Surg Endosc (1997) 11: 393396

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Initial experience with breast biopsy utilizing the advanced breast biopsy instrumentation (ABBI)
G. S. Ferzli, J. B. Hurwitz
Department of Surgery, Staten Island University Hospital, 78 Cromwell Avenue, Staten Island, NY 10304, USA Received: 23 August 1996/Accepted: 14 October 1996

Abstract. The Advanced Breast Biopsy Instrumentation (ABBI) system combines a cylindrical single-use biopsy device with digital stereotactic imaging that achieves targeting of radiographic lesions to 1 mm. This allows complete removal of specimens in a one-step procedure that does not involve separate trips to radiology and then surgery. The ABBI system improves on core needle biopsy and fine-needle aspiration and may reduce the need for open biopsy. The authors initial 34 cases utilizing the ABBI system were reviewed. The accuracy of specimen targeting, the success rate of lesion removal, and operative complications were some of the issues assessed. Six cases were not suitable for the procedure: the mammographic lesion was not visualized in four, and the breast was too thin on compression in two. There was successful removal of the lesion in 27 of the remaining 28 cases. There were no local wound complications, and patient satisfaction was high in all completed biopsies. The ABBI system is an effective new form of minimally invasive breast surgery. It provides complete excision of mammographic abnormalities. Its use of the most direct path to these lesions allows for minimal removal of adjacent normal tissue. In this study there were no complications and very little patient pain. Key words: Advanced Breast Biopsy Instrumentation ABBI

sion of suspicious lesions. The Advanced Breast Biopsy Instrumentation (ABBI) uses stereotactic technology to achieve total excision of abnormalities rather than sampling. This provides superior accuracy and reliability of histologic diagnosis while maintaining the benefits of minimally invasive methods. Initial experience, technique, and early results are reported. Materials and methods Preoperative evaluation
Stereotactic-guided biopsy with the ABBI is offered to patients who require evaluation of suspicious nonpalpable mammographic lesions. All films are thoroughly evaluated by the surgeon and an experienced radiologist, frequently with magnification and spot compression views, and often with adjunctive ultrasonography. Since the prone position with arms at the side and neck turned must be maintained for 20 to 30 min, all candidates are questioned regarding neck and back problems. Weight, chronic obstructive pulmonary disease (COPD), and a history of psychiatric problems are also determinants of eligibility for the procedure. No preoperative lab testing is required unless the patient is on an anticoagulant. Patients are asked to stop taking aspirin 1 week prior to the procedure. It is advantageous to screen the patient with digital stereotactic views prior to surgery to confirm visualization of the abnormality and to assess the thickness of the breast on compression. The procedure is performed by the surgeon and one technician.

Technique Needle-guided open biopsy is the current standard for the evaluation of nonpalpable mammographic abnormalities. Available minimally invasive localization techniques include fine-needle aspiration and large core biopsy. Neither of these, however, is capable of achieving complete exciThe patient is placed in the prone position on a dedicated stereotactic biopsy table; the breast hangs freely through the top aperture. The breast is then compressed. The front plate has a 5 5 inch opening through which the biopsy is performed. After a scout image confirms the lesion within the front aperture, stereo views are obtained. The images are acquired digitally and reconstructed on a computer monitor. Successful targeting and biopsy are possible only if the lesion is visualized from both angles. The abnormality is then acquired in each of the stereo images and the software calculates the horizontal (x), vertical (y), and depth (z) coordinates of the targeted lesion. The skin is prepped and draped in sterile fashion and the ABBI gun (Fig. 1) of appropriate size is chosen. Sizes vary from 5 to 20 mm in diameter. The smallest instrument that will allow complete removal with

Correspondence to: G. S. Ferzli

394

Fig. 1. ABBI gun. Above: outer canula. Below: inner needle. Fig. 2. The lesion is targeted and its horizontal, vertical, and depth coordinates are calculated and displayed at the top right of the monitor ( 1.0, y 27.0, z 23.0). Fig. 3. The needle of the ABBI gun is advanced into the target. A pair of stereo views confirms good alignment of the needle tip with the lesion. Fig. 4. The T of the ABBI gun is deployed to anchor the needle. Stereo views confirm the alignment of the T with the target. Fig. 5. The cannula of the ABBI is advanced into the breast to 15 mm beyond the T. A pair of stereo views again confirms good alignment of the lesion with the needle, T, and cannula. Measurements allow verification that the z of cannula tip is indeed greater than the z of the T. This ensures that the snare wire will transect breast tissue beyond the T. Fig. 6. Specimen and breast radiographs confirming the removal of the target lesion.

good margins is chosen, and it is then seated on the gun holder in the home position. The horizontal and vertical coordinates of the lesion are downloaded from the computer to a motor that moves the gun holder to the corresponding position (Fig. 2). The skin is infiltrated with local anesthesia

at the insertion site and an incision is made to accommodate the finder needle. This needle is then inserted to the predetermined depth coordinate (z). This can be done manually or mechanically. A pair of stereo digital images is then acquired (Fig. 3). The needle tip should be perfectly aligned

395 with the lesion. Occasionally the lesion may no longer be seen due to the lidocaine injection within the breast parenchyma. In this case, the coordinates of the tip of the needle are measured and compared to the original lesion coordinates. If both coordinates are equal, then the procedure may continue. A T is now deployed at the end of the needle which anchors the needle in the breast. Stereo views are again acquired, because sometimes the patient pulls away or the breast moves and the lesion may no longer align with the calculated coordinates (Fig. 4). Additional lidocaine is now injected to accommodate the cannula of the ABBI gun. The skin incision is extended to allow insertion of whatever size gun has been previously selected. A motorized oscillating blade now advances a circular blade into the breast tissue through the cylindrical cannula. Skin edges should be lifted and kept free from the canula while advancing the blade. Cutting continues to a depth (z) 15 mm greater than that of the deployed T. The knife blade is withdrawn from the cylinder and the outer cannula is advanced an additional 2 mm to reach the depth that was occupied by the blade when it protruded beyond the sheath. The cannula has within its tip a snare wire that, upon activation, will cut through the deepest border of the cylinder of breast tissue. A pair of stereo views is again obtained both to ensure that the suspicious tissue is fully within the cannula and to confirm the depth of the wire snare, which must exceed the depth of the T (Fig. 5). The cannula, deployed T, and lesion should be perfectly aligned. The snare is then fired with the proximal wire attached to electrocautery. The instrument is then removed from the incision site in its entirety. The wound is inspected for bleeding and further electrocautery or packing may be used as necessary. Compression is then released, and a scout image of the breast is obtained with any packing removed to document the complete removal of the abnormality. Further confirmation is obtained with a specimen radiograph (Fig. 6). During this time the patient is returned to the supine position to allow irrigation and closure of the wound, followed by a pressure dressing and compressive bra. Patients may return immediately to full activity; some will require mild analgesics.

sure. One of the inadequate margins was due to procedural error, one was due to improper calibration, and the last was due to patient movement during the procedure. In one patient with cancer a positive margin was not revealed until permanent tissue sections were reviewed by pathology. This patient ultimately opted for a modified radical mastectomy. Patients were routinely given a compressive bra postoperatively and were examined 1 week later. No hematomas, seromas, or wound infections were seen. A few patients developed minimal ecchymosis around the suture line, but this never exceeded 12 cm. Four patients were given prescriptions for acetaminophen and codeine, but did not take the medication.

Discussion When the radiologic assessment of nonpalpable breast abnormalities suggests or is inconclusive for malignancy, patients are historically offered excisional biopsy following localization using the Kopans hook wire technique [10]. This traditional needle-directed open biopsy is a two-step, time-consuming, invasive, and costly procedure, with general anesthesia often required. The radiologist may fail to place the localization wire through or near the target lesion, and the surgeon may not always successfully remove it. A failure rate varying from 0.2 to 22% has been reported [6, 11, 18, 19]. Stereotactic fine-needle aspiration (SFNA) cytologic study of occult breast lesions was introduced in the United States in 1986 in an attempt to reduce the number of open procedures and improve the positive predictive value of mammography [1, 4]. This technique, however, even though minimally invasive and cost-effective, requires specialized breast cytopathologists. It has an inadequate sampling rate of 1026% and a false-negative rate of 514% [2, 3, 9, 14]. Aspiration cytology has been extended to large core biopsy in an attempt to improve diagnostic specificity. Parker et al. initially reported their experience with an automated needle core biopsy device used with stereotactic mammography, using 18-gauge, then 16-gauge, and finally 14-gauge needles [15]. They later reported on 102 patients biopsied with 14-gauge needles, interpreted by the one cytologist [16]. Core biopsy was followed by excisional biopsy, and histologic results from the gun biopsy and the surgical biopsy specimens were in agreement in 96%. A subsequent multi-institutional study reporting on 6,152 lesions concluded that percutaneous large core breast biopsy is a reproducible and reliable alternative to surgical biopsy [17]. The problem of specimen insufficiency encountered with SFNA had been virtually eliminated, and sampling errors could be decreased by taking multiple cores. Needle track seeding of tumor has never been demonstrated. Lieberman et al. reported on 145 lesions sampled with an automated gun and a 14-gauge needle and concluded that stereotactic 14 gauge core biopsy achieved a 99% diagnostic yield with five specimens for masses, but additional specimens may be necessary to diagnose some calcified lesions [13]. These numbers are supported both by Fine [5], who recommends four to five cores for nodular densities and nine to twelve for microcalcifications, and Parker,

Results ABBI was offered as a diagnostic modality to 34 consecutive nonselected patients who presented with nonpalpable abnormal findings on mammography (24 nodules, ten microcalcifications) between April 12 and July 30, 1996. Six patients were excluded from the study; the abnormality could not be digitally imaged in four cases, and the breast was too thin on compression (less than 30 mm) in two. In the remaining 28 patients, 27 specimens were removed successfully (18 nodules, nine microcalcifications). The procedure was aborted in one patient with a history of psychiatric illness because of phobia and discomfort. Patient ages ranged from 32 to 75 (mean 49). The duration of the procedure varied from 20 to 45 min (mean 30). Lidocaine usage varied from 1.5 to 30 cc (mean 14). Three patients were given 10 mg of Valium the night before operation and again on the morning of the procedure. Twentymillimeter cannulae were used in ten patients; 15 mm in eight; and 10 mm in nine. In four patients the snare wire did not activate properly and Metzenbaum scissors were used to complete the excision of the partially divided breast tissue. In two cases the stereotactic unit was not calibrated correctly, resulting in improper centering of the lesion by 4 mm on the y axis. One 75-year-old woman with multiple medical problems fainted 15 min postoperatively and required admission to the hospital for overnight monitoring. Pathology was benign in 24 specimens and malignant in three. In three patients we realized that the margins were inadequate immediately after removal of the specimen. Proper margins were then obtained by further excising the involved margin; once with the breast tissue under compression and twice after turning the patient prior to wound clo-

396

who suggests ten or more for calcifications. Other advantages of the large core biopsy are detailed in a report by Israel [7]. These include the ability to perform biopsies on lesions seen in only one view, multiple lesions, and breasts that contain implants. A substantial decrease in cost in comparison to open biopsy is also realized. An important limitation of large core biopsies is the finding of atypical ductal hyperplasia, which raises concern about the concomitant presence of carcinoma in the lesion sampled. Lieberman [12] found carcinoma in 11 of 21 cases identified as atypical ductal hyperplasia on stereotactic core biopsy and therefore recommended routine open biopsy for this finding. Similarly, Jackman [8] found carcinoma in nine of sixteen such cases. In all cases of large core biopsy, the lesion is not completely removed and follow-up mammography is routinely recommended. The ABBI combines the time and cost savings of core biopsy with the complete specimen removal that has heretofore only been possible with open needle localization biopsy. The average operative time in this series compares with that of core biopsies, and the institutional cost for an ABBI is $1,200, compared to $2,700 for an open biopsy. Most of this savings is due to not needing to use an operating room. ABBI is more expensive than core biopsy, but confirmed complete removal of the lesion may eliminate the need for a short-term follow-up mammogram. There is also no longer an issue over how many cores to perform to adequately sample a lesion. Specimen orientation is excellent and makes further excision of tissue for inadequate margins straightforward. The one cosmetic disadvantage of ABBI is that it precludes the use of a circumareolar incision. This is potentially offset by removing less tissue and by always taking the most direct route to the lesion. Contraindications to ABBI include weight greater than 300 pounds; the inability to lie still for 2030 min (COPD, asthmatic, psychiatric disorders); thin, ptotic breasts (less than 30 mm after compression); and lesions close to the chest wall that cannot be visualized with the stereotactic unit. Patients with bloody discharge and retro-areolar superficial lesions are not proper candidates. The range of cannula sizes from 5 to 20 mm allows great flexibility and matching to lesion size. These cannulae are in need of some mechanical refinement, however, because the snare wire did not activate in one case, and it partially cut through the breast tissue in four other cases. The surgeon must always check that the equipment is properly calibrated, since this may affect proper centering of the cannula, as demonstrated in two of our cases. The only morbidity requiring hospitalization involved a 75-year-old woman with multiple medical problems who fainted 10 min after the operation and needed observation overnight. The add-on units used originally for stereotactic biopsy were associated with vasovagal episodes, which were markedly reduced by the use of prone tables. In retrospect, however, this elderly patient with multiple medical problems should have been biopsied in an operating room setting. Conclusion The Advanced Breast Biopsy Instrumentation single-use minimally invasive biopsy device used by the authors

proved itself to be highly successful. If patients meet the requirements, this procedure provides fast, accurate lesion removal, without complications. Biopsy of nonpalpable mammographically detected lesions using the ABBI is more aggressive and slightly more costly than automated needle core biopsy; however, it provides complete removal of the lesion and avoids the cost and invasiveness of the traditional open needle localization biopsy. While ABBI will certainly not completely replace needle localization biopsy, it should definitely assume a significant role in the diagnosis of breast disorders. References
1. Azavedo E, Suane G, Aver G (1989) Stereotactic fine-needle biopsy in 2594 mammographically detected non-palpable lesions. Lancet 1: 10331036 2. Costa MJ, Tadros T, Hilton G, Birdsong G (1993) Needle aspiration cytology: utility as a screening tool for clinically palpable lesions. Acta Cytol 37: 461471 3. Dowlatshahi K, Gent HJ (1989) Nonpalpable breast tumor diagnosis with stereotactic localization and fine-needle aspirations. Radiology 170: 424433 4. Dowlatshahi K, Jokich PM, Schmidt R, Bibbo M, Dawson PJ (1987) Cytologic diagnosis of occult breast lesions using stereotaxic needle aspiration. Arch Surg 122: 13431346 5. Fine RE, Boyd BA (1996) Stereotactic breast biopsy: a practical approach. Am Surg 62: 96102 6. Hasselgren P, Hummel RP, Georgian-Smith O, Fieler M (1993) Breast biopsy with needle localization: accuracy of specimen x-ray and management of missed lesions. Surgery 114: 836842 7. Israel PQ, Fine RE (1995) Stereotactic needle biopsy for occult breast lesions: a minimally invasive alternative. Am Surg 61: 8791 8. Jackman RJ, Nowels KW, Shepard MJ, Finkelstein SI, Marzoni FA (1994) Stereotaxic large core needle biopsy of 450 non-palpable breast lesions with surgical correlation in lesions with cancer or atypical hyperplasia. Radiology 193: 9195 9. Jackson VP, Bassett LW (1990) Stereotactic fine-needle aspiration biopsy for non-palpable breast lesions. AJR Am J Roentgenol 154: 11961197 10. Kopans DB, Lindfers K, McCarthy KA, Meyer JE (1985) Spring hookwire breast lesion localizer: use with rigid-compression mammographic systems. Radiology 157: 537538 11. Landercasper J, Gundersen SB, Gundersen AL, Cogbill TH, Travelli R, Strutt P (1987) Needle localization and biopsy of non-palpable lesions of the breast. Surg Gynecol Obstet 164: 399403 12. Lieberman L, Cohen MA, Dershaw DD, Abramson AF, Hann LE, Rosen PP (1995) Atypical ductal hyperplasia diagnosed at stereotaxic core biopsy of breast lesions: an indication for surgical biopsy. AJR Am J Roentgenol 164: 11111113 13. Lieberman L, Dershaw DD, Rosen PP, Abramson AF, Deutch BM, Hann LE (1994) Stereotaxic 14 gauge breast biopsy: how many core biopsy specimens are needed? Radiology 192: 793795 14. Lofgren M, Anderson I, Lindholm K (1990) Stereotactic fine-needle aspiration for cytologic diagnosis of non-palpable breast lesions. AJR Am J Roentgenol 154: 11911195 15. Parker SH, Lovin JD, Jobe WE, Leuthke JM, Hopper KD, Yakes WF, Burke BJ (1990) Stereotactic breast biopsy with a biopsy gun. Radiology 176: 741747 16. Parker SH, Lovin JD, Jobe WE, Burke BJ, Hopper KD, Yakes WP (1991) Non-palpable breast lesions: stereotactic automated large core biopsies. Radiology 180: 403407 17. Parker SH, Burbank F, Jackman RJ, Aucreman CJ, Cardenosa G, Link TM, Cosciage JL, Eklund GW, Evans III WP, Garver PR (1994) Percutaneous large core breast biopsy: a multi-institutional study. Radiology 193: 359364 18. Petrovich JA, Ross DS, Sullivan JW, Lake TP (1989) Mammographic wire localization in diagnosis and treatment of occult carcinoma of breast. Surg Gynecol Obstet 168: 239243 19. Yanaskas BC, Knelson MH, Abernethy ML, Cutino JT, Clark RL (1988) Needle localization biopsy of occult lesions of the breast: experience in 199 cases. Invest Radiol 23: 2933

Surg Endosc (1997) 11: 359361

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Preoperative pneumatic dilatation represents a risk factor for laparoscopic Heller myotomy
M. Morino, F. Rebecchi, V. Festa, C. Garrone
Department of Surgery, Clinica Chirurgica I, University of Turin, C/so A.M. Dogliotti 14, 10126 Turin, Italy

Abstract Background: The development of minimally invasive surgery has renewed interest in the surgical therapy of achalasia. Methods: 21 patients with esophageal achalasia underwent Hellers laparoscopic myotomy with anterior fundoplication between August 1991 and March 1996. Results: There were two intraoperative perforations of the mucosa sutured laparoscopically with no postoperative sequelae; both complications occurred in patients previously treated with pneumatic dilatation; no perforations occurred in the 14 patients who had not been submitted to pneumatic dilatation (28% vs 0%). There were no surgical mortalities and no postoperative morbidities. Complete relief of dysphagia and modifications of radiological and manometric patterns were achieved in all patients. All patients remain asymptomatic at follow-ups ranging from 3 to 55 months after surgery. Conclusions: Laparoscopic Hellers myotomy is as effective as traditional surgery in treating symptoms and has all the advantages of pneumatic dilatation in terms of short hospital stay, quick recovery, and low cost; preoperative pneumatic dilatation is a risk factor for intraoperative mucosal perforation. Key words: Achalasia Esophagomyotomy Laparoscopy Fundoplication

pneumatic dilatation the treatment of choice for achalasia. However, the recent development of minimally invasive surgery has renewed interest in surgical therapy for achalasia [5]. Laparoscopic Hellers myotomy with either a full or partial or no fundoplication combines the short hospital stay and rapid patient recovery associated with forced dilatation with the high success rate and low recurrence rate typical of conventional surgery [15]. This report describes a 4-year clinical experience in the laparoscopic management of esophageal achalasia and evaluates pneumatic dilatation with respect to the surgical procedure.

