felix JP, Priest BT, Brochu RM, Dai K, Hoyt SB, London C, Tang YS, Duffy JL, Parsons WH, Kaczorowski GJ, Garcia ML. Characterization of a new class of potent inhibitors of the voltage-gated sodium channel Nav1.7. Biochemistry 2007;46:14693703 17. Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 1983;16:109 10 18. Kim SH, Chung JM. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 1992;50:355 63 19. Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL. Quantitative assessment of tactile allodynia in the rat paw. J Neurosci Methods 1994;53:55 63 20. Gold MS. Na() channel blockers for the treatment of pain: context is everything, almost. Exp Neurol 2008;210:1 6 21. Black JA, Dib-Hajj S, McNabola K, Jeste S, Rizzo MA, Kocsis JD, Waxman SG. Spinal sensory neurons express multiple sodium channel alpha-subunit mRNAs. Brain Res Mol Brain Res 1996;43:11731 22. Waxman SG, Kocsis JD, Black JA. Type III sodium channel mRNA is expressed in embryonic but not adult spinal sensory neurons, and is reexpressed following axotomy. J Neurophysiol 1994;72:466 70 23. Kim CH, Oh Y, Chung JM, Chung K. The changes in expression of three subtypes of TTX sensitive sodium channels in sensory neurons after spinal nerve ligation. Brain Res Mol Brain Res 2001;95:153 61 24. Kim CH, Oh Y, Chung JM, Chung K. Changes in three subtypes of tetrodotoxin sensitive sodium channel expression in the axotomized dorsal root ganglion in the rat. Neurosci Lett 2002;323:125 8 25. Lai J, Gold MS, Kim CS, Bian D, Ossipov MH, Hunter JC, Porreca F. Inhibition of neuropathic pain by decreased expression of the tetrodotoxin-resistant sodium channel, NaV1.8. Pain 2002;95:14352 26. Gold MS, Weinreich D, Kim CS, Wang R, Treanor J, Porreca F, Lai J. Redistribution of Na(V)1.8 in uninjured axons enables neuropathic pain. J Neurosci 2003;23:158 66 27. Joshi SK, Mikusa JP, Hernandez G, Baker S, Shieh CC, Neelands T, Zhang XF, Niforatos W, Kage K, Han P, Krafte D, Faltynek C, Sullivan JP, Jarvis MF, Honore P. Involvement of the TTXresistant sodium channel Nav 1.8 in inflammatory and neuropathic, but not post-operative, pain states. Pain 2006;123:75 82 28. Akopian AN, Souslova V, Engl S, Okuse K, Ogata N, Ure J, Smith A, Kerr BJ, McMahon SB, Boyce S, Hill R, Stanfa LC, Dickenson AH, Wood JN. The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways. Nat Neurosci 1999;2:541 8 29. Kerr BJ, Souslova V, McMahon SB, Wood JN. A role for the TTX-resistant sodium channel Nav 1.8 in NGF-induced hyperalgesia, but not neuropathic pain. Neuroreport 2001;12:3077 80 30. Jarvis MF, Honore P, Shieh CC, Chapman M, Joshi S, Zhang XF, Kort M, Carroll W, Marron B, Atkinson R, Thomas J, Liu D, Krambis M, Liu Y, McGaraughty S, Chu K, Roeloffs R, Zhong C, Mikusa JP, Hernandez G, Gauvin D, Wade C, Zhu C, Pai M, Scanio M, Shi L, Drizin I, Gregg R, Matulenko M, Hakeem A, Gross M, Johnson M, Marsh K, Wagoner PK, Sullivan JP, Faltynek CR, Krafte DS. A-803467, a potent and selective Nav1.8 sodium channel blocker, attenuates neuropathic and inflammatory pain in the rat. Proc Natl Acad Sci USA 2007;104:8520 5 31. Black JA, Liu S, Tanaka M, Cummins TR, Waxman SG. Changes in the expression of tetrodotoxin-sensitive sodium channels within dorsal root ganglia neurons in inflammatory pain. Pain 2004;108:237 47 32. Gould HJ III, Engl JD, Soignier RD, Nolan P, Minor LD, Liu ZP, Levinson SR, Paul D. Ibuprofen blocks changes in Na v 1.7 and 1.8 sodium channels associated with complete Freunds adjuvant-induced inflammation in rat. J Pain 2004;5:270 80
957
33. Nassar MA, Stirling LC, Forlani G, Baker MD, Matthews EA, Dickenson AH, Wood JN. Nociceptor-specific gene deletion reveals a major role for Nav1.7 (PN1) in acute and inflammatory pain. Proc Natl Acad Sci USA 2004;101:12706 11 34. Nassar MA, Levato A, Stirling LC, Wood JN. Neuropathic pain develops normally in mice lacking both Nav1.7 and Nav1.8. Mol Pain 2005;1:24 35. Cummins TR, Howe JR, Waxman SG. Slow closed-state inactivation: a novel mechanism underlying ramp currents in cells expressing the hNE/PN1 sodium channel. J Neurosci 1998;18:960719 36. Brochu RM, Dick IE, Tarpley JW, McGowan E, Gunner D, Herrington J, Shao PP, Ok D, Li C, Parsons WH, Stump GL, Regan CP, Lynch JJ Jr, Lyons KA, McManus OB, Clark S, Ali Z, Kaczorowski GJ, Martin WJ, Priest BT. Block of peripheral nerve sodium channels selectively inhibits features of neuropathic pain in rats. Mol Pharmacol 2006;69:82332 37. Stummann TC, Salvati P, Fariello RG, Faravelli L. The antinociceptive agent ralfinamide inhibits tetrodotoxin-resistant and tetrodotoxin-sensitive Na currents in dorsal root ganglion neurons. Eur J Pharmacol 2005;510:197208
38. Devor M, Wall PD, Catalan N. Systemic lidocaine silences ectopic neuroma and DRG discharge without blocking nerve conduction. Pain 1992;48:261 8 39. Xiao WH, Bennett GJ. C-fiber spontaneous discharge evoked by chronic inflammation is suppressed by a long-term infusion of lidocaine yielding nanogram per milliliter plasma levels. Pain 2008;137:218 28 40. Morinville A, Fundin B, Meury L, Jureus A, Sandberg K, Krupp J, Ahmad S, ODonnell D. Distribution of the voltage-gated sodium channel Na(v)1.7 in the rat: expression in the autonomic and endocrine systems. J Comp Neurol 2007;504:680 9 41. Catterall WA, Goldin AL, Waxman SG. International Union of Pharmacology. XLVII. Nomenclature and structure-function relationships of voltage-gated sodium channels. Pharmacol Rev 2005;57:397 409 42. Davies PS, Galer BS. Review of lidocaine patch 5% studies in the treatment of postherpetic neuralgia. Drugs 2004;64:937 47 43. Janig W. What is the mechanism underlying treatment of pain by systemic application of lidocaine? Pain 2008;137:5 6 44. Rowbotham MC, Davies PS, Fields HL. Topical lidocaine gel relieves postherpetic neuralgia. Ann Neurol 1995;37:246 53
958
Cytokine Gene Expression After Total Hip Arthroplasty: Surgical Site versus Circulating Neutrophil Response
Asokumar Buvanendran, MD* Kendall Mitchell, PhD Jeffrey S. Kroin, PhD* Michael J. Iadarola, PhD
BACKGROUND: After surgery, cytokines and chemokines are released at the surgical wound site, which can contribute to postoperative pain, local inflammation, and tissue repair. Multiple cell types are present that can release cytokines/chemokines at the wound site and, thus, the exact cellular source of these molecules is unclear. We sought to better understand the contribution of neutrophils to cytokine/chemokine gene expression at the surgical wound site during the initial postsurgery phase of total hip arthroplasty (THA). METHODS: Hip drain fluid was collected at 24 h postsurgery from six patients undergoing standardized THA. In addition, venous blood was collected presurgery and 24 h postsurgery. Neutrophils were isolated, total RNA extracted, and a biotinylated cRNA probe generated. The probes were hybridized with a cDNA microarray containing approximately 100 oligonucleotide sequences representing various human cytokines/chemokines or receptor genes. Changes in gene expression seen in the microarray were verified by reverse transcription polymerase chain reaction. RESULTS: In the microarray analysis of hip drain neutrophils, interleukin-1 receptor antagonist (IL1RN), interleukin-18 receptor 1 (IL18R1), macrophage migration inhibitory factor (MIF), and macrophage inflammatory protein 3 (CCL20) were upregulated, whereas interleukin-8 receptor (IL8RB/CXCR2) was consistently downregulated, compared with presurgery blood neutrophils. All of these changes were confirmed by reverse transcription polymerase chain reaction. CONCLUSION: There is a distinct cytokine gene expression profile in neutrophils at the THA surgical wound site at 24 h postsurgery when compared with that found in presurgery circulating neutrophils. Understanding these changes may allow us to knowledgeably manipulate neutrophil activity to reduce postoperative pain and inflammation without impairing wound healing.
(Anesth Analg 2009;109:959 64)
espite advances in both surgical and anesthetic treatments, several studies demonstrate that approximately 80% of postoperative patients experience moderate or severe pain postsurgery.1 Surgery induces a state of inflammatory response and understanding the pathophysiology of this response can potentially yield new targets for therapy. After hip replacement surgery, factors contributing to postoperative pain include prostaglandin E2 and cytokines, both of which are increased in wound site exudates.2 The cellular origin of such immuno- or neuroactive molecules in
From the *Department of Anesthesiology, Rush University Medical Center, Chicago, Illinois; and Neurobiology and Pain Therapeutics Section, National Institute of Dental and Craniofacial Research, NIH, Bethesda, Maryland. Accepted for publication March 9, 2009. Supported by the Intramural Research Program, NIDCR, NIH, DHHS, and University Anesthesiologists S.C., Chicago, IL. Address correspondence to Dr. Asokumar Buvanendran, Department of Anesthesiology, Rush University Medical Center, 1653 W. Congress Parkway, Chicago, IL 60612. Address e-mail to asokumar@aol.com. Reprints will not be available from the author. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ac1746
wound exudates has not yet been determined because they may arise from a variety of cell types and from plasma extravasates. In the classic model, delineating the phases of healing in cutaneous wounds, there is an initial stage in which blood platelets release clotting factors as well as growth factors and cytokines, such as plateletderived growth factor and TGF-.3 This is followed by an inflammatory stage in which polymorphonuclear neutrophils cells and soon afterward macrophages migrate into the wound site.4 6 Typically, at 24 h after a cutaneous injury, the neutrophil population is at its maximum at the wound site and the activity of these neutrophils may play a critical role in recovery.5 At the same time, several studies have suggested that various leukocyte populations may secrete a variety of proalgesic compounds (e.g., interleukin [IL]-6) and analgesic compounds (e.g., IL-10) that can either sensitize or reduce the excitability of nociceptive primary afferent nerve endings, respectively, and modulate nociceptive input into the central nervous system.7,8 Matching the cell type(s) to the released molecules may provide insight into our understanding of inflammatory and nociceptive processes and wound healing in the periphery.
959
At present, the local cellular response to deep surgical trauma, such as hip replacement surgery, has not been well characterized. Exudates from hip drains (HDs) obtained after surgery contain leukocytes that have undergone emigration from the bloodstream to the wound site. In this study, we showed that at 24 h almost all of these leukocytes in the HD are neutrophils. Thus, the HD fluid at 24 h allows for clarifying the contribution of one population of leukocytes to tissue damage and wound healing without contamination from other cell types. This study analyzed multiple cytokine messenger RNAs from 24 h HD and circulating blood neutrophils to identify which cytokines and cytokine receptors are altered after major surgery.
off and washed twice in Hanks balanced salt solution. A cell count and analysis of the percentage of each type of white blood cell was performed, and the cell pellet frozen at 80C. The frozen pellet was then sent to the National Institutes of Health for analysis.
Microarray Analysis of Gene Expression

Total RNA was isolated from the frozen neutrophil pellets using TRIzol reagent (Invitrogen, San Diego, CA) and was further purified by the RNeasy Mini kit (Qiagen, Valencia, CA) with an additional step of DNase treatment. Neutrophil pellets from all samples were processed in an identical manner. RNA was quantified fluorometrically by the RiboGreen reagent (Molecular Probes, Eugene, OR). Biotinylated cRNA was generated using total RNA as a template with a microarray cRNA synthesis kit (SuperArray, Frederick, MD). The labeled cRNA probe was hybridized to an inflammatory cytokine-targeted microarray containing genes encoding cytokines and ILs associated with the inflammatory responses and their receptors, representing various human cytokine and cytokine receptor genes and control genes (Oligo GEArray, OHS-011, 113 oligonucleotide sequences; SuperArray). Chemiluminescence images were captured by a charge-coupled device camera (AlphaImager, Alpha Innotech, San Leandro, CA) and analyzed using ImageQuant 5 software (Molecular Dynamics, Piscataway, NJ). Eighteen membranes were used to obtain gene expression for presurgical circulating neutrophils (B0), postsurgical circulating neutrophils (B24), and HD neutrophils for the six patients. Per membrane, genes were normalized to -actin expression (also located on the membrane) and the normalized values were put in three groups: B0, B24, and HD. Transcripts demonstrating a twofold change9 and a P 0.05 (using Wilcoxons signed rank test) between the B0 and HD groups were chosen for further evaluation.
METHODS
Patient Selection
After IRB approval (April 18, 2005) from Rush University Medical Center, patients scheduled for total hip arthroplasty (THA) signed informed consent forms and were enrolled in the study. Patients were 80-yr-old and were without recent trauma or systemic infection within 3 mo of surgery date. Patients who had used corticosteroid medications within 3 mo of the surgery date were also excluded.
Study Protocol
All nonsteroidal antiinflammatory therapy was discontinued 14 days before surgery (routine clinical practice to avoid perioperative bleeding). At the preoperative visit, demographic data were collected and all preoperative medications, including the dose, route, and duration, were recorded. A standardized surgical technique of noncemented THA was performed through an anterolateral approach in all patients. All patients underwent standardized surgical management with combined spinal/epidural anesthesia.2
Reverse Transcription Polymerase Chain Reaction

To confirm the results of array screening, transcripts identified as altered in the gene microarray were further analyzed by reverse transcription polymerase chain reaction (RT-PCR) using the extracted RNA obtained from each patient (see above). Per gene, B0, B24, and HD samples from all patients were examined simultaneously. RT-PCR was performed using the Access RT-PCR system (Promega, Madison, WI). The PCR primer pairs and product sizes of genes that were successfully amplified are listed in Table 1. The RT-PCR analysis was performed according to the manufacturers instruction in 25 L reaction mixture containing exactly 8 ng of RNA. RT-PCR steps were 1 cycle of 45 min at 45C for reverse transcription, 1 cycle of 2 min at 94C for inactivation of transcriptase, 28 35 cycles of 30 s at 94C for denaturation, 1 min at 55C for annealing, 2 min at 68C for extension, and final extension at 68C for 7 min. For each gene, all patient samples were amplified the same time for the
Sample Collection
Venous blood (10 mL) was collected in tubes with lithium heparin before surgical incision and at 24 h after the start of surgery. When the hip replacement was completed, a standard drain was placed in the deepest portion of the wound, in proximity to the newly replaced joint. Drain exudates were collected over a 60-min period, 2324 h from start of surgery, in a 400 mL capacity Hemovac reservoir. The approximate volume of fluid obtained over this 60-min period was 10 mL, and no patient had excessive bleeding. Preliminary analysis with cell slide smears demonstrated that, at this time point, 95%98% of HD leukocytes are neutrophils. Cells from either blood or HD fluid were fractionated by placing the sample over a separation gel (1-Step Polymorphs, Accurate Chemicals, Westbury, NY) and centrifuging at 500g for 30 min. A lower band of polymorphonuclear neutrophil cells appears in the gel, and this fraction was aspirated
960
Postsurgical Local Neutrophil Response
Table 1. Reverse Transcription Polymerase Chain Reaction (RT-PCR) Primer Pairs and Length of PCR Products Gene Primer pairs Product (bp)
299 251 250 250 234 266
-Actin CTCCTGAGCGCAAGTACTCC GTCACCTTCACCGTTCCAGT IL1RN CTCCTGGGGGTTCTTTCTTC TAGGGAACTTTGCACCCAAC IL8RB ATTCTGGGCATCCTTCACAG TGAGGCTTGGAATGTGACTG IL18R1 GAAGAACGCCGAGTTTGAAG ATTTTCTTCCCCGAACATCC MIF AGAACCGCTCCTACAGCAAG ATTTCTCCCCACCAGAAGGT CCL20 CTGGCCAATGAAGGCTGT GACAAGTCCAGTGAGGCACA
same number of cycles: 28 cycles for IL1RN, 30 cycles for IL18R1, MIF, IL8RB, and CCL20, and 23 cycles for -actin. The RT-PCR products were separated by electrophoresis on 2% agarose/ethidium bromide gels and images were acquired by an AlphaImager chargecoupled device camera. The relative intensities of the RT-PCR products, as visualized on the gel, were analyzed quantitatively using ImageQuant 5 software. The results were normalized to -actin. Comparisons of gene expression from B0, B24, and HD samples were made by repeated measures mixed model with Tukey-Kramer post hoc test (P 0.05 for significance).
Figure 1. Microarray analysis for genes associated with cytokines, chemokines, and their receptors. The oligo GEArray membrane containing approximately 100 genes was hybridized with biotin-labeled cRNA generated with RNAs from neutrophils from presurgical blood (left) and from hip drain neutrophils of the same patient (right). Transcripts which were consistently altered in patients are indicated by an arrow. Table 2. Genes Upregulated or Downregulated in Hip Drain Neutrophils After Total Hip Arthroplasty Using Microarray Analysis Symbol
IL8RB MIF IL18R1 IL1RN CCL20 CCR3 CX3CR1 CCR5 LTB
Gene name
Interleukin-8 receptor /CXCR2 Macrophage migration inhibitory factor Interleukin-18 receptor 1 Interleukin-1 receptor antagonist LARC/MIP-3 CC-CKR-3/CKR3 CCRL1/CMKBRL1 CC-CKR-5 Cytokine P33
Fold change
0.23 2.7 3.1 3.2 5.3 0.19 0.35 0.43 0.26
RESULTS
Samples from six patients undergoing THA were analyzed. Patient ages ranged from 46 to 80 yr and were equally distributed between genders. For both blood and HD samples, 96% 3% of the leukocytes in the cell pellet were neutrophils. Lymphocytes, eosinophils, and monocytes were always 10%. In the microarray analysis, genes significantly upregulated in the HD neutrophils when compared with the presurgery blood sample neutrophils were IL-1 receptor antagonist (IL1RN), IL-18 receptor 1 (IL18R1), macrophage migration inhibitory factor (MIF), and macrophage inflammatory protein 3 (CCL20), whereas IL-8 receptor (IL8RB/CXCR2), CCR3, CX3CR1, CCR5, and LTB were downregulated (Fig. 1, Table 2). All of the genes expressed in the microarray are tabulated in the online supplement table (see Supplemental Digital Content 1, available at: http://links.lww.com/A1335), which shows that most cytokines or chemokines did not show a twofold change or a statistically significant change in HD neutrophils when compared with the presurgery blood sample neutrophils. Changes seen in the gene microarray were confirmed by RT-PCR (Fig. 2) for ILIRN, IL18R1, MIF, and CCL20 for which expression in the HD neutrophil population was upregulated compared with those in circulating presurgery blood. In addition, IL18R1 expression was also increased in 24 h blood compared with presurgery blood. RT-PCR also confirmed that IL8RB/CRCX2 expression in HD was downregulated
Genes that show more than a twofold upregulation or downregulation in 24 h hip drain neutrophils versus presurgery blood neutrophils with P 0.05.
compared with presurgery blood. IL8RB/CXCR2 was also decreased in 24 h blood but not to the same extent as in the HD population. As a control, Figure 2 also demonstrates that -actin expression did not differ between HD and blood neutrophils. We also attempted to examine the expression of CCR3, CCR5, CXCR1, and LTB via RT-PCR. Genes CCR5 and CX3CR1 had very weak signals in the microarray and could not be amplified during RTPCR. Although LTB had an adequate signal on the microarray, it could not be amplified appropriately with two different primer pairs in the RT-PCR. CCR3 also gave a low signal on the arrays and did not show downregulation in the RT-PCR gel. Based on the low baseline signals for these four genes and/or the difficulty in amplifying them (e.g., LTB), we consider them marginally or nonexpressed in this cell population.
DISCUSSION
A complex array of cytokines and chemokines are found in the extracellular milieu after surgeryinduced tissue damage.2,10 In such a complex cellular
961
Figure 2. Left, Gel electrophoresis of reverse transcription polymerase chain reaction (RT-PCR) expression of neutrophil mRNA showing genes that were significantly upregulated or downregulated in the gene microarray. Results from two representative patients are displayed for presurgery blood (B0), 24 h blood (B24), and 24 h postsurgery hip drain (HD). -Actin was the internal control. Right, RT-PCR analysis of mRNA levels, relative to -actin, showing four genes upregulated (MIF, IL18R1, IL1RN, CCL20) and one gene downregulated (IL8RB/CXCR2) in 24 h postsurgery HD neutrophils versus presurgery blood neutrophils. Data presented as mean sem; *P 0.05, ***P 0.001. In the graph, -actin expression of B0, B24, and HD neutrophils were normalized to the B0 average. milieu, identifying the sources of individual cytokines/ chemokines is difficult because multiple cell types at the wound site express these molecules. By investigating HD exudates from patients who had surgical hip replacement, we were able to obtain an almost pure population of neutrophils enabling us to better assess the contribution of this cell type to cytokine/ chemokine production in the extracellular space. This approach demonstrated that neutrophils collected at 24 h after THA express altered transcript levels in a distinct set of cytokines compared with presurgical circulating neutrophils. One of the cytokines that is induced in HD neutrophils, but not in 24-h circulating neutrophils, after THA encodes the proinflammatory cytokine MIF. Functionally, MIF can induce the expression of proinflammatory cytokines and cyclooxygenase 2.11 MIF has also been suggested to amplify neurogenic inflammation induced by the pronociceptive neuropeptide substance P.12 The secretion of MIF protein in patients suffering from endometriosis has been shown to correlate with pain,13,14 suggesting the possibility that this molecule is involved with sensitization of nociceptive nerve terminals. MIF can also activate macrophages and inhibit apoptosis of these cells thereby sustaining macrophage activity during inflammation.15,16 Macrophages, which are generally recruited to the wound site after neutrophils, may in-turn release factors that can sensitize primary afferent fibers and contribute to inflammation. It may be interesting then to determine whether MIF produced by HD neutrophils is involved in the recruitment of macrophages and the reported increase in cytokines and prostaglandin.2 At the site of inflammation, it is common to observe the induction of both proinflammatory and antiinflammatory cytokines.10 The observed induction of the
962
antiinflammatory cytokine IL1RN in the HD neutrophils is consistent with this duality. IL1RN functions to antagonize the binding of the proinflammatory cytokine IL-1 to its receptor.17 IL-1 protein was previously reported to be elevated in the HD after THA, has well-documented roles in inflammation, and like MIF, has been demonstrated to sensitize nerve terminals to pain.2,18 The increase in IL1RN after THA may thus be a natural process for attenuating the effects of IL-1, suggesting that neutrophils at this point may also participate in postoperative repair. Concordant with this hypothesis, chronic treatment with IL1RN or anakinra, a recombinant analog of IL1RN, has been demonstrated to reduce basal nociceptive sensitivity in mice.19,20 Moreover, anakinra has been used clinically to reduce joint inflammation in rheumatoid arthritis patients.21 In this context, it would be interesting to determine whether further supplementation with exogenous IL1RN after THA could reduce pain and inflammation without having an adverse affect on wound healing. We also detected significant alterations in message levels for two cytokine/chemokine receptors. The gene encoding IL18R1, which is the receptor for the proinflammatory cytokine IL-18, is significantly increased in HD neutrophils. Increased IL18R1 mRNA has also been reported in synovial fluid neutrophils of patients suffering from rheumatoid arthritis.22 It has been reported that IL-18 recruits neutrophils to the inflamed site via IL18R1,22 which is consistent with the massive, sustained infiltration of neutrophils after THA. IL-18 also activates neutrophils through this receptor, causing them to release various cytokines and chemokines.22 Regarding nociception, intraplantar injection of IL-18 causes mechanical hyperalgesia,23 although the exact role of the neutrophils in
this process needs further examination. Finally, the fact that IL18R1 is also induced in 24-h circulating neutrophils suggests that this receptor may have been involved in the emigration from the blood to the wound site. We also observed that IL8RB/CXCR2, a high affinity receptor for IL-8/CXCL8, is strongly downregulated in HD neutrophils. IL-8, which promotes chemoattraction and activation of neutrophils,24,25 has been reported to be increased in HD exudates2 after THA. Thus, it is possible that IL8RB-expressing neutrophils are recruited to the wound site via IL-8. IL-8 can also regulate the expression of IL8RB, because in vitro studies have shown that this molecule causes a rapid downregulation of its receptors.26,27 It is possible then that the downregulation of IL8RB in HD neutrophils may be caused by the increased levels of IL-8 in the HD exudate. Given that the HD neutrophils have reached their final destination site, a mechanism for decreasing receptors such as IL8RB seems likely in these cells. Understanding why receptors with similar functions (e.g., IL18R1) are not downregulated suggests a multiplicity of regulatory processes and provides different targets for manipulation. The transcript encoding macrophage inflammatory protein 3 (MIP 3/CCL20) was increased in HD neutrophils at 24 h after THA. The encoding molecule is a strong chemoattractant for lymphocytes, while weakly attracting neutrophils. Thus, a release of CCL20 from the neutrophils may be a signal for recruitment of cells other than neutrophils. Regarding pain, CCL20 has not been thoroughly investigated, although it is increased in conditions in which pain is elevated. For example, neutrophils in synovial fluid of some patients with rheumatoid arthritis show increased CCL20 mRNA, but CCL20 expression in the rheumatoid study was not detectable in circulating blood neutrophils.28 CCL20 has also been reported to be altered in oral wound healing models.29 Although this study focused on the most consistently altered genes within a 113-gene microarray, it cannot be excluded that additional genes are altered in HD neutrophils. Moreover, the 24-h time point may not be the optimal time for detecting alterations in other genes, although 24 h is the time point when neutrophils predominate at wound sites.4 6 Although it would be of interest to analyze cytokine and chemokine mRNA in neutrophils at later time points (e.g., Days 2 4), the risk of infection makes it unsafe to maintain a drain beyond the 24-h recovery period. It is also important to note that a lack of alteration in neutrophil mRNA does not mean that the corresponding protein is not secreted by another cell type. Thus, cytokines that were reported in HD fluid in our previous study,2 IL-1, IL-6, and IL-8, may have originated, for example, from tissue macrophages surrounding the HD catheter. A limitation of this study is that we did not show that MIF and CCL20 are also produced and released into the HD exudates after
THA as would be suggested by the upregulation of the encoding mRNAs. The expression of ILIRN, IL18R1, and IL8RB proteins in isolated HD neutrophils remains to be evaluated using methods such as Western blots or flow cytometry30 as well as the functional consequence in ex vivo experiments. In contrast to the presurgery neutrophils, the neutrophils taken from the HD and circulating blood at 24 h were obtained while the patients were receiving epidural anesthesia. Thus, it is possible that anesthesia may have influenced the cytokine/chemokine profile in our study. However, a previous study failed to detect a difference in plasma IL-6 and tumor necrosis factor- levels after THA with general versus regional spinal/epidural anesthesia, suggesting that the cell types involved in this study may not have been affected by the anesthetic.31 The data obtained here demonstrate that opponent processes (anti- and proinflammatory) occur in HD neutrophils after THA. Deciphering their roles in events such as pain processing, inflammation, and tissue repair could aid in our approach to treating patients pre- and postoperatively. REFERENCES
1. Apfelbaum JL, Chen C, Mehta SS, Gan TJ. Postoperative pain experience: results from a national survey suggest postoperative pain continues to be undermanaged. Anesth Analg 2003;97: 534 40 2. Buvanendran A, Kroin JS, Berger RA, Hallab NJ, Saha C, Negrescu C, Moric M, Caicedo MS, Tuman KJ. Upregulation of prostaglandin E2 and interleukins in the central nervous system and peripheral tissue during and after surgery in humans. Anesthesiology 2006;104:40310 3. Diegelmann RF, Evans MC. Wound healing: an overview of acute, fibrotic and delayed healing. Front Biosci 2004;9:2839 4. Englelhardt E, Toksoy A, Goebeler M, Debus S, Brocker EB, Gillitzer R. Chemokines IL-8, GROalpha, MCP-1, IP-10, and Mig are sequentially and differentially expressed during phasespecific infiltration of leukocyte subsets in human wound healing. Am J Pathol 1998;153:1849 60 5. Gillitzer R, Goebeler M. Chemokines in cutaneous wound healing. J Leukoc Biol 2001;69:51321 6. Park JE, Barbul A. Understanding the role of immune regulation in wound healing. Am J Surg 2004;187:11S16S 7. Cunha FQ, Ferreira SH. Peripheral hyperalgesic cytokines. Adv Exp Med Biol 2003;521:2239 8. Rittner HL, Machelska H, Stein C. Leukocytes in the regulation of pain and analgesia. J Leukoc Biol 2005;78:121522 9. Schena M, Shalon D, Heller R, Chai A, Brown PO, Davis RW. Parallel human genome analysis: microarray-based expression monitoring of 1000 genes. Proc Natl Acad Sci USA 1996;93: 10614 9 10. Yang HY, Mitchell K, Keller JM, Iadarola MJ. Peripheral inflammation increases Scya2 expression in sensory ganglia and cytokine and endothelial related gene expression in inflamed tissue. J Neurochem 2007;103:1628 43 11. Calandra T, Roger T. Macrophage migration inhibitory factor: a regulator of innate immunity. Nat Rev Immunol 2003;3:791 800 12. Meyer-Siegler KL, Vera PL. Intraluminal antibodies to macrophage migration inhibitory factor decrease substance P induced inflammatory changes in the rat bladder and prostate. J Urol 2004;172:1504 9 13. Akoum A, Metz CN, Al-Akoum M, Kats R. Macrophage migration inhibitory factor expression in the intrauterine endometrium of women with endometriosis varies with disease stage, infertility status, and pelvic pain. Fertil Steril 2006;85:1379 85
963
14. Morin M, Bellehumeur C, Therriault MJ, Metz C, Maheux R, Akoum A. Elevated levels of macrophage migration inhibitory factor in the peripheral blood of women with endometriosis. Fertil Steril 2005;83:86572 15. Hudson JD, Shoaibi MA, Maestro R, Carnero A, Hannon GJ, Beach DH. A proinflammatory cytokine inhibits p53 tumor suppressor activity. J Exp Med 1999;190:1375 82 16. Mitchell RA, Liao H, Chesney J, Fingerle-Rowson G, Baugh J, David J, Bucala R. Macrophage migration inhibitory factor (MIF) sustains macrophage proinflammatory function by inhibiting p53: regulatory role in the innate immune response. Proc Natl Acad Sci USA 2002;99:34550 17. Arend WP, Guthridge CJ. Biological role of interleukin 1 receptor antagonist isoforms. Ann Rheum Dis 2000;59(suppl 1):i60 4 18. Ferreira SH, Lorenzetti BB, Bristow AF, Poole S. Interleukin-1 beta as a potent hyperalgesic agent antagonized by a tripeptide analogue. Nature 1988;334:698 700 19. Wolf G, Yirmiya R, Goshen I, Iverfeldt K, Holmlund L, Takeda K, Shavit Y. Impairment of interleukin-1 (IL-1) signaling reduces basal pain sensitivity in mice: genetic, pharmacological and developmental aspects. Pain 2003;104:471 80 20. Baamonde A, Curto-Reyes V, Juarez L, Meana A, Hidalgo A, Menendez L. Antihyperalgesic effects induced by the IL-1 receptor antagonist anakinra and increased IL-1beta levels in inflamed and osteosarcoma-bearing mice. Life Sci 2007;81:673 82 21. Furst DE. Anakinra: review of recombinant human interleukin-I receptor antagonist in the treatment of rheumatoid arthritis. Clin Ther 2004;26:1960 75 22. Leung BP, Culshaw S, Gracie JA, Hunter D, Canetti CA, Campbell C, Cunha F, Liew FY, McInnes IB. A role for IL-18 in neutrophil activation. J Immunol 2001;167:2879 86
23. Verri WA Jr, Schivo IR, Cunha TM, Liew FY, Ferreira SH, Cunha FQ. Interleukin-18 induces mechanical hypernociception in rats via endothelin acting on ETB receptors in a morphine-sensitive manner. J Pharmacol Exp Ther 2004;310:710 7 24. Holmes WE, Lee J, Kuang WJ, Rice GC, Wood WI. Structure and functional expression of a human interleukin-8 receptor. Science 1991;253:1278 80 25. Murphy PM, Tiffany HL. Cloning of complementary DNA encoding a functional human interleukin-8 receptor. Science 1991;253:1280 3 26. Samanta AK, Oppenheim JJ, Matsushima K. Interleukin 8 (monocyte-derived neutrophil chemotactic factor) dynamically regulates its own receptor expression on human neutrophils. J Biol Chem 1990;265:1839 27. Chuntharapai A, Kim KJ. Regulation of the expression of IL-8 receptor A/B by IL-8: possible functions of each receptor. J Immunol 1995;155:258794 28. Schlenk J, Lorenz HM, Haas JP, Herrmann M, Hohenberger G, Kalden JR, Ro llinghoff M, Beuscher HU. Extravasation into synovial tissue induces CCL20 mRNA expression in polymorphonuclear neutrophils of patients with rheumatoid arthritis. J Rheumatol 2005;32:2291 8 29. McGrory K, Flaitz CM, Klein JR. Chemokine changes during oral wound healing. Biochem Biophys Res Commun 2004;324: 31720 30. van Eeden SF, Klut ME, Walker BA, Hogg JC. The use of flow cytometry to measure neutrophil function. J Immunol Methods 1999;232:23 43 31. Hgevold HE, Lyberg T, Ka hler H, Haug E, Reikers O. Changes in plasma IL-1beta, TNF-alpha and IL-6 after total hip replacement surgery in general or regional anaesthesia. Cytokine 2000;12:1156 9
964
The Neuraxial Effects of Intraspinal Amitriptyline at Low Concentrations

Fernanda B. Fukushima, MD, PhD* Guilherme A. M. Barros, MD, PhD* Maria ngela E. A. Marques, MD, PhD Edison I. O. Vidal, MD, MPH Eliana M. Ganem, MD, PhD*
BACKGROUND: As a result of amitriptylines vast array of actions, it could potentially be used as an intraspinal adjuvant in neuraxial anesthesia and/or in the treatment of refractory neuropathic pain. None of the previous studies examining the safety profile of intraspinal single doses of amitriptyline found signs of toxicity at concentrations below 15.4 mM/L (0.5%) and the current hypothesis regarding the pathophysiology of amitriptyline toxicity suggests it might be safe at low concentrations while still having relevant clinical effects. Hence, we conducted this study to assess the clinical and histological toxicity of intraspinal amitriptyline at the lowest dosages previously known to be effective. METHODS: Twenty-one dogs were randomized to receive a 1-mL single intraspinal dose of one of the three solutions: saline (0.9%), amitriptyline (0.15%), or amitriptyline (0.3%). The dogs were evaluated clinically 1 h after awakening from anesthesia and 21 days later. At 21 days, all animals were killed, and histological sections of the spinal cord and surrounding meninges were retrieved for analysis. RESULTS: All dogs recovered motor function, anal sphincter tone and sensibility. With the exception of one dog in the 0.15% amitriptyline group, all animals in both amitriptyline groups had marked adhesive arachnoiditis, which was absent in the control group. No evidence of direct neural damage was found on histological sections stained by glial fibrillary acidic protein technique in any of the study animals. CONCLUSION: The intraspinal administration of amitriptyline to dogs even in low concentrations is strongly associated with the development of intense meningeal adhesive arachnoiditis and is not safe even at low concentrations for which there was no previous evidence of toxicity.
(Anesth Analg 2009;109:96571)
n the past decades, there has been considerable progress in the knowledge of the physiopathology of neuropathic pain.13 However, the treatment of neuropathic pain syndromes remains an important therapeutic challenge, and the quest for an ideal drug with high levels of efficacy and low toxicity is far from being fulfilled.4 6 Tricyclic antidepressant drugs are among the most effective drugs used in the treatment of neuropathic pain and correspond to the third most prescribed drugs for the treatment of nononcological pain syndromes.7 Besides their action in the recaptation of norepinephrine and serotonin, tricyclic antidepressants also have antiinflammatory activity, block
From the Departments of *Anesthesiology and Pathology, Sa o Paulo State UniversityUNESP, Botucatu/SP; Department of Preventive and Social Medicine, Campinas State UniversityUNICAMP, Campinas/SP; and Albert Einstein Hospital, Sa o Paulo/SP, Brazil. Accepted for publication March 24, 2009. Supported by Brazilian Federal Agency for the Support and Evaluation of Graduate Education (CAPES). Address correspondence and reprint requests to Fernanda Bono Fukushima, MD, PhD, Rua Itororo , 427 Vila Galo Americana-SP CEP 13466-240, Brazil. Address e-mail to fernandafukushima@ gmail.com. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ad581e
Na, K, and Ca2 ion channels, -2 adrenergic, nicotinic, muscarinic, N-methyl-d-aspartate, gammaaminobutyric-B, and histaminergic receptors.8 17 Amitriptyline is the prototype drug of this class of drugs and has been used for more than 40 yr in the management of neuropathic pain.18 Nevertheless, its use is frequently limited by several factors, such as anticholinergic side effects. As a result of amitriptylines vast array of actions, it could potentially be used as an intraspinal adjuvant in neuraxial anesthesia and/or in the treatment of refractory neuropathic pain. Theoretically, in selected cases, the intraspinal administration of amitriptyline could enable higher efficacy with fewer doses and systemic side effects. Previous studies analyzed many aspects of the safety/toxicity of amitriptyline.10,14,16,19 35 When administered IV, it had less cardiovascular and central nervous system toxicity than bupivacaine.33 A few studies analyzed the neurotoxicity profile of intraspinal amitriptyline in single doses20,24,34,36 and under continuous infusions.35 Even though continuous intrathecal infusions seemed to be toxic in low concentrations,35 none of the studies of a single intraspinal dose found signs of neurotoxicity for concentrations below 15.4 mM/L (0.5%).20,24,34,36 However, all the studies restricted their analyses to short-term
965
Vol. 109, No. 3, september 2009
(7 days or less) clinical and/or histological evaluations, e.g., in most cases, animals without clinical signs of neurologic deficits were not submitted to histological examination.24,34 Therefore, the possibility of either true clinical safety or later occurring toxicity at low concentrations was not properly assessed.37,38 Furthermore, evidence suggests that, according to the currently proposed mechanism of amitriptyline toxicity, there might be no neurotoxicity for low concentrations of this drug29 below 10 mM/L (0.3%). Hence, an experimental study was conducted to assess whether small single doses of intraspinal amitriptyline are associated with clinical and histological signs of neurotoxicity after 21 days of its administration.
METHODS
Animals
After receiving approval from the local ethics committee on animal experimentation, 21 adult mongrel dogs of both sexes were obtained from the Experimental Animal Center at the State University of Sa o Paulo at Botucatu Campus. The weights of the animals varied from 7 to 16 kg, and the length of their vertebral column ranged from 53 to 70 cm. All dogs were kept under clinical observation in individual cells for at least 40 days before they were defined as healthy and could be randomized for the experiment. All animals were assessed before spinal puncture and had normal neurologic function. All tests were performed in accordance with the guidelines of the International Association for the Study of Pain.39 The sample size was calculated according to Fleiss et al.40 estimating a proportion of histological neurotoxicity of 1% and 80% in the control and amitriptyline groups, respectively, so as to obtain a power value of 90% while setting the one-sided level for statistical significance at 0.05.
submitted to antisepsis with a 2% chlorhexidine gluconate solution, and sterile fields were appropriately positioned. Subarachnoid puncture was performed through the median line approximately 45 to the skin with a 22-gauge Quincke needle. Any difficulties during the procedure and the color of the cerebrospinal fluid (CSF) were recorded. In any circumstance, when a traumatic spinal tap was identified, as defined by the presence of hemorrhagic CSF or the need for more than one puncture, the animal was immediately excluded from the study and no solution was administered into the subarachnoid space. Only two dogs were excluded due to these criteria. Once the needle was properly located and clear CSF could be identified, the 1-mL randomized solution was injected for approximately 10 s through 1-mL disposable syringes.
Solution Specications
The amitriptyline solutions administered to Groups 2 and 3 composed of 0.15% and 0.3% hydrochloride amitriptyline, respectively (Vashuda Pharma Chem Limited, Andhra Pradesh, India), at pH 5.5 and free of chemical preservatives. The solutions were supplied in individual ampules by a compounding pharmacy certified in the formulation of parenteral medications. Solutions were stored at 4C, protected from direct exposure to light, and administered about 36 h after synthesis. The 0.9% saline solution (Baxter Healthcare Corp., Sa o Paulo, Brazil) administered to the control group had a pH of 5.0.
Evaluation and Outcomes

The study animals were evaluated at two different times: 1 h after intraspinal administration of the solution after recovery from anesthesia and 21 days later. Each animal was assessed regarding the following secondary outcomes: motor deficit, anal sphincter tonus, and painful sensibility. The primary outcome of the study was histological analysis of the spinal cord and meninges of the dogs 21 days after subarachnoid administration of study solutions. Motor deficit was determined by the ability to walk, jump, and sustain the tail in an upward position. Anal sphincter relaxation was ascertained through visual inspection. Nociception was assessed by reaction to painful pressure and thermal stimuli. To control for possible interference due to visual perception of the stimuli by the animals, one researcher was responsible for masking the animals with a nontransparent cloth comfortably positioned around their neck. Pressure nociceptive stimuli were elicited by the bilateral pinch of a skin fold over sacral, lumbar and thoracic dermatomes, and interdigital membranes of hindlimbs. Thermal painful stimuli were provided by a thermo algometer set at 50C for 10 s touching the interdigital membranes of the hindlimbs. The presence of pain
Experimental Groups
The dogs were randomized to one of three experimental groups according to the type of solution to be administered into the subarachnoid space. Seven dogs were allocated to each group. Group 1 was defined as the control group, in which 0.9% normal saline solution was injected intraspinally. Groups 2 and 3 received 0.15% and 0.3% amitriptyline solutions, respectively. All three solutions had the same 1-mL volume. The researchers responsible for spinal administration of the solution were blinded to its content.
Spinal Puncture Protocol