Materials and methods

From August 1991 to March 1996, 21 patients (6 male and 15 female; mean age 43 years, range 1271) with esophageal achalasia underwent Heller laparoscopic cardiomyotomy with anterior fundoplication (Dor procedure) at the Department of Surgery of the University of Turin. All patients had dysphagia; 14 complained of periodic regurgitation, three patients complained of heartburn, and nine reported significant (>5 kg) weight loss in the 6 months prior to surgery. The mean duration of symptoms was 59 months (range 4 months10 years). Upper gastrointestinal endoscopy excluded malignancy in all patients; all patients had had a barium esophagogram with changes consistent with achalasia. Esophageal manometry was suggestive of achalasia in all cases, showing 100% synchronous postdeglutitory contractions with aperistalsis of the esophageal body and an uncoordinated or unrelaxing lower esophageal sphincter. Seven patients (33%) had been previously treated with pneumatic balloon dilatation (three patients once, two patients twice, two patients four times); all of them had recurrent dysphagia after a mean of 15 months (range 2 months10 years). Two patients had had previous upper abdominal surgery. The surgical technique reproduced by laparoscopic approach the classical laparotomic cardiomyotomy and has been previously described [11]. In all cases an anterior fundoplication was performed.

Achalasia of the esophagus is a disease of unknown origin that affects motor functions throughout the esophagus. It is characterized by failure of relaxation of the lower esophageal sphincter after swallowing with aperistalsis of the esophageal body. The available treatments for achalasia include pneumatic dilatation and esophageal myotomy, first performed by Heller in 1913 [8]. In spite of the results of the only randomized trial [4] comparing these two treatments, most authors consider
Correspondence to: M. Morino

Results All the procedures were completed by laparoscopy with no need for conversion. Mean operative time was 155 min (range 100210). Blood loss was negligible and never passed 100 cc. The only operative complication was a small perforation

360

Fig. 1. Mucosal perforations in patients submitted or not submitted to preoperative pneumatic dilatation.

of the mucosa at the level of the cardias in two patients (9.5%); both patients had previously submitted to pneumatic dilatation (2 and 4 dilatations, respectively); the perforation was laparoscopically sutured in both cases and had no postoperative sequelae. No perforations occurred in the 14 patients who had not been submitted to pneumatic dilatation (Fig. 1). Postoperative course was uneventful for all patients, with a mean hospital stay of 3.4 days (range 26). Return to normal activities was possible within 10 days after surgery for 11 patients, between 10 and 20 days for nine patients, and after more than 20 days for the last patient. At present, with a follow-up ranging from 3 to 55 months, none of our patients has complained of recurrent dysphagia and only one patient developed an asymptomatic gastro-esophageal reflux. Seventeen patients have been studied 6 months after surgery by manometry and pH monitoring. Manometric measurements revealed a significant decrease in gastro-esophageal sphincter pressure when compared to preoperative data: mean preoperative pressure 38.1 mmHg, mean postoperative pressure 15.3 mmHg. No patients showed abolition of or zero sphincter pressure. Acid reflux tests were positive in one patient; this patient remains asymptomatic at 23 months from surgery.

Discussion Before the advent of laparoscopic surgery the choice of treatment for patients suffering from esophageal achalasia was pneumatic dilatation or surgical myotomy. Csendes [4] in 1989 reported the only randomized controlled trial comparing pneumatic dilatation and surgical myotomy. After a 5-year follow-up, 95% of the patients who underwent surgery were asymptomatic and 2% had mild heartburn, while 65% of the dilated patients were asymptomatic and none had heartburn. While these results suggest that surgery may be more efficacious than pneumatic dilatation in treating achalasia, gastroenterologists still favor pneumatic dilatation because it requires only an overnight hospitalization, because the patient recovers in 1 week, and because the average cost is one-quarter the cost of surgery [13]. The advent of minimally invasive surgical treatment for achalasia may change this situation. The major advantage of the laparoscopic operation compared to the laparotomic one is the postoperative course, which is characterized by all the

benefits of minimal access surgery: minimum pain, no ileus, short hospital stay, optimum cosmetic results, and early return to normal activity [16, 17]. Though numbers are small in this study and in the literature, laparoscopic Heller myotomy seems to be a safe procedure with a morbidity rate ranging from 0% to 14% and no reported mortality [16, 17]. Therefore, operative mortality and morbidity of laparoscopic myotomy compare favorably to the results reported in the literature for traditional Heller myotomy [27], where mortality rates vary from 0 to 4% and morbidity from 4 to 18%. The most significant intraoperative or immediately postoperative complication of esophageal myotomy is the perforation of the esophageal or gastric mucosa. This complication can occur intraoperatively during the open Heller operation as frequently as 18% of the time [2], while the only consistent published series on laparoscopic myotomy reports two mucosal perforation out of 13 procedures (14%) [17]. In the present series we had two mucosal tears (9.5%) that were sutured laparoscopically with no postoperative sequelae. The incidence of mucosal perforation was significantly higher in the group of patients who underwent multiple sessions of pneumatic dilatation (28% vs 0%) before surgical treatment. Pneumatic dilatation causes submucosal microhemorrages that will heal with a covering of fibrous tissue; as a consequence the dissection of the correct plane between muscular layers and submucosal can be difficult, increasing the risk of causing an inadvertent mucosal perforation. If a direct perforation should occur, it is crucial to identify it during surgery. In our experience the perfect view and magnification obtained by laparoscopy permitted a thorough evaluation of the myotomy and an immediate diagnosis of the perforation. Some authors [117] have suggested using air insufflation or methylene blue injection through the nasogastric tube at the end of the myotomy; others have proposed intraoperative (I.O.) use of flexible endoscopy [15]. As shown for open Heller procedures the myotomy must be extensive in length and complete in depth to avoid recurrent dysphagia. It should ascend at least 5 cm along the thoracic esophagus and include 12 cm of gastric muscular layers. In order to verify that the myotomy has been completed some authors [10] propose to insert a Blakemore catheter and to inflate it in order to distend the esophageal mucosa and in case residual muscle fibers. This maneuver presents the risk of damaging the mucosa and we do not suggest its use, especially in laparoscopy, as the magnification of laparoscopic view is such that single circular muscle fibers are easily identified and sectioned. More recently Pellegrini [12] proposed I.O. esophagoscopy to accurately gauge the extent of the myotomy. We routinely use I.O. manometry to verify the radicality of the myotomy. The manometric probe is inserted at the end of the myotomy before placing the antireflux valve. The gradient of pressure between the stomach and the esophagus should be neutralized at the end of a complete myotomy. If this is not the case the operative field should be checked, looking for persisting circular fibers. In our experience I.O. manometry influenced our surgical technique in one-third of the cases: in six out of 18 patients a positive I.O. manometry led the surgeon to

361

find residual circular muscular fibers and to complete the section. We believe that I.O. manometry has played a major role in the present series in which no cases of postoperative dysphagia were encountered. The most significant late complication of myotomy is gastroesophageal reflux. Overall incidence of symptomatic reflux reported after laparotomic myotomy is usually less than 10% [18], but studies with larger follow-up have reported reflux rates as high as 52% [9]. The high frequency of late postoperative symptomatic gastroesophageal reflux is a strong argument in favor of performing an antireflux procedure as a complement to the myotomy [24]. Among the different antireflux procedures we prefer to perform an anterior hemi-Nissen or Dor technique because the procedure permits one to obtain the three above-mentioned advantages without necessitating any further dissection of the esophagus. We feel that it has now been demonstrated that Heller myotomy may be performed laparoscopically with minimal morbidity and a short hospital stay. Long-term evaluation of the efficacy of this new procedure is needed, but results should not be different from those observed after Heller myotomy and Dor fundoplication carried out by open surgery; 80% of our patients have been controlled 6 months postoperatively by manometry and 24-h pH monitoring: There were no manometric signs of recurrence and only in one patient were acid-reflux tests positive; this patient remains asymptomatic 23 months after surgery. These functional results compare favorably with those of traditional surgery. In our view laparoscopic esophageal myotomy appears to be the ideal treatment for esophageal achalasia: It is as effective as traditional surgery in treating symptoms and has all the advantages of pneumatic dilatation in terms of short hospital stay, quick recovery, and low cost. Therefore we believe that laparoscopic myotomy will rapidly become the treatment of choice for esophageal achalasia for the vast majority of patients. In our opinion the clinical role of pneumatic dilatation should be reconsidered not only with respect to the high percentage of recurrence and reflux [3, 14] but also because it represents a risk for laparoscopic myotomy.

References
1. Ancona E, Peracchia A, Zaninotto G, Rossi M, Bonavina L, Segalin A (1993) Heller laparoscopic cardiomyotomy with antireflux anterior fundoplication (Dor) in the treatment of esophageal achalasia. Surg Endosc 7: 459461 2. Black J, Vorbach AN, Collis JL (1976) Results of Hellers operation for achalasia of the oesophagus. The importance of hiatal repair. Br J Surg 63: 949953 3. Csendes A, Braghetto I, Mascaro J, Henriquez A (1988) Late subjective and objective results of oesophagomyotomy in 100 patients with achalasia of the oesophagus. Surgery 104: 469475 4. Csendes A, Braghetto I, Henriquez A, Cortes C (1989) Late results of a prospective randomized study comparing forceful dilatation and oesophagomyotomy in patients with achalasia. Gut 30: 229304 5. Cuschieri A (1992) The spectrum of laparoscopic surgery. World J Surg 16: 10891097 6. Cuschieri A, Berci G (1992) Laparoscopic biliary surgery. 2nd ed. Blackwell Scientific, London 7. Ellis FH, Drozier RE, Watkins E (1984) The operation for the oesophageal achalasia: results of oesophagomyotomy without antireflux operation. J Thorac Cardiovasc Surg 88: 344347 8. Heller E (1914) Extramukose Cardiaplastik beim chronischen Cardiospasmus mit Dilatation des Oesophagus. Mitt Grenzgeb Med Chir 27: 141149 9. Jara FM, Toledo-Pereira LH, Lewis JH, Mulligan DJ (1979) Longterm results of oesophagomyotomy for achalasia of the oesophagus. Arch Surg 114: 935940 10. Mailard JN, Hay JM (1987) Les myotomies oesophagiennes. Enciclopedie Medico Chirurgicale (Paris), Techniques chirurgicales appareil digestif 11: 18 11. Morino M, Rebecchi F, Festa V, Garrone G (1995) Laparoscopic Heller cardiomyotomy with intraoperative manometry in the management of oesophageal achalasia. Int Surg 80: 332335 12. Pellegrini CA, Leichter R, Patti M, Somberg K, Ostroff JW, Way L (1993) Thoracoscopic esophageal myotomy in the treatment of achalasia. Ann Thorac Surg 56: 680682 13. Richter JE (1991) Motility disorders of the oesophagus. In: Yamada T (ed) Textbook of gastroenterology. JB Lippincott, Philadelphia, p 1083 14. Richter JE (1991) Achalasia: whether the knife or balloon? Not such a difficult question. Am J Gastroenterol 86: 810811 15. Robertson GSM, Lloyd DM, Wicks ACB, Caestecker DE, Veitch PS (1995) Laparoscopic Hellers cardiomyotomy without an antireflux procedure. Br J Surg 82: 957959 16. Shimi S, Nathanson LK, Cuschieri A (1991) Laparoscopic cardiomyotomy for achalasia. J R Coll Edinb 36: 152154 17. Swanstrom L, Pennings J (1995) Laparoscopic esophagomyotomy for achalasia. Surg Endosc 9: 286292 18. Vantrappen G, Hellemans J (1980) Treatment of achalasia and related motor disorders. Gastroenterol 79: 144150

Surg Endosc (1997) 11: 362365

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Surgical ultrasound in suspected acute appendicitis

A. Zielke,1 C. Hasse,1 H. Sitter,2 O. Kisker,1 M. Rothmund1
1 2

Department of General Surgery, Philipps-University of Marburg, Baldinger Strasse, PO-Box 100, 35043 Marburg, Germany Institute of Theoretical Surgery, Philipps-University of Marburg, Baldinger Strasse, 35043 Marburg, Germany

Received: 16 July 1996/Accepted: 19 August 1996

Abstract Background: Ultrasonography (US) by acknowledged experts enhances the diagnostic performance and reduces the rate of negative laparotomies in patients with suspected acute appendicitis (AA). Methods: The diagnostic accuracy and clinical impact of routine US performed by surgical residents was prospectively studied in 504 unselected patients admitted for AA. Clinical and US findings were correlated with laparotomy findings and pathological outcome in 135 patients (113 cases with proven AA, prevalence 22.4%) and clinical as well as follow-up data were compared in the remainder. Results: The overall accuracy, sensitivity, and specificity of the clinical diagnosis of AA were 84.9%, 51.3%, and 94.6% and those of US were 93.6%, 83.1%, and 96.6%. Joint evaluation of the results from clinical evaluation and US further improved diagnostic performance (accuracy 93.4%, sensitivity 84.1%, specificity 96.2) and significantly reduced the rate of diagnostic errors to 3.4% (p < 0.001) and unnecessary laparotomies to 9.6% (p < 0.01) in patients with suspected AA. Conclusions: Ultrasonographic evaluation of the patient with suspected AA is considered to be of value in surgical practice. Key words: Ultrasound Acute appendicitis Prospective trial Medical decision making

clock diagnostic tool. Therefore, we conducted a prospective observational trial to evaluate the efficacy of US in diagnosing this condition. The study aims to answer the following questions: Can surgical residents achieve results comparable to those of expert sonographers? Does routine use of US provide additional useful information with regard to the decision to perform a laparotomy? Does it reduce the rate of unnecessary laparotomies?

Patients and methods

Over a 15-month period, all patients admitted to our hospital for suspected acute appendicitis were included in this study. History and physical examination followed standardized patient interview pathways, similar to those used by the European Community Study Group on Acute Abdominal Pain [3]. A provisional clinical diagnosis was established by a senior registrar or consultant, and patients were allocated to one of three clinically defined categories, which allowed for a grading of the clinical suspicion of AA; group A: acute appendicitis very likely: urgent laparotomy; group B: acute appendicitis equivocal: in-hospital observation; group C: acute appendicitis very unlikely: follow as outpatient. Subsequent to classification, US was performed by one of five sonographically trained surgeons using a Siemens Sonoline SL1 ultrasound device and 3.5- and 5-MHz curved arrays and 7.5-MHz linear small-part array transducers (Siemens AG, Erlangen, Germany). Examination of the entire abdomen, the retroperitoneum, and the pelvis was followed by a detailed exploration of the right lower quadrant with graded compression as described by Puylaert [16, 17]. A target-like appearance of the appendix in transverse view with a diameter of 6 mm, lack of compressability and peristalsis, and the demonstration of a blind-ending tip in longitudinal view were considered positive diagnostic criteria of acute appendicitis. Nonvisualization of the appendix was regarded a negative result. Patients were reclassified according to the outcome of US: Final therapeutic decisions were made according to the decision tree presented in Fig. 1. Whenever unequivocally positive US criteria for acute appendicitis were revealed, patients were scheduled for urgent laparotomy, irrespective of their clinical presentation, because of the consistently high accuracy of US for this condition. Also, in selected cases, patients of group A were not operated upon if US excluded AA and offered a valid differential diagnosis. Specimens from all patients undergoing surgery were subject to thorough histological examination. Positive histopathologic criteria of acute appendicitis required demonstration of granulocyte infiltration of the appendix wall (minimum requirement: infiltration of the mucosa and demonstration of mucosal affections, e.g., superficial ulcerous lesions). These criteria were defined as a standard of reference for final diagnoses. Appendiceal perforation was diagnosed during laparotomy. For patients who were subject to conservative treatment, final

Recent prospective studies have convincingly shown that ultrasonography (US) in suspected acute appendicitis (AA) improves the diagnostic accuracy in experienced hands [8, 16, 7, 19]. The clinical significance of this method has been outlined by its ability to reduce delay in diagnosis as well as unnecessary laparotomies in patients suspected to have AA [17, 19]. However, it is entirely unclear whether surgeons can effectively use this diagnostic modality as a round the
Correspondence to: A. Zielke

363 Table 1. Results of clinical and ultrasonographic evaluation of 504 patients with suspected acute appendicitisa Total patients 504 113 94 378 13 19 Group B 79 58 51 17 4 7 C 188 42 36 139 7 6 237 13 7 222 2 6

A Number of patients Acute appendicitis Ultrasonography TP Ultrasonography TN Ultrasonography FP Ultrasonography FN

a

T/F: true, false; N/P: negative, positive

Table 2. Final diagnoses for 504 patients admitted for suspected acute appendicitis of 504 patients with suspected acute appendicitisa Final diagnoses Proven acute appendicitis Mesenteric lymphadenitis, ileitis Other gastrointestinal conditions Gynecological conditions Urologic conditions Other conditions Nonspecific abdominal pain
a

Patients 113 68 35 28 9 33 218

Ultrasound TP 88 55 28 17 4 20 *

TP: true positive; * no ultrasonographic abnormalities in these patients

Fig. 1. Flow chart of diagnostic workup and therapeutic options. Patients with unequivocal US signs of acute appendicitis undergo surgery, irrespective of the clinical presentation. Patients of group A may not be operated upon if US clearly refutes acute appendicitis and offers a valid differential diagnosis.

diagnoses were substantiated by evaluation of the clinical course and of all subsequent examinations including follow-up observations. Clinical, US, and results after reclassification were compared by calculating the respective sensitivity, specificity, predictive value, likelihood ratio, post-test odds, and usefulness index (UI) according to published methods [3, 4]. The likelihood ratio of a positive test (LR+) expresses how many times greater the probability of a test result among patients with the condition is as opposed to those without. It should always be >1 and 10 signifies a good test. The UI was calculated according to [d(d r)], where d is the incidence of the condition in the disease (sensitivity) and r is the incidence of the finding in a reference population (1 specificity). A test is regarded useful with a UI of 0.35 [10]. The rate of negative appendectomies and the negative laparotomy rate were calculated according to the definition given by deDombal [5]. Bad diagnostic errors address the number of patients in whom a nonsurgical condition was diagnosed where in fact urgent surgery would have been needed [3]. Differences were tested for significance by 2 tests and p < 0.05 accepted to identify significance.