The animals were fasted for 12 h before the procedure, with water ad libitum. All dogs were submitted to the same anesthetic technique, such as IV etomidate (2 mg/kg) and fentanyl (0.005 mg/kg). A 10-cm area around the site of the spinal puncture at L6-7 intervertebral space level was washed with water and soap, followed by hair removal and skin cleansing with 0.9% normal saline. Finally, the naked skin was
966
Neuraxial Administration of Amitriptyline
was defined by the following: limb withdrawal, vocalization, and facial expression. All three secondary outcomes (motor deficit, anal sphincter relaxation, and nociception) were classified dichotomously into absent or present. If the slightest deficit in any of these dimensions were observed at clinical assessment, the animal would be classified as positive for deficit according to the dichotomous classification used. After IV anesthesia with sodium pentobarbital, animals were killed by electroshock. Thereafter, the lumbar and sacral segments of the spinal cord with the surrounding meninges were quickly removed in 3 min to minimize the risk of injuries to those tissues from ischemia and apoptosis. The anatomical pieces were fixed in formalin 10% solution for 7 days before histological sections were prepared encompassing from about 10 cm above the level of the spinal puncture to the end of the cauda equina. The histological sections were stained by hematoxylin and eosin, Masson trichrome, and glial fibrillary acidic protein (GFAP) techniques and examined by optical microscopy. Three researchers (FBF, EMG, and MEAM) blinded to the intervention and experienced in histological neurotoxicity assessment classified by consensus each of the sections according to the presence or absence of histological injury. If any kind of lesion was identified, it was further specified, and to verify a possible dose-related gradient effect, injuries were stratified according to severity and extent as ascertained by consensus. All clinical and histological evaluations were performed by researchers blinded to the solutions administered to each animal.
Table 1. Results of the Assessment of Motor Blockade, Anal Sphincter Relaxation, Thermal, and Pressure Nociception 1 h After Intraspinal Administration of the Randomized Solution According to Experimental Groupa
Motor blockade Control group AMT 0.15% AMT 0.3% 0 7* 7* Anal sphincter relaxation 0 7* 7* Thermal nociception 7 1 2 Pressure nociception 7 1 1
AMT amitriptyline. a Control Group is the reference group for comparison. * P 0.0006. P 0.0047. P 0.021.
To evaluate the effectiveness of the randomization procedure and the comparability of the three study groups, one-way analysis of variance was performed for animals weights and length of their vertebral columns. One-sided Fishers exact test was selected to compare the frequencies of the findings on the primary and the secondary outcomes between the amitriptyline and the control groups. The KruskalWallis test was performed for the comparison among treatment groups regarding the stratification of histological injuries. As previously reported, the level for statistical significance was set at 0.05. The R software (Version 2.7.1) was used for the performance of statistical analysis.41
Control group animals had no impaired motor function, anal sphincter relaxation or decreased nociception at 1 h or at 21 days after the subarachnoid administration of 0.9% normal saline. Twenty-one days after the administration of the amitriptyline solutions, none of the animals in Groups 2 and 3 displayed signs of motor impairment, anal sphincter relaxation or abnormal reactions to painful stimulation. Table 1 lists motor function, anal sphincter tone, and sensibility to painful stimuli 1 h after intraspinal administration of the randomized solution. The assessments of histological sections are summarized in Table 2. None of the animals in the three groups had direct injuries to neural tissue, as ascertained by GFAP staining; however, in the two amitriptyline groups marked adhesive, arachnoiditis was observed, as shown in Figure 1. To identify a possible dose-related injury gradient, histological sections of the amitriptyline groups were further stratified for four dimensions: presence of meningeal thickening, fibrosis, meningeal adherence, and inflammatory lymphoplasmocytic infiltrate. Each of these dimensions was classified into a 4-point scale, yielding a score from 0 to 12 quantitatively describing the extent and severity of the arachnoiditis. Figure 2 depicts the total meningeal toxicity score of the two amitriptyline groups. The comparison between the two amitriptyline groups by means of the KruskalWallis test demonstrated a statistically significant association between severity and extensive histological injury and higher doses of amitriptyline (P 0.012).
DISCUSSION
In the pursuit of the perfect local anesthetic, one marked by low toxicity and high-potency profile, amitriptyline has become the focus of several investigations due to its potent Na channel-blocking effect.14,42 The wide array of direct local actions on the nervous system by amitriptyline, such as N-methyl-daspartate, gamma-aminobutyric acid-b, and histaminergic receptor blockade,13,43,44 calls attention to its potential use as a direct drug or adjuvant drug for neuraxial therapy in resistant cases of neuropathic
RESULTS
One-way analysis of variance revealed all three groups were similar with respect to weight (P 0.37) and length of the vertebral column (P 0.65). There were no deaths among the study animals. During the 21 days of observation, none of them exhibited abnormal behavior.
967
Table 2. Abnormal Findings Observed by Means of Examination of the Histological Sectionsa Direct injury to neural tissue
Control group AMT 0.15% AMT 0.3% 0 0 0
Blood vessels,vascular wall thickening

0 6* 7
Meninges, brous thickening with inammatory inltration and formation of trabeculae

0 6* 7
Control Group received 0.9% normal saline. AMT amitriptyline. a Group 1 is the reference group for comparison. * P 0.002. P 0.001.
Figure 1. A, Low magnification (1) glial fibrillary acidic protein (GFAP) stained section. Arrow: exuberant meningeal
thickening. B, Hematoxylin- and eosin-stained section. Arrow: meningeal adhesion. C, Hematoxylin- and eosin-stained section. Long arrow: vascular cuffing; short arrow: lymphoplasmocitic infiltrate. D, Masson trichome-stained section. Arrow: fibrous thickening of meninges with collagen deposition and meningeal adhesion. E, Masson trichome-stained section. Arrow: collagen deposition around the vessel. F, GFAP-stained section demonstrating meningeal thickening with underlying normal nervous tissue. All sections are from animals of the 0.3% amitriptyline group.
Figure 2. Each bar represents the toxicity

score ascertained for each animal by consensus. Comparison among groups by KruskalWallis test revealed statistically significant difference (P 0.012).
pain syndromes and perhaps even as an opioidsparing drug in spinal anesthesia. Other studies have investigated some aspects of the safety and toxicity of parenteral and single intrathecal doses of amitriptyline.19 26,30,3236,45 47 It is interesting to note
968
that none of the studies of single intraspinal doses of amitriptyline demonstrated evidence of neurotoxicity at concentrations below 15.4 mM/L (0.5%). This is consistent with the hypothesis by Kitagawa et al.29 that the main mechanism of cellular toxicity induced
by amitriptyline derives from its amphiphilic properties, which cause cell membrane solubilization and rupture at concentrations above its molecular selfaggregation level. According to that hypothesis, amitriptyline concentrations below a given selfaggregation level (0.3% for the intraspinal model) could be safe. To our knowledge, this was the first investigation to specifically examine both clinically and histologically the neurotoxicity profile of single intraspinal doses of amitriptyline in such small concentrations with reference to outcomes for longer than 1 wk. This study design allowed us to observe the presence of adhesive arachnoiditis without neural tissue injury, a finding secondary to intraspinal administration of this drug. This observation would have possibly gone undetected if the histological analysis had been performed earlier because the development of adhesive arachnoiditis usually occurs at a slower pace, weeks to months after spinal puncture.38,48 There were no signs of clinical neurologic deficits until final clinical evaluation, 21 days after spinal puncture. Other authors, analyzing the neurotoxicity of subarachnoid administration of different doses of the usual local anesthetics have also noted lack of association between histological and clinical findings related to adhesive arachnoiditis.49 51 This observation might be explained by the unfeasibility of performing concomitant long-term clinical follow-up of one animal and serial histological sections of spinal cord segments in the same animal at successive times. Nevertheless, because of the progressive nature of the meningeal changes described, it is likely that some animals would develop clinically relevant neurologic symptoms or deficits, such as pain or even cauda equina syndrome. It is not clear whether the present findings can be explained by the membrane solubilization hypothesis previously cited,29 not only because amitriptyline was administered at and below the reported membrane solubilization-inducing concentration but also because histological damage was restricted to the meninges (neural injury was excluded by GFAP staining). The current observation of adhesive arachnoiditis suggests either the presence of other inflammation-related toxicity mechanisms as proposed by others27,30,52 or a lower self-aggregation/membrane solubilization threshold. Our results suggest that this is a dosedependent mechanism because the KruskalWallis test disclosed statistical significance between the higher amitriptyline concentration experimental group and more severe arachnoiditis. One important question that must be addressed is whether arachnoiditis is a specific pathological finding of intrathecal amitriptyline because commonly used local anesthetics also have been reported to induce spinal cord toxicity.49,50 The mechanisms underlying those injuries are not well understood; however, the reported typical pattern of histological damage induced by local anesthetics is neuronal (axonal
swelling and degeneration, macrophage infiltration, demyelination, subpial vacuolization, and central necrosis) and not arachnoiditis or other meningeal changes. Other researchers34,35 studying intraspinal amitriptyline in different concentrations and using different techniques found meningeal lesions and neuronal damage. Therefore, the present findings suggest that arachnoiditis is specific to amitriptyline. Perhaps, further studies examining the differences between the pharmacologic properties of local anesthetics and amitriptyline and the differential pattern of histological damage associated with those drugs might lead to better understanding of their own toxicity mechanisms. This study was a blinded, randomized, controlled trial, in which potential confounding due to lesions induced by the spinal puncture procedure in the amitriptyline groups was controlled by comparison with the control group, in which normal saline was intraspinally administered in the same total volume and with a similar pH as the amitriptyline solution. The experimental model selected of single-dose intraspinal administration is similar to the usual spinal anesthesia procedures used in humans and displays less risk of complications than other models in which implantable intrathecal catheters are used. The drug solutions were free of chemical preservatives so as to exclude the possibility that the histological findings were reactions to a substance other than amitriptyline. The doses used were the lowest doses reported to have clinical effects. It seems unlikely that the present findings could be explained by unmeasured hemodynamic changes in the amitryptiline groups because other researchers20 using larger intrathecal doses of amitriptyline did not observe hemodynamic changes (spinal blood flow or mean arterial blood pressure). Similarly, fast removal and fixation of the anatomical piece, as well as the comparison with controls submitted to the same procedure, make our findings very unlikely due to cord extraction-related ischemic injuries or other procedure-related mechanisms. This study has limitations. The specific osmolality of the amitriptyline solutions was not measured. However, other researchers29 have described solutions of amitriptyline in normal saline at similar concentrations (0.1%, 0.3%, and 0.5%) displaying osmolalities ranging from 280 to 350 mOsm/kg, whereas the osmolality of CSF ranges from 280 to 300 mOsm/kg, which would not be expected to cause histological damage given the diluting volume of the CSF and the velocity of injection (10 s). A dichotomous classification was used instead of a scalar measure for the comparison of motor function, anal sphincter tone, and sensibility to painful stimulation; however, because the classification scheme used was quite stringent (the slightest deficit would be classified as positive), it is likely that the comparisons were even more rigorous than scalar comparisons. The assessment of anal
969
sphincter tone was made exclusively by visual inspection without use of a sphincter manometer. Even though the observation of sphincter relaxation is usually straightforward, the assessment method used would not be able to detect minor deficits regarding anal sphincter pressure. On the basis of the present results, intraspinal administration of amitriptyline solutions is markedly associated with the development of intense adhesive arachnoiditis even at low doses, for which there were no previous reports of toxicity. Even if no evidence of histological damage to the arachnoid membrane of dogs had been found, the therapeutic index of this drug would remain undetermined and its intrathecal use would not be recommended. Further studies are necessary to elucidate the mechanism of the meningeal toxicity observed. REFERENCES
1. Flor H, Kalso E, Dostrovsky JO. International Association for the Study of Pain: Proceedings of the 11th World Congress on Pain. Seattle: IASP Press, 2006 2. Flor H. IASP Scientific Program Committee: Pain 2005: an updated review: refresher course syllabus. Seattle: IASP Press, 2005 3. Devor M. Sodium channels and mechanisms of neuropathic pain. J Pain 2006;7:S312 4. Scadding JW. Treatment of neuropathic pain: historical aspects. Pain Med 2004;5(suppl 1):S3 8 5. Woolf CJ, Decosterd I. Implications of recent advances in the understanding of pain pathophysiology for the assessment of pain in patients. Pain 1999;suppl 6:S1417 6. Rowbotham MC. Treatment of neuropathic pain: perspective on current options. In: Justins DM, ed. Pain 2005: an updated review: refresher course syllabus. Seattle: IASP Press, 2005:10719 7. Max MB, Gilron IH. Antidepressants, muscle relaxants, and N-methyl-d-aspartate receptor antagonists. In: Loeser JD, ed. Bonicas management of pain. 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 2001:1710 26 8. Choi A, Mitchelson F. Variation in the affinity of amitriptyline for muscarine receptor subtypes. Pharmacology 1994;48: 293300 9. Galeotti N, Ghelardini C, Bartolini A. Involvement of potassium channels in amitriptyline and clomipramine analgesia. Neuropharmacology 2001;40:75 84 10. Ghelardini C, Galeotti N, Bartolini A. Antinociception induced by amitriptyline and imipramine is mediated by alpha2Aadrenoceptors. Jpn J Pharmacol 2000;82:130 7 11. Gumilar F, Arias HR, Spitzmaul G, Bouzat C. Molecular mechanisms of inhibition of nicotinic acetylcholine receptors by tricyclic antidepressants. Neuropharmacology 2003;45: 964 76 12. Lavoie PA, Beauchamp G, Elie R. Tricyclic antidepressants inhibit voltage-dependent calcium channels and Na()-Ca2 exchange in rat brain cortex synaptosomes. Can J Physiol Pharmacol 1990;68:1414 8 13. McCarson KE, Duric V, Reisman SA, Winter M, Enna SJ. GABA(B) receptor function and subunit expression in the rat spinal cord as indicators of stress and the antinociceptive response to antidepressants. Brain Res 2006;1068:109 17 14. Pancrazio JJ, Kamatchi GL, Roscoe AK, Lynch C, 3rd. Inhibition of neuronal Na channels by antidepressant drugs. J Pharmacol Exp Ther 1998;284:208 14 15. Reynolds JL, Ignatowski TA, Gallant S, Spengler RN. Amitriptyline administration transforms tumor necrosis factor-alpha regulation of norepinephrine release in the brain. Brain Res 2004;1023:11220 16. Sawynok J, Reid AR, Esser MJ. Peripheral antinociceptive action of amitriptyline in the rat formalin test: involvement of adenosine. Pain 1999;80:4555
17. Watanabe Y, Saito H, Abe K. Tricyclic antidepressants block NMDA receptor-mediated synaptic responses and induction of long-term potentiation in rat hippocampal slices. Neuropharmacology 1993;32:479 86 18. McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ, Moore RA. A systematic review of antidepressants in neuropathic pain. Pain 1996;68:21727 19. Barnet CS, Louis DN, Kohane DS. Tissue injury from tricyclic antidepressants used as local anesthetics. Anesth Analg 2005;101:1838 43 20. Cerda SE, Tong C, Deal DD, Eisenach JC. A physiologic assessment of intrathecal amitriptyline in sheep. Anesthesiology 1997;86:1094 103 21. Eisenach JC, Gebhart GF. Intrathecal amitriptyline. Antinociceptive interactions with intravenous morphine and intrathecal clonidine, neostigmine, and carbamylcholine in rats. Anesthesiology 1995;83:1036 45 22. Estebe JP, Myers RR. Amitriptyline neurotoxicity: dose-related pathology after topical application to rat sciatic nerve. Anesthesiology 2004;100:1519 25 23. Fridrich P, Eappen S, Jaeger W, Schernhammer E, Zizza AM, Wang GK, Gerner P. Phase Ia and Ib study of amitriptyline for ulnar nerve block in humans: side effects and efficacy. Anesthesiology 2004;100:1511 8 24. Gerner P, Haderer AE, Mujtaba M, Sudoh Y, Narang S, Abdi S, Srinivasa V, Pertl C, Wang GK. Assessment of differential blockade by amitriptyline and its N-methyl derivative in different species by different routes. Anesthesiology 2003;98:1484 90 25. Gerner P, Mujtaba M, Khan M, Sudoh Y, Vlassakov K, Anthony DC, Wang GK. N-phenylethyl amitriptyline in rat sciatic nerve blockade. Anesthesiology 2002;96:1435 42 26. Gerner P, Mujtaba M, Sinnott CJ, Wang GK. Amitriptyline versus bupivacaine in rat sciatic nerve blockade. Anesthesiology 2001;94:6617 27. Haller I, Lirk P, Keller C, Wang GK, Gerner P, Klimaschewski L. Differential neurotoxicity of tricyclic antidepressants and novel derivatives in vitro in a dorsal root ganglion cell culture model. Eur J Anaesthesiol 2007;24:702 8 28. Heughan CE, Allen GV, Chase TD, Sawynok J. Peripheral amitriptyline suppresses formalin-induced Fos expression in the rat spinal cord. Anesth Analg 2002;94:42731 29. Kitagawa N, Oda M, Nobutaka I, Satoh H, Totoki T, Morimoto M. A proposed mechanism for amitriptyline neurotoxicity based on its detergent nature. Toxicol Appl Pharmacol 2006;217:100 6 30. Lirk P, Haller I, Hausott B, Ingorokva S, Deibl M, Gerner P, Klimaschewski L. The Neurotoxic Effects of Amitriptyline Are Mediated by Apoptosis and are Effectively Blocked by Inhibition of Caspase Activity. Anesth Analg 2006;102:1728 33 31. McCaslin PP, Yu XZ, Ho IK, Smith TG. Amitriptyline prevents N-methyl-d-aspartate (NMDA)-induced toxicity, does not prevent NMDA-induced elevations of extracellular glutamate, but augments kainate-induced elevations of glutamate. J Neurochem 1992;59:4015 32. Sawynok J, Reid AR, Liu XJ, Parkinson FE. Amitriptyline enhances extracellular tissue levels of adenosine in the rat hindpaw and inhibits adenosine uptake. Eur J Pharmacol 2005;518:116 22 33. Srinivasa V, Gerner P, Haderer A, Abdi S, Jarolim P, Wang GK. The relative toxicity of amitriptyline, bupivacaine, and levobupivacaine administered as rapid infusions in rats. Anesth Analg 2003;97:915, table of contents 34. Sudoh Y, Desai SP, Haderer AE, Sudoh S, Gerner P, Anthony DC, De Girolami U, Wang GK. Neurologic and histopathologic evaluation after high-volume intrathecal amitriptyline. Reg Anesth Pain Med 2004;29:434 40 35. Yaksh TL, Tozier N, Horais KA, Malkmus S, Rathbun M, Lafranco L, Eisenach J. Toxicology profile of N-methyl-daspartate antagonists delivered by intrathecal infusion in the canine model. Anesthesiology 2008;108:938 49 36. Chen YW, Huang KL, Liu SY, Tzeng JI, Chu KS, Lin MT, Wang JJ. Intrathecal tri-cyclic antidepressants produce spinal anesthesia. Pain 2004;112:106 12 37. Burton CV. Lumbosacral arachnoiditis. Spine 1978;3:24 30 38. Woods W, Franklin R. Progressive adhesive arachnoiditis following spinal anesthesia. Calif Med 1951;75:196 8 39. Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 1983;16:109 10
970
40. Fleiss JL, Levin BA, Paik MC. Statistical methods for rates and proportions. 3rd ed. Hoboken, NJ: Wiley, 2003 41. R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN3-900051-07-0, URL: http://www.R-Project.org. 42. Sudoh Y, Cahoon EE, Gerner P, Wang GK. Tricyclic antidepressants as long-acting local anesthetics. Pain 2003;103:49 55 43. Bryson HM, Wilde MI. Amitriptyline. A review of its pharmacological properties and therapeutic use in chronic pain states. Drugs Aging 1996;8:459 76 44. Sawynok J, Reid A. Peripheral interactions between dextromethorphan, ketamine and amitriptyline on formalin-evoked behaviors and paw edema in rats. Pain 2003;102:179 86 45. Gerner P, Luo SH, Zhuang ZY, Djalali AG, Zizza AM, Myers RR, Wang GK. Differential block of N-propyl derivatives of amitriptyline and doxepin for sciatic nerve block in rats. Reg Anesth Pain Med 2005;30:344 50 46. Gerner P. Tricyclic antidepressants and their local anesthetic properties: from bench to bedside and back again. Reg Anesth Pain Med 2004;29:286 9
47. Eisenach JC, Gebhart GF. Intrathecal amitriptyline acts as an N-methyl-d-aspartate receptor antagonist in the presence of inflammatory hyperalgesia in rats. Anesthesiology 1995;83:104654 48. Hurst EW. Adhesive arachnoiditis and vascular blockage caused by detergents and other chemical irritants: an experimental study. J Pathol Bacteriol 1955;70:16778 49. Rosen MA, Baysinger CL, Shnider SM, Dailey PA, Norton M, Curtis JD, Collins M, Davis RL. Evaluation of neurotoxicity after subarachnoid injection of large volumes of local anesthetic solutions. Anesth Analg 1983;62:802 8 50. Ready L, Plumer M, Haschke R, Austin E, Sumi S. Neurotoxicity of intrathecal local anesthetics in rabbits. Anesthesiology 1985;63:364 70 51. Ganem EM, Vianna PT, Castiglia YMM, Marques M, Vane LA. Neurotoxicity of subarachnoid hyperbaric bupivacaine in dogs. Reg Anesth Pain Med 1996;21:234 8 52. Chung EY, Shin SY, Lee YH. Amitriptyline induces early growth response-1 gene expression via ERK and JNK mitogenactivated protein kinase pathways in rat C6 glial cells. Neurosci Lett 2007;422:43 8
971
Regional Anesthesia
Section Editor: Terese T. Horlocker
Sciatic Nerve Catheter Placement: Success with Using the Raj Approach
Christopher Robards, MD R. Doris Wang, MD Steven Clendenen, MD Beth Ladlie, MD Roy Greengrass, MD
BACKGROUND: Continuous regional analgesia has increased in popularity and is becoming standard of care for many painful surgical procedures. Various approaches of sciatic catheter insertion have been proposed, each with attributes and disadvantages. We investigated whether the Raj approach that uses a simple midpoint landmark between the ischial tuberosity and greater trochanter will facilitate sciatic catheter placement. METHODS: After informed consent, 20 patients were recruited to receive sciatic catheter placement using the Raj approach. An insulated Tuohy needle was inserted perpendicular to skin at the midpoint of a line between the ischial tuberosity and greater trochanter. After sciatic nerve stimulation, a catheter was inserted 2 4 cm past the end of the needle and secured. The catheters were then incrementally injected with 30 mL of 1.5% mepivacaine. Twenty minutes after local anesthetic injection, sensory block was assessed using cold and pinprick tests, whereas motor block was assessed using a modified Bromage score. Complications and side effects were recorded. RESULTS: In all instances, blocks were easy to perform and were successful. No major side effects or complications were noted. CONCLUSION: Use of a simple landmark between easily identifiable bony structures enhances the simplicity and placement of a sciatic nerve catheter and is recommended for use in clinical practice.
(Anesth Analg 2009;109:9725)
ostoperative pain associated with total joint replacements, particularly total knee arthroplasty, is often severe and refractory to IV opioids. Advances in needle and catheter systems have enabled continuous regional analgesia to become increasingly popular with the primary advantage of providing analgesia well into the postoperative period. Single injection sciatic nerve block and catheter techniques have been performed utilizing different approaches at many levels after the nerve exits the sciatic notch in the pelvis.15 One of the most commonly used techniques of sciatic nerve blockade is the Labat approach during which the patient is placed in the Sims position. Because the Euclidian Geometry required to identify a surface landmark for needle insertion using this technique can be difficult, particularly in an obese patient population, alternative approaches have been suggested.6,7 Although these alternative approaches may appear promising, the soft tissue landmarks identified using these techniques can still be misleading. In 1975, Raj described a technique of blocking the sciatic nerve at the point where it lies between the greater trochanter of the femur and ischial tuberosity.8 The
From the Department of Anesthesiology, the Mayo Clinic Florida, Jacksonville, Florida. Accepted for publication April 9, 2009. Address correspondence and reprint requests to Christopher B. Robards, MD, Mayo Clinic Florida, 4500 San Pablo Rd., Jacksonville, FL 32224. Address e-mail to robards.christopher@mayo.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae0ee7
accuracy and simplicity of this technique relies on the fact that these bones are easily palpable and consistent landmarks requiring solely their identification and the midpoint between them before needle insertion. In addition to the simplicity of needle insertion site identification, this technique is desirable because moving the patient into a Sims position is not required and the sciatic nerve is more superficial at this level compared with transgluteal approaches. Although there is some evidence demonstrating success with sciatic nerve catheters using another subgluteal approach,2,9 there are no studies in which a Raj approach was used for catheter placement. This study was designed to determine whether the simplicity of the single injection technique could be reproduced using a catheter placement technique. This pilot study was performed to determine the reliability, feasibility, and success rate of sciatic catheters placed via the Raj approach.
METHODS
After institutional review board approval, 20 patients, who met inclusion criteria (patients undergoing primary unilateral knee arthroplasty or ankle surgery, 18 yr-of-age, ASA physical status of I, II, or III, competent and able to provide informed consent), were recruited for the study. Exclusion criteria included patients 18 yr of age, pregnant or lactating patients, patients who are unwilling or unable to provide written informed consent, patients who refuse
972
Figure 2. Surface landmarks and needle insertion site for