Results A consecutive and unselected series of 504 patients was entered into this study. These comprised 211 male patients (median 21 years of age, range 286) and 293 female patients (median 21 years of age, range 190). The results of clinical and US evaluation as well as the final diagnoses are given in Tables 13; 113 of 504 patients had histologically proven AA (prevalence 22.4%). These comprised 58 male and 55 female patients with a median age of 26 (range 1065) and 23 (range 181) years of age, respectively. On admission, 58 (51%) of 113 patients with AA were allocated into urgent-surgery group A, 42 (37%) were placed into the

equivocal group B, and the remaining 13 (12%) went into group C (sensitivity 51.3%). Twenty-one patients were classified false-positive on initial physical examination (specificity 94.6%, accuracy 84.9%). Ultrasonography enabled us to diagnose AA in 94 of the 113 patients (sensitivity 83.1%); 36 of the 55 patients with acute appendicitis presenting with equivocal or highly unsuspected signs were correctly diagnosed by ultrasonography. In 19 cases of proven appendicitis (16.8%) US did not detect the disease, including seven group A patients. Falsepositive sonographic results were recorded in 13 cases (specificity 96.6%, accuracy 93.6%). Of these, four group A and three group B patients underwent surgery (Table 1). Of the 21 patients who were clinically false positive (group A), US was true negative in 17 (80%). Nine of these patients did not undergo surgery because US excluded appendiceal inflammation and diagnosed a different, nonsurgical condition. Among those differential diagnoses with a typical US morphology, mesenteric lymphadenitis, ileitis and ileocolitis in both instances were the most common. Diagnostic US results were obtained in 124 of 173 (72%) patients with a confirmed alternative condition: mesenteric lymphadenitis (9), ileocecitis (16), a variety of gastrointestinal disorders including intussusception, ileus, perforated peptic ulcer, diverticulitis (28), gynecological (17) and urological conditions (4), and miscellaneous diagnoses like acute cholecystitis, choledocholithiasis, liver metastases, and omental infarction a.o. (20). In 218 patients no definite diagnosis was establishedthese were considered patients with nonspecific abdominal pain (Table 2). During 22 of the 135 operations an innocent appendix was removed (16.2%). Nine patients revealed other surgical conditions requiring urgent surgery (acute cholecystitis 2, intussusceptions 2, perforated colonic diverticulitis 1, per-

364 Table 3. Performance parameter and indices of clinical evaluation, ultrasound and reclassification of 504 patients with suspected acute appendicitisa Accuracy Clinical evaluation Ultrasonography Reclassification Clinical evaluation Ultrasonography Reclassification 84.9 93.6 93.4 LR+ 9.55 25.01 21.95 Sensitivity 51.3 83.1 84.1 PTodds 2.76 7.23 6.33 Specificity 94.6 96.6 96.2 UI 0.235 0.664 0.674 PPV 73.4 87.8 86.4 Diagnostic errors 55/504 (10.9%) 19/504 (3.7%) 18/504 (3.5%) NPV 87.1 95.2 95.2

a PPV/NPV: positive/negative predictive value, LR+: likelihood ratio of a positive test, PTodds: post-test odds for acute appendicitis, UI: usefulness index

forated colonic adenocarcinoma 1, perforated Meckels diverticulum 1, aggressive retroperitoneal lipomatosis 1, infarction of the omentum 1). At the time of operation the disease had progressed to perforation in 23 patients (20.3%: perithyphlitic abscesses 3, perforating appendicitis with localized peritonitis 17, and generalized peritonitis 3). While only 14 of these patients (60.1%) had been allocated into group A, US made the diagnosis in 18 cases (78.3%) immediately. Following reclassification, the final diagnosis of an AA was true positive in 95/113 (84.1%) patients and true negative in 376/391 (96.2%) cases (Table 3). The overall accuracy was 93.4%. This resulted in a reduction of the rate of bad diagnostic errors from 10.9% to 3.5% (p < 0.001, 2). Without US the negative laparotomy rate would have been at least 21/141 (14.9%), whereas after reclassification it was calculated as 13/135 (9.6%, p < 0.01, 2). The clinicopathological performance indices are summarized in Table 3. US demonstrated superior sensitivity and specificity compared to clinical evaluation alone, resulting in an increase in the overall diagnostic accuracy from 84.9% for the provisional clinical decision reached to 93.6% for evaluation by US. Likewise the positive predictive value rose from 0.734 to 0.878. This improvement was also reflected in the likelihood ratios, as well as the utility index, both showing a sharp increase for US. However, the joint evaluation of clinical and US findings yielded the highest usefulness index (0.674). Discussion Negative laparotomy rates of 25% on average make acute appendicitis one of those diseases that suffers from one of the highest degrees of surgical misdiagnosis [4, 5, 11, 12, 17]. Delay in diagnosis and treatment may result in a significant increase in appendiceal rupture and subsequent morbidity and mortality [1, 7, 11, 12, 15]. US is a newly used imaging modality in AA and results from experts are impressive [2, 69, 17, 19]. Currently, it is entirely unclear whether US should be included in the surgical workup of these patients [7, 11, 13, 14]. We therefore evaluated the performance of surgeons routinely using US for patients with a possible diagnosis of AA. This prospective trial establishes that US can effectively be used in the surgical workup of these patients. The clinical diagnosis of an acute appendicitis in this study carried an overall accuracy of 85%, which lies well within the range of other prospective trials [1, 5, 9, 11, 15]. However, using US, the overall accuracy for the diagnosis

of AA rose to 93.6%. The ultrasonographically trained surgical residents achieved an overall sensitivity of 83.1% and specificity of 96.6%. These results are quite similar to those of previous investigations by known experts of ultrasonography, reporting of a sensitivity ranging between 75 and 89% and a specificity of 94 to 100% [9, 14, 1619], concurring with other reports demonstrating the feasibility of a round-the-clock US service by surgeons trained in ultrasonography [9, 14, 18]. From this study, it appeared that patients presenting with equivocal signs of AA would benefit most from ultrasonography. The majority of patients contributing to the impressive increase in diagnostic performance comprised patients of the clinically equivocal group B, including several patients with perforated disease. Patients of this group carried a 22.3% risk of an AA and US yielded a 71% diagnostic gain with respect to an immediate diagnosis. Patients clinically highly unlikely for AA (group C) had a 5.5% chance of AA and US made the diagnosis in 54% of these cases. In the remaining patients US was negative but raised suspicion to a level that led to hospital admission. All of these patients would otherwise have been inadequately followed as outpatients. Moreover, several patients with clinically highly suspected appendicitis were not operated upon because US was able to exclude AA as the cause of abdominal pain and establish a differential diagnosis not warranting surgery. Thus, use of US in patients with suspected AA resulted in a distinct improvement of clinical performance. Accordingly, we observed a significant reduction of bad diagnostic errors, as well as of negative laparotomies. The reduction of the rate of unnecessary laparotomies, which until now had only been reported by experts from nonsurgical departments, exemplifies the clinical importance of this method [9, 19]. The increase in performance was also reflected in the likelihood ratios, posttest odds, and usefulness index: US met the criteria for a good diagnostic test in all of the parameters, whereas clinical evaluation offered an acceptable positive likelihood ratio only. We therefore conclude that sonographically trained surgical residents can achieve results comparable to those of expert sonographers, that routine use of US in patients with a possible diagnosis of acute appendicitis improves the diagnostic performance, and that the use of US reduces the rate of unnecessary laparotomies. References
1. Berry J Jr, Malt RA (1984) Appendicitis near its centenary. Ann Surg 200: 567575

365 2. Borushok KF, Jeffrey RB, Laing FC, Townsend RR (1990) Sonographic diagnosis of perforation in patients with acute appendicitis. AJR Am J Roentgenol 154: 275278 3. Clarke JR, Hayward CZ (1990) A scientific basis for surgical reasoning I. Sensitivity, specificity, prevalence and predictive values. Theor Surg 5: 129132 4. Clarke JR, Hayward CZ (1990) A scientific basis for surgical reasoning II. Probability revisionodds ratios, likelihood ratios and Bayes theorem. Theor Surg 5: 206210 5. de Dombal FT (1992) Objective medical decision making in acute abdominal pain. Final Report. COMAC-BME, Project No: MR4*/034/ UK 6. Gaensler EHL, Jeffrey RB, Laing FC, Townsend RR (1989) Sonography in patients with suspected acute appendicitis. Value in establishing alternative diagnoses. AJR Am J Roentgenol 152: 4951 7. Hoffmann J, Rasmussen OO (1989) Aids in the diagnosis of acute appendicitis. Br J Surg 76: 774779 8. Jeffrey RB, Laing FC, Townsend RR (1988) Acute appendicitis: sonographic criteria based on 250 cases. Radiology 167: 327329 9. Jess P, Bierregaard B, Brynitz S, Holst-Christensen J, Kalaja E, LundKirstensen J. (1981) Acute appendicitis. Prospective trial concerning diagnostic accuracy and complications. Am J Surg 141: 232234 10. Levelle SM, Kanagaratoam B (1990) The information value of clinical data. Int J Biomed Comput 26: 203209 11. Lewis FR, Holcroft JW, Boey J, Dunphy JE (1975) Appendicitis. A critical review of diagnosis and treatment in 1000 cases. Arch Surg 110: 677684 12. Malt RA (1986) The perforated appendicitis. New Engl J Med 24: 15461547 13. Nauta RJ, Magnant C (1986) Observation versus operation for abdominal pain in the right lower quadrant. Am J Surg 151: 746748 14. Pearson RH (1988) Ultrasonography for diagnosing acute appendicitis. Br Med J 297: 309310 15. Pieper R, Kager L, Nu sman P (1982) Acute appendicitis. A clinical study of 1018 cases of emergency appendectomy. Acta Chir Scand 148: 5155 16. Puylaert JBCM (1986) Acute appendicitis: ultrasound evaluation using graded compression. Radiology 158: 355360 17. Puylaert JBCM, Rutgers PH, Lalisang RI, de Vries BC, van der Werft SDJ, Do rr JPJ, Block RAPR (1987) A prospective study of ultrasonography in the diagnosis of appendicitis. New Engl J Med 317: 666 669 18. Rossi P, Covarelli P, Mosci F, Biscacci R, Sensi B, Moggi L (1996) Ultrasonography in the management of acute appendicitis. Surg Endosc 10: 619621 19. Schwerk WB, Wichtrup B, Ru schoff J, Rothmund M (1990) Acute and perforated appendicitis: current experience with ultrasound-aided diagnosis. World J Surg 14: 271276

News and notices

Surg Endosc (1997) 11: 406408

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

New Address for the European Association for Endoscopic Surgery (E.A.E.S.)
Effective January 1, 1997, the new correspondence, telephone, and fax numbers of the E.A.E.S. office are: E.A.E.S. Office, c/o Mrs. Ria Palmen Luchthavenweg 81 Unit 1.42 5657 EA Eindhoven The Netherlands or: P.O. Box 335 5500 AH Veldhoven The Netherlands Tel: +31 40 2525288 Fax: +31 40 2523102

For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

Advanced Endoscopic Skills Surgical Skills Unit University of Dundee Scotland, UK

Each month Professor Cuschieri Surgical Skills Unit offers a 412 day course in Advanced Endoscopic Skills. The course is intensely practical with hands on experience on a range of simulated models. The program is designed for experienced endoscopic surgeons and covers advanced dissection techniques, extracorporeal knotting techniques, needle control, suturing, internal tying technique, stapling, and anastomotic technique. Individual workstations and a maximum course number of 10 participants allows for personal tuition. The unit offers an extensive collection of surgical videos and the latest books and publications on endoscopic surgery. In addition, participating surgeons will have the opportunity to see live advanced laparoscopic and/or thoracoscopic procedures conducted by Professor Cuschieri and his team. The course is endorsed by SAGES. Course fee including lunch and course materials is $1850. For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

Fellowship in Minimally Invasive Surgery George Washington Medical Center Washington, DC USA
A one-year fellowship is being offered at the George Washington University Medical Center. Interested candidates will be exposed to a broad range of endosurgical Education and Research Center. Active participation in clinical and basic science research projects is also encouraged. For further information, please contact: Carole Smith 202-994-8425 or, send curriculum vitae to: Dr. Jonathan M. Sackier Director of Endosurgical Education and Research George Washington University Medical Center Department of Surgery 2150 Pennsylvania Avenue, N.W. 6B-417 Washington, DC 20037, USA

Essentials of Laparoscopic Surgery Surgical Skills Unit University of Dundee Scotland, UK

Under the direction of Professor A. Cuschieri the Surgical Skills Unit is offering a three-day practical course designed for surgeons who wish to undertake the procedures such as laparoscopic cholecystectomy. This intensely practical program develops the necessary operating skills, emphasizes safe practice, and highlights the common pitfalls and difficulties encountered when starting out. Each workshop has a maximum of 18 participants who will learn both camera and instrument-manipulation skills in a purpose-built skills laboratory. During the course there is a live demonstration of a laparoscopic cholecystectomy. The unit has a large library of operative videos edited by Professor Cuschieri, and the latest books on endoscopic surgery are on display in our Resource area. Course fee including lunch and course materials is $860.

The Practical Aspects of Laparoscopic Fundoplication Surgical Skills Unit University of Dundee Scotland, UK
A three-day course, led by Professor Cuschieri, designed for experienced laparoscopists wishing to include fundoplication in their practice. The course covers the technical details of total and partial fundoplication using small group format and personal tuition on detailed simulated models. There will be an opportunity to observe one of these procedures live during the course. Maximum course number is six. Course fee including lunch is $1850. For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

407

Courses at George Washington University Endosurgical Educational and Research Center

George Washington University Endosurgical Educational and Research Center is proud to offer a wide range of surgical endoscopy courses. These courses include advanced laparoscopic skills such as Nissen fundoplication, colon resection, common bile duct exploration, suturing, as well as subspecialty courses. Individual surgeons needs can be met with private tuition. The Washington D.C. area is a marvelous destination to visit for recreational pursuits which can be arranged by the facility to suit your personal agenda. For further details please contact: Carole Smith: Department of Surgery 2150 Pennsylvania Avenue NW 6B Washington, DC 20037, USA Tel: (202) 994-8425

nical nuances, and troubleshooting; visual perception problems and solutions; magnified eye-hand coordination; and two-handed (ambidextrous) technique. Courses are offered year-round by individual arrangement. The MOET Institute is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide continuing medical education for physicians and designates these CME activities for 2040 credit hours in Category 1 of the Physicians Recognition Award of the American Medical Association. These programs are also endorsed by the Society of Gastrointestinal Endoscopic Surgeons (SAGES). For further information, please contact: Wanda Toy, Program Administrator Microsurgery & Operative Endoscopy Training (MOET) Institute 153 States Street San Francisco, CA 94114, USA Tel: (415) 626-3400 Fax: (415) 626-3444

Courses at the Royal Adelaide Centre for Endoscopic Surgery

Basic and Advanced Laparoscopic Skills Courses are conducted by the Royal Adelaide Centre for Endoscopic Surgery on a regular basis. The courses are limited to six places to maximize skill development and tuition. Basic courses are conducted over two days for trainees and surgeons seeking an introduction to laparoscopic cholecystectomy. Animal viscera in simulators is used to develop practical skills. Advanced courses are conducted over four days for surgeons already experienced in laparoscopic cholecystectomy who wish to undertake more advanced procedures. A wide range of procedures are included, although practical sessions can be tailored to one or two procedures at the participants request. Practical skills are developed using training simulators and anaesthetised pigs. Course fees: $A300 ($US225) for the basic course and $A1,600 ($US1,200) for the advanced course. For further details and brochure, please contact: Dr. D. I. Watson or Professor G. G. Jamieson The Royal Adelaide Centre for Endoscopic Surgery Department of Surgery Royal Adelaide Hospital Adelaide SA 5000 Australia Tel: +61 8 224 5516 Fax: +61 8 232 3471

IIIrd European Workshop on Digestive Surgery March 1314, 1997 Brussels, Belgium
The IIIrd European Workshop on Digestive Surgery is focused on live operative demonstrations performed and narrated by European experts. Panel discussions, quizzes, and selected communications will take place during the sessions. The aims are to confront alternative procedures and to provide young surgeons with an overview of selected fields. Topics will be: functional anorectal surgery, proctology, colorectal surgery, and updates in laparoscopic surgery. Course direction: J. J. Houben, MD. For further information, please contact: Administrative Secretariat Conference Services s.a. Avenue de lObservatoire 3, bte 17 B-1180 Brussels, Belgium Tel: +32 2 375 16 48 Fax: +32 2 375 32 99

Medicine and the Law for Junior Hospital Doctors April 11, 1997 Middlesbrough, UK
For further information, please contact: Miss Welsh ENTER North Riding Infirmary Newport Road Middlesbrough TS1 5JE UK

Advanced Laparoscopic Suturing and Surgical Skills Courses MOET Institute San Francisco, CA, USA
This intensive hands-on training program is intended to help the surgeon develop proficiency in the essential laparoscopic surgical techniques. A sequence of progressively challenging exercises has been designed to enable the surgeon to improve his or her laparoscopic dexterity, efficiency, and creativity. Exact and meticulous technique is emphasized so that the surgeon can apply these skills with confidence. Personal instruction is provided by Zoltan Szabo, Ph.D., F.I.C.S., Director of the MOET Institute, and surgeons are allowed to progress their own pace. Each participant has sole use of a laparoscopic training station equipped with high-quality clinical laparoscopic equipment and instrumentation. Inanimate, animal tissue, and optional live animal models are utilized. Features of these program include: fluently choreographed instrument movements; economy of movement and flawless technique; needle and suture handling skills (2-0 to 7-0); precision suturing, knotting, ligature, and anastomosis techniques; atraumatic, hemostatic tissue handling and dissection; optimal angles of approach (coaxial alignment of setup and geometry of port positioning); laparoscopic surgical strategy, tech-

First International Baltic Conference of Videosurgery of the Viscerosynthesis Section of the Association of Polish Surgeons April 2427, 1997 Gdansk, Poland
For further information, please contact: Organizing Secretariat Second Department of Surgery Medical University of Gdansk 1 Prof. Z Kieturakisa Street 80-742 Gdansk, Poland Tel/Fax: (0 048 58) 31 87 75

408

4th International Meeting on Laparoscopic Surgery May 17, 1997 Berne, Switzerland
Main topic: Acute appendicitis: Standard treatment and the role of laparoscopic surgery For further information, please contact: Mrs. Caroline Zrcher Klinik fr Viszerale und Transplantationschirurgie Universitt Bern Inselspital CH-3010 Bern, Switzerland Tel: +41 31 632 97 22 Fax: +41 31 632 97 23

SETUR Congress Department Cumhuriyet Cad. No. 107 80230 Elmadag Istanbul, Turkey Tel: (90.212) 23003 36 Fax: (90.212) 240 82 37

9th International Meeting Society for Minimally Invasive Therapy July 1416, 1997 Kyoto, Japan
Scientific program to include: Plenary, Parallel, Poster, and Video sessions. Host Chairman: Professor Osamu Yoshida, Department of Urology, Kyoto University, 54 Shogoin Kawahara-sho, Sakyo, Kyoto 606, Japan. Phone: +81 75 751-3328, Fax: +81 75 751-3740. This meeting coincides with the Gion Festival in Kyoto, one of the greatest festivals in Japan. For further information, please contact: Secretariat of SMIT 9th Annual International Meeting c/o Academic Conference Planning 383 Murakami-cho Fushimika, Kyoto 612 Japan Tel: +81 75 611-2008 Fax: +81 75 603-3816

European Course on Laparoscopic Surgery (French language) May 1316, 1997 (English language) November 1821, 1997 Brussels, Belgium
Course director: G.B. Cadiere For further information, please contact: Administrative Secretariat Conference Services s.a. Avenue de lObservatoire, 3 bte 17 B-1180 Bruxelles Tel: (32 2) 375 16 48 Fax: (32 2) 375 32 99

6th World Congress of Endoscopy Surgery Roma 98 6th International Congress of European Association for Endoscopic Surgery June 36, 1998 Rome, Italy
The program will include: the latest, original high quality research; symposia; plenary lectures; abstract presentations (video, oral, and posters); EAES and SAGES postgraduate courses, OMED postgraduate course on therapeutic endoscopy; working team reports; educational center and learning corner; meeting of the International Society of Nurses and Associates; original and non original scientific reports; and a world expo of new technology in surgery. For further information, please contact: Congress Secretariat: Studio EGA Viale Tiziano, 19 00196 Rome, Italy Tel: +39 6 322-1806 Fax: +39 6 324-0143

Joint Euro Asian Congress of Endoscopic Surgery 5th Annual Congress of the European Association for Endoscopic surgery (EAES) 3rd Asian-Pacific Congress of Endoscopic Surgery June 1721, 1997 Instanbul, Turkey
The Congress will include a joint postgraduate course EAES/SAGES/ ELSA on June 17th. For information and registration:

Surg Endosc (1997) 11: 405

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The author replies

My coauthors and I wish to thank you for furnishing us with a copy of the letter by Dr. Franklin and the opportunity to reply. Dr. Franklins major criticism of the data is a valid one; the procedures in the two treatment groups were not the same. However, this experimental design was not deliberate. We had hoped our volume of laparoscopic colon surgery and patient acceptance of the study would be high enough to successfully recruit two more-uniform groups of patients for prospective comparison. The funds to perform the study were granted with a time limitation, and therefore we reported those procedures which were able to be performed. No attempt was made to hide the fact that the procedures were not comparable. Emphasis of this weakness in the study is given in the Discussion. We also did not attempt to hide the fact that our operating times for laparoscopic surgery were lengthy. They were further increased by the analysis of the data in an intention-to-treat fashion, thereby including one patient converted from laparoscopic to open surgery in the laparoscopic group. As mentioned in the Discussion, analysis of the data with this patient in the Open group would have changed the conclusions of the study. Clearly we would have liked to have had more patients, more uniformity of procedures, and shorter operating times. We also would have liked to have included data on gastric emptying studies, which was excluded due to the high incidence of patients in the Open group refusing a solid-meal gastric-emptying study on the second postoperative day. Our conclusions are based on the data obtained and presented. We do not disagree with Dr. Franklins opinion that perhaps, with shorter operating times, the recovery of postoperative gastrointestinal function could be shorter after laparoscopic colectomy. Perhaps even just larger numbers of patients in each group would have shown a significant difference. However, the data in this study did not demonstrate that conclusion clearly. Studies by us in dogs have shown a greater difference in recovery, but again not vastly different recoveries. The Discussion of the paper emphasizes that scientific validation of a faster recovery from postoperative ileus in humans after laparoscopic vs open colon resection is still lacking. Dr. Franklin admits in his letter that his measurement of return of GI function in his and other large series of this procedure has not been done scientifically. We would urge him to do so, in order that the benefits of his surgical skills and short operating times already conferred to his patients could be documented as validating the beneficial use of a laparoscopic approach to colon resection for benign disease. B. Schirmer
Department of Surgery University of Virginia Health Sciences Center Charlottesville, VA 22908 USA

Original articles
Surg Endosc (1997) 11: 326330

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Laparoscopic pancreaticoduodenectomy in the porcine model

D. B. Jones,1 J. S. Wu,1,2 N. J. Soper1,2
1 2

Department of Surgery, Washington University School of Medicine, One Barnes Plaza, St. Louis, MO 63110, USA Institute for Minimally Invasive Surgery, Washington University School of Medicine, One Barnes Plaza, St. Louis, MO 63110, USA

Received: 22 April 1996/Accepted: 10 July 1996

Abstract Background: Minimally invasive techniques offer theoretical advantages for treating resectable periampullary neoplasms. Laparoscopic pancreaticoduodenectomy (LPD) was first reported in 1992 and has been performed clinically despite lack of animal data to support the operation. The purpose of this study was to develop LPD in an acute porcine model and to assess safety and efficacy before considering clinical trials. Methods: LPD was initiated in six domestic pigs under general anesthesia. Once pneumoperitoneum was created, five 10-mm access ports were placed (one central and two in each flank). After cholecystectomy, the duodenum was mobilized and the proximal jejunum was divided distal to the ligament of Treitz. The neck of the pancreas was separated from the superior mesenteric vein, and the midstomach was divided by a stapler. Pancreaticojejunostomy (PJ), choledochojejunostomy (CDJ), and gastrojejunostomy (GJ) were performed using interrupted sutures. The animals were immediately sacrificed and the operative site was examined. Results: LPD was aborted in three animals due to complications: intestinal perforation with fecal contamination (one) and prolonged resection time 2.5 h (two). LPD was completed in three animals (operative time ranged from 5.0 to 7.5 h, blood loss < 200 cc); however, at sacrifice one PJ and two CDJs had small posterior leaks. The efferent loop of the GJ was narrowed by the staple line in one pig. All animals had extensive ecchymosis of the jejunal serosa due to excessive manipulation. Conclusion: Despite a significant number of anastomotic leaks in the immediate postoperative period, laparoscopic pancreaticoduodenectomy is feasible in a porcine model. Further studies and technical development are necessary before laparoscopic pancreatic resection can be performed on a more widespread basis.

Key words: Laparoscopic surgery Pancreaticoduodenectomy Whipple procedure Choledochojejunostomy Gastrojejunostomy Pancreaticojejunostomy

In the United States pancreatic cancer is the fourth most common type of cancer, accounting for over 25,000 deaths each year [13]. Only a minority of patients are deemed candidates for a curative resection, while nearly 90% have unresectable lesions and a bleak prognosis. Advances in laparoscopic techniques for diagnosis, staging, and treatment of pancreatic cancer are only now beginning to be critically evaluated. The first successful laparoscopic pancreaticoduodenectomy (LPD) in a human was performed in 1992; the hope has been that by avoiding a laparotomy, patients would experience less pain and recuperate sooner [6]. However, because of the prolonged operative time, postoperative morbidity, and daunting technical aspects of LPD, few surgeons have adopted this new procedure. Although initial experience with laparoscopic surgery for cholecystectomy and, subsequently, other intraabdominal operations held the promise that technically advanced procedures, such as a pancreaticoduodenectomy, might be feasible and beneficial, animal studies were needed to confirm these clinical applications of laparoscopy. Most importantly, surgeons must prove that this approach will not compromise the traditional open operation. The aim of this study was to assess the feasibility of performing laparoscopic pancreaticoduodenectomy in an acute animal model. Materials and methods Operative technique

Presented in poster format at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), Philadelphia, Pennsylvania, 1417 March 1996 Correspondence to: N. J. Soper

Animals were prepared with a 24-h clear liquid diet and one bottle of magnesium citrate. Animals received a general anesthetic, including premeditation with intramuscular atropine 0.04 mg/kg, ketamine 20 mg/kg, and acepromazine 0.5 mg/kg. An endotracheal tube was inserted and connected to a ventilator with a tidal volume of 500 ml at a respiratory rate of

327

Fig. 1. Port placement for LPD in the porcine model. Five 1011 mm ports were placed (one central and two in each flank). One 18-mm port was used for introduction of an endoscopic stapler. Fig. 2. The free end of the jejunum was anastomosed to the distal pancreatic remnant. The cut end of the pancreas was invaginated into the jejunum and tacked with 3-0 absorbable sutures. Fig. 3. An end-to-side choledochojejunostomy was performed 10 cm distal to the PJ. The anastomosis was performed in a single layer using four 3-0 silk stay sutures and eight 4-0 absorbable interrupted sutures using intracorporeal suturing. Fig. 4. Gastrojejunostomy was performed 30 cm distal to the choledochojejunostomy. Fig. 5. Completed laparoscopic pancreaticoduodenectomy.

16 breaths/min with isofluorane titrated to effect. In the supine position, the abdomen was prepared with Betadine and draped in sterile fashion. An orogastric tube was placed for gastric decompression and a CO2 pneumoperitoneum was established with a Veress needle to a pressure of 15 mmHg. A 1011 mm trocar and sheath were inserted into the peritoneal cavity inferior to the umbilicus, through which a 10-mm video laparoscope was placed. The pig was secured in a 15 left lateral decubitus position on the operating table. Four additional 10-mm ports were placed with two in each flank. One 18-mm port was used for introduction of an endoscopic stapler (Fig. 1). The operating table was placed in 15 of axial rotation to the left with gravity displacing the spleen, stomach, and large bowel. A fan retractor through the superior right port was used to elevate the liver, and the exposure was maintained by stabilizing the retractor with a mechanical arm (Leonard Medical, Inc., Huntington Valley, PA). The common bile duct was isolated with a right-angle dissector and clipped distally with a large clip applier to allow subsequent dilation of the common bile duct (Ethicon-Endosurgery, Inc., Cincinnati, OH).

With the gallbladder retracted over the liver, the hepatocystic triangle was identified, and the cystic duct and cystic artery were isolated and clipped with an endoscopic multiple clip applier/divider LD320 (EthiconEndosurgery, Inc.). The gallbladder was dissected from the hepatic fossa using monopolar electrocautery and removed from the abdominal cavity. The anterior wall of the portal vein was freed from the posterior surface of the pancreas using blunt dissection. The midstomach was divided with a linear cutting stapler. The gastroduodenal artery was double clipped and divided. The common bile duct was then divided just above the cystic duct insertion. The jejunum was divided with a linear cutting stapler 15 cm distal to the ligament of Treitz and mobilized to reach the right upper quadrant without tension. Vessels emanating from the head of the pancreas were doubly clipped and divided. The inferior pancreaticoduodenal artery was doubly clipped and divided. The neck of the pancreas overlying the portal vein was divided with a linear stapler. The specimen (duodenum, bile duct, head of the pancreas) was bagged in a plastic pouch and placed in the pelvis for retrieval at the conclusion of the operation.

328 The proximal end of the jejunum was anastomosed to the cut end of the pancreatic remnant in an end-to-side fashion. This remnant was invaginated 12 cm into the jejunum and tacked with four 3-0 Vicryl (Ethicon, Inc., Somerville, NJ) sutures. Six additional sutures secured the seromuscular layer of jejunum to the capsule of the pancreas (Fig. 2). The loop of jejunum was brought up to the common bile duct without tension. The common bile duct was transected horizontally. A 0.5-cm incision was made along the antimesenteric aspect of the jejunum 10 cm distal to the pancreaticojejunostomy. Four 3-0 silk stay sutures (anterior and posterior centrally, two laterally) approximated the common bile duct and jejunum for suturing. Clips temporarily fastened suture strands together until the knots were tied. An end-to-side anastomosis was constructed in a single-layer using eight additional interrupted Vicryl sutures (Fig. 3). The four stay sutures were then tied. Gastrojejunostomy was performed 30 cm distal to the choledochojejunostomy. The anastomosis was performed using a single firing of a 60-mm linear endoscopic stapler (Fig. 4). The suture lines were inspected, and the common enterotomy was closed using another application of the stapler. After completion of the entire procedure (Fig. 5), the abdominal cavity was copiously lavaged with normal saline until the aspirate was clear. Table 1. Autopsy findings Anastomosis Choledochojejunostomy Pancreaticojejunostomy Gastrojejunostomy Problem Small posterior leak Small posterior leak Narrowing of efferent limb n (%) 2 (67%) 1 (33%) 1 (33%)

jejunostomy in one animal (33%) was significantly narrowed (Table 1). Discussion The application of laparoscopic techniques has revolutionized general surgery in the 1990s [16]. Randomized trials have confirmed the benefit of laparoscopic cholecystectomy [2], laparoscopic appendectomy [1, 5], and laparoscopic herniorrhaphy [18] compared to traditional open surgery. Enthusiasm for diagnostic laparoscopy has resulted in surgeons mastering staging laparoscopy with laparoscopic ultrasound exploration [8]. At many institutions, including our own, patients with cancer of the pancreas are routinely staged by laparoscopic examination and ultrasonography to assess resectability before considering a laparotomy for curative resection. At laparoscopy, 40% of patients will have peritoneal metastasis and are spared the morbidity of a laparotomy [19]. For the subset of patients with obstruction of the duodenum, laparoscopic gastrojejunostomy may be performed [12, 17]. Laparoscopic biliary bypass, performed with or without gastrojejunostomy, may provide long-term relief of biliary obstruction [4, 9, 14]. For those patients deemed resectable for cure, it is unknown whether laparoscopic pancreaticoduodenectomy might lessen morbidity and improve quality of life compared to its open counterpart. Potential problems with LPD are its technical difficulty and the appropriateness of its application as an oncologic resection. The current study attempted to address the former question. Recently, Gagner and Pomp have reported laparoscopic pancreaticoduodenectomy in two patients [7]. In the first case, a 30-year-old woman presented with chronic pancreatitis. The patient had previously undergone an open cholecystectomy and transduodenal sphincteroplasty for pancreas divisum and chronic pancreatitis localized in the pancreatic head. Due to intractable abdominal pain, laparoscopic pancreaticoduodenectomy was performed with an operative time of 10 h. Her postoperative course was notable for delayed gastric emptying secondary to presumed edema of the gastrojejunostomy. This resolved after bowel rest and intravenous hyperalimentation for 20 days. In the second case, a 72-year-old patient presented with painless jaundice and an ampullary adenocarcinoma <2 cm in size. The patient underwent an 8-h laparoscopic pylorus-preserving pancreaticoduodenectomy. The pathological specimen confirmed the diagnosis with adequate resection margins and negative lymph node involvement. Postoperatively, the patient developed a minor pancreatic leak that was managed conservatively with drainage, bowel rest, somatostatin, and antibiotics. The patient was discharged after 2 months for social reasons. There have been no signs of recurrence during 9 months of follow-up. Our animal model, limited in its attempt to mimic the biliary dilatation seen clinically in patients with periampul-

Animal experience
All operations and studies were conducted in accordance with the guidelines established by the Committee for Humane Care of Laboratory Animals at Washington University. The technique of laparoscopic pancreaticoduodenectomy was developed in six animals with immediate sacrifice and examination of the operative field after completing the pancreaticojejunostomy, choledochojejunostomy, and gastrojejunostomy. At the conclusion, a laparotomy was performed to examine the peritoneal cavity for evidence of hemorrhage or anastomotic leaks.

Data analysis
Operative time, intraoperative complications, and postoperative findings at autopsy were reported. Summary values in the text are presented as mean SEM.