sciatic nerve block using Raj approach.
Figure 1. Proper leg positioning for sciatic nerve block using Raj approach. regional anesthesia, patients who have a contraindication to regional anesthesia (i.e., coagulopathy, bleeding diathesis), and patients who have a known allergy to amide local anesthetics. After informed consent, standard monitors (noninvasive blood pressure cuff, ECG, pulse oximetry) were placed on each patient. Supplemental oxygen was administered via nasal canula at 23 L/min. Intravenous midazolam was administered in increments of 1 mg for anxiolysis and fentanyl was administered in increments of 50 g for pain associated with the procedure. The patient was positioned supine with the operative extremity flexed at the hip and flexed 90 at the knee (Fig. 1). The greater trochanter and ischial tuberosity were identified and marked (Fig. 2). The needle insertion site was at the midpoint of a line joining the greater trochanter and ischial tuberosity at the level of the gluteal crease. After sterile prep and drape, and subcutaneous local anesthetic injection, an 18-G, 4-in. (100 mm) insulated Tuohy needle (B. Braun, Bethlehem, PA) attached to a nerve stimulator (Stimuplex Dig RC, B. Braun) with an initial setting of 1.5 mA, 2 Hz was inserted perpendicular to the skin with the bevel directed cephalad (toward the gurney). On successful sciatic nerve stimulation (plantar flexion/dorsiflexion)
at a current of 0.5 mA, a 20-G multiport, closed tip polyamide nylon catheter (B. Braun) was advanced through the needle to lie approximately 2 4 cm past the tip. If difficulty was encountered with catheter advancement (inability to easily advance catheter on first attempt), 10 mL of sterile saline were injected through the needle injection side port, and an attempt to advance the catheter was repeated. An anesthesiologist present during the block procedure documented the number of needle passes and the time taken to perform the block procedure. After negative aspiration and negative test dose injection (3 mL 1.5% lidocaine with 1:200,000 epinephrine), 30 mL of preservative free mepivacaine 1.5% was injected through the sciatic catheter in 5 mL increments. The catheters were fixed to the skin with Steri-Strips (3M, St. Paul, MN) and a Tegaderm (3M). Sensory and motor blockade were assessed 20 min after the completion of injection by another anesthesiologist who was blinded to the regional block technique performed. Sensory blockade was assessed in the plantar and dorsal surfaces of the foot using a blunt tip needle. Motor block of the tibial and common peroneal branches of the sciatic nerve were assessed using a modified Bromage score (Appendix). Successful block was defined as the presence of sensory block and motor block (Bromage scale 2) in either the common peroneal or tibial nerve distribution, 20 min after local anesthetic injection through the catheter.
Data Collection
The following information was collected from 20 patients in this observational study: age, sex, height, weight, time from needle insertion to completion of catheter insertion, total number of skin punctures made during block, total number of redirections made during block, lowest current achieved during block, difficulty during catheter insertion, need for saline dilation, distance from skin to sciatic nerve, depth of catheter, presence of blood in catheter, fentanyl dose, midazolam dose, Bromage score in the tibial nerve distribution, Bromage score in the peroneal nerve distribution, and
973
Table 1. Patient Characteristics Variable

Age Sex (male) BMI
Numerical data are summarized with the sample mean (SD). BMI body mass index.
Table 2. Surgery/Block Characteristics Summary (N 20)

69 (9) 10 (50%) 31 (4)
Variable
Side of surgery (right) Time taken to place needle and catheter (min) Number of skin punctures made during block 1 2 3 Number of redirections made during block Lowest current achieved during block (mA) Nerve distribution with motor responsea Tibial nerve Peroneal nerve Difficulty during catheter insertion Saline dilation Distance from skin to sciatic nerve (cm) Depth of catheter (cm) Fentanyl dose (g) Midazolam dose (mg)
Summary (N 20)
9 (45%) 2.8 (1.6) 15 (75%) 4 (20%) 1 (5%) 1 (04) 0.46 (0.09) 15 (79%) 4 (21%) 2 (10%) 4 (20%) 5.9 (2.0) 6.5 (2.1) 128 (62) 2.9 (1.1)
sensory block assessment at the bottom of the foot, heel, anterior and lateral aspect of the foot.
Numerical data was summarized with the sample median and range. Categorical data was summarized with number and percent of patients. An exact binomial 95% confidence interval was used to estimate the proportion of successful blocks.
RESULTS
Twenty patients were recruited for the study. All 20 patients completed the study. All patients tolerated positioning and catheter insertion well using moderate sedation. Patient characteristics are presented in Table 1. All 20 patients (100%) had a successful block. Most patients (75%) required one needle puncture site, only one required more than two puncture sites. Average time for catheter placement was 2.8 min with 1.2 needle redirections. Difficult catheter advancement was encountered in two patients (10%) but was easily managed with saline injection through the needle before catheter readvancement. Two other patients had saline injection through the needle before catheter advancement because of the anesthesiologists preference. Average distance of the sciatic nerve from skin was 5.9 cm. Tibial nerve stimulation was observed in 16 patients and peroneal nerve stimulation in four patients (Table 2). Success rate of the Raj sciatic catheter is summarized in Table 3. All patients had sensory anesthesia in the distribution of both the tibial and peroneal nerves. At the 20 min time point, 16 patients had significant weakness in both terminal nerve distributions, whereas four patients had significant weakness in the peroneal distribution only. None of the patients demonstrated any signs of local anesthetic toxicity during the procedure.
Numerical data are summarized with the sample mean (SD) and the number of redirections given as a median with range. a Not available for one patient.
Table 3. Patient Outcomes Variable

Sensory block Plantar aspect of the foot Dorsum of the foot Motor block Plantar flexion weakness (Bromage T 2) Dorsi flexion weakness (Bromage P 2) Successful block
Summary (N 20)
20 (100%) 20 (100%) 16 (80%) 20 (100%) 20 (100%)
DISCUSSION
Continuous sciatic nerve blockade using a subgluteal approach has been described in the past with excellent results.9 12 However, we know of no reports of sciatic nerve catheter placement using a Raj approach. The benefits of using this approach over other approaches are that patient repositioning to a Sims position is not required and that dependence on difficult to identify landmarks is eliminated. Easy to identify landmarks are particularly important in an obese patient population where block failure is more likely to occur.13 All patients in our study population (100%) had a successful block as defined by sensory
974
Raj Approach for Sciatic Nerve Catheters
loss to pinprick on the dorsum or plantar surface of the foot at 20 min (in actuality both plantar and dorsum sensory blockade was present in all patients). All blocks were quickly performed (average, 2.8 min) with minimal needle redirections (average, 1 redirect). Most catheters (90%) were easily advanced on the first attempt. The remaining two catheters (10%) were easily advanced after the injection of 10 mL of saline. Although data were not compiled regarding analgesia in the postoperative period, because the local anesthetic was administered through the peripheral nerve catheter, a functional catheter is assumed. Pain associated with total knee arthroplasty is variable, particularly in the posterior knee. In fact some patients, although it is the minority, do not require the use of continuous sciatic nerve blockade.14 This has led some practitioners to perform single injection sciatic nerve blockade or forego sciatic nerve blockade altogether in favor of IV patient-controlled analgesia. However, because the majority of patients do in fact have at least some degree of posterior knee pain following total knee arthroplasty,14 an easy to perform, reliable approach to sciatic nerve catheter placement is particularly attractive for patients in whom narcotic analgesia is best avoided (allergy, history of opioid associated nausea/vomiting, obstructive sleep apnea, cognitive
dysfunction)15,16 or is likely to be inadequate (chronic opioid dependence).17 With the increasing prevalence of obesity, an approach to sciatic nerve blockade that is both easily performed and reliable is ideal. By flexing the leg at the hip in the supine position, the Raj approach to sciatic nerve blockade accentuates the two bony landmarks necessary for identification and simplifies nerve blockade. Furthermore, it requires little cooperation on the part of the patient by allowing them to remain in a supine position. It should be noted that a subgluteal approach for sciatic nerve blockade may not provide adequate posterior thigh anesthesia or analgesia because of a proximal branching of the posterior cutaneous nerve of the thigh.6 There are clearly some limitations to this pilot study. First of all, although all patients in our study population were considered overweight, few of them were obese, and there were no patients considered morbidly obese by body mass index criteria. Because the aim of our study was to demonstrate the feasibility of sciatic nerve catheter placement and subsequent blockade with through the catheter injection of local anesthetic using the Raj approach, patients with a normal or slightly elevated body mass index were not excluded. A follow-up study to test the utility of this approach over the classic Labat approach in an obese patient population should recruit morbidly obese patients and compare the two approaches directly. Secondly, because sensory and motor testing data in the postoperative period were not included, we are unable to ascertain the functionality of these catheters in terms of postoperative analgesia. However, the catheters were used for the initial dosing, and 100% of patients had a sensory deficit and 80% of patients had a motor deficit in both tibial and common peroneal distributions before the surgical procedure, making some degree of functionality implicit. Furthermore, the employed method of dosing sciatic catheters in the postoperative period at our institution (due in part to surgeon concern for foot drop and desire for participation in rehabilitation) is one of the intermittent boluses (4 6 mL every 4 6 h). This is identical except in terms of volume to the initial (and 100% successful) block. Because all patients studied underwent total knee arthroplasty and received either spinal anesthesia or general anesthesia we cannot determine whether or not the blocks ever attained true surgical anesthesia. Again, comprehensive sensory and motor testing indicated that all blocks were successful. In summary, our data suggest that a functional peripheral nerve catheter can be easily placed using the Raj approach to sciatic nerve blockade with a high degree of success. Although further randomized controlled studies are necessary to draw extensive conclusions from this data, the implication is that this approach is an easy alternative to other previously described approaches to sciatic nerve catheter placement and blockade.
APPENDIX
Description of the Bromage score adapted to the tibial nerve: grade criteria I Full capacity for plantar flexion of the foot II Just able to plantar flex the foot III Unable to plantar flex the foot but with free movement of the toes IV Unable to move the foot Description of the Bromage score adapted to the peroneal nerve: grade criteria I Full capacity for dorsiflexion of the foot II Just able to dorsiflex the foot III Unable to dorsiflex the foot but with free movement of the toes IV Unable to move the foot
REFERENCES
1. Di Benedetto P, Casati A, Bertini L, Fanelli G. Posterior subgluteal approach to block the sciatic nerve: description of the technique and initial clinical experiences. Eur J Anaesthesiol 2002;19:682 6 2. Di Benedetto P, Bertini L, Casati A, Borghi B, Albertin A, Fanelli G. A new posterior approach to the sciatic nerve block: a prospective, randomized comparison with the classic posterior approach. Anesth Analg 2001;93:1040 4 3. Chelly JE, Delaunay L. A new anterior approach to the sciatic nerve block. Anesthesiology 1999;91:1655 60 4. Morris GF, Lang SA, Dust WN, Van der Wal M. The parasacral sciatic nerve block. Reg Anesth 1997;22:223 8 5. Fanelli G, Sansone V, Nobili F, Pedotti E, Aldegheri G. [Locoregional anesthesia for surgical arthroscopy of the knee]. Minerva Anestesiol 1992;58:1215 6. Franco CD, Choksi N, Rahman A, Voronov G, Almachnouk MH. A subgluteal approach to the sciatic nerve in adults at 10 cm from the midline. Reg Anesth Pain Med 2006;31:21520 7. Franco CD. Posterior approach to the sciatic nerve in adults: is euclidean geometry still necessary? Anesthesiology 2003;98:723 8 8. Raj PP, Parks RI, Watson TD, Jenkins MT. A new single-position supine approach to sciatic-femoral nerve block. Anesth Analg 1975;54:489 93 9. Taboada M, Rodriguez J, Valino C, Vazquez M, Laya A, Garea M, Carceller J, Alvarez J, Atanassoff V, Atanassoff PG. A prospective, randomized comparison between the popliteal and subgluteal approaches for continuous sciatic nerve block with stimulating catheters. Anesth Analg 2006;103:244 7 10. Di Benedetto P, Casati A, Bertini L. Continuous subgluteus sciatic nerve block after orthopedic foot and ankle surgery: comparison of two infusion techniques. Reg Anesth Pain Med 2002;27:168 72 11. Di Benedetto P, Casati A, Bertini L, Fanelli G, Chelly JE. Postoperative analgesia with continuous sciatic nerve block after foot surgery: a prospective, randomized comparison between the popliteal and subgluteal approaches. Anesth Analg 2002;94:996 1000 12. Cappelleri G, Aldegheri G, Ruggieri F, Mamo D, Fanelli G, Casati A. Minimum effective anesthetic concentration (MEAC) for sciatic nerve block: subgluteus and popliteal approaches. Can J Anaesth 2007;54:2839 13. Cotter JT, Nielsen KC, Guller U, Steele SM, Klein SM, Greengrass RA, Pietrobon R. Increased body mass index and ASA physical status IV are risk factors for block failure in ambulatory surgeryan analysis of 9,342 blocks. Can J Anaesth 2004;51:810 6 14. Ben-David B, Schmalenberger K, Chelly JE. Analgesia after total knee arthroplasty: is continuous sciatic blockade needed in addition to continuous femoral blockade? Anesth Analg 2004;98:7479 15. Aubrun F, Marmion F. The elderly patient and postoperative pain treatment. Best Pract Res Clin Anaesthesiol 2007;21:109 27 16. Cullen DJ. Obstructive sleep apnea and postoperative analgesiaa potentially dangerous combination. J Clin Anesth 2001;13:835 17. Carroll IR, Angst MS, Clark JD. Management of perioperative pain in patients chronically consuming opioids. Reg Anesth Pain Med 2004;29:576 91
975
Regional Anesthesia for Vascular Access Surgery

Elizabeth B. Malinzak, BS Tong J. Gan, MHS, MB, FRCA
BACKGROUND: Approximately 25% of initial arteriovenous fistula (AVF) placements will fail as a result of thrombosis or failure to develop adequate vessel size and blood flow. Fistula maturation is impacted by patient characteristics and surgical technique, but both increased vein diameter and high fistula blood flow rates are the most important predictors of successful AVFs. Anesthetic techniques used in vascular access surgery (monitored anesthesia care, regional blocks, and general anesthesia) may affect these characteristics and fistula failure. METHODS: We performed a literature search using key words in the PubMed/ MEDLINE database. Seven articles that related to the effects of anesthesia on AVF construction, including sympathetic block, vein dilation, blood flow, adverse outcomes, or patency rates, comprised the sources for this review. RESULTS: Significant vasodilation after regional block administration is seen in both the cephalic and basilic veins. These vasodilatory properties may assist with AVF site selection. In the intraoperative and postoperative periods, use of a regional block, compared with other anesthetic techniques, resulted in significantly increased fistula blood flow. The greater sympathetic block contributed to vessel dilation and reduced vasospasm. Use of regional techniques in AVF construction yielded shorter maturation times, lower failure rates, and higher patency rates. CONCLUSION: Use of regional blocks may improve the success of vascular access procedures by producing significant vasodilatation, greater fistula blood flow, sympathectomy-like effects, and decreased maturation time. However, a largescale, prospective, clinical trial comparing the different anesthetic techniques is still needed to verify these findings.
(Anesth Analg 2009;109:976 80)
nd-stage renal disease (ESRD) currently affects more than 350,000 Americans, and each year, this population increases by approximately 7%.1,2 Creation of permanent vascular access though surgical construction of an arteriovenous fistula (AVF) is preferred for ESRD patients receiving chronic hemodialysis. However, about one quarter of these initial attempts will fail as a result of stenosis or failure of the vessels to develop adequate blood flow.3 Interventions to maintain the access site cost dialysis patients more than $600 million per year.4 Consequently, many studies have attempted to determine the effect of patient characteristics, anesthetic management, or surgical techniques on graft or fistula failure rate. In this investigation, we performed a review of the literature to determine how the use of regional anesthesia can influence outcomes in AVF construction.
From the Department of Anesthesiology, Duke University Medical Center, Duke University School of Medicine, Durham, North Carolina. Accepted for publication April 3, 2009. Supported by Medical Student Anesthesia Research Fellowship, Foundation for Anesthesia Education and Research. Address correspondence and reprint requests to Dr. Tong J. Gan, Department of Anesthesiology, Duke University Medical Center, Box 3094, Durham, NC 27710. Address e-mail to tjgan@duke.edu. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181adc208
METHODS
The literature search for this review was conducted in December 2007 using the PubMed/MEDLINE database. Two key words, one from List A and one from List B, were joined with the term and in all possible combinations for the literature search. Key words from List A included arteriovenous fistula, AVF, vascular access, arteriovenous graft, AVG, dialysis, dialysis access, end-stage renal disease, ESRD, chronic kidney disease, and CKD. List B included regional anesthesia, brachial plexus block, BPB, supraclavicular, infraclavicular, axillary, and interscalene. The results were limited to full text articles published in the English language. After the initial search, five articles were excluded because they did not pertain to the use of regional anesthesia for creation of vascular access for dialysis in adults. Of the remaining articles, seven related to the effects of regional anesthesia on AVF construction, including sympathetic block, vein dilation, blood flow, adverse outcomes, or patency rates (Tables 1 and 2). These articles comprised the basic sources for the purposes of this review. The other 16 articles described or compared various regional block approaches for vascular access placement in terms of onset, duration of action, and/or pharmacokinetics of local anesthetics. These articles exceeded the intended scope of this review but were helpful sources for background information.
976
Table 1. Studies Comparing Regional Anesthesia with Other Anesthetic Techniques in Vascular Access Surgery Author
Mouquet et al.
12
Subjects and methods

36 AVF or AVG subjects receiving BPB (n 9), MAC (n 9), ISO (n 9), or HAL (n 9)
Main outcome(s)
Brachial artery blood flow and AVF blood flow (mL/min)
Results
During anesthesia and in the immediate postoperative period, blood flow in the BPB group was significantly greater than the HAL group. In the late postoperative period, BPB and both general anesthesia groups had significantly higher blood flow than MAC. As compared with GA, BPB causes significant venous dilation and a significantly lower PI during and after the surgery. Use of a SGB resulted in significantly higher postoperative AVF blood flow, significantly higher average peak velocity, and significantly lower mean maturation time than LA alone. There was no significant association between anesthetic technique and any adverse outcome.
Shemesh et al.27
36 AVF or AVG subjects receiving BPB (n 31) or GA (n 5)
PI (measure of sympathectomy), basilic vein diameter (mm) AVF blood flow (mL/ min), peak radial artery velocity (cm/s), maturation time (d)
Yildirim et al.32
50 AVF subjects receiving SGB LA (n 25) or LA (n 25)
Solomonson et al.26
469 AVF subjects receiving LA, GA, or BPB
Morbidity and mortality
AVF arteriovenous stula; AVG arteriovenous graft; BPB brachial plexus block; MAC monitored anesthesia care; ISO isourane; HAL halothane; PI pulsatility index; GA general anesthesia; SGB stellate ganglion block; LA local anesthesia.
Table 2. Studies Reporting the Effects of Regional Anesthesia on Outcomes in Vascular Access Surgery Author
Shemesh et al.
8
Subjects and methods