Results Successful laparoscopic pancreaticoduodenectomies were performed in three of the six animals. Laparoscopic pancreaticoduodenectomy was aborted in the remaining three animals due to prolonged dissection time (>2.5 h; n 2) and intestinal perforation with peritoneal spillage with inadequate visualization (n 1). These animals were excluded from further analyses. We aborted the operation in two pigs after the dissection took longer than 2.5 h because of the undue struggle secondary to technical difficulties and to the anatomy of the porcine model (see Discussion section). Adequate exposure by retracting the five lobes of the porcine liver was difficult and contributed to all of the aborted cases. The intestinal perforation was caused by a 10-mm Babcock clamp during retraction of the bowel. Of the three pancreaticoduodenectomies completed without intraoperative complications, the mean operative time was 6.5 h (5.0, 7.0, and 7.5 h, respectively). At autopsy, four of the nine anastomoses in these animals were unsatisfactory due to leakage or stenosis, all of which were unrecognized during the operation. Specifically, two choledochojejunostomies and one pancreaticojejunostomy had small posterior leaks (66% and 33%, respectively); gastro-

329

lary tumors, probably was not ideal. The absence of a tumor mass or periductal inflammation in the porcine model rendered such a resection less difficult than in actual human cases, whereas the lack of pancreatic induration rendered pancreatojejunostomy more challenging. Furthermore, porcine anatomy differs from the human anatomy in that there are five hepatic lobes of similar proportions, making adequate retraction and exposure difficult. The duodenal, pancreatic, and midcolonic anatomy also differ from the human. In addition, the common bile duct and the pancreatic duct drain into the duodenum through separate openings. Despite these differences, we believed that the model could still serve its purpose in our attempt to evaluate the feasibility of laparoscopic pancreaticoduodectomy specifically, the ability to create multiple watertight anastomoses of adequate caliber [3]. We aborted the operation in two pigs of our series after the dissection took longer than 2.5 h because of the undue struggle secondary to technical difficulties and to the anatomy of the porcine model. The primary surgeon (N.J.S.) has had many experiences with open pancreaticoduodenectomies in human patients prior to this attempt in pigs. Nevertheless, these two animals were the first two in our series of six pigs, and there indeed was a learning curve associated with this procedure. We believe that perhaps if we had more experience with the pig anatomy and if we had continued in our struggle with the dissection, the procedure could have been completed. Had these two cases been completed, this procedure in this model would be considered very feasible indeed. Our major obstacle, or limitation, was the technically difficult laparoscopic suturing required to perform multiple anastomoses. Anastomotic leaks, especially of the pancreaticojejunostomy, are not uncommon following open pancreaticoduodenectomy. We attempted to duplicate anastomotic principles of open surgery by eliminating tension across the anastomosis while achieving a watertight mucosa-to-mucosa apposition and to ensure adequate blood supply to both ends. Intracorporeal suturing techniques permitted delicate tissue handling and precise microsurgical knotting of the pancreaticojejunostomy (PJ) and choledochojejunostomy (CDJ) [11]. Gastrojejunostomy was rapidly completed using a linear stapler. At the time of completing LPD in three animals, the anastomoses appeared to be intact. However, at autopsy pancreatic and biliary leaks as well as narrowing of the gastrojejunostomy were discovered. Failure to adequately evaluate the posterior aspects of the anastomoses after their completion hindered the laparoscopic approach. While advanced laparoscopic procedures including pancreaticojejunostomy [6], choledochojejunostomy [9], and gastrojejunostomy [15, 17] can technically be performed individually similar to their open counterparts, the combined operation necessary for a LPD resulted in a significant rate of anastomotic leakage. This study showed that the main problems were the time-consuming effort to perform the three anastomoses with intracorporeal knotting and the difficulty in assuring that the anastomoses were patent and watertight. New stapling devices, biologic glues, and threedimensional video systems [10] may eventually facilitate performing these anastomoses more quickly and effectively.

We conclude that the porcine model is helpful for gaining laparoscopic expertise in gastric, pancreatic, and biliary surgery. While gastrojejunostomies and choledochojejunostomies have been shown to be feasible in animal models and in patient care, our animal data suggest that combining these procedures with pancreaticojejunostomies is also feasible. In the few cases actually completed, the greatest problem was the quality of the anastomosis. On initial inspection, this would seem to be a clear limitation of the procedure. However, the findings in the immediate postsurgical autopsy may or may not have been clinically significant if the model used had a longer postoperative course prior to autopsy. Small leaks after Whipple procedures are common and often not of major morbidity if adequately drained intraoperatively. Our findings support the conclusion of the technical feasibility of laparoscopic pancreaticoduodenectomy in humans reported only by one group so far [6]. Further studies and technical development are necessary before laparoscopic pancreatic resection can be performed on a more widespread basis.
Acknowledgment. The authors gratefully acknowledge the Washington University Institute for Minimally Invasive Surgery as funded by an educational grant from Ethicon-Endosurgery, Inc. Product support was donated by Olympus, Karl Storz, and Ethicon, Inc. The authors also thank Jerome Brewer, RN, Deanna L. Dunnegan, RN, Donna R. Luttmann, RN, and Thomas A. Meininger for their technical assistance.

References
1. Attwood SEA, Hill ADK, Murphy PG, Thornton J, Stephens RB (1992) A prospective randomized trial of laparoscopic versus open appendectomy. Surgery 112: 497501 2. Barkun JS, Barkun AN, Sampalis JS, Fried G, Taylor B, Wexler MJ, Goresky CA, Meakins JL and the McGill Gallstone Treatment Group (1992) Randomized controlled trial of laparoscopic vs. minicholecystectomy. Lancet 340: 116119 3. Cameron BH, ORegan PJ, Anderson DL (1994) A pig model for advanced laparoscopic biliary procedures. Surg Endosc 8: 14231424 4. Fletcher DR, Jones RM (1992) Laparoscopic cholecystojejunostomy as palliation for obstructive jaundice in inoperable carcinoma of pancreas. Surg Endosc 6: 147149 5. Frazee RC, Roberts JW, Symmonds RE, Snyder SK, Hendricks JC, Smith RW, Custer MD, Harrisan JB (1994) A prospective randomized trial comparing open versus laparoscopic appendectomy. Ann Surg 219: 725731 6. Gagner M (1994) Laparoscopic duodenopancreatectomy. In: Steichen FM, Welter R (eds) Minimally invasive surgery and new technology. Quality Medical Publishing, St Louis, pp 192199 7. Gagner M, Pomp A (1994) Laparoscopic pylorus preserving pancreaticoduodenectomy. Surg Endosc 8: 408410 8. John TG, Grieg JD, Crosbie JL, Miles WFA, Garden OJ (1994) Superior staging of liver tumors with laparoscopy and laparoscopic ultrasound. Ann Surg 220: 711719 9. Jones DB, Brewer JD, Meininger TA, Soper NJ (1995) Sutured or fibrin-glued laparoscopic choledochojejunostomy. Surg Endosc 9: 10201027 10. Jones DB, Brewer JD, Soper NJ (1995) Next-generation 3D videosystems may improve laparoscopic task performance. In: Morgan K (ed) Interactive technology and the new paradigm for healthcare. IOS Press, Amsterdam, pp 152160 11. Jones DB, Soper NJ (1996) Laparoscopic suturing and knot-tying technique. In: Ponsky JL, McFadyen BV (eds) Operative thoracoscopy and laparoscopy. Lippincott-Raven, Philadelphia, 1996, p. 125143

330 12. Mouiel J, Katkhouda N, Whites S, Dumas R (1992) Endolaparoscopic palliation of pancreatic cancer. Surg Laparosc Endosc 2: 241243 13. Parker SL, Tong T, Bolden S, et al. (1996) Cancer statistics, 1996. CA Cancer J Clin 46: 527 14. Schob O, Schlumpf R, Kunz M, Uhlschmid GK, Largiader F (1993) Technique of laparoscopic cholecystojejunostomy with a Roux-en-Y loop. Surg Laparosc Endosc 3 (5): 386390 15. Soper NJ, Brunt LM, Brewer JD, Meininger TA (1994) Laparoscopic Billroth II gastrectomy in the canine model. Surg Endosc 8: 13951398 16. Soper NJ, Brunt LM, Kerble K (1994) Laparoscopic general surgery. N Engl J Med 330: 409419 17. Sosa JL, Zaleski M, Puenti I (1994) Laparoscopic gastrojejunostomy technique: case report. J Laparoendosc Surg 4: 215220 18. Stoker DL, Spiegelhalter DJ, Singh R, Wellwood JM (1994) Laparoscopic versus open inguinal hernia repair: randomized prospective trial. Lancet 343: 12431245 19. Warshaw AL, Fernandez Del Castillo C (1992) Pancreatic carcinoma. N Engl J Med 326: 455465

Letters to the editor

Surg Endosc (1997) 11: 402

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

False aneurysm of a hepatic artery branch

A recent issue of Surgical Endoscopy contains an article describing the postoperative development of a false aneurysm of a hepatic artery branch after laparoscopic cholecystectomy [1]. In the text, the authors indicate, As far as we could ascertain, however, this type of complication has not been described before. An addendum to the references includes similar reports by Lennard et al. and Bergey et al. Without getting into the posture of who reported what and when, I would like to suggest that the authors of the present article review a missed publication in the very same journal from 1994 (copy enclosed). My co-authors, Dr. Genyk and Dr. Keller, and I congratulate Dr. Porte and his colleagues for the nice contribution to what seems to be a slowly increasing body of literature related to hepatic artery injuries during laparoscopic cholecystectomy. Reference
1. Porte RJ, Coerkamp EG, Koumans RKJ (1996) False aneurysm of a hepatic artery branch and a recurrent subphrenic abscess. Surg Endosc 10: 6163

N. B. Halpern
The University of Alabama at Birmingham 417 Kracke Building 1922 Seventh Avenue South Birmingham, AL 35294-0007 USA

Surg Endosc (1997) 11: 387389

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Laparoscopic repair of a colonoscopic perforation

M. A. Velez,1,2 D. S. Riff,3 J. M. Mule4
1 2

Department of Surgery, Martin Luther Hospital, 1830 W. Romneya Drive, Anaheim, CA 92801 USA Harbor-UCLA Medical Center, Torrance, CA 90509 USA 3 Gastrointestinal Laboratory and Endoscopy Unit, Anaheim Memorial Hospital, Anaheim, CA 92801 USA 4 Anaheim Memorial Hospital, Anaheim, CA 92801 USA Received: 15 September 1995/Accepted: 16 January 1996

Abstract. We report a case of laparoscopic repair of a diagnostic colonoscopic perforation. No other such reports were noted in the literature. The management of colonoscopic perforations has become controversial. Operative vs nonoperative treatment is continually debated. The morbidity of operative management is significant. Colostomy is often performed. Laparoscopy should allow early evaluation of operative patients and primary repair of those with minimal contamination and no residual pathology. The benefits of minimally invasive surgery, such as shortened hospitalization and rapid return to full activities, including work, were realized in our patient. Laparoscopy should be considered in the selective management of colonoscopic perforations. Key words: Laparoscopy Colonoscopy Perforation

2-0 silk running Lembert suture. A 10-mm JP drain was left in the pelvis after the abdomen was copiously irrigated with saline. Cultures were taken prior to the repair. His drain and nasogastric tube were removed on postop day 2 (after return of bowel function), and a clear fluid diet was started. Claforan had been started preoperatively and was changed to p.o. Augmentin on day of discharge, postop day 3. His temperature and WBCs remained normal. His laparoscopy wounds (210 mm and 25 mm) healed well without infection. Within 2 weeks he was back to work with no restrictions. Eight months later he presented with gallstone pancreatitis. Laparoscopic cholecystectomy with operative cholangiogram was performed. No adhesions were seen in the abdomen or pelvis (Fig. 3).

Discussion Colonoscopy is a remarkably safe procedure with a low incidence of major complications such as bleeding and perforation [1, 3, 4, 6]. Bleeding can almost always be treated nonoperatively but perforations usually require exploration. Although some perforations are now selectively observed, those that require operative repair cause significant morbidity and carry a major risk of litigation against endoscopists [2]. The incidence of perforations has been reported as between 0.045 and 3% [3, 4]. Diagnostic colonoscopy carries a lower risk of perforation, ranging from 0.05 to 0.8%. Therapeutic colonoscopy carries a risk of perforation ranging from 0.073 to 3% [35, 7]. Perforation during diagnostic colonoscopy can be caused by mechanical or pneumatic pressures. Mechanical perforations can occur via excessive stretching by the shaft of the colonoscope during advancement of the instrument or by means of rotational motion with the side of the scope [1, 3, 4]. These perforations most frequently occur in the antimesenteric side of the sigmoid colon and often result in a large longitudinal tear into the peritoneal cavity. However, they may be localized or contained in the retroperitoneum. Perforation during therapeutic colonoscopy usually occurs during hot biopsy or polypectomy. These perforations are caused by thermal injury at the operative site and result in a smaller injury with less contamination than tears from diagnostic colonoscopy [4]. Conservative vs surgical management of colonoscopic

Case report
A 56-year-old man underwent diagnostic colonoscopy in October 1994 for evaluation of a change in bowel habits and tenesmus. The examination was performed without difficulty and found no abnormalities. Postprocedure he complained of diffuse abdominal pain with radiation to his shoulders. His abdomen was distended but soft with mild diffuse tenderness and no guarding or rebound. An X-ray revealed a large pneumoperitoneum. Because of the diffuse tenderness and radiation of his pain operative intervention was recommended. Laparoscopy was discussed and the patient consented to repair with possible colostomy. A large perforation, measuring approximately 5 cm, was found in the antimesenteric side of the distal sigmoid colon (Fig. 1). The laceration was repaired laparoscopically in two layers: inner layer with 2-0 chromic running suture (Fig. 2) and outer layer with

Correspondence to: M. A. Velez, 1801 W. Romneya Drive, Suite 202, Anaheim, CA 92801, USA

388

Fig. 1. Approximately 5-cm laceration in the antimesenteric border of the distal sigmoid colon.

Fig. 2. Mucosal and muscular layers approximated with running 2-0 chromic suture.

perforations is controversial. Kavin et al. [3] recommended conservative treatment for silent perforations and in patients with mild or localized symptoms and signs diagnosed within 48 h of injury. Perforations diagnosed late may also be managed conservatively if infection is confined. Antibiotics should be given to all patients as soon as a perforation is suspected. Lo and Beaton recommend operative treatment in perforations that occur during diagnostic colonoscopy because these perforations are generally large. Others [3, 57] recommend a selective approach dependent upon repeated reassessment of the patient. Patients with generalized peritonitis, worsening clinical signs, or failure to improve on conservative management should receive operative in-

tervention. Primary suture of the perforation is possible provided the colon has been cleansed, there is minimal contamination, and there is no residual pathology such as tumor or stricture [1, 3, 4, 6]. The morbidity associated with colonoscopic perforations is significant. Lo and Beaton [4] reported a retrospective review of 26,708 consecutive colonoscopic procedures. There were 12 perforations. All five perforations from diagnostic colonoscopy and one of seven perforations from therapeutic colonoscopy were treated operatively. Four of the five patients treated operatively recovered fully after an average hospitalization of 20.4 days (range 1242). The other patient died of sepsis on the 28th hospital day. All patients treated nonoperatively recovered without further complications and were discharged after an average hospitalization of 7.1 days (range 512). A colostomy was performed in five of the six patients treated operatively. Our patient underwent laparoscopic repair (primary) and was discharged after a hospitalization of 3 days. He was able to return to work with no restrictions within 2 weeks of his injury. Subsequent endoscopy revealed no adhesions. Depending on the operative findings and the skill and experience of the surgeon laparoscopic repair may be attempted. If endoscopic repair is possible the benefits associated with other minimally invasive procedures may be obtained. Because the wounds are small they are unlikely to become infected or herniate. Bowel function should return more rapidly. Reduced discomfort should allow early ambulation and decreased pulmonary and embolic complications. Early discharge from the hospital and rapid return to work without restrictions can be accomplished. Early intervention laparoscopically may also decrease the need for colostomy and reoperation for its takedown. Laparoscopy should be considered in the management of colonoscopic perforations.

389

Fig. 3. Repeat laparoscopy for gallstone pancreatitis showing area of previous repair in sigmoid colon. Note lack of adhesions in pelvis.