157 AVF or AVG subjects receiving a supraclavicular BPB
Results
AVF: 57.3% of vascular accesses placed AVF immediate failure: 0% AVF early failure: 6.8% AVF 1 yr-assisted primary patency rate: 81.8% AVF 18 mo secondary patency rate: 98.6% Average basilica and cephalic vein diameters significantly increased after BPB 30% of subjects had modification of original operative plan from AVG to AVF or from proximal to distal AVF site No significant different in patency rate between the group that maintained the original operative plan and the group with changes Venodilatation: post-BPB versus tourniquet: Distal cephalic 42% Midcephalic 19% Midbasilic 26%
Comments
Description of successful algorithm designed to achieve DOQI goals using BPBs. AVFs have a low early failure rate and a high secondary patency rate.
Laskowski et al.31
26 AVF or AVG subjects receiving an infraclavicular BPB
Use of regional anesthesia can lead to improved site selection and increased opportunity for AVF creation because of significant vasodilation, without adverse effects on patency.
Hingorani et al.19
34 AVF or AVG subjects receiving axillary, interscalene, and/or infraclavicular BPB
Venodilatation with BPB is augmented compared with tourniquet.
AVF arteriovenous stula; AVG arteriovenous graft; BPB brachial plexus block; DOQI Kidney Disease Outcomes Quality Initiative.
Types of Vascular Access

A functional and mature access site is able to be cannulated by two dialysis needles and achieve flow
rates more than 350 mL/min, usually beginning 3 4 mo after placement. Early failure of a fistula occurs between 1 wk and 1 mo after a surgical procedure to
977
establish access. Primary patency is the interval from access placement until any intervention to the inflow artery, venous outflow, or central veins designed to maintain or reestablish functionality and maturation. Interventions can include angioplasty, vein ligation, thrombectomy or thrombolysis, or surgical revision.3,57 Two methods are currently used to connect an artery and vein to establish vascular access: an autogenous AVF or a nonautogenous arteriovenous graft using a prosthetic or biograft conduit. Arteriovenous grafts were used more commonly because of their ease of placement, low early thrombosis rate, and short interval to cannulation. However, AVFs require fewer interventional procedures, have better long-term patency, and a lower rate of complications leading to failure.2,8 Therefore, the Kidney Disease Outcomes Quality Initiative guidelines recommend the construction of autogenous radiocephalic AVFs in the nondominant upper extremity as the first option for patients with ESRD.9,10*
Table 3. Factors That Increase Risk of AVF Failure Factor

Patient characteristics
33
Specics
Age 65 yr Female34,35 Tobacco use6,34 History of AVF (especially ipsilateral)6,26,36 Dialysis36 Diabetes mellitus6,34,35 Peripheral vascular disease33 Coronary artery disease33 Hypertension6 Cephalic vein diameter 2 mm6 Radial artery diameter 1.6 mm6 Distal location35 Anastomosis: side-to-side end-toside,37 accomplished by suture (versus clips/staples)16,34 Use of venous loops (versus clamps)36 No intraoperative heparin36 Limited surgical experience34,35 Postanastomosis blood flow 160 mL/min13 AVF blood flow 350 mL/min5 Cephalic vein diameter 4 mm14 Lack of access surveillance34 Limited use of antiplatelet agents and/or calcium channel blockers34
Comorbidities
Preoperative Surgical/ intraoperative
Failure of AVFs
Although they are the first choice for vascular access, radiocephalic AVFs have a primary failure range of 24%35% because of thrombosis or inadequate blood flow.11 After completion of the anastomosis, turbulent blood flow activates platelets and endothelial cells, which can result in thrombosis and reduction of fistula blood flow, usually in the 3 4 cm of the vein from the anastomosis.3,5 Primary AVF failure rates may be affected by certain patient characteristics and surgical factors (Table 3). During and after surgery, maintaining a high blood flow helps to prevent thrombosis and failure in the postoperative period.12 AVFs with flow rates more than 160 mL/min after completion of the anastomosis and AVFs with flow rates more than 350 mL/min 2 wk postprocedure have higher patency and maturation rates.5,6,13 Vessel size more than 0.4 cm also improves maturation rates by allowing for decreased resistance and increased blood flow.14 However, after AVF construction, the vessel dilation created by nitric oxide and other vasodilators is limited by the wall shear stress produced by inflammation and intimal hyperplasia.15,16 Even a year after construction, the shear stress does not return to baseline.17
Postoperative
AVF arteriovenous stula.
Anesthesia in Vascular Access Surgery

Three anesthetic techniques are commonly used for vascular access surgery: monitored anesthesia care (MAC), regional anesthesia, and general anesthesia. MAC, which combines sedation with local anesthesia, is a simple, well-tolerated technique but can be difficult to maintain for a long duration. It may require repeated injections, which entails additional pain and
*NKF-F/DOQI Clinical Practice Guidelines for Vascular Access: Update 2000. National Kidney Foundation. Available at: www. kidney.org/professionals/kdoqi/guidelines_updates/doqi_uptoc. html#va. Accessed November 7, 2007.
sedation as well as an increased risk of intravascular injection, potentially leading to central nervous system or cardiac toxicity.18 20 Furthermore, this method does not offer motor block, and deleterious vasospasm is more common and severe with this technique.5,19 For prolonged or complex vascular access procedures, general anesthesia is commonly used. As compared with local anesthesia, this method does not fail to provide adequate anesthesia and also increases AVF blood flow.21 However, many patients requiring vascular access have severe comorbidities, including cardiovascular disease, chronic lung disease, metabolic disease, neuropathy, and/or immunosuppression, which can lead to changes in the patients hemodynamics, stress response, and potential drug interactions during general anesthesia.5,19,20,22 Regional anesthesia may attenuate the side effects of general anesthesia. For example, in patients undergoing AVF placement, the use of a BPB bypasses the stress of induction and avoids the hemodynamic disturbances and systemic drug effects seen in general anesthesia patients with severe comorbidities.21,23,24 BPB are more often used for AVF creation in diabetic, cardiac, and elderly patients than general anesthesia or MAC.25 Regional anesthesia provides better postoperative analgesia and faster recovery from anesthesia.19,26 28 There are, however, risks with this technique, including unintentional damage to the surrounding anatomy, neuropathy from nerve injury, hematoma, infection, and injection of local anesthetic in vessels leading to central nervous system and cardiac toxicity.29 There also can be a longer latency
978
Anesthesia for Vascular Access Surgery
between administration and anesthesia, and a small failure rate between 1% and 3% depending on the experience of the anesthesiologist.5,21 The hyperdynamic circulation in ESRD patients promotes absorption of local anesthetics into the bloodstream, resulting in high plasma concentrations.30 However, despite high mepivacaine plasma concentrations after BPB in 10 ESRD patients, there were no reports of serious systemic toxicity.21
Inuence of Regional Anesthetic Techniques on AVF Outcomes

Few studies have examined how the use of regional anesthesia affects the construction of AVFs regarding vasodilation, blood flow, sympathetic block, and patency (Table 1). Vasodilation after regional block administration is seen in both the cephalic and basilic veins. In one study, compared with application of a tourniquet, use of a block increased venodilation by 42% in the distal cephalic vein, 19% in the midcephalic vein, and 26% in the midbasilic vein.19 Significant basilic venous dilation was confirmed in a prospective study of 36 patients, with more than 1.5 mm average dilation measured after regional block.27 These vasodilatory properties may assist with site selection. After administration of a regional block and the resulting vasodilation, surgical plans were altered (i.e., graft to fistula or proximal to more distal site) in 30% of patients with no change in patency rates.31 The vasodilatory properties of regional block may help to maintain a high blood flow rate in the fistula both in the intraoperative and postoperative periods. Mouquet et al.12 measured and calculated fistula blood flow in four anesthetic groups: MAC, BPB, isoflurane, and halothane. After administration of anesthesia, there was a significant increase in brachial artery blood flow in the BPB group because of vasodilatation, decreased forearm vascular resistance, and greater blood velocity. The general anesthetic groups also had enhanced blood flow but only as a result of improved blood velocity. In the immediate postoperative period, the BPB subjects had significantly higher AVF blood flow than halothane and MAC subjects. At postoperative Days 3 and 10, all groups but the MAC cohort had a high blood flow with decreased forearm vascular resistance. Overall, the blood flow increased most significantly and with minimal hemodynamic changes in the BPB group as a result of its vasodilatory effect.12 Sympathetic block may have a direct effect on the vein to produce dilation. Alternatively, the sympathetic block may cause arterial dilation, which augments venous return and consequently produces venodilation.31 In one AVF study, the diminished sympathetic tone generated by preemptive stellate ganglion blockade improved arterial dilation and prevented radial artery spasm by lessening the reactivity of the arterial muscle that normally results from
surgical manipulation.32 Sympathectomy-like effects of a BPB in AVF surgery were illustrated by Shemesh et al. by calculating the pulsatility index (PI) ratio, a ratio of arterial resistances and blood flows that reflects the magnitude of the sympathetic change. In the regional block subjects, PI decreased and remained low for 5 h after block administration, reflecting a sympathectomy-like effect. In the general anesthesia patients, the PI diminished after induction but increased immediately after the patients regained consciousness.27 A greater sympathectomy-like effect, in combination with enlarged vessel diameter and increased blood flow, may also enhance to the patency of the fistula. The use of the preemptive stellate ganglion blockade also resulted in increased average fistula flow, greater average peak radial artery velocity, and slower mean AVF maturation time.32 Additionally, the combination of a tourniquet and a BPB yielded AVFs that had low failure rates (0% immediate failure and 6.8% failure within 1 mo) and high patency rates (81.8% at 1 yr and 98.6% at 2 yr).8
CONCLUSIONS
It is evident that the successful creation and maturation of AVFs is affected by a number of factors. Although preoperative planning and variations in the surgical procedure might affect the success of the procedure, additional factors in the perioperative period, including choice of anesthetic technique, may affect the physiologic response in the patient and the fistula. Use of regional blocks may likely improve the success of vascular access procedures. They have been shown to allow for significant vasodilatation, higher fistula blood flow, and sympathectomy-like effects. They have the potential to affect site and vessel selection for the AVF, as well as fistula maturation. However, without a large-scale, prospective, clinical trial, it still remains unclear whether the prevailing anesthetic techniques are associated with different surgical outcomes. REFERENCES
1. Schulman G, Peale C, Himmelfarb J. Hemodialysis. In: Brenner BM, Rector FC, eds. Brenner & Rectors the kidney. 8th ed. Philadelphia, PA: Saunders, 2008:195799 2. Tolkoff-Rubin N. Treatment of irreversible renal failure. In: Cecil RL, Goldman L, Ausiello DA, eds. Cecil textbook of medicine. 23nd ed. Philadelphia, PA: Saunders, 2008:936 40 3. Kian K, Vassalotti JA. The new arteriovenous fistula: the need for earlier evaluation and intervention. Semin Dial 2005;18:37 4. Patel ST, Hughes J, Mills JL Sr. Failure of arteriovenous fistula maturation: an unintended consequence of exceeding dialysis outcome quality Initiative guidelines for hemodialysis access. J Vasc Surg 2003;38:439 45; discussion 45 5. Konner K, Nonnast-Daniel B, Ritz E. The arteriovenous fistula. J Am Soc Nephrol 2003;14:1669 80 6. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M Jr, Miller A, Scher L, Trerotola S, Gregory RT, Rutherford RB, Kent KC. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002;35:60310 7. Wolford HY, Hsu J, Rhodes JM, Shortell CK, Davies MG, Bakhru A, Illig KA. Outcome after autogenous brachial-basilic upper arm transpositions in the post-National Kidney Foundation Dialysis Outcomes Quality Initiative era. J Vasc Surg 2005;42:951 6
979
8. Shemesh D, Zigelman C, Olsha O, Alberton J, Shapira J, Abramowitz H. Primary forearm arteriovenous fistula for hemodialysis accessan integrated approach to improve outcomes. Cardiovasc Surg 2003;11:35 41 9. Jindal K, Chan CT, Deziel C, Hirsch D, Soroka SD, Tonelli M, Culleton BF; Canadian Society of Nephrology Committee for Clinical Practice Guidelines. Hemodialysis clinical practice guidelines for the Canadian Society of Nephrology. J Am Soc Nephrol 2006;17:S127 10. Hemodialysis Adequacy 2006 Work Group. Clinical practice guidelines for hemodialysis adequacy, update 2006. Am J Kidney Dis 2006;48(suppl 1):S290 11. Rodriguez JA, Armadans L, Ferrer E, Olmos A, Codina S, Bartolome J, Borrellas J, Piera L. The function of permanent vascular access. Nephrol Dial Transplant 2000;15:402 8 12. Mouquet C, Bitker MO, Bailliart O, Rottembourg J, Clergue F, Montejo LS, Martineaud JP, Viars P. Anesthesia for creation of a forearm fistula in patients with endstage renal failure. Anesthesiology 1989;70:909 14 13. Won T, Jang JW, Lee S, Han JJ, Park YS, Ahn JH. Effects of intraoperative blood flow on the early patency of radiocephalic fistulas. Ann Vasc Surg 2000;14:468 72 14. Robbin ML, Chamberlain NE, Lockhart ME, Gallichio MH, Young CJ, Deierhoi MH, Allon M. Hemodialysis arteriovenous fistula maturity: US evaluation. Radiology 2002;225:59 64 15. Dixon BS. Why dont fistulas mature? Kidney Int 2006;70: 141322 16. Lin PH, Bush RL, Nguyen L, Guerrero MA, Chen C, Lumsden AB. Anastomotic strategies to improve hemodialysis access patencya review. Vasc Endovascular Surg 2005;39:135 42 17. Dammers R, Tordoir JH, Kooman JP, Welten RJ, Hameleers JM, Kitslaar PJ, Hoeks AP. The effect of flow changes on the arterial system proximal to an arteriovenous fistula for hemodialysis. Ultrasound Med Biol 2005;31:132733 18. Strichartz GR, Berde CB. Local anesthetics. In: Miller RD, Fleisher LA, eds. Millers anesthesia. 6th ed. Philadelphia, PA: Elsevier Churchill Livingstone, 2005:57399 19. Hingorani AP, Ascher E, Gupta P, Alam S, Marks N, Schutzer RW, Multyala M, Shiferson A, Yorkovich W, Jacob T, SallesCunha S. Regional anesthesia: preferred technique for venodilatation in the creation of upper extremity arteriovenous fistulae. Vascular 2006;14:23 6 20. Viscomi CM, Reese J, Rathmell JP. Medial and lateral antebrachial cutaneous nerve blocks: an easily learned regional anesthetic for forearm arteriovenous fistula surgery. Reg Anesth 1996;21:25 21. Rodriguez J, Quintela O, Lopez-Rivadulla M, Barcena M, Diz C, Alvarez J. High doses of mepivacaine for brachial plexus block in patients with end-stage chronic renal failure. A pilot study. Eur J Anaesth 2001;18:171 6 22. Eldredge SJ, Sperry RJ, Johnson JO. Regional anesthesia for arteriovenous fistula creation in the forearm: a new approach. Anesthesiology 1992;77:1230 1
23. Weissman C. The metabolic response to stress: an overview and update. Anesthesiology 1990;73:308 27 24. Seltzer JL. Is regional anesthesia preferable to general anesthesia for outpatient surgical procedures on an upper extremity? Mayo Clin Proc 1991;66:544 7 25. Alsalti RA, el-Dawlatly AA, al-Salman M, Jommaa S, Amro K, Dweiri MA, Jasser MT. Arteriovenous fistula in chronic renal failure patients: comparison between three different anesthetic techniques. Middle East J Anesthesiol 1999;15:30514 26. Solomonson MD, Johnson ME, Ilstrup D. Risk factors in patients having surgery to create an arteriovenous fistula. Anesth Analg 1994;79:694 700 27. Shemesh D, Olsha O, Orkin D, Raveh D, Goldin I, Reichenstein Y, Zigelman C. Sympathectomy-like effects of brachial plexus block in arteriovenous access surgery. Ultrasound Med Biol 2006;32:81722 28. Leonard IE, Chinappa V. Vascular access procedures for haemodialysispotential hazard of regional anaesthesia. Anaesthesia 2001;56:917 8 29. Wedel DJ, Horlocker TT Nerve blocks. In: Miller RD, Fleisher LA, eds. Millers anesthesia. 6th ed. Philadelphia, PA: Elsevier Churchill Livingstone, 2005:1686 95 30. McEllistrem RF, Schell J, OMalley K, OToole D, Cunningham AJ. Interscalene brachial plexus blockade with lidocaine in chronic renal failurea pharmacokinetic study. Can J Anaesth 1989;36:59 63 31. Laskowski IA, Muhs B, Rockman CR, Adelman MA, Ranson M, Cayne NS, Leivent JA, Maldonado TS. Regional nerve block allows for optimization of planning in the creation of arteriovenous access for hemodialysis by improving superficial venous dilatation. Ann Vasc Surg 2007;21:730 3 32. Yildirim V, Doganci S, Yanarates O, Saglam M, Kuralay E, Cosar A, Erdal Guzeldemi M. Does preemptive stellate ganglion blockage increase the patency of radiocephalic arteriovenous fistula? Scand Cardiovasc J 2006;40:380 4 33. Lok CE, Allon M, Moist L, Oliver MJ, Shah H, Zimmerman D. Risk equation determining unsuccessful cannulation events and failure to maturation in arteriovenous fistulas (REDUCE FTM I). J Am Soc Nephrol 2006;17:3204 12 34. Gibbons CP. Primary vascular access. Eur J Vasc Endovasc Surg 2006;31:5239 35. Ernandez T, Saudan P, Berney T, Merminod T, Bednarkiewicz M, Martin PY. Risk factors for early failure of native arteriovenous fistulas. Nephron Clin Pract 2005;101:c39 44 36. Feldman HI, Joffe M, Rosas SE, Burns JE, Knauss J, Brayman K. Predictors of successful arteriovenous fistula maturation. Am J Kidney Dis 2003;42:1000 12 37. Miller PE, Tolwani A, Luscy CP, Deierhoi MH, Bailey R, Redden DT, Allon M. Predictors of adequacy of arteriovenous fistulas in hemodialysis patients. Kidney Int 1999;56:275 80
980
Anesthesia for Vascular Access Surgery
An Anatomical Study of the Transversus Abdominis Plane Block: Location of the Lumbar Triangle of Petit and Adjacent Nerves
Zorica B. Jankovic, MD, PhD* Frances M. du Feu Patricia McConnell, BSc, PhD
BACKGROUND: The transversus abdominis plane (TAP) block is a new technique for providing analgesia to the anterior abdominal wall. Most previous studies have used the lumbar triangle of Petit as a landmark for the block. In this cadaveric study, we determined the exact position and size of the lumbar triangle of Petit and identified the nerves affected by the TAP block. METHODS: The position of the lumbar triangle of Petit was assessed unilaterally in 26 cadaveric specimens relative to reliably palpable surface landmarks. In addition, a series of dissections were performed to explore the course of the nerves blocked by the TAP. RESULTS: The mean distance from the midaxillary line along the iliac crest to the center of the base of the lumbar triangle of Petit at the level of the subcutaneous tissue and over the skin surface was 6.9 cm (range, 4.59.2 cm) and 9.3 cm (range, 4 15.1 cm), respectively. The center of the lumbar triangle of Petit was 1.4 cm above the iliac crest. The depth of the TAP at the lumbar triangle of Petit position was 0.5 4 cm and at the midaxillary line it was 0.52 cm. The average size of the lumbar triangle of Petit was 2.3 cm 3.3 cm 2.2 cm, with an average area of 3.63 1.93 cm2. The three cadaveric specimens we explored showed the nerves blocked by TAP passed lateral to the triangle. An incidental finding was that in 66% of specimens the lumbar triangle of Petit contained small branches of the subcostal artery. CONCLUSIONS: The lumbar triangles of Petit found in the specimens in this study were more posterior than the literature suggests. The position of the lumbar triangle of Petit varies largely and the size is relatively small. The relevant nerves to be blocked had not entered the TAP in the specimens in this study at the point of the lumbar triangle of Petit. At the midaxillary line, however, all the nerves were in the TAP.
(Anesth Analg 2009;109:9815)
he transversus abdominis plane (TAP) is an anatomical space between the internal oblique and the transversus abdominis muscle and spans the abdomen wherever these two muscles exist. The TAP block is a new, rapidly expanding regional anesthesia technique that involves a single large bolus injection of local anesthetic into this anatomical compartment to saturate somatic afferents before they leave the TAP to supply the anterior abdominal wall from T8 to L1 dermatomes.1 Initially, the TAP block was described as easy to perform and without major complications.1 However, with increasing use, different techniques
From the *Department of Anaesthesia, St Jamess University Hospital, Leeds, UK; and Institute of Membrane and Systems Biology, University of Leeds, Leeds, UK. Accepted for publication April 3, 2009. Address correspondence and reprint requests to Dr. Zorica B. Jankovic, Department of Anaesthesia, Lincoln Wing, St Jamess University Hospital, Beckett St., Leeds LS9 7TF, UK. Address e-mail to Zorica.Jankovic@leedsth.nhs.uk. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181ae0989
emerged, serious complications occurred, and high failure rates were reported.25 The lumbar triangle of Petit is an anatomical area that in theory provides a reliable reference point for inserting the needle into the TAP compartment.6 8 The triangle is formed posteriorly by the lateral border of the latissimus dorsi muscle and anteriorly by the posterior free border of the external oblique muscle, with the iliac crest as the base. The floor of the triangle from superficial to deep is formed by subcutaneous tissue, internal oblique muscle, and transversus abdominis muscle, respectively.9 However, the precise location of the lumbar triangle of Petit remains controversial. This study defined the exact position and size of the triangle relative to palpable surface landmarks. In addition, a series of dissections were explored to show the course of the nerves that would be affected by the TAP block.
METHODS
Twenty-six cadaveric specimens (14 women and 12 men), age 72102 years and mean height 161.8 9.9
981
Figure 1. Specimen 1: Identification of the lumbar triangle of Petit (relative to bony landmarks). (S superior; I inferior; L
lateral; M media). 1: midaxillary point at the iliac crest to the lateral base of the lumbar triangle of Petit; 2: base of the lumbar triangle of Petit; 3: lateral side of the lumbar triangle of Petit (the posterior attachment of the external oblique to the iliac crest); 4: medial side of the lumbar triangle of Petit (the anterior attachment of the latissimus dorsi muscle to the iliac crest); 5: medial base of the lumbar triangle of Petit; 6: anterior superior iliac spine to posterior superior iliac spine along the iliac crest.
cm, were dissected. The cadavers were preserved using 75 L of embalming fluid (consisting of 756 mL industrial methyl, 125 mL phenol, 40 mL formaldehyde, 22 mL glycerol, and 57 mL water per liter of fluid) introduced into the common carotid artery by gravity feed and were then refrigerated for at least 4 wk. All cadavers were in the prone position for the quantitative study and in the supine position for the qualitative study.
5. Medial base of the lumbar triangle of Petit to posterior superior iliac spine. 6. Anterior superior iliac spine to posterior superior iliac spine along the iliac crest. In 24 of the specimens, the superficial position of the triangle was also measured from the outside of the skin and subcutaneous tissue and from the midaxillary line to the middle of the lumbar triangle of Petit base along the iliac crest. We calculated the distance between the midaxillary line and the middle of the lumbar triangle of Petit base along the iliac crest by adding half the length of the lumbar triangle of Petit base to the distance between the midaxillary line at the level of the iliac crest and the lateral base of the lumbar triangle of Petit. We calculated the height of the lumbar triangle of Petit triangle corresponding to its base for all cadavers (Fig. 2).
Quantitative Study
To measure the precise position of the lumbar triangle of Petit, pins were inserted at the anterior superior iliac spine, the posterior superior iliac spine, the midaxillary point at the iliac crest, and at the angles of the lumbar triangle of Petit as shown in Figure 1. A flexible tape measure was used to measure the distances between the points to the nearest millimeter. The measurements taken from the lumbar triangle of Petit position are as follows. 1. Distance from the midaxillary point at iliac crest to the lateral base of the lumbar triangle of Petit. 2. Base of the lumbar triangle of Petit. 3. Lateral side of the lumbar triangle of Petit (the posterior attachment of external oblique muscle to the iliac crest). 4. Medial side of the lumbar triangle of Petit (the anterior attachment of latissimus dorsi to the iliac crest).
982
Anatomy of Lumbar Triangle of Petit
Qualitative Study
Specimens 1, 2, and 3, women aged 97, 92, and 83 years, respectively, were dissected to examine the position of the lumbar triangle of Petit in relation to the iliohypogastric and ilioinguinal nerves.
Statistical Tests
A paired t-test was used to determine whether there was any significant difference in the position of
Figure 2. Diagram of lumbar triangle of Petit depicting landmarks.
Table 1. Lumbar Triangle of Petit: Measurements in 26 Cadavers Standard Range Mean deviation (cm) (cm) (cm)
Midaxillary to lateral base 2.59.0 of lumbar triangle of Petit (n 26) Base of lumbar triangle of 1.24.5 Petit (n 26) Lateral side of lumbar 1.77.5 triangle of Petit (n 26) Medial side of lumbar 0.87.5 triangle of Petit (n 26) Medial base of lumbar 6.813.5 triangle of Petit to posterior superior iliac spine (n 26) Anterior superior iliac 20.030.0 spine to posterior superior iliac spine along iliac crest (n 26) Superficial position of lumbar triangle of Petit Midaxillary line to 4.015.1 middle of the triangle at the base of the triangle (n 24) 5.8 2.3 3.3 2.2 9.7 1.56 1.03 1.36 1.38 1.96
25.2
2.8
Figure 3. Position of the most anterior lumbar triangle of

Petit (Specimen 3) (2.5 cm posterior to the midaxillary line) shown in a lateral view marked on the surface of the skin. 9.3 2.46
the lumbar triangle of Petit between male and female cadavers.
RESULTS
Quantitative Results
There were no significant differences in any measured variables between male and female cadavers (Table 1). A wide variation in the distance from the midaxillary line at the iliac crest to the center of the base of the lumbar triangle of Petit was found both deep to the subcutaneous tissue (mean, 6.9 cm; range, 4.59.2 cm) and over the skin surface (mean, 9.3 cm; range, 4 15.1 cm).
The center of the lumbar triangle of Petit was a mean of 1.1 cm (range, 0.23.6 cm) above the iliac crest at the subcutaneous level and 1.4 cm (0.3 4.5 cm) at the skin surface level. The depth of the TAP at the lumbar triangle of Petit position was 0.5 4 cm and at the midaxillary line it was 0.52 cm. The surface position and dimensions of the most anterior lumbar triangle of Petit found in this study (Specimen 3) were drawn onto a live subject (Fig. 3). The iliac crest was drawn onto the surface, anterior superior iliac crest to posterior superior iliac crest, and the midaxillary point on the iliac crest was marked and used to show the lumbar triangle of Petit 4 cm posterior to the midaxillary line. The lumbar triangle of Petit measurements presented in Table 1 were used to produce a diagram of lumbar triangle of Petit shapes and sizes found in the cadavers (Fig. 4). The mean size of the lumbar triangle of Petit was 2.3 cm at the base, 3.3 cm at the lateral
983
side, and 2.2 cm at the medial side. The average area of the triangle was 3.63 1.93 cm2.
Qualitative Results
The iliohypogastric, subcostal, and intercostals nerves had a constant course in the TAP in relation to the midaxillary line in the three cadavers studied (Fig. 5). At this point, the nerves have not yet branched extensively. In Part A of Figure 6, the first lumbar nerve (in purple) branches into the iliohypogastric (blue) and ilioinguinal (pink) nerves, whereas in Part B of Figure 6, the first lumbar nerve is already divided as it exits the psoas major compartment. In all three dissections, the nerve branches do not enter the TAP until lateral to the lumbar triangle of Petit. An incidental finding was that in 16 of 24 specimens the lumbar triangle of Petit contained small blood vessel branches of the subcostal artery.
Figure 4. Comparison of sizes and shapes of the lumbar triangle of Petit and the distance of each posterior to the midaxillary line.
DISCUSSION
The TAP block technique was developed recently as a result of the clinical need for a simple and efficient analgesic technique for abdominal procedures, including inguinal hernia repair, hysterectomy, cesarean delivery, colectomy, and suprapubic prostatectomy.6,8 10 Lumbar triangle of Petit has been used as a landmark for TAP block in many studies because the TAP is directly accessible through the lumbar triangle of Petit. This study found the center of the lumbar triangle of Petit to be an average of 6.9 cm (4.59.2 cm) posterior to the midaxillary point at the iliac crest and 9.3 cm (4 15.1 cm) posterior to the midaxillary line when skin surface position is measured, which is considerably more posterior than is suggested by the literature.1,6 The center of the lumbar triangle of Petit is 1.4 cm above the iliac crest at skin level. Because of tissue shrinkage in the cadavers, this distance could be even longer in the patient population. The location of the lumbar triangle of Petit in patients has been identified differently in several
Figure 5. Lateral view of the nerves in the left transversus abdominis plane (TAP). The internal oblique muscle has been reflected to show the transversus abdominis muscle and nerves in the TAP. The intercostal nerves (T9, T10, and T11) are shown in yellow, the subcostal nerve (T12) is shown in green, and the iliohypogastric nerve is shown in blue.
Figure 6. (A). Position of the lumbar

triangle of Petit relative to the first lumbar nerve (purple), the iliohypogastric nerve (blue), and the ilioinguinal nerve (pink) as they enter the transversus abdominis plane (TAP) lateral to the lumbar triangle of Petit; the angles of the triangle are indicated by the blue pins. (S: superior; I: inferior; L: lateral; M: medial.) (B) Specimen 2 (supine). The first lumbar nerve has branched before exiting the psoas major compartment; the ilioinguinal and iliohypogastric nerves are shown entering the TAP at the lateral edge of the lumbar triangle of Petit.
984
Anatomy of Lumbar Triangle of Petit
clinical studies using the lumbar triangle of Petit as a landmark for TAP block. Rafi2 states that the point of needle insertion should be at the dip of the lateral border of the latissimus dorsi muscle along the iliac crest. Rafi found this position to be close to the midaxillary line at the level of the L3 4 intervertebral space. McDonnell et al.1 located the lumbar triangle of Petit by palpating the iliac crest from anterior to posterior until the latissimus dorsi muscle was felt and the lumbar triangle of Petit was taken to be just anterior. However, Petit lumbar hernias are described to be medial to a vertical line between the end of the 12th rib and the iliac crest.11 It is possible that the great variability in the distance from the midaxillary line to the center of the lumbar triangle of Petit found in this study (4 15.1 cm along iliac crest and 1.4 cm above iliac crest) can explain the difficulty in determining the lumbar triangle of Petit position. The depth of TAP at the position of the lumbar triangle of Petit was between 0.5 and 4.0 cm, dependent on adipose tissue. When performing the block, the variation in depth needed to reach the TAP should be considered. The iliohypogastric, subcostal, and intercostals nerves had a constant course in the TAP in relation to the midaxillary line in the three cadavers studied. At this point, the nerves have not yet branched extensively, as has been confirmed in a recent study by Rozen et al.12 The iliohypogastric nerves were not in the TAP at the lumbar triangle of Petit. However, lumbar triangle of Petit is an access point for injecting local anesthetic into the compartment and therefore the iliohypogastric nerve should be effectively blocked by TAP block. The optimal volume of local anesthetic cannot be inferred from this study. In more than half of the specimens, the lumbar triangle of Petit contained small vessels entering the TAP. The other blood vessels (the subcostal artery and the ascending branch of the deep circumflex iliac artery) in the area were found to be along the iliac crest in the TAP. As shown in Figure 4, the lumbar triangles of Petit in this study vary in angle, shape, and size among individuals, making identification of the lumbar triangle of Petit difficult. The average size of the lumbar triangle of Petit is relatively small, and the presence of adipose tissue makes lumbar triangle of Petit identification even more difficult. Thus, the lumbar triangle
of Petit is a misleading landmark for anterior abdominal wall anesthesia. These results presented here are from adult cadavers and the conclusions drawn from this study should not be applied to children. In conclusion, the lumbar triangle of Petit is more posterior than the literature suggests. The lumbar triangle of Petit varies greatly in its position and its size is relatively small; the presence of adipose tissue significantly changes the position. As a result, it is difficult to find the lumbar triangle of Petit solely on palpation. The posterior position of the lumbar triangle of Petit would make the TAP block less convenient to perform on supine patients. REFERENCES
1. McDonnell JG, ODonnell BD, Farrell T, Gough N, Tuite D, Power C, Laffey JG. Transversus abdominis plane block: a cadaveric and radiological evaluation. Reg Anesth Pain Med 2007;32:399 404 2. Rafi AN. Abdominal field block: a new approach via the lumbar triangle [Correspondence]. Anaesthesia 2001;56:1024 6 3. McDonnell JG, ODonnell BD, Curley G, Heffernan A, Power C, Laffey JG. The analgesic efficacy of TAP block after abdominal surgery: a prospective randomised controlled trial. Anesth Analg 2007;104:1937 4. Farooq M, Carey M. A case of liver trauma with a blunt regional anesthesia needle while performing transversus abdominis plane block. Reg Anesth Pain Med 2008;33:274 5 5. Jankovic Z, Ahmad N, Ravishankar N, Archer F. Transversus abdominis plane block: how safe is it? Anesth Analg 2008; 107:1758 9 6. Kuppuvelumani P, Jaradi H, Delilkan A. Abdominal nerve blockade for postoperative analgesia after caesarean section. Asia Oceania J Obstet Gynaecol 1993;19:1659 7. McDonnell JG, Curley G, Carney J, Benton A, Costello J, Maharaj CH, Laffey JG. The analgesic efficacy of transversus abdominis plane block after cesarean delivery: a randomised controlled trial. Anesth Analg 2008;106:186 91 8. Carney JJ, McDonnell JG, Bhinder R, Maharaj CH, Laffey JG. Efficacy of transversus abdominis plane block using ropivacaine in multimodal postoperative pain relief in total abdominal hysterectomy surgery. Reg Anesth Pain Med 2007;32:137 9. ODonnell BD: The transversus abdominis plane (TAP) block in open retropubic prostatectomy [Letter to the Editor]. Reg Anesth Pain Med 2006;31:91 10. Carney JJ, McDonnell JG, Bhinder R, Maharaj CH, Laffey JG. Ultrasound guided continuous transversus abdominis plane block for postoperative pain relief in abdominal surgery. Reg Anesth Pain Med 2007;32:24 11. Bhasin SK, Khan AB, Sharma S. Bilateral Petits triangle hernia. J Med Educ 2006;8:163 4 12. Rozen WM, Tran TMN, Ashton MW, Barrington MJ, Ivanusic JJ, Taylor GI. Refining the course of the thoracolumbar nerves: a new understanding of the innervation of the anterior abdominal wall. Clin Anat 2008;21:32533
985
Brief Report
Unilateral Anesthesia Does Not Affect the Incidence of Urinary Retention After Low-Dose Spinal Anesthesia for Knee Surgery
Wolfgang G. Voelckel, MD* Lukas Kirchmair, MD* Peter Rehder, MD Ivo Garoscio, MD Dietmar Krappinger, MD Thomas J. Luger, MD*
We evaluated whether unilateral low-dose spinal anesthesia may reduce the likelihood of postoperative urinary retention. Forty patients scheduled for knee arthroscopy randomly received bilateral (n 20) or unilateral (n 20) spinal anesthesia with 6-mg hyperbaric bupivacaine 0.5%. The incidence of urinary retention (500 mL) assessed with an ultrasound device (Bladderscan) and subsequent temporary catherization was 7/20 patients in the bilateral versus 6/20 in the unilateral group (not significant). We concluded that unilateral low-dose spinal anesthesia does not further decrease the likelihood of urinary retention. Our results demonstrate the value and necessity of monitoring bladder volume postoperatively.
(Anesth Analg 2009;109:986 7)
ostoperative urinary retention and bladder overdistention may cause permanent detrusor damage,1 thus leading to persistent bladder dysfunction. Recognized risk factors for postoperative bladder distension are age 60 yr, duration of surgery (120 min), and spinal anesthesia.2 Excessive perioperative IV fluids and adrenergic medication needed for hemodynamic stabilization may further increase the incidence and severity of urinary retention.3 Low dose plain or hyperbaric 0.5% bupivacaine has been associated with minimal hemodynamic changes and decreases the duration of spinal anesthesia.4 Theoretically, unilateral spinal blockade will only partially affect bladder innervation thus allowing spontaneous voiding. In previous studies, the likelihood of urinary retention ranged between 0% and 2% after low-dose unilateral spinal anesthesia but bladder volume was not assessed.4,5 This study compared the incidence of urinary retention (500 mL), monitored with a portable ultrasound device, in patients after a bilateral spinal anesthesia versus patients maintained in a lateral
From the *Department of Anesthesiology and Critical Care Medicine, Medical University, Innsbruck; Department of Anesthesiology and Critical Care Medicine, AUVA Trauma Center, Salzburg; Department of Neuro-Urology, Medical University, Innsbruck, Austria; Department of Anesthesiology and Critical Care Medicine, Klinikum Amberg, Germany; and Department for Trauma Surgery and Sports Medicine, Medical University, Innsbruck, Austria. Accepted for publication April 17, 2009. Reprints will not be available from the author. Address correspondence to Wolfgang Voelckel, MSc, DEAA, Department of Anesthesiology and Critical Care Medicine, AUVA Trauma Center Salzburg, Dr. Franz-Rehrl-Platz 5, 5020 Salzburg, Austria. Address e-mail to wolfgang.voelckel@auvi.at. Copyright 2009 International Anesthesia Research Society
DOI: 10.1213/ane.0b013e3181af406e
decubitus position for 20 min after injection of 6-mg hyperbaric 0.5% bupivacaine.
METHODS
With approval of the Ethics Committee, 40 ASA Class I patients scheduled for elective knee arthroscopy in spinal anesthesia were enrolled with the study. Informed consent was obtained and patients were monitored with electrocardiogram, noninvasive arterial blood pressure, and pulse oximetry. If baseline bladder volume assessed with an ultrasound scanner (Bladderscan, Diagnostic Ultrasound, Bladderscan BVI 3000, Verathon Inc., Bothell, WA) exceeded 100 mL, patients were prompted to void. IV infusion of 7 mL kg1 h1 Ringers lactate solution was started and maintained until the end of surgery, followed by an additional 4 mL kg1 h1 postoperatively. The patients were placed in lateral decubitus position with the operative limb to be operated in the dependent position. Dural puncture was performed at the L3 4 interspace with a 25-gauge pencil point needle (Portex, Germany). Immediately after successful puncture, patients were assigned according to computerized random list to the bilateral or unilateral group. For bilateral anesthesia, 6-mg hyperbaric 0.5% bupivacaine was injected with the needle bevel cranially directed, and the patients were immediately turned in the supine position. In the unilateral spinal anesthesia group, the needle orifice was directed toward the dependent side and 6-mg hyperbaric 0.5% bupivacaine was slowly injected within 30 s; lateral decubitus position was subsequently maintained for 20 min. The motor blockade was evaluated using a modified Bromage scale (0 no motor block; 3 unable to move limb). Urinary volume was assessed by bladder scan
986
Table 1. Demographic Data of Patients Receiving Unilateral Versus Bilateral Low-Dose Spinal Anesthesia Unilateral group N 20
Age (yr) Body mass index (kg/m2) Gender (M/F) Duration of surgery (min) Volume infused (mL) Until end of surgery Until 60 min postoperatively 54.5 11.5 27.2 4.8 11/9 27 9 622 160 950 280
Bilateral group N 20
56.5 16.5 29 4.2 13/7 29 8 655 165 980 300
There were no signicant differences between the groups.
at the end of surgery and every 60 min thereafter until the patient was able to void spontaneously or until the urinary bladder volume exceeded 500 mL. In the latter situation, the patient was prompted to void. If the patient failed to do so, urinary catherization was performed to empty the bladder. SPSS 15.0 (SPSS, Chicago, IL) was used for statistical analysis. For independent samples, a t-test or a nonparametric MannWhitney test was performed. The KolmogorovSmirnov test was used for determination of the distribution form. For the analysis of categorical data, a Fishers exact test was performed. The probability level was set at P 0.05.
RESULTS
No differences in the demographic variables, duration of surgery, and fluid volume infused were observed between the groups (Table 1). The maximum sensory level on the operated side was T9 in the bilateral group versus T10 in the unilateral group. In the unilateral group, four patients had a sensory level (impaired discrimination of temperature) of L1 and a Bromage score between 1 and 2 on the contralateral side. Incidence of urinary retention and subsequent temporary catheterization was 35% in the bilateral versus 30% in the unilateral group (P 0.05). A bladder volume 300 mL was assessed in 4/7 patients in the bilateral and 5/6 patients in the unilateral group immediately at the end of surgery. None of these patients was able to void spontaneously, and a temporary catheter had to be inserted when the urinary bladder volume exceeded 500 mL. In the unilateral group, occurrence of a partial motor block was not associated with urinary retention. The mean sd time (h:min) until patients were able to void spontaneously was not statistically different in the bilateral versus unilateral group (4:16 1:13 vs 3:36 1:07) as was the time of complete neurologic recovery defined as the time until ambulation in both groups (3:41 1:04 vs 3:21 0:52).
spinal anesthesia. The incidence of bladder overdistention was not decreased by maintaining the patients in a lateral decubitus position for 20 min after injection. In addition, the time until complete recovery from spinal anesthesia defined as unimpeded ambulation and mean time until spontaneous voiding did not differ significantly. However, one major finding of our study was the observation that a bladder volume 300 mL immediately after surgery was a strong indicator for postoperative urinary retention. Our results demonstrate that even when unilateral spinal anesthesia is achieved, bladder innervation will be compromised and consequently, strict monitoring of the bladder volume is important. A lower incidence of urinary retention in previous studies4,5 may be simply explained by the fact that an ultrasound was not used. Postoperative urinary retention is a significant problem,6 and repeated assessment of the bladder volume with an ultrasound device has been shown to be accurate and effective in detecting bladder overdistension.2,7 When a bladder scan is used, routine catheterization after low-dose spinal anesthesia for minor knee surgery is not warranted5,8 but close monitoring is mandatory in all patients.9 In conclusion, even unilateral low-dose spinal anesthesia does not prevent the risk of postoperative urinary retention. Our results suggest the standardized use of a portable ultrasound device to evaluate bladder volume in patients at risk of urinary retention. REFERENCES
1. Tammela T, Kontturi M, Lukkarinen O. Postoperative urinary retention. II. Micturition problems after the first catheterization. Scand J Urol Nephrol 1986;20:257 60 2. Lamonerie L, Marret E, Deleuze A, Lembert N, Dupont M, Bonnet F. Prevalence of post-operative bladder distension and urinary retention detected by ultrasound measurement. Br J Anaesth 2004;92:544 6 3. Rosseland LA, Stubhaug A, Breivik H. Detecting postoperative urinary retention with an ultrasound scanner. Acta Anaesthesiol Scand 2002;46:279 82 4. Fanelli G, Borghi B, Casati A, Bertini L, Montebugnoli M, Torri G. Unilateral bupivacaine spinal anesthesia for outpatient knee arthroscopy. Italian Study Group on Unilateral Spinal Anesthesia. Can J Anaesth 2000;47:746 51 5. Esmaoglu A, Karaoglu S, Mizrak A, Boyaci A. Bilateral vs. unilateral spinal anesthesia for outpatient knee arthroscopies. Knee Surg Sports Traumatol Arthrosc 2004;12:155 8 6. Pappalardo GB, Aranzulla F, Di Santo S, Barrera A, Beltrutti D. Complications of super-selective subarachnoid anesthesia (SSA) with hyperbaric bupivacaine: experiences with 355 patients in general and orthopaedic surgery. Panminerva Med 1997;39:415 7. Rosseland LA, Stubhaug A, Breivik H, Medby PC, Larsen HH. [Postoperative urinary retention]. Tidsskr Nor Laegeforen 2002;122:902 4 8. Ng KO, Tsou MY, Chao YH, Mui WC, Chow LH, Chan KH. Urinary catheterization may not be necessary in minor surgery under spinal anesthesia with long-acting local anesthetics. Acta Anaesthesiol Taiwan 2006;44:199 204 9. Luger TJ, Garoscio I, Rehder P, Oberladstatter J, Voelckel W. Management of temporary urinary retention after arthroscopic knee surgery in low-dose spinal anesthesia: development of a simple algorithm. Arch Orthop Trauma Surg 2008;128:60712
DISCUSSION
The observed incidence of postoperative urinary retention and subsequent catheterization was approximately 30% after both low-dose bilateral and unilateral
987
Section Editor: Lawrence Saidman
Letters to the Editor