References
1. Baillie J (1994) Complications of endoscopy. Endoscopy 26: 185203 2. Gerstenberger PD, Plumeri PA (1993) Malpractice claims in gastrointestinal endoscopy: analysis of an insurance industry data base. Gastrointest Endosc 39(2): 132138 3. Kavin H, Sinicrope F, Esker AH (1992) Management of perforation of the colon at colonoscopy. Am J Gastroenterol 87(2): 161167

4. Lo AY, Beaton HL (1994) Selective management of colonoscopic perforations. J Am Coll Surg 179(3): 333337 5. Macrae FA, Tan KG, Williams CB (1983) Towards safer colonoscopy: a report on the complications of 5000 diagnostic or therapeutic colonoscopies. Gut 24: 7683 6. Thomson SR, Fraser M, Stupp C, Baker LW (1994) Iatrogenic and accidental colon injurieswhat to do? Dis Colon Rectum 37(5): 496 502 7. Waye JD, Lewis BS, Yessayan S (1992) Colonoscopy: a prospective report of complications. J Clin Gastroenterol 15(4): 347351

Surg Endosc (1997) 11: 376380

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Intracranial pressure
Effects of pneumoperitoneum in a large-animal model
R. J. Rosenthal,1 J. R. Hiatt,1,3 E. H. Phillips,1 W. Hewitt,1 A. A. Demetriou,1,3 M. Grode2
1

Division of General Surgery, Department of Surgery, Cedars-Sinai Research Institute, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles, CA 90048, USA 2 Division of Neurosurgery, Department of Surgery, Cedars-Sinai Research Institute, Cedar-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles, CA 90048, USA 3 Department of Surgery, UCLA School of Medicine, UCLA Medical Plaza, Los Angeles, CA 90024, USA Received: 19 March 1996/Accepted: 24 May 1996

Abstract Background: The effects of pneumoperitoneum on intracranial pressure (ICP) have received relatively little attention. This study was undertaken to investigate the changes in ICP occurring as a result of increased intraabdominal pressure (IAP) and positioning in animals with normal and elevated ICP. Method: Five pigs (average weight 60 lb) were studied. A subarachnoid screw was placed for ICP monitoring. End tidal CO2 was monitored. Ventilation was performed to keep PCO2 between 30 and 50 mmHg. Measurements of arterial blood gases, mean arterial blood pressure, and ICP were recorded at four different levels of intraabdominal pressure (IAP 0, 8, 16, and 24 mmHg), both in the supine and Trendelenburg positions. A Foley catheter was introduced into the subarachnoid space to elevate the intracranial pressure, and the same measurements were performed. Results: There was a significant and linear increase in ICP with increased IAP and Trendelenburg position. The combination of increased IAP of 16 mmHg and Trendelenburg position increased ICP 150% over control levels. Conclusions: Patient positioning and level of IAP should be taken into consideration when performing laparoscopy on patients with head trauma, cerebral aneurysms, and other conditions associated with increased ICP. Key words: Pneumoperitoneum Intracranial pressure Hemodynamic response Contraindications

physiologic effects of pneumoperitoneum. Studies from the 1950s and more recent research have examined the effects of pneumoperitoneum on stress response, stress hormones, and hemodynamic and cardiorespiratory parameters [13, 17, 19, 20, 28]. Relatively few investigations have analyzed effects on cerebrovascular circulation and intracranial pressure [8, 16]. Intracranial effects have immediate clinical relevance, as diagnostic laparoscopy has been advocated as an important modality in early management of trauma. Reports have shown the efficacy of the procedure in patients with penetrating thoracoabdominal trauma [23]. Diagnostic laparoscopy has also been used successfully in evaluating patients with blunt abdominal trauma, but it remains unclear whether this approach is safe in patients with possible brain injuries or other intracranial processes [16]. The aim of our study was to investigate changes in ICP occurring as a result of increased intraabdominal pressure (IAP) and positioning in the setting of normal and elevated ICP. We conducted these investigations under ICP conditions which would simulate current clinical indications for use of laparoscopic procedures: normal ICP, as would occur for elective laparoendoscopic procedures, and increased ICP, as would occur in patients with head injury. For the latter, baseline ICP was artificially raised using a standard model [30], and ICP values were recorded at various levels of intraabdominal pressure. Materials and methods

The performance of therapeutic laparoscopic procedures, enabled by prolonged carbon dioxide (CO2) pneumoperitoneum, has stimulated a renewal of interest in the pathoCorrespondence to: E. Phillips

All experimental procedures and protocols were reviewed and approved by the appropriate institutional committees. This study utilized five male and female farm pigs with an average weight of 60 lb. Each animal served as its own control. The experiments performed in each animal were divided in two phases, and in addition, each phase was conducted with animals in supine (A) and Trendelenburg (B) positions. During phase I, animals had normal baseline ICP values; in phase II, baseline ICP values were artifi-

377 Table 1. IAP and ICPa Position ICP/supine/no block ICP/Trend/no block ICP/supine/IC block ICP/Trend/IC block
a

IAP 0 16.0 6* 21.8 6* 30.2 14* 34.4 14*

IAP 8 17.8 7* 25.2 8* 33.8 17* 37.8 14*

IAP 16 21.0 6* 27.8 9* 35.6 16* 40.0 14*

IAP 24 24.6 8* 30.6 12* 38.6 16* 42.4 14*

IAP: intraabdominal pressure/ICP: intracranial pressure/Trend: Trendelenburg * Mean SD with p < 0.05: IAP 0 mmHg vs 8 mmHg, 16 mmHg and 24 mmHg

IB. The abdomen was deflated, and baseline measurements were made after placing the animal in a 20 Trendelenburg (head-down) position. Pneumoperitoneum was then restored to levels of 8, 16, and 24 mmHg of IAP, and all measurements were repeated.

Experimental procedurephase II: elevated ICP

IIA. By inflating the Foley catheter balloon with 2 ml of saline, the intracranial pressure was elevated, and baseline measurements of HR, ICP, MAP, and ABG were obtained with the animal placed in the supine position. The abdomen was then reinflated to 8, 16, and 24 mmHg of IAP, and all parameters described in phase IA were measured. Fig. 1. Changes in intracranial pressure related to increase in intraabdominal pressure and positioning. IIB. With the intracranial balloon inflated, the animal was placed into a 20 Trendelenburg position, and the same measurements as in phase IB were performed. Statistical analysis of the data was performed using a paired Students t-test. A p value of less than 0.05 was considered significant.

cially elevated. ICP, mean arterial pressure (MAP), heart rate (HR), and arterial blood gases (ABG) were measured at various time points during each phase. Measurements were carried out in the supine (phases IA and IIA) and Trendelenburg positions (phases IB and IIB) at 0, 8, 16, and 24 mmHg of IAP after creating pneumoperitoneum with CO2 insufflation. Animals were fasted overnight prior to surgery. Preoperative medication with sedatives/anesthetics included acepromazine (0.6 mg/kg IM), ketamine (20 mg/kg IM), and atropine (0.05 mg/kg IM). After anesthesia was induced, the airway was protected with an oral endotracheal tube, and animals were mechanically ventilated with a tidal volume of 500 ml/kg aimed at maintaining baseline PaCO2 at 3545 mmHg. Anesthesia was maintained with an isofluorane/O2 mixture (1.55%). For monitoring of ABGs and MAPs, the right femoral artery was dissected and cannulated with a disposable catheter, which was attached to 78304A monitor (Hewlett Packard, Boise, ID) with a disposable pressure transducer (Baxter, Melrose Park, IL). For continuous monitoring of ICP, a twist drill hole was made in the skull through a 1-cm incision in the right prefrontal area, and a Camino fiberoptic ICP transducer system was inserted intracerebrally and connected in the standard manner to a Camino V420 monitor (Camino Laboratories, San Diego, CA). An additional burr hole was placed in the left prefrontal area, where a Foley catheter (8 F) was introduced into the subdural space. The catheter was used to artificially increase the ICP in phases IB and IIB.

Results Effect of IAP on ICP There was an immediate, linear, and significant increase in ICP when the IAP was increased above 8 mmHg in both phases of the study (Table 1, Fig. 1). Trendelenburg positioning produced a further increase in ICP in both phases of the study at all levels of IAP. When the baseline ICP was raised (phase II), increasing the IAP above 8 mmHg produced the same pattern of (additional) ICP increase as seen in phase I.

Effect of IAP on PaCO2 Experimental procedurephase I: normal ICP There was an increase in PaCO2 in all phases of the study after pneumoperitoneum was induced, but the increase was not statistically significant (Table 2, Fig. 2). All animals were hyperventilated in order to maintain PaCO2 between 35 and 45 mmHg. This was done to neutralize the effects of the elevated blood CO2 level on the cerebrovascular system and ICP and to reproduce the clinical setting.

IA. The animal was placed in the supine position. A 1-cm incision was made in the navel area, and a 12-mm Veress needle (Karl Storz, Culver City, CA) was inserted into the abdominal cavity. Carbon dioxide was introduced by means of a laparoflator (Karl Storz Gmbh, Culver City, CA), and a pneumoperitoneum was created to a pressure of 15 mmHg. A 10 11-mm trocar (Endopath, Ethicon Endo Surgery, Cincinnati, OH) was inserted into the abdominal cavity, and the abdomen was explored using a 10-mm Hopkins optic attached to a microchip camera (Karl Storz, Culver City, CA). After all measuring sites were in place and calibrated, the abdomen was deflated, and the animal was kept in the supine position. Baseline measurements of ICP, MAP, HR, AIP, and ABG were taken. The abdomen was then insuflated to IAP levels of 8, 16, and 24 mmHg, and all measurements were repeated at each level.

Effect of IAP on MBP and HR After pneumoperitoneum was induced to a pressure of 8 mmHg, there was an increase in MBP and a decrease in HR.

378 Table 2. IAP and PaCO2a Position PaCO2/supine/no block PaCO2/Trend/no block PaCO2/supine/IC block PaCO2/Trend/IC block
a

IAP 0 24.6 9* 33.0 8* 31.46 6* 34.70 7*

IAP 8 27.1 6* 28.0 8* 35.4 6* 54.7 5*

IAP 16 29.1 8* 31.8 6* 36.5 9* 33.4 10*

IAP 24 30.4 7* 32.6 6* 37.6 6* 37.4 10*

IAP: Intraabdominal pressure/PaCO2: arterial pressure of CO2/Trend: Trendelenburg * Mean SD with p < 0.05: IAP 0 mmHg vs 8 mmHg, 16 mmHg and 24 mmHg

Fig. 2. Changes in PaCO2 related to increase in intra-abdominal pressure and positioning.

However, none of the changes was statistically significant or linear in pattern (Table 3, Figs. 3 and 4). Changes in all measured parameters (ICP, MAP, HR, PACO2) returned to baseline values after the abdominal cavity was deflated and the IAP was normalized (Table 4).

Discussion Elevations of IAP occur in diverse clinical settings including intestinal obstruction, ascites, intraperitoneal hemorrhage, tension pneumoperitoneum following a perforated viscus, and during laparoscopic procedures. In the present investigation, establishment of pneumoperitoneum produced a significant and immediate increase in ICP. Other investigators have reported similar observations in animals [11] and humans [14]. In 1994, Fuji et al. [8] reported that intraperitoneal CO2 insufflation caused an increase in cerebral blood flow in patients undergoing laparoscopic cholecystectomy and concluded that these changes were the result of an increase in arterial CO2 tension (PaCO2). In the same year, Josephs et al. [16] observed that establishment of pneumoperitoneum in an animal model produced a significant increase in intracranial pressure (ICP). We postulate that this ICP increase is mediated by two mechanisms: early (mechanical) and late (chemical). In the early, mechanical stage, establishment of elevated IAP compresses the large vessels of the abdomen, including the inferior vena cava, and produces an increase in central venous pressure (CVP) which reduces venous drainage from the CNS and lumbar plexus and increases the pressure in the CSF [21]. Luz et al. observed in dogs that an increase in IAP from 10 to 15 mmHg resulted in a significant elevation of the wedge pressure from 6 to 8 cmH2O without

PEEP [19]. The increase in ICP caused by an increase in CVP can also be explained by the Monroe-Kellie hypothesis. If one of the four compartments of the CNS (vascular, parenchymal, osseous, or CSF) expands rapidly, there is no time for the other compartments to buffer those changes and the ICP will rise. In Joseph et al.s study [16], the increase in ICP was measured only at an IAP of 15 mmHg of CO2. The effects of low-pressure or higher-pressure pneumoperitoneum were not described. Our study shows a significant increase in ICP at low (8 mmHg) and high (up to 24 mmHg) pressures of pneumoperitoneum. These findings suggest that a low-pressure pneumoperitoneum would not represent a safer alternative for laparoscopic diagnostic procedures in patients with suspected head trauma and intracranial hypertension. The late stage (chemical) is mediated by hypercapnia produced from the absorption of CO2 in the abdominal cavity. The CO2 which is not removed by ventilation results in an increase in PaCO2 and reflex vasodilatation of the CNS vasculature, which results in increased ICP [4, 11]. Levels of PaCO2 in our study were slightly increased after pneumoperitoneum was created, because all animals were hyperventilated to prevent a PaCO2 effect on the cerebrovascular circulation. However, there is sufficient evidence in the literature supporting a direct effect of elevated PaCO2 on ICP. In 1954 Westlake and Kaye investigated emphysematous patients with high PaCO2 and acute respiratory infections and found high CSF pressures [31]. It was noted that such patients may exhibit signs and symptoms of elevated ICP (headache, blurred vision, and papilledema), as described by Newton and Bone [22], Fuji et al. concluded that creation of CO2 pneumoperitoneum in patients undergoing laparoscopic cholecystectomy produces hypercapnia and reflex vasodilatation of the CNS with increased flow through the middle cerebral artery [8]. He observed statistically significant effects and showed a correlation between the increased PaCO2 and relative cerebral blood flow. The increased cerebral blood flow demonstrated by Fuji et al. reflects a correspondingly large increase in total cerebral blood flow which in turn results in an increase in ICP, as stated by Monroe-Kelly [10]. In contrast to the immediate increase in ICP seen after the IAP was increased in our own study, Fuji showed that cerebral blood flow increased 10 min after peritoneal insulation was begun and PaCO2 elevated. The blood flow returned to baseline values 10 min after peritoneal deflation and PaCO2 returned to baseline. These observations were also corroborated by Liu et al. [18] in a similar study. Another important aspect which may be related to the increase in ICP is the hemodynamic effect of pneumoperitoneum. Our study shows an increase in MBP and a de-

379 Table 3. IAP and HRa Position HR/supine/no block HR/Trend/no block HR/supine/IC block HR/Trend/IC block
a

IAP 0 125.4 32* 119.6 19* 102.4 9* 103.6 15*

IAP 8 123.6 23* 114.8 19* 104 10* 101.2 16*

IAP 16 119.4 19* 113.2 19* 100.6 14* 101.8 18*

IAP 24 120.4 20* 111.8 19* 103.2 15* 105.6 22*

IAP: Intraabdominal pressure/HR: Heart rate/Trend: Trendelenburg * Mean SD with p < 0.05: IAP 0 mmHg vs 8 mmHg, 16 mmHg and 24 mmHg

Fig. 3. Changes in heart rate related to increase in intra-abdominal pressure and positioning.

Fig. 4. Changes in mean blood pressure related to increase in intraabdominal pressure and positioning.

Table 4. IAP and MBPa Position MBP/supine/no block MBP/Trend/no block MBP/supine/IC block MBP/Trend/IC block
a

IAP 0 86 31* 105 20* 115.2 20* 117 15*

IAP 8 100.4 33* 98.2 25* 118.2 16* 118.6 9*

IAP 16 103.4 23* 115.8 38* 119 15* 118.2 6*

IAP 24 94 26* 112.2 26* 119 16* 123.2 8*

IAP: intraabdominal pressure/MBP: mean blood pressure/Trend: Trendelenburg * Mean SD with p < 0.05: IAP 0 mmg vs. 8 mmg, 16 mmhg and 24 mmg

crease in HR after pneumoperitoneum was induced, but the change was not statistically significant. Numerous studies showed significant effects [2, 12, 25, 26]. Melville et al. demonstrated that pneumoperitoneum induces release of catecholamines and vasopressin [20]. In his study, it was hypothesized that the increased IAP stimulated intraabdominal and intrathoracic receptors that in turn activate release of vasopressin [24]. It has long been recognized that the brain plays a critical role in regulating arterial pressure [15]. Cushing [5] first defined the quantitative nature of this effect, demonstrating that an acute elevation of intracranial pressure resulted in a elevation of the systemic blood pressure and a fall of the heart rate [3, 29]. Cushing also showed that this response was graded and occurred when the pressure within the cranium exceeded that of the systolic blood pressure. This vasomotor reaction is an ischemic response of the CNS to a rise in CSF pressure. In order to avoid this reaction, there is an increase in MAP higher than the CSF pressure keeping the arteries open and maintaining blood flow to the CNS [9]. Further studies by Ducker et al. [7] in chimpanzees showed that this response might be released not only by the isch-

emia but also by the increased ICP itself. Doba et al. and other authors reached similar conclusions [5, 6, 32]. In addition Drucker also showed that the increase in MAP was accompanied by an increase in CVP and TPR (total peripheral resistance). The increased TPR can also be explained by the Cushing reaction, because the latter is mainly a sympathetic stimulus. These observations were confirmed by Brown [1], who shows that there is an increase in venous tone following transient increases in ICP due to sympathetic discharge and release of catecholamines. All of these observations indicate that the increased CVP, MAP, and TPR found during laparoscopic procedures are due not only to a mechanical effect of the IAP on the large vessels of the abdomen but also due to an additional complex neurohormonal response which may be mediated by the CNS and the response to increased ICP. This emphasizes the need for careful intraoperative monitoring and appropriate patient selection for laparoscopic procedures. Extra care should be taken in the case of high-risk patients [13], which would include those with space-occupying intracranial masses such as aneurysms and tumors, and pulmonary processes which impair CO2 ex-

380

change, such as emphysema. Since we have shown that pneumoperitoneum produces an immediate and significant increase in ICP, head injury must be considered a relative contraindication to laparoscopy under pneumoperitoneum. Other methods of exposure such as abdominal wall lift devices [27] may be used, or ICP monitoring should be considered. References
1. Brown FK (1956) Cardiovascular effects of acutely raised intracranial pressure. Am J Physiol 185: 510514 2. Carroll BJ, Chandra M, Phillips EH, Margulies DR (1993) Laparoscopic cholecystectomy in critically ill cardiac patients. Am Surg 59: 783785 3. Cooper A (1824) Lectures on the principles and practice of surgery with additional notes and cases by Frederick Tyrell. T & G Underwood, London, 1: 282313 4. Cullen LK, Steffey EP, Bailey CS, Kortz G, Da Silva Curiel J, Belhorn RW, Woliner MJ, Elliot AR, Jarvis KA (1990) Effect of high Paco2 and time on cerebrospinal fluid and intraocular pressure in halothane anesthetized horses. Am J Vet Res 51(2): 300304 5. Cushing H (1902) Some experimental and clinical observations concerning states of increased intracranial tension. Am J Med Sci 124: 375400 6. Doba N, Reis DJ (1972) Localization within the lower brainstem of a receptive area mediating the pressor response to increased intracranial pressure (the Cushing response). Brain Res 47: 487491 7. Ducker TB, Simmons RL, Anderson RW, Kempe LG (1968) Hemodynamic cardiovascular response to raised intracranial pressure. Med Annal District Columbia 37(10): 523526 8. Fujii Y, Tanaka H, Tsurukoa S, Toyooka H, Amaha K (1994) Middle cerebral arterial blood flow velocity increases during laparoscopic cholecystectomy. Anesth Analg 78: 8083 9. Guyton AC (1991) Nervous regulation of the circulation and rapid control of arterial pressure. In: Guyton AC (ed) Phisiology. 8th ed. WB Saunders, Philadelphia, Chapter 18, pp 202203 10. Hansen N, Stonestreet B, Rosenkrantz T (1983) Validity of Doppler measurements of anterior cerebral artery blood flow velocity; correlation with cerebral blood flow in piglets. Pediatrics 72: 526531 11. Hargreaves DM (1990) Hypercapnia and raised cerebrospinal fluid pressure. Anesthesia 45(12): 712 12. Hashimoto S, Hashikura Y, Munakata Y, Kawasaki S, Makuuchi M, Hayashi K, Yanagisawa K, Numata M (1993) Changes in the cardiovascular and respiratory systems during laparoscopic cholecystectomy. J Laparoendosc Surg 3(6): 535539 13. Ho HS, Gunther RA, Wolfe BM (1992) Intraperitoneal carbon dioxide insufflation and cardiopulmonary functions. Arch Surg 127: 928933 14. Irgau I, Koyfman Y, Tikelis JI (1995) Elective intraoperative intracranial pressure monitoring during laparoscopic cholecystectomy. Arch Surg 130: 10111013