Airway Topical Anesthesia
To the Editor: Although we appreciate the reference to our recent article describing topical anesthesia to the airway in morbidly obese patients using 2% vs 4% lidocaine administered via an atomizer,1 we believe that Xue et al.2 have misrepresented our findings. Contrary to their assertion, excellent intubating conditions were achieved in the majority of our cases. The gagging observed was mild in all but one instance (1/27). Although Xue et al. conclude that they achieved superior intubating conditions compared with those in our study, they nonetheless report a 61.5%73.1% incidence of grimacing and coughing during intubation. We, however, observed mild gagging during intubation in 42% and 18% of patients receiving 2% and 4% lidocaine, respectively. This being said, it is difficult to compare intubating conditions in the two studies because of differences in the degree of airway manipulation during application of the topical anesthetic and intubation, patient populations studied, and in the grading systems that were employed. Given these caveats, it is difficult to support their conclusion that the spray-asyou-go-technique offers superior intubating conditions compared with our technique using an atomizer. We agree that the lidocaine plasma concentrations are lower with the spray-as-you-go-technique, but this advantage is achieved at the expense of a considerably longer time required for topical airway anesthesia (approximately 23 min compared with 4.5 min). This may be unacceptable in certain critical situations. We have successfully used the atomizer technique to achieve topical airway anesthesia for awake fiberoptic intubation in hundreds of cases and have invariably found this to be rapid, effective, and safe. The technique described by
Xue et al.2 is another strategy to achieve a similar end, and, in fact, may be combined with the atomizer technique.
Paul M. Wieczorek, MDCM Steven B. Backman, MDCM, PhD
Department of Anesthesia McGill University Montreal, Quebec, Canada paul.wieczorek@gmail.com
REFERENCES
1. Wieczorek PM, Schricker T, Vinet B, Backman SB. Airway topicalisation in morbidly obese patients using atomized lidocaine: 2% compared with 4%. Anaesthesia 2007;62:984 8 2. Xue FS, Liu HP, He N, Xu YC, Yang QY, Liao X, Xu XZ, Guo XL, Zhang YM. Spray-as-you-go airway topical anesthesia in patients with a difficult airway: a randomized, double blind comparison of 2% and 4% lidocaine. Anesth Analg 2009; 108:536 43
DOI: 10.1213/ane.0b013e3181add3b0
In Response: We agree with Wieczorek et al.1 that the comparison of the intubating conditions between our and their studies is difficult because of differences in the study subjects, observed methods, and airway manipulation during application of the topical anesthetic and intubation.2,3 In our study, 61%73% of patients did display grimacing and coughing responses during awake fiberoptic orotracheal intubation, but most of these responses were slight and did not significantly impede fiberscopy or tracheal intubation. Our results also showed that the range of patient reaction and coughing scores were only 1.9 2.0 and 1.71.9, respectively. According to evaluation criteria of tracheal intubating conditions used in our study, excellent and acceptable intubating conditions were obtained in 27% and 73% of patients, respectively, in the 2% lidocaine group and 35% and 65% of patients, respectively, in the 4% lidocaine group. Therefore, we concluded that under adequate sedation with fentanyl and midazolam, both 2% and 4% lidocaine administered to the
airway by a spray-as-you-go technique can provide clinically acceptable (not excellent) intubating conditions for awake fiberoptic orotracheal intubation.2 One main disadvantage of this technique is that it is time-consuming because it requires repeated lidocaine spraying of the different targeted areas. Because the primary aims of our study were to compare safety and efficacy of 2% and 4% lidocaine during airway topical anesthesia with a spray-as-you-go technique, we allowed a 35-min waiting period after each lidocaine spray to provide adequate penetration of local anesthetic into the airway mucosa for maximal effect.4 Therefore, our protocol required 2124 min from the first application of lidocaine to intubation of the trachea. One major advantage of the technique of Wieczorek et al. using atomized lidocaine is the rapidity with which the airway can be anesthetized.3 However, although they have successfully used this technique to achieve airway anesthesia for awake fiberoptic intubation in hundreds of cases, we remain concerned that the relatively large doses of lidocaine (800 1600 mg) administered over a short time (5 min) may increase the risk of local anesthetic toxicity. Practically speaking, to improve effectiveness of topical airway anesthesia, a combination of various methods (e.g., applying a gel or ointment, gargling a viscous solution, depositing local anesthetic droplets via anatomizer or nebulizer, utilizing a transtracheal injection, or direct deposit of a local anesthetic to the airway with the spray-as-you-go technique through a fiberoptic bronchoscope) is often required.58 Overall, providing topical anesthesia to the nasal and/or oral mucosa in combination with a method to anesthetize the laryngeal/ tracheal structures is the most effective and the most commonly chosen plan. Because patient safety is always of paramount importance
991
in management of difficult airway and each case is different, the anesthesiologist must weigh the risks and benefits of all the methodologies to arrive at the plan that is optimal for that particular patient.9
Fu S. Xue, MD
Department of Anesthesiology, Plastic Surgery Hospital Chinese Academy of Medical Sciences and Peking Union Medical College Beijing, Peoples Republic of China fruitxue@yahoo.com.cn
Levosimendan for Calcium Channel Blocker Poisoning in Humans

To the Editor: For several reasons, we question the appropriateness of using levosimendan for treatment of calcium channel blocker (CCB) toxicity as recently reported by Varpula et al.1 First, despite the intensive conventional therapy reportedly utilized in their cases, subtherapeutic insulin dosing was used. In a previous report of life-threatening CCB poisoning, insulin therapy was effective at 0.51.0 U kg1 h12 in contrast to the 20 and 10 U/h in the two patients reported by Varpula et al. Second, hyperinsulinemia-euglycemia therapy is a safe and effective treatment for CCB overdose3,4 and is suggested as an adjunct therapy in current resuscitation guidelines.5 Third, the authors refer to a study by Graudins et al.6 and correctly stated that the levosimendan group had improved hemodynamics over placebo but omitted they did worse than those receiving calcium alone. Finally, a recent study investigating the effect of levosidmendan in verapamil toxicity in dogs demonstrated trends toward more pronounced bradycardia and earlier death.7 Levosimendan has not been demonstrated to be an effective therapy for CCB poisoning in animals and we question its use in human cases of CCB poisoning.
Tamara R. Espinoza, MD
Department of Emergency Medicine Cook County-Stroger Hospital Chicago, Illinois tespinoza@ccbh.org
REFERENCES
1. Varpula T, Rapola J, Sallisalmi M, Kurola J. Treatment of serious calcium channel blocker overdose with levosimendan, a calcium sensitizer. Anesth Analg 2009;108: 790 2 2. Boyer EW, Shannon M. Treatment of calcium-channel-blocker intoxication with insulin infusion. N Engl J Med 2001;344: 17212 3. Yuan TH, Kerns WP II, Tomaszewski CA, Ford MD, Kline JA. Insulin-glucose as adjunctive therapy for severe calcium channel antagonist poisoning. J Toxicol Clin Toxicol 1999;37:46374 4. Kline JA, Tomaszewski CA, Schroeder JD, Raymond RM. Insulin is a superior antidote for cardiovascular toxicity induced by verapamil in the anesthetized canine. J Pharmacol Exp Ther 1993;267:744 50 5. Mottram AR, Erickson TE. Toxicology in emergency cardiovascular care. In Field J, ed. The textbook of emergency cardiovascular care and CPR. Philadelphia: Lippincott Williams & Wilkins, 2009:443 67 6. Graudins A, Najafi J, Rur-SC MP. Treatment of experimental verapamil poisoning with levosimendan utilizing a rodent model of drug toxicity. Clin Toxicol (Phila) 2008;46:50 6 7. Abraham MK, Scott SB, Meltzer A, Barrueto F Jr. Levosimendan does not improve survival time in a rat model of verapamil toxicity. J Med Toxicol 2009; 5:37
DOI: 10.1213/ane.0b013e3181add57d
Nong He, MD
Department of Anesthesiology Shougang Hospital, Peking University Beijing, Peoples Republic of China
He P. Liu, MD
Department of Anesthesiology Third Afliated Hospital, Xinxiang Medical College Xinxiang, Henan, Peoples Republic of China
REFERENCES
1. Wieczorek PM, Backman SB. Airway tropical anesthesia. Anesth Analg 2009;109:991 2. Xue FS, Liu HP, He N, Xu YC, Yang QY, Liao X, Xu XZ, Guo XL, Zhang YM. Spray-as-you-go airway topical anesthesia in patients with a difficult airway: a randomized, double blind comparison of 2% and 4% lidocaine. Anesth Analg 2009;108:536 43 3. Wieczorek PM, Schricker T, Vinet B, Backman SB. Airway topicalisation in morbidly obese patients using atomised lidocaine: 2% compared with 4%. Anaesthesia 2007;62:984 8 4. Morris IR. Pharmacologic aids to intubation and the rapid sequence induction. Emerg Med Clin North Am 1988;6:753 68 5. Sanchez A, Iyer RR, Morrison DE. Preparation of the patient for awake intubation. In: Hagberg CA, ed. Benumofs airway management. Principles and practice. 2nd ed. St. Louis: Mosby-Year Book Inc, 2007: 26377 6. Williams KA, Barker GL, Harwood RJ, Woodall NM. Combined nebulization and spray-as-you-go topical local anaesthesia of the airway. Br J Anaesth 2005; 95:549 53 7. Kundra P, Kutralam S, Ravishankar M. Local anaesthesia for awake fibreoptic nasotracheal intubation. Acta Anaesthesiol Scand 2000;44:511 6 8. Sutherland AD, Williams RT. Cardiovascular responses and lidocaine absorption in fiberoptic-assisted awake intubation. Anesth Analg 1986;65:389 91 9. Simmons ST, Schleich AR. Airway regional anesthesia for awake fiberoptic intubation. Reg Anesth Pain Med 2002; 27:180 92
DOI: 10.1213/ane.0b013e3181add3cb
Allan R. Mottram, MD
Department of Emergency Medicine Cook County-Stroger Hospital Toxikon Consortium Chicago, Illinois
Sean M. Bryant, MD
Department of Emergency Medicine Cook County-Stroger Hospital Toxikon Consortium Illinois Poison Center Chicago, Illinois
In Response: Espinoza et al.1 question the rationale of using levosimendan in severe calcium channel poisoning as described in our case report. Previously published experimental models of calcium channel blocker poisoning have investigated levosimendan as a solitary treatment2 or together with calcium substitution.3 In practice, these severely ill patients receive many treatments that have an effect on overall hemodynamics. Besides levosimendan, our patients were treated with large doses of catecholamines and vasopressin to maintain blood pressure. Because vasodilatation was counteracted with vasoconstrictors, improved cardiac function after levosimendan was seen, which has also been demonstrated in an experimental setting.2 Convincing evidence of the effect of interventions is difficult to achieve in such life-threatening clinical scenarios. This is also the case with hyperinsulinaemia/euglycaemia therapy in calcium channel blocker poisoning. The exact mechanism is
992
poorly defined and current evidence consists of few isolated cases or small series.4,5 Hyperinsulinaemia/euglycaemia therapy is recommended by some guidelines, whereas others, including our national guidelines, do not recognize it. Our first patients did receive insulin at a dose of approximately 0.35 IU kg1 h1, which is 70% of the dose recommended.5
Tero Varpula, MD, PhD
Department of Anesthesiology and Intensive Care Medicine Jorvi Hospital, Helsinki University Hospital Espoo, Finland tero.varpula@hus.
Janne Rapola, MD, PhD

Department of Cardiology Helsinki University Hospital Helsinki, Finland
Marko Sallisalmi, MD
Department of Anesthesiology and Intensive Care Medicine Helsinki University Hospital Helsinki, Finland
Jouni Kurola, MD, PhD

Intensive Care Unit Kuopio University Hospital Kuopio, Finland
REFERENCES
1. Espinoza TR, Mottram AR, Bryant SM. Levosimendan for calcium channel blocker poisoning in humans. Anesth Analg 2009;109:992 2. Abraham MK, Scott SB, Meltzer A, Barrueto F Jr. Levosimendan does not improve survival time in a rat model of verapamil toxicity. J Med Toxicol 2009;5:37 3. Graudins A, Najafi J, Rur-SC MP. Treatment of experimental verapamil poisoning with levosimendan utilizing a rodent model of drug toxicity. Clin Toxicol (Phila) 2007;46:50 6 4. Boyer EW, Shannon M. Treatment of calcium-channel-blocker intoxication with insulin infusion. N Engl J Med 2001;344: 17212 5. Yuan TH, Kerns WP II, Tomaszewski CA, Ford MD, Kline JA. Insulin-glucose as adjunctive therapy for severe calcium channel antagonist poisoning. J Toxicol Clin Toxicol 1999;37:46374
DOI: 10.1213/ane.0b013e3181add5eb
Limitations in Ultrasound Imaging Techniques in Anesthesia: Obesity and Muscle Atrophy?

To the Editor: Ota et al.1 concludes that, when using ultrasound guidance (US), the anterior approach to sciatic nerve
block is performed as easily and successfully as the posterior approach. The authors stated that, in elderly individuals, sciatic nerve identification is less successful because of muscle atrophy in which fascia may not be distinguishable with US imaging. Additionally, the authors implied that in obese patients, the sciatic nerve is not clearly visualized because of its deep anatomic location. On the basis of above considerations, I would like to comment on a few issues. Ultrasound imaged muscle bundles are seen as hypoechoic zones, whereas the perimysium and aponeurosis are seen as hyperechoic structures. In the case of muscle atrophy, hypoechoic muscle bundles degenerate, whereas perimisium and aponeurosis remain intact. The atrophic muscles, depicted as hyperechoic structures, reflect US energy, thus decreasing the ability of the US beam to penetrate in deeper tissues.2,3 In obese patients, because of deep anatomic location of nerves, the US beam travels a greater distance, resulting in beam attenuation. In addition, other factors may affect imaging quality through fat, which are as follows: 1) Exaggerated attenuation, i.e., the adipose tissue, has a nonlinear relationship to frequency as opposed to the usually assumed linear relationship in most biological tissues; 2) Phase aberration of the sound field because of uneven speed of sound in the irregularly-shaped adipose layers. This is due to differing speeds of sound in the overlying, nonhomogeneous tissue above the focus of the transducer; and 3) Reflection because of mismatch of acoustic impedance at the fat/muscle interfaces. When the US beam crosses a boundary between muscle layer and fat, a portion of energy is reflected back to the transducer because of different acoustic velocity between the two tissues (pure fat 1450 m/s and muscle 1580 m/s).4 6 In these cases, image quality may be improved by using different technical approaches which reduce
speckling, clutter, or other acoustic artifacts.7 Advanced US imaging techniques, such as compound and harmonic imaging, improve the image because of a reduction of these artifacts. For example, harmonic imaging reduces phase aberration artifacts from overlying tissue and compound imaging reduces similar artifacts by averaging multiple scan lines from different directions.7,8 Additionally, compression of fat, location of the fat in the focus of the transducer, and large beam width of the US signal may improve imaging quality through fat.4,9 Among the major US innovations of recent years, 3D US is the ideal tool to avoid the limitations affecting the diffusion and reliability associated with traditional US.10 However, many studies are required to ascertain its utility in the imaging of nerve structures. In conclusion, obesity and muscle atrophy mainly increase the number of reflective interfaces not only leading to more echoes but also decreasing incident sound available to penetrate deeper tissues, such as nerves, vessels, or other targeted structures.
Theodosios Saranteas, PhD
2nd Department of Anesthesiology, School of Medicine University of Athens, Attikon Hospital Athens, Greece, EU saranteas@ath.forthnet.gr
REFERENCES
1. Ota J, Sakura S, Hara K, Saito Y. Ultrasound-guided anterior approach to sciatic nerve block: a comparison with the posterior approach. Anesth Analg 2009;108:660 65 2. Chhem R, Kaplan P, Dussault R. Ultrasonography of the musculoskeletal system. Radiol Clin North Am 1994;32: 275 89 3. Saranteas T, Chantzi C, Iatrou C, Kostopanagiotou G, Dimitriou V. Ultrasound and regional anaesthesia techniquesis there any limitation? Reg Anesth Pain Med 2007;32:546 7 4. Fiegler W, Felix R, Langer M, Schultz E. Fat as a factor affecting resolution in diagnostic ultrasound: possibilities for improving picture quality. Eur J Radiol 1985;5:304 9 5. Feigenbaum H. Physics and instrumentation. In Feingenbaum H, Armstrong WF, Rayan T, eds. Feigenbaums echocardiography. Philadelphia: Lippincott William and Wilkins, 2005:125
993
6. Shmulewitz A, Teefey SA, Robinson BS. Factors affecting image quality and diagnostic efficacy in abdominal sonography: a prospective study of 140 patients. J Clin Ultrasound 1993;21:62330 7. Entrekin RR, Porter BA, Sillesen HH, Wong AD, Cooperberg PL, Fix CH. Real time spatial compound imaging: applications to breast, vascular and musculoskeletal ultrasound. Semin Ultrasound CT MR 2001;22:6577 8. Shapiro RS, Wagreich J, Parsons RB, Stancato-Pasik A, Yeh HC, Lao R. Tissue harmonic imaging sonography: evaluation of image quality compared with conventional sonography. AJR Am J Roentgenol 1998;171:1203 6 9. Browne JE, Watson AJ, Hoskins PR, Elliott AT. Investigation of the effect of subcutaneous fat on image quality performance of 2D conventional imaging and tissue harmonic. Imaging Ultrasound Med Biol 2005;31:957 64 10. Cimmino M, Grassi W. What is new in ultrasound and magnetic resonance imaging for musculoskeletal disorders? Best Pract Res Clin Rheumatol 2008;22:1141 8
DOI: 10.1213/ane.0b013e3181ae09a4
Surrogate Outcomes: They Dont Get It

To the Editor: The recent Letter to the Editor by Kranke et al.1 regarding postoperative nausea and vomiting (PONV) dispute[s] the notion that PONV should be described as a surrogate outcome, as suggested by frequently cited editorials, only because quality of life measures and scores for patient satisfaction are not consistently affected by preventive measures in an unselected patient population. As the author of both frequently cited editorials,2,3 I counter Kranke et al. with an example of the ultimate antiemetic, pancuronium plus midazolam. Paralysis assures that vomiting is absent, whereas sedation prevents assessment of nausea. The outcome measure accepted by Kranke and co-workers4 and common to PONV studies, counting the number of emetic episodes, would yield remarkable success for such a regimen. Kranke et al. rightly criticize my proposal as absurd and irrelevant: antiemesis would have been accomplished at enormous cost and risk and without benefiting the patient. Many antiemetic/PONV articles cite a study by Gold et al.5 showing that PONV increased recovery
994
room stay and led to unplanned hospital admission. Unfortunately, these studies either did not measure recovery room stay and the incidence of unplanned hospital admission or failed to demonstrate that antiemetics influenced these outcomes. Such studies remain to be done. In the present cost-containment environment, therapeutic decisions should be evidence based. One obvious true outcome is patient satisfaction. Scuderi et al.6 evaluated the effect of antiemetics on patient satisfaction; prophylactic treatment improved satisfaction only 4%. Based on that value, 25 patients must be treated to create one additional happy customer; in turn, when Kranke et al. administer antiemetics prophylactically, they should consider this calculation in assessing cost-benefit ratio. Krankes claim that quality of life measures and scores for patient satisfaction are not consistently affected by preventive measures in an unselected patient population is anecdotal. If Kranke et al. are convinced that failure to demonstrate improved patient satisfaction results from the patient populations being unselected, I encourage them (or others) to perform a prospective randomized clinical trial using a selected patient population. Perhaps this might put the issue to rest. Meanwhile, I prefer that clinicians make decisions based on meaningful outcomes.
Dennis Fisher, MD
P Less Than San Francisco, California sher@plessthan.com
similar clinical efficacy against both nausea and vomiting. Anaesthesia 2009;64: 14751 5. Gold BS, Kitz DS, Lecky JH, Neuhaus JM. Unanticipated admission to the hospital following ambulatory surgery. JAMA 1989;262:3008 10 6. Scuderi PE, James RL, Harris L, Mims GR. Antiemetic prophylaxis does not improve outcomes after outpatient surgery when compared to symptomatic treatment. Anesthesiology 1999;90:360 71
DOI: 10.1213/ane.0b013e3181b08193
In Response: We fully support the notion espoused by Fisher1 that clinicians (should) make decisions based on meaningful outcomes. Dr. Fisher suggests that nausea and vomiting are not meaningful outcomes because pancuronium and midazolam reduce nausea and vomiting to zero. True. However, his reductio ad absurdum argument could be made about many meaningful outcomes. Is pain the fifth vital sign or merely a surrogate? Like nausea and vomiting, pain scores could be reduced to zero from a combination of midazolam and pancuronium. Does that make pain not a meaningful outcome? We contend that nausea and vomiting are meaningful outcomes. Throwing up is universally considered unpleasant. We will not offer a reference, but invite anyone who enjoys vomiting to disprove us by counterexample. Additionally, 1. Anesthesiologists consider postoperative nausea and vomiting (PONV) a relevant outcome2 and have dubbed it our big little problem.3 2. PONV matters to our patients.4,5 Patients are willing to pay US$56 US$100 out of their own pocket for a totally effective antiemetic.6 3. Similar to the occurrence of severe postoperative pain, inadequate management of PONV may imply medico-legal consequences because PONV is considered at least as troublesome as postoperative pain.4,7 4. The forceful expulsion of gastric contents in the course of severe
REFERENCES
1. Kranke P, Roewer N, Smith AF, Piper SN, Wallenborn J, Eberhart LHJ. Postoperative nausea and vomiting: what are we waiting for? Anesth Analg 2009;108: 1049 50 2. Fisher DM. Surrogate end points, are they meaningful? Anesthesiology 1994;81:795 6 3. Fisher DM. Surrogate outcomes: meaningful not! Anestheiolosgy 1999;90: 355 6 4. Jokela RM, Cakmakkaya OS, Danzeisen O, Korttila KT, Kranke P, Malhotra A, Paura A, Radke OC, Sessler DI, Soikkeli A, Roewer N, Apfel CC. Ondansetron has
vomiting or PONV may lead to life-threatening complications.8 5. There may be economic consequences of PONV, especially in an outpatient setting, because the occurrence of PONV is a leading factor for unplanned hospital admissions.9,10 Eliminating the big little problem3 may be our big little chance.11
Peter Kranke
Department of Anaesthesia and Critical Care University Hospitals of Wu rzburg Wu rzburg, Germany kranke_p@klinik.uni-wuerzburg.de
8. Schumann R, Polaner DM. Massive subcutaneous emphysema and sudden airway compromise after postoperative vomiting. Anesth Analg 1999;89:796 9. Kokinsky E, Thornberg E, Ostlund AL, Larsson LE. Postoperative comfort in paediatric outpatient surgery. Paediatr Anaesth 1999;9:24351 10. Blacoe DA, Cunning E, Bell G. Paediatric day-case surgery: an audit of unplanned hospital admission Royal Hospital for Sick Children, Glasgow. Anaesthesia 2008; 63:610 5 11. Kranke P, Roewer N, Smith AF, Piper SN, Wallenborn J, Eberhart LH. Postoperative nausea and vomiting: what are we waiting for? Anesth Analg 2009;108:1049 50
DOI: 10.1213/ane.0b013e3181b081aa
Andrew F. Smith
Department of Anaesthesia and Lancaster Patient Safety Research Unit Lancaster, UK
Swen N. Piper
Department of Anesthesiology and Intensive Care Hospital of Ludwigshafen/Rhine Germany
Decrease in Bispectral Index While Correcting Hyperglycemia and an Increase in Bispectral Index with Correction of Hypoglycemia
To the Editor: The Bispectral Index (BIS), a parameter derived from the electroencephalogram, is currently used to assess depth of anesthesia. Severe hypoglycemia may induce a decrease in electroencephalogram in either diabetic or nondiabetic patients.1,2 This report describes a sudden decrease in BIS with hypoglycemia and an increase in BIS with the correction of the hypoglycemia. A 74-yr-old 74-kg woman was admitted on the day of surgery for repair of an internal iliac aneurysm. She had Type 2 diabetes mellitus, hypertension, and stable angina, which was treated medically. Medications included metformin (500 mg bid), glibenclamide (2.5 mg/d), metoprolol (25 mg bid), aspirin (100 mg/d), and atorvastatin (20 mg/d). Her blood sugar level was 14.7 mmol/L before induction. She had omitted her oral hypoglycemics on the day of surgery. Anesthesia was induced with 2-mg midazolam, 120-mg propofol, and 100-g fentanyl, and muscle relaxation was accomplished with 40-mg rocuronium. The patients trachea was intubated with a Size 7 endotracheal tube. The BIS was continuously monitored using an A-2000 XP monitor (Aspect MS, Natick, MA). Anesthesia was maintained
Jan Wallenborn
Department of Anesthesiology and Intensive Care University Hospitals of Leipzig Leipzig, Germany
Norbert Roewer
Department of Anaesthesia and Critical Care University Hospitals of Wu rzburg Wu rzburg, Germany
Leopold H. J. Eberhart
Department of Anesthesiology and Intensive Care University Hospitals of Marburg and Gieen Marburg, Germany
REFERENCES
1. Fisher DM. Surrogate outcomes: they dont get it. Anesth Analg 2009;109:994 2. Macario A, Weinger M, Truong P, Lee M. Which clinical anesthesia outcomes are both common and important to avoid? The perspective of a panel of expert anesthesiologists. Anesth Analg 1999;88:108591 3. Kapur PA. The big little problem. Anesth Analg 1991;73:2435 4. Macario A, Weinger M, Carney S, Kim A. Which clinical anesthesia outcomes are important to avoid? The perspective of patients. Anesth Analg 1999;89:652 8 5. Eberhart LH, Mauch M, Morin AM, Wulf H, Geldner G. Impact of a multimodal anti-emetic prophylaxis on patient satisfaction in high-risk patients for postoperative nausea and vomiting. Anaesthesia 2002;57:10227 6. Gan T, Sloan F, Dear Gde L, El-Moalem HE, Lubarsky DA. How much are patients willing to pay to avoid postoperative nausea and vomiting? Anesth Analg 2001;92:393 400 7. van Wijk MG, Smalhout B. A postoperative analysis of the patients view of anaesthesia in a Netherlands teaching hospital. Anaesthesia 1990;45:679 82 Vol. 109, No. 3, September 2009
with O2/air and end-tidal sevoflurane 1.1%. The BIS was in the range of 38 46. Actrapid 6 iu (purified human neutral insulin) was given because the blood sugar was 14.7 mmol/L and over the next 8 12 min a rapid decrease in BIS to 1216 was observed. During this time, there had been no change in ventilatory or hemodynamic parameters or sevofurane concentration. A finger stick blood glucose measurement showed severe hypoglycemia (2.3 mmol/L). Glucose 35 mL 50% over 1 min was administered. A rapid increase in BIS to 3539 over 4 8 min was noted. As before, there were no changes in ventilation, hemodynamic parameters, or anesthetic depth. Glucose measurement revealed a blood sugar of 7.2 mmol/L. A rapid change in BIS with correction of hypoglycemia has been reported in a patient in an intensive care unit setting wherein the patient was sedated with midazolam and sufentanil.3 However, this appears to be the first report of a decrease in BIS related to correction of hyperglycemia followed by a rapid increase in the BIS with the correction of the hypoglycemia. This observation suggests that a sudden decrease in BIS could be an indication of severe hypoglycemia and should prompt a measurement of blood sugar concentration.
Manu Narayanaswamy, MBBS, FANZCA
Department of Anesthesia Gosford Hospital, Gosford New South Wales, Australia manu.swamy@gmail.com
REFERENCES
1. Gilbert TT, Wagner MR, Halukurike V, Paz HL, Garland A. Use of bispectral electroencephalogram monitoring to assess neurologic status in unsedated, critically ill patients. Crit Care Med 2001; 29:19962000 2. Pramming S, Thorsteinsson B, Stigsby B, Binder C. Glycaemic threshold for changes in electroencephalograms during hypoglycemia in patients with insulin dependent diabetes. BMJ 1988;296:6657 3. Vivien B, Olivier Langeron O, Bruno Riou B. Increase in bispectral index (BIS) while correcting a severe hypoglycemia. Anesth Analg 2002;95:1824 5
DOI: 10.1213/ane.0b013e3181adf919
995
Stroke Volume Calculation by Esophageal Doppler Integrates Velocity Over Time and Multiplies This Area Under The Curve by the Cross Sectional Area of the Aorta
To the Editor: In our recent review1 we wish to acknowledge an error that was astutely observed by Dr. Archer. In describing the calculation of stroke volume by means of esophageal Doppler, we state that the area under the curve of the velocity-time graph is computed mathematically as the integral of the derivative of velocity over time (dV/dt) from T0 to T1 (where T0 is the start of aortic blood flow and T1 is the end of flow). This is not accurate. The area under the curve for the velocity-time graph should be described as the integral of the velocity curve over time, not the integral of its derivative. The area under this curve is the distance traveled by blood during systole, also called the stroke distance, measured in cm. Stroke volume is then obtained by multiplying stroke distance by the cross sectional area of aorta (cm2) to obtain stroke volume (cm3). We thank Dr. Archer for pointing out this error and apologize for any confusion this may have caused.
Thomas L. Archer, MD, MBA
University of California San Diego, California tarcher@ucsd.edu
Ultrasound-Guided Intercostal Approach to Thoracic Paravertebral Block

To the Editor: In the first report of ultrasoundguided thoracic paravertebral block (USG-TPVB),1 the authors imaged the transverse process and the thoracic paravertebral space (TPVS) in a longitudinal parasagittal plane, the needle was inserted out-of-plane to the probe using a conventional technique, and loss-of-resistance to saline was used as the end point
for needle placement rather than ultrasonographic visualization of needletip position. Our USG-TPVB technique is typically performed using in-plane technique that utilizes direct visualization of needle-tip position and local anesthetic spread as the end point. USG-TPVB is performed using a high-frequency linear array probe. A convex probe is also useful in obese patients. The patient is placed in a lateral decubitus position with the side to be blocked uppermost or in a prone position. After aseptic preparation of the skin and the probe, the probe is placed on the rib
Figure 1. Ultrasound image of the thoracic paravertebral space at the level of T3. TP transverse process; TPVS apex of thoracic paravertebral space; IICM internal intercostal membrane; EICM external intercostal muscle; PL pleura.
Duane J. Funk, MD, FRCP(C)

University of Manitoba Winnipeg, Manitoba, Canada
Eugene Moretti, MD Tong J. Gan, MD, FRCA

Department of Anesthesia Duke University Medical Center Durham, North Carolina
REFERENCE
1. Funk DJ, Moretti EW, Gan TJ. Minimally invasive cardiac output monitoring in the perioperative setting. Anesth Analg 2009; 108:88797
DOI: 10.1213/ane.0b013e3181ae901c
Figure 2. Ultrasound image of the needle-tip placement into the thoracic paravertebral space. TP transverse process; EICM external intercostal muscle; N Tuohy needle; PL pleura.
996
Figure 3. Ultrasound image of local anesthetic spread. N Tuohy needle; EICM external intercostal muscle; LA local anesthetic; PL pleura. at the selected level with the medial edge of probe in contact with the spinous process, so that the horizontal view of the rib is visualized as a hyperechoic line with posterior acoustic shadowing. The probe is then moved caudally into the intercostal space between adjacent ribs. The inferior part of transverse process is visualized as a hyperechoic convex line with posterior acoustic shadowing. The apex of TPVS is visualized as a wedge-shaped hypoechoic space surrounded by the hyperechoic line of the pleura below and the internal intercostal membrane above (Fig. 1). The apex of TPVS is laterally continuous with the intercostal space2 (Fig. 2). The internal intercostal ligament is medially continuous with the superior costotransverse membrane; these two membranes cannot be distinguished by ultrasonography. A 20gauge Tuohy needle is inserted in a lateral-to-medial direction from the outer edge of probe with the bevel facing the probe using an in-plane approach and advanced until the needle tip penetrates through the internal intercostal membrane. After a negative aspiration test for blood, 1520 mL of local anesthetic is injected into the TPVS slowly. The pleura is seen being pressed ventrally during local anesthetic injection (Fig. 3). The technique we describe has been modified from that reported by Kappis3 in 1912. A 10-cm needle was introduced three finger breadths from the midline at an angle of 45 to the skin and advanced into the TPVS until the needle tip was in
contact with the posterolateral aspect of the vertebral body. Kappiss technique was eventually abandoned because of the risk of needle penetration through the intervertebral foramen resulting in intrathecal injection or spinal cord injury. With the ultrasound-guided technique, however, needle contact with the vertebral body is not necessary, and the complications associated with Kappiss technique can be avoided. Moreover, Tuohy needle insertion and advancement tangential to the pleura can lessen the risk of pleural and intercostal vascular puncture. The safety and reliability of USG-TPVB await future confirmation.
Yasuyuki Shibata, MD Kimitoshi Nishiwaki, MD, PhD
Department of Anesthesiology Nagoya University, School of Medicine Nagoya, Japan yshiba@tuba.ocn.ne.jp
REFERENCES
1. Hara K, Sakura S, Nomura T, Saito Y. Ultrasound guided thoracic paravertebral block in breast surgery. Anaesthesia 2009;64:2235 2. Burns DA, Ben-David B, Chelly JE, Greensmith JE. Intercostally placed paravertebral catheterization: an alternative approach to continuous paravertebral blockade. Anesth Analg 2008; 107:339 41 3. MacIntosh RR. Paravertebral block. In: MacIntosh RR, Brycle-Smith R, eds. Local analgesia: abdominal surgery. Baltimore, MD: Williams and Wilkins Co., 1953:60 3
DOI: 10.1213/ane.0b013e3181af7e7b
997
Section Editor: Paul White
Book, Multimedia, and Meeting Reviews

Anesthesia Crash Course
Horton CN. New York: Oxford University Press, 2009. ISBN-10: 0195371879, ISBN-13: 978-0195371871. 192 pages. $19.95. or newcomers to the field of anesthesiology, the first few days of training can be a daunting experience, requiring rapid acquisition of basic knowledge of a broad range of topics requiring an understanding of physiology and pharmacology. Anesthesia Crash Course is a unique book that provides incoming students (e.g., residents, interns, and externs [medical students]) with a concise and easy-to-read overview of the field of anesthesiology. While its format and scope are similar to the now outof-print Anesthesia for the Uninterested by Birch and Tolmie, Anesthesia Crash Course is clearly directed at a different audiencethose who are actually interested in the practice of anesthesiology. Anesthesia Crash Course is divided into 14 bite-sized chapters, many of which have pithy titles such as Better Living through Chemistry and Dont Launch Your Lunch. By and large, the chapters are logically sequenced, starting with a basic introduction to the anesthesia machine and checkout procedures, progressing through basic pharmacology and preoperative preparation of the patient, and culminating with technical skills, such as airway management, vascular access, and regional anesthetic techniques. Each chapter is designed to guide a complete novice, with little or no previous knowledge of the topic, to a point where he or she is comfortable with basic principles and terminology. For example, the chapter on Sharp Objects devotes 7 pages to starting an IV; it begins with a bill of materials and includes suggestions regarding how to affix a tourniquet, how to locate a suitable vein, how to avoid pulling the needle out while advancing the catheter, and how to tape the finished product. Two of the chapters of Anesthesia Crash Course deviate from this step-by-step approach. The first of these, Less Filling, Tastes Great, provides point-counterpoint discussions of controversial topics that beginning anesthesiologists are likely to encounter during patient-based discussions with their mentors. Examples include, How much fluid should I give my patients? Should we use low flow anesthetic techniques when possible? and What is the best way to secure an endotracheal tube? The chapter entitled, What IF? A Brief Guide to Various Situations provides useful guidance on some common (e.g., What if I cant start the IV), not so common (e.g., What if the patient wont wake up at the end of the case?) and downright rare (What if the hospital catches fire or there is some other hospital-wide emergency?) problems which a new resident might encounter in the operating room. Anesthesia Crash Course does have some notable shortcomings. While the book is designed to provide a broad overview, it is oversimplified at times. Valuable clinical information, such as drug dosing, is not included making it a less valuable resource while in the operating room, despite its pocket-sized format. There is minimal information on respiratory or cardiac physiology and little emphasis on the difficult airway algorithm. There are numerous technical errors; while these are not critical for the absolute novice, they will lead to incorrect responses on standardized examinations. Examples include a statement that remifentanil is metabolized by plasma cholinesterases (p.38, should be nonspecific tissue esterases); a statement that anesthetic gases are typically analyzed by mass spectrometry (p.4, should be infrared analysis); and a recommendation that patients older than 55 years should receive a preoperative chest radiograph (p.55, contradicting the current ASA guidelines). In summary, we recommend Anesthesia Crash Course to individuals who require a rapid overview of the basic concepts in anesthesiology, such as medical students or residents who are beginning their first real exposure to the specialty. The writing is concise, and the simple, conversational tone makes it easy to read the book from cover-to-cover in a couple of hours. The author uses basic, easy-to-understand terminology and does not assume that the reader is familiar with abbreviations and jargon. The pictures and drawings are simple, well labeled, and easy to understand. However, while Anesthesia Crash Course provides a useful introduction, it is not a substitute for a more advanced introductory textbook, such as Stoelting & Millers Basics of Anesthesia or Barash, Cullen & Stoeltings Handbook of Clinical Anesthesia. For new anesthesiology residents beginning their training with differing degrees of previous exposure to the specialty, reading Anesthesia Crash Course prior to the start of residency will help to level the playing field. Therefore, we recommend sending a copy of Anesthesia Crash Course to all incoming residents as required reading before the start of their training. Christopher Beyus, MD
CA-1 Anesthesiology Resident
Jeffrey B. Gross, MD
Professor of Anesthesiology and Pharmacology University of Connecticut School of Medicine Farmington, CT gross@neuron.uchc.edu
A Practical Approach to Regional Anesthesia, 4th ed.

Mulroy MF, Bernards CM, McDonald SB, Salinas FV. Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins, 2009. ISBN-13: 978-0-7817-6854-2, ISBN10: 0-7817-6854-3. 356 pages. $85.00 (paperback). ulroys recent publication of A Practical Approach to Regional Anesthesia, Fourth Edition is a revised and updated version of Regional Anesthesia: An Illustrated Procedure Guide, Third Edition (2002). The text is part of the practical approach series by Lippincott Williams & Wilkins. This edition boasts many changes from the third edition, including color illustrations and the addition of ultrasound techniques and images. In addition, the book has also expanded the number of contributing authors, and all chapters now follow an outline format. Organizationally, the book maintains the same chapter content and progression as the third edition, beginning logically with chapters on local anesthetics, premedication and equipment, followed by regional anesthetic block techniques and specialty procedures. A Practical Approach to Regional Anesthesia is an excellent introductory book for beginning regional anesthesia practitioners looking for information on specific block techniques. Each chapter is clearly written and the outline format introduces standard regional topics in a concise and logical fashion. The authors also discuss controversial topics involving the use of local anesthetics and adjuvants for regional anesthesia, including fentanyl use in epidurals, the mixing of local anesthetics, and the use of depo-local anesthetic preparations. The chapter content is supported by frequent illustrations, tables, and graphs that reinforce the authors point and facilitate reader comprehension. Each
998
Book and Multimedia Reviews
chapter is well referenced, and the key references are bolded. Despite its many positive attributes for the beginning anesthesiologist, some limitations should be pointed out. With regard to its sections on ultrasoundguided regional anesthesia, many of the recommendations for probe use and technique are subjective with little or no data to support one approach over another. Although the ultrasound images are generally of good resolution and acceptable quality, they are fairly small and lack the depth markings that are extremely useful for teaching purposes. In addition, some of the illustrations of probe positioning lack sufficient detail to help inexperienced practitioners with surface placement of the probe. As with any text that transitions from single to multiple authors, some internal inconsistency was found. For example, the quality and clarity of illustrations varies
between chapters, and, in some instances, authors differ in their views on the clinical utility of specific regional anesthetic practices (e.g., the addition of bicarbonate to local anesthetics). For the serious book collector, the quality of the published product is average at best. The book is published in a glossy soft-cover, 8 14 11 14 size format, and the binding is of the modern adhesive variety. It contains 356 pages, and the print size is large enough for easy reading. The paper weight is fairly light, such that figures and tables from opposite and subsequent pages are transparent (although this does not significantly impair the readability of the text material). Mulroys A Practical Approach to Regional Anesthesia, Fourth Edition is a relatively inexpensive and useful introductory textbook on peripheral nerve blockade. It provides the reader with an
easily comprehended discussion of important issues related to regional anesthesia in an outline format that is well-referenced and includes many practical discussion points regarding specific regional anesthetic techniques. Although the ultrasound information and images could be improved, the book does provide a good initial exposure to ultrasound use in regional anesthesia. Overall, this book is a handy reference source for anesthesia providers who are looking for an overview of regional anesthetic medications, equipment, and techniques. Christopher M. Duncan, MD
duncan.christopher@mayo.edu
Hugh M. Smith, MD, PhD

Department of Anesthesia Mayo Clinic Rochester, MN
999

Anesthesia Analgesia September 2009

Încărcat de

Informații document

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Anesthesia Analgesia September 2009

Încărcat de

Drepturi de autor:

Formate disponibile

Editorial

Spend Time on Patients and Families or on Documentation?

Vol. 109, No. 3, September 2009