15. Ivankovich AD, Miletich DJ, Albrecht DR, Heyman HJ, Bonnet RF (1975) Cardiovascular effects of intraperitoneal insufflation with carbon dioxide and nitrous oxide in the dog. Anesthesiology 42: 281287 16. Josephs LG, Este McDonald JR, Birkett DH, Hirsch EF (1993) Diagnostic laparoscopy increases intracranial pressure. J Trauma 36(6): 815819 17. Le Roith D, Bark H, Nyksa M, Glick SM (1982) The effect of abdominal pressure on plasma antidiuretic hormone levels in the dog. J Surg Res 32: 6569 18. Liu SY, Leighton T, Davis I (1991) Prospective analysis of cardiopulmonary responses to laparoscopic cholecystectomy. J Laparoendosc Surg 1: 241246 19. Luz CM, Polarz H, Bohrer H, Hundt G, Dorsam J, Martin E (1994) Hemodynamic and respiratory effects of pneumoperitoneum and PEEP during laparoscopic pelvic lymphadenectomy in dogs. Surg Endosc 8: 2527 20. Melville RJ, Frizis HI, Forsling ML, Le Quesne LP (1985) The stimulus of vasopressin release during laparoscopy. Surg Gynecol Obstet 161: 253256 21. Milhorat TH (1975) The third circulation revisited. J Neurosurg 42: 628645 22. Newton DAG, Bone I (1979) Papilloedema and optic atrophy in chronic hypercapnia. Br J Dis Chest 73: 399404 23. Ortega AE, Tang E, Froes ET, Asensio JA, Katkhouda N, Demetriades D (1996) Laparoscopic evaluation of penetrating thoracoabdominal traumatic injuries. Surg Endosc 10: 1922 24. Punnonen R, Vinamaki O (1982) Vasopressin release during laparoscopy. Lancet 1: 175176 25. Reis DJ, Ruggiero DA, Morrison SF (1989) The C1 of the rostral ventrolateral medulla oblongata. A critical brainstem region for control of resting and reflex integration of arterial pressure. Am J Hypertens 2(12): 363374 26. Smith I, Benzie RJ, Gordon NLM, Kelman GR, Swapp GH (1971) Cardiovascular effects of peritoneal insufflation of carbon dioxide for laparoscopy. Br Med J 14: 410411 27. Smith RS, WR Fry, RH Koehler (1993) Gasless laparoscopy in the evaluation of trauma. In: Smith RS, Organ CH (eds) Gasless laparoscopy with conventional instruments. The next phase in minimally invasive surgery. Norman, San Francisco, 4: 95113 28. Soliz Herruzo JA, Castellano G, Larrodera L, Morillas JD, Moreno Sanchez D, Provencio R, Munoz-Yague MT (1989) Plasma arginine vasopressin concentration during laparoscopy. Hepatogastroenterology 36: 499503 29. Spencer W, Horsely V (1892) On the changes produced in the circulation and respiration by increase of the intracranial pressure or tension. Philos Trans 182: 201254 30. Sullivan HG, Miller JD, Becker DP (1977) The physiological basis of intracranial pressure change with progressive epidural brain compression. J Neurosurg 47: 532534 31. Westlake EK, Kaye M (1954) Raised intracranial pressure in emphysema. Br Med J 1: 302304 32. Yesnick L, Gelhorn E (1939) Studies on increased intracranial pressure and its effects during anoxia and hypoglycemia. Am J Physiol 128: 185194

Surg Endosc (1997) 11: 347350

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Early laparoscopic cholecystectomy for acute cholecystitis

S. M. Garber, J. Korman, J. M. Cosgrove, J. R. Cohen
Division of Minimally Invasive Surgery, Department of Surgery, Long Island Jewish Medical Center, New Hyde Park, NY 11042, USA Received: 28 March 1996/Accepted: 12 September 1996

Abstract Background: The timing of laparoscopic cholecystectomy for acute cholecystitis remains controversial. Methods: One hundred ninety-four patients with acute cholecystitis were reviewed. The conversion rates for the various number of days of symptoms before surgery were analyzed. The conversion rate dramatically increased from 3.6% for those patients with 4 days of symptoms to 26% for those patients with 5 days of symptoms. The mean number of days of symptoms prior to surgery in those patients who underwent successful laparoscopic cholecystectomy was 4.1 as compared to 8.0 in those patients who required open cholecystectomy (p < 0.0001). Based on this data the patients were divided into two groups. Group 1 consisted of 109 patients who underwent laparoscopic cholecystectomy within 4 days of onset of symptoms and group 2 consisted of 85 patients who underwent laparoscopic cholecystectomy after more than 4 days following onset of symptoms. Results: The conversion rate from laparoscopic to open cholecystectomy was 15%. The conversion rate for group 1 was 1.8% as compared to 31.7% for group 2 (p < 0.0001). Indications for conversion were inability to identify the anatomy secondary to inflammatory adhesions (68%), cholecystoduodenal fistula (18%), and bleeding (14%). The major complication rate for group 1 was 2.7% as compared to 13% for group 2 (p 0.007). The mortality rate for all patients with attempted laparoscopic cholecystectomy for acute cholecystitis was 1.5%. The average procedure time for group 1 was 100 37 min vs 120 55 min in group 2. The average number of postoperative hospital days in group 1 was 5.5 2.7 days as compared to 10.8 2.7 days in group 2. Conclusions: We advocate early laparoscopic cholecystectomy within 4 days of onset of symptoms to decrease major complications and conversion rates. This decreased conversion rate results in decreased length of procedure and hospital stay.

Key words: Acute cholecystitis Early laparoscopic cholecystectomy Complications

Laparoscopic cholecystectomy is the accepted treatment of symptomatic gallstones and chronic cholecystitis. The role of laparoscopic cholecystectomy in treating acute cholecystitis has also been well established [4, 7, 8, 12, 13, 15, 18, 20, 23, 26, 29, 30]. Acute inflammation can distort the biliary ductal anatomy and result in a higher incidence of major complications [3, 15]. The rate of conversion to open surgery is high during an episode of acute cholecystitis, and this high incidence of conversion decreases the economic advantages and early return to activity after laparoscopic cholecystectomy [4, 8, 13, 15, 23, 26, 30]. The timing of laparoscopic cholecystectomy for acute cholecystitis remains controversial. Open cholecystectomy for acute cholecystitis within 2 to 4 days of the onset of symptoms to decrease complications and hospital stay has been extensively reported in the literature [1, 5, 11, 25]. The performance of early vs delayed laparoscopic cholecystectomy has been shown to decrease morbidity, conversion rates, and hospital stay [17]. The results of early laparoscopic cholecystectomy for acute cholecystitis are not widely reported. This study was undertaken to determine the optimal time from the onset of symptoms to perform early laparoscopic cholecystectomy. Methods
This study reviewed 641 consecutive patients who underwent laparoscopic cholecystectomy at the Long Island Jewish Medical Center between January 1993 and December 1994. Two hundred twenty-four patients (35%) were diagnosed with acute cholecystitis. All patients were admitted on an emergency basis with a diagnosis of acute cholecystitis defined as: (1) positive Murphys sign, (2) fever and/or leukocytosis, (3) positive HIDA scan, and/or ultrasound findings consistent with acute cholecystitis (thickened gallbladder wall, gallstones, and pericholecystic fluid collection). Open cholecystectomy was performed in 30 patients secondary to large palpable gallbladders in septic patients and in patients with previous major upper abdominal operations. One hundred ninety-four patients who underwent attempted laparoscopic cholecystectomy during the initial hospital-

Presented at Society of American Gastrointestinal Endoscopic Surgeons (SAGES), Philadelphia, Pennsylvania, USA, 16 March 1996 Correspondence to: J. M. Cosgrove

348 Table 1. Demographic and laboratory data 4 days of symptoms (group 1) n 109 Age (years) Sex (female) Previous upper abdominal surgery Previous lower abdominal surgery WBC
a

>4 days of symptoms (group 2) n 85 54.8 18.8 52 (61%) 3 (3.5%) 30 (35%) 12 4.3

Statistical significancea NS NS NS NS NS

50.7 17.4 78 (72%) 2 (2%) 41 (38%) 11.7 4.4

Fig. 1. This graph illustrates the dramatic increase in conversion rates for those patients whose surgery occurs more than 4 days from the onset of symptoms.

NS not significant

Table 2. Histology ization for acute cholecystitis were reviewed. The timing of laparoscopic cholecystectomy was analyzed. The patients were divided into two groups based on analysis of the days of symptoms when conversion rates dramatically increase (Fig. 1). The change in conversion rates achieved statistical significance between 4 and 5 days of symptoms. Group 1 consisted of 109 patients who underwent laparoscopic cholecystectomy within 4 days of onset of symptoms (mean 2.7 1.0). Group 2 consisted of 85 patients who underwent laparoscopic cholecystectomy more than 4 days after the onset of symptoms (mean 6.9 2.7). The mean number of inpatient days prior to undergoing surgery was 1.9 1.4 in group 1 and 2.8 4.8 in group 2. The demographic data and laboratory results between the two groups were comparable (Table 1). The histological findings for group 1 and group 2 were similar (Table 2). Laparoscopic cholecystectomy was performed by 18 different surgeons. The majority (150 operations) were performed by eight of those surgeons. The distribution of surgeons between the two groups was similar. A Hasson (open) technique was used in 90 patients, based on surgeon preference or in patients with previous abdominal surgery, and a Verres (closed) needle technique was used in 104 patients. Intraoperative cholangiogram was performed in 45 patients (23 patients in group 1 and 22 patients in group 2) based on specific indications, including increased liver function enzymes, history of jaundice or gallstone pancreatitis, and unclear anatomy at the time of operation. The number of days of symptoms prior to surgery, conversion rates, reasons for conversion, complications, length of procedure, postoperative days, and total hospital stay were analyzed. All variables are expressed as mean standard deviation (SD) and compared by the unpaired t-test or chi-square test with Yates correction. A probability level of less than 0.05 was considered statistically significant. Group 1 Chronic cholecystitis Acute cholecystitis Acute suppurative cholecystitis Acute & chronic cholecystitis Acute gangrenous cholecystitis Acute necrotizing cholecystitis
a

Group 2 20 36 0 25 4 0

Statisticsa NS NS NS NS NS NS

25 43 4 28 6 3

NS not significant

Table 3. Results of group 1 vs group 2 Group 1 Conversion rates Major complication rates Procedure time (minutes) Length of stay (days)
a

Group 2 31.7% 13% 120 55 10.8 4.6

Statisticsa p < 0.0001 p 0.007 NS NS

1.8% 2.7% 100 37 5.5 2.7

NS not significant

Results Conversion rates The conversion rate from laparoscopic to open cholecystectomy for all patients with acute cholecystitis was 15% as compared to 5.1% for all patients who underwent elective laparoscopic cholecystectomy (p < 0.0001). The conversion rate for group 1 was 1.8% as compared to 31.7% for group 2 (p < 0.0001) (Table 3). Indications for conversion in patients with acute cholecystitis were inability to identify anatomy secondary to inflammatory adhesions (68%), cholecystoduodenal fistula (18%), and bleeding (14%).

(p < 0.05). The incidence of major complications for patients who underwent attempted laparoscopic cholecystectomy for acute cholecystitis was 7.2%. The major complication rate in group 1 was 2.7% as compared to 13% for group 2 (p 0.007) (Table 3). Major complications directly related to laparoscopic cholecystectomy for patients with acute cholecystitis included bleeding requiring transfusion or conversion to open cholecystectomy (3%) in one patient in group 1 and five patients in group 2, enterotomy (1%) in one patient in group 1 and one patient in group 2, collection in gallbladder fossa (1%) in two patients in group 2, and cystic duct leak (0.5%) in one patient in group 2. The mortality rate was 1.5% including one patient in group 1 and two patients in group 2. Bleeding was secondary to cystic artery bleeding in two patients and bleeding from the liver bed in the remaining four patients. Other complications included respiratory failure and pneumonia (4%), wound infection (3%), prolonged ileus (1%), pseudomembranous colitis (1%), deep vein thrombosis (0.5%), and atrial fibrillation (0.5%). Procedure time The average procedure time for group 1 was 100 37 min as compared to 120 55 min in group 2 (NS) (Table 3).

Complications The complication rate for patients with acute cholecystitis was 19.6% compared to 10.7% for elective cholecystectomy

349

Length of stay The average number of postoperative hospital days in group 1 was 3.5 3.5 as compared to 4.9 5.4 in group 2 (NS). The total length of stay for group 1 was 5.5 2.7 days as compared to 10.8 4.6 days for group 2 (NS) (Table 3).

Discussion Essenhigh was the first to advocate early operation for an acutely inflamed gallbladder, in 1966 [6]. Multiple prospective, randomized trials have shown that early cholecystectomy is the preferred treatment of acute cholecystitis [2, 14, 16, 19, 21, 28]. Early cholecystectomy is recommended in order to operate on patients during the edema phase of acute inflammation, as opposed to later on when hypervascularity, abscess formation, necrosis, and scarring occur. Sianesi et al. recommended open cholecystectomy within 72 h of the onset of symptoms because of a low incidence of positive bile cultures, a negligible percentage of postoperative complications and mortality, and a short hospitalization associated with lower costs [25]. Gardner et al. recommend open cholecystectomy for acute cholecystitis within 48 h of admission because it effectively reduces morbidity and the length of hospital stay [11]. Our study is consistent with the literature on early cholecystectomy for acute cholecystitis. The histological diagnosis of acute cholecystitis did not always correlate with the clinical diagnosis of acute cholecystitis. The diagnosis of acute cholecystitis is best based on clinical grounds, as done in this paper. Acute cholecystitis had been considered a contraindication to laparoscopic cholecystectomy in the past [24]. Since then, there have been many reports in the literature advocating laparoscopic cholecystectomy for acute cholecystitis [4, 7, 8, 12, 13, 15, 18, 20, 23, 26, 29, 30]. Postoperative recovery after laparoscopic cholecystectomy for an inflamed gallbladder has been shown to be quicker than that after open surgery, the faster recovery being attributed to the less invasive approach of laparoscopy [15]. The increased conversion rates in group 2 led to longer hospital stays. Although our data did not achieve statistical significance secondary to the high standard deviations there was a definite trend toward longer hospital stays in group 2 with a higher conversion rate. The complication rates for laparoscopic cholecystectomy have been reported to be comparable to those for open cholecystectomy [27]. The increased incidence of complication for acute cholecystitis has been shown in the past [15]. Our overall complication rate of 19.6% is comparable to a report by Zucker et al., who showed a complication rate of 16.8% [30]. A high incidence of bile leak (3%) has been shown, as compared to 0.5% in our study [4]. The major complication rate for those patients with 4 days of symptoms was significantly lower than for those patients who underwent laparoscopic cholecystectomy with greater than 4 days of symptoms. The most common major complication was bleeding. A conversion rate of 15% for laparoscopic cholecystectomy for acute cholecystitis vs 5% for elective laparoscopic

cholecystectomy is consistent with the finding that acute cholecystitis is a risk factor for conversion [10]. Conversion rates in the literature range from 6.5% to 35% [4, 15, 20, 23, 27, 29, 30]. However, in some series the diagnosis of acute cholecystitis was not clearly elucidated. The most common reason for conversioninflammatory adhesions obscuring the anatomyis compatible with other studies on laparoscopic cholecystectomy for acute cholecystitis [18, 20, 22]. Performing surgery within 4 days of onset of symptoms results in less inflammatory adhesions and well-defined edema planesand, thus, in safer dissection. Early laparoscopic cholecystectomy has been advocated during initial hospitalization [17]. Twenty-seven percent of patients who have their cholecystectomies delayed are readmitted for emergency cholecystectomy prior to their scheduled elective procedure [9]. Another 20% of patients fail conservative treatment during the initial hospitalization for acute cholecystitis and require surgery [9]. Hawasli advocated early laparoscopic cholecystectomy within 2448 h of symptoms secondary to the ease of removal during the edema phase of the acute inflammation as opposed to the difficulty in removing the severely scarred gallbladder after waiting 46 weeks of conservative treatment [12]. Delayed surgery resulted in longer operative time and increased morbidity. Cox et al. studied 98 patients with acute cholecystitis and advocated laparoscopic cholecystectomy within 72 h of admission [4]. Using the interval between admission and surgery as a measure is not completely adequate in addressing the severity of acute cholecystitis. Many patients do not present to the hospital at the onset of an attack of acute cholecystitis; often there is a delay. Our study looked at the number of days of symptoms prior to surgery, which we consider a more accurate reflection of the onset of inflammation. A large number of patients in group 2 underwent delayed surgery secondary to presenting to the hospital late during an attack of acute cholecystitis. The policy at our institution is to attempt laparoscopic cholecystectomy as soon as possible after the patient presents to the hospital with an attack of acute cholecystitis. This study supports performing early laparoscopic cholecystectomy in patients with acute cholecystitis as soon as the diagnosis is made. The decision as to whether to perform early vs late laparoscopic cholecystectomy many weeks after an acute attack of cholecystitis for those patients that present late to the hospital after 4 days of symptoms was not reviewed in this study. Data in the literature has shown that delayed or interval cholecystectomy leads to increased morbidity, conversion rates, and hospital stay [17]. Conclusion We advocate early laparoscopic cholecystectomy within 4 days of onset of symptoms to decrease major complications and conversion rates. The decreased conversion rate results in decreased length of procedure and hospital stay. References
1. Addison NV, Finan PJ (1988) Urgent and early cholecystectomy for acute gallbladder disease. Br J Surg 75: 141143

350 2. Alinder G, Herlin P, Lindgren B, Holmin T (1985) The costeffectiveness of early or delayed surgery in acute cholecystitis. World J Surg 9: 329334 3. Cameron JC, Galacz TR (1991) Laparoscopic cholecystectomy. Ann Surg 214: 251 4. Cox MR, Wilson TG, Luck AJ, Jeans PL, Padbury RT, Toouli J (1993) Laparoscopic cholecystectomy for acute inflammation of the gallbladder. Ann Surg 218: 630634 5. Edlund Y, Eldh J, Kock NG (1972) Acute cholecystitis. Timing in non-emergency cases. Acta Chir Scand 138: 176178 6. Essenhigh DM (1966) Management of acute cholecystitis. Br J Surg 53: 10321038 7. Fabre JM, Fagot H, Domergue J, Guillon F, Balmes M, Zaragosa C, Baumel H (1994) Laparoscopic cholecystectomy in complicated cholelithiasis. Surg Endosc 8: 11981201 8. Flowers JL, Bailey RW, Scovill WA, Zucker KA (1991) The Baltimore experience with laparoscopic management of acute cholecystitis. Am J Surg 161: 388392 9. Fowkes FGR, Gunn AA (1990) The management of acute cholecystitis and its cost. Br J Surg 67: 613617 10. Fried GM, Barkun JS, Sigman HH, Joseph L, Clas D, Garzon J, Hinchey EJ, Meakins JL (1994) Factors determining conversion to laparotomy in patients undergoing laparoscopic cholecystectomy. Am J Surg 167: 3539 11. Gardner B, Masur R, Fujimoto J (1973) Factors influencing the timing of cholecystectomy in acute cholecystitis. Am J Surg 125: 730733 12. Hawasli A (1994) Timing of laparoscopic cholecystectomy in acute cholecystitis. J Laparoendosc Surg 4: 916 13. Jacobs M, Verdeja JC, Goldstein HS (1991) Laparoscopic cholecystectomy in acute cholecystitis. J Laparoendosc Surg 1: 175177 14. Jarvinen HJ, Hastbacka J (1980) Early cholecystectomy for acute cholecystitis: a prospective randomized study. Ann Surg 191: 501505 15. Kum CK, Goh PMY, Isaac JR, Tekant Y, Ngoi SS (1994) Laparoscopic cholecystectomy for acute cholecystitis. Br J Surg 81: 1651 1654 16. Lahtinen J, Alhava EM, Aukee S (1978) Acute cholecystitis treated by early and delayed surgery: a controlled clinical trial. Scand J Gastroenterol 13: 673678 17. Lo CM, Liu CL, Lai ECS, Fan ST, Wong J (1996) Early versus delayed laparoscopic cholecystectomy for treatment of acute cholecystitis. Ann Surg 233: 3742 18. Lujan JA, Parilla P, Robles R, et al. (1995) Laparoscopic cholecystectomy in the treatment of acute cholecystitis. J Am Coll Surg 181: 7577 19. McArthur P, Cuschieri A, Sells RA, Shields R (1975) Controlled clinical trial comparing early with interval cholecystectomy for acute cholecystitis. Br J Surg 62: 850852 20. Miller RE, Kimmelstiel FM (1993) Laparoscopic cholecystectomy for acute cholecystitis. Surg Endosc 7: 296299 21. Norby S, Herlin P, Holmin T (1983) Early or delayed cholecystectomy in acute cholecystitis? A clinical trial. Br J Surg 70: 163165 22. Peters JH, Krailadsiri W, Incarbone R, Bremner CG, Froes E, Ireland AP (1994) Reasons for conversion from laparoscopic to open cholecystectomy in an urban teaching hospital. Am J Surg 168: 555558 23. Rattner DW, Ferguson C, Warshaw AL (1993) Factors associated with successful laparoscopic cholecystectomy for acute cholecystitis. Ann Surg 217: 233236 24. Schirmer BD, Edge SB, Dix J, Hyser MJ, Hanks JB, Jones RS (1991) Laparoscopic cholecystectomy. Treatment of choice for symptomatic cholelithiasis. Ann Surg 213: 665677 25. Sianesi M, Ghirarduzzi A, Percudani M, Dellanna B (1984) Cholecystectomy for acute cholecystitis: timing of operation, bacteriologic aspects, and postoperative course. Am J Surg 148: 609612 26. Singer JA, Mckeen RV (1994) Laparoscopic cholecystectomy for acute or gangrenous cholecystitis. Am Surg 60: 326328 27. Unger SW, Rosenbaum G, Unger HM, Edelman DS (1993) A comparison of laparoscopic and open treatment of acute cholecystitis. Surg Endosc 7: 408411 28. Vanderlinden W, Sunzel H (1970) Early versus delayed operation for acute cholecystitis: a controlled clinical trial. Am J Surg 120: 713 29. Wilson RG, Macintyre IMC, Nixon SJ, Saunders JH, Varma JS, King PM (1992) Laparoscopic cholecystectomy as a safe and effective treatment for severe acute cholecystitis. BMJ 305: 394396 30. Zucker KA, Flowers JL, Bailey RW, Graham SM, Buell J, Imbembo AL (1993) Laparoscopic management of acute cholecystitis. Am J Surg 165: 508514

Surg Endosc (1997) 11: 331335

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Laparoscopic colorectal surgery

Do we get faster?
F. Agachan, J. S. Joo, M. Sher, E. G. Weiss, J. J. Nogueras, S. D. Wexner
Department of Colorectal Surgery, Cleveland Clinic Florida, 3000 W. Cypress Creek Road, Ft. Lauderdale, FL 33309-1743, USA Received: 14 March 1996/Accepted: 31 July 1996

Abstract Background: A variety of parameters can affect the outcome of laparoscopic colorectal surgery. All consecutive laparoscopic colorectal procedures (LCP) were analyzed in an attempt to define an operative time curve for different categories of procedures. Additionally, impacts of case number and procedure type on length of procedure were assessed. Methods: Our computerized data system was reviewed for all patients who underwent LCP in a 4-year period. Parameters reviewed included age, sex, surgical indications, procedures performed, length of procedure, intraoperative and postoperative complications, incidence and causes for conversion, duration of postoperative ileus, and hospital stay. Results: Between August 1991 and December 1995, 175 patients with a mean age of 48.4 (range 1588) years underwent LCP. Patients were divided chronologically into five consecutive groups. Procedures were classified as either basic or complex. Complex procedures were those in which there was either a fixed tumor, an abscess or fistula, or extensive intraabdominal adhesions from prior surgery. Complex procedures performed each year ranged from 37% to 66%. As well, the percentage of patients with adhesions increased from 17% in 1991 to 29% in 1995. Despite increased difficulty, the intraoperative complication rate fell significantly from 29% in 1991 to 8% in 1995 (p < 0.005). Additionally, the operative length decreased from a mean of 201 min in 1991 to a mean of 141 min in 1995 (p < 0.05). Conclusion: The rapid improvement in these parameters may reflect both ascents in the learning curve and change in type of procedure. Adhesions, due to prior surgery or inflammation making dissection tedious, is the most important technical factor which effects operation time (p < 0.001). However, despite increased complexity, operating time de-

creased, reflecting improved skills. Thus, the experienced laparoscopic surgeon can increase the spectrum of applications with expectations of shorter operations and lower complication rates. Key words: Laparoscopy Colorectal surgery Operative time Learning curve

As laparoscopic technology has improved over the past 8 years, so too, have the indications for its application [4, 5, 6]. Early reports included initial limited experience and discussed feasibility of laparoscopic approaches to disease of the colon and rectum. Much controversy existed regarding postoperative pain, postoperative ileus, and length of hospitalization. Moreover, early criticism focused on the increased length of the procedure and increased complication rates. These problems may be due to lack of experience. Therefore, the aims of this study were to analyze the impact of experience and case type on operative time and morbidity.

Materials and methods

In a period of 52 months, all consecutive laparoscopic colorectal procedures (LCP) were prospectively evaluated. Data regarding patient demographics, operative course, and intraoperative and postoperative complications were prospectively recorded in a computerized database. Procedures performed in this series ranged from simple ones such as stoma creation to restorative proctocolectomy and creation of an ileoanal reservoir. Procedures were chronologically divided into five consecutive groups of 35 patients each for further evaluation. Group I represents the first 35 cases and group V the last 35 cases. Procedures were also categorized based on their complexity. Complex procedures were defined as those in which there was either a fixed tumor, an abscess or fistula, or extensive intraabdominal adhesions from prior surgery. The number of cases per surgeon was also assessed.

Presented at the 1996 Scientific Session of the 5th World Congress Endoscopic Surgery, Philadelphia, Pennsylvania, USA, 1317 March 1996 Correspondence to: S. D. Wexner

332 Table 1. Diagnosis Diagnosis Inflammatory bowel disease (n74) Crohns disease Mucosal ulcerative colitis Diverticulitis Neoplasia of colon and anorectum (n58) Colonic polyps Familial adenomatous polyposis Small bowel lymphoma Colonic carcinoma Rectal carcinoma Anal carcinoma Kaposis sarcoma Functional bowel disorders (n30) Fetal incontinence Constipation Rectal prolapse Other (n 13) Sigmoid volvulus Severe anal stenosis Irradiation proctitis Small-bowel laceration Ischemic colitis Chronic abdominal pain Colonic tuberculosis Rectal tuberculosis Total GI 1 9 1 1 4 1 7 3 1 3 2 2 35 GII 5 10 4 5 3 1 1 3 1 1 1 35 GIII 7 3 2 7 1 2 2 6 1 1 1 1 1 35 GIV 10 6 5 1 4 1 4 2 1 1 35 GV 10 6 6 2 1 1 3 3 1 1 1 35 Total 33 22 19 24 9 1 11 9 3 1 15 12 3 2 2 2 1 2 2 1 1 175

Definitions and criterion for surgery

All patients were assessed for the possible use of laparoscopic surgery, and patient suitability for the procedure was decided by the surgeon. Previously, operative techniques for all procedures were defined [3, 4, 8]. A laparoscopic-assisted procedure includes a laparoscopic phase in which the bowel is fully mobilized laparoscopically, after which an extracorporeal phase includes the resection and the anastomosis. The extracorporeal phase is performed through an incision generally 35 cm long. In a complete laparoscopic procedure all phases of mobilization, vascular control, and anastomosis are undertaken through ports rather than through an incision. A laparoscopic procedure was considered converted to an open procedure if any unplanned incision was made or if any planned incision was made either sooner than planned or larger than planned. Conversion was also considered to have occurred during an assisted segmental procedure if any incision larger than 5 cm was made except for Pfannesteils incision used during restorative proctocolectomy. Conversion may be due to complications or due to judgment of the surgeon.

Statistical analysis
Statistical analysis was performed using a one-way analysis of variance to reject the null hypothesis that all the mean operative times were equal for all chronological groups. The Student-Newman-Keuls multiple comparison procedure was used to isolate comparisons between chronological group means that were responsible for or contributed to the rejection of the null hypothesis. One-way analysis of variance, Students t-test for independent samples, the Pearson chi-square statistic, and correlation analysis were used to determine what other variables may have influenced operative time and whether such variables were independent of the chronological groups, as appropriate (Instat Graphpad Software Inc., 1993, San Diego, CA). The level of statistical significance was set at p 0.05.

Laparoscopic status of the performed procedures was summarized in Table 2. There were no statistically significant differences in mean age (p > 0.05) or sex ratio (p > 0.05) among the five chronological groups (Table 3). The mean operative time for the entire group was 161 min. A one-way analysis of variance demonstrated that all chronological group means and procedure means were not equal (p < 0.0001) (Table 4) Student-Newman-Keuls multiple comparison procedure revealed that the mean operative times of groups I and II (201, 179 min) were greater than the mean operative times of each of the other chronological groups (139, 145, 141) (p < 0.001) (Fig. 1) Among the remaining groups, the differences in mean operative times were not significant, as reflected by the flattening of the learning curve. Although converted procedures were found to be more time consuming than were laparoscopic or laparoscopic-assisted procedures (p < 0.05), laparoscopic procedures were completed earlier than laparoscopic assisted ones (Figs. 2 and 3). Complexity of the procedures was increased in groups IV and V (p < 0.05). Similarly, the density of adhesions in groups III, IV, and V were higher than in groups I and II (p < 0.001). Although complexity index and density of adhesions were increased, there was no difference for conversion rate among the groups (p > 0.05) (Table 4). Between 1991 and 1995 length of hospitalization (Fig. 4), length of postoperative ileus (Table 3), and overall morbidity rate consistently decreased (p < 0.05) (Fig. 5). Discussion Excluding a variety of parameters that can affect operating room time, such as anesthesia and setup time, expertise of the operating room team, type of procedure, and severity of the disease, operative time alone may be a useful parameter for the definition of the learning curve [13, 7].

Results Between August 1991 and December 1995, 175 LCPs were performed on patients of a mean age of 48.4 (range 1588) years. There were 102 females and 73 males. Indications for surgery included benign and malignant diseases (Table 1).

333 Table 2. Profile of procedures GI Right hemicolectomy Left hemicolectomy Sigmoid colectomy Diverting ileostomy Diverting colostomy Small-bowel resection Adhesiolysis Total abdominal colectomy with ileoanal reservoir Total abdominal colectomy with ileorectal anastomosis Total abdominal colectomy with end ileostomy Hartmanns creation or closure Abdominoperineal resection Rectopexy Low anterior resection Ileorectal anastomosis with ileostomy closure Total 6 4 2 1 14 3 2 1 2 35 GII 7 3 1 1 12 3 2 3 2 1 35 GIII 9 1 4 7 1 2 2 4 2 1 2 35 GIV 9 1 6 8 3 2 1 1 1 2 1 35 GV 12 2 7 7 1 3 1 1 1 35 Total 43 4 24 23 7 2 5 28 8 2 11 5 4 5 4 175

Table 3. Clinical characteristics GI Patients (n) Age (years) Gender (M/F) Surgeon Aa n Ba C D Intraoperative morbidity (n-%) Postoperative morbidity (n-%) Overall morbidity (n-%) Mortality (%) Clear diet (days) Bowel movements (days) Length of hospitalization (days)
a

GII 35 44.1 14/21 19 15 1 0 10/35 (29%) 11/35 (31%) 19/35 (54%) 0 3.1 4.1 7.9

GIII 35 51.4 14/21 0 35 0 0 1/35 (3%) 6/35 (17%) 7/35 (20%) 0 2.4 3.5 7

GIV 35 48.1 17/18 0 31 1 3 4/35 (11%) 7/35 (20%) 8/35 (23%) 0 2.3 3.3 6.9

GV 35 50.1 16/19 0 34 0 1 3/35 (8%) 2/35 (6%) 4/35 (11%) 0 1.8 3 5.9

Total 175 48.4 73/102 47 122 2 4 28/175 (16%) 38/175 (22%) 55/175 (31%) 0 2.6 3.6 7.3

35 48.2 12/23 28 7 0 0 10/35 (29%) 12/35 (34%) 17/35 (48%) 0 3.4 4.3 8.6

During periods I and II, surgeons A and B performed cases together

Table 4. Variables related to operating time Parameters Complexity (n-%) Intraoperative complications (n-%) Adhesions (n-%) Conversion rates (n-%) Mean length of operative time (min) GI 13/35 (37%) 10/35 (29%) 7/35 (20%) 8/35 (23%) 201 GII 20/35 (57%) 10/35 (29%) 6/35 (17%) 7/35 (20%) 179 GIII 14/35 (40%) 1/35 (3%) 9/35 (26%) 7/35 (20%) 139 GIV 23/35 (66%) 4/35 (11%) 11/35 (31%) 10/35 (29%) 145 GV 22/35 (63%) 3/35 (8%) 10/35 (29%) 7/35 (20%) 141 Total 92/175 (53%) 28/175 (16%) 43/175 (25%) 39/175 (22%) 161

The data for assessment of the learning curve for laparoscopic cholecystectomy are well established and have demonstrated that operative times can be diminished to levels comparable to those reported for open cholecystectomy. Similar data are only now beginning to become available for laparoscopic colorectal procedures. Recently, Senagore et al. [7] reported a three-surgeon series of 60 laparoscopicassisted colectomies divided into the first 20, second 20, and third 20. They reported similar mean operative time for the

first two groups, which decreased by approximately 25 min for the third group. Wishner et al. [9] reported a learning curve for six chronological groups of a total of 150 laparoscopic-assisted colectomies. They observed a significant decrease in mean operative time, from 250 min to 156 min over the first 3550 cases after which there was no further time reduction. The length of surgery plateaued at 140 min for the remaining groups. Interestingly, they could not find any significant differences for the conversion rates and

334

Fig. 1. Mean length of operating time. Fig. 4. Mean length of hospitalization.

Fig. 5. Comparison of morbidity rates and postoperative complications for each group.

Figs. 2 and 3. Length of operating time and number of procedures for laparoscopic status of each group.

length of hospital stay among the chronological groups although 76% of the procedures were segmental resections. The term conversion refers to the critical decision point during which the surgeon realizes that further operative progression utilizing laparoscopic maneuvering may be inappropriate for the reasons enumerated earlier in this manuscript. The data in this study have shown that if the decision to convert is made early in the procedure the overall operative time is not significantly increased. Thus conversion must be considered a safe alternative rather than failure. Fowler et al. [1] analyzed 60 laparoscopic colon resections. Although they did not report the learning curve with operative time results for all procedures, they noted that operating time for segmental resections had gradually decreased with experience. In our experience, operating time was directly related to

chronological group. Although complexity increased up to 66% and density of adhesions to 31%, intraoperative complications decreased down to 8% while conversion rates stabilized at 22% and mean operative time decreased to 141 min. This decrease in time was significant for the first two groups, suggesting that 70 cases were required before the learning curve flattened. At this point operative time leveled off at 140 min for the remainder of cases. The number of cases which constituted the learning curve was almost identical to that reported by Wishner et al., [9], and the actual length of operative time after the plateau was identical at 140 min. It is interesting that similar results were achieved whether the surgeons were eight general surgeons performing many types of laparoscopic procedures and some colorectal surgery or two colorectal surgeons performing exclusively colorectal surgery, some of it laparoscopically. The additional 20 cases required in our learning curve may again be attributable to the high number of total abdominal colectomies performed on patients with mucosal ulcerative colitis early in our experience [8]. Table 3 analyzes the number of cases done per surgeon within each group. During groups I and II, surgeons A and B worked together and performed 35 and 34 procedures in these two groups, respectively. Therefore, the majority of the learning curve was surmounted in tandem between these two surgeons working as a team. As can be seen, after group II, 95% of the laparoscopic procedures were performed by a single surgeon who had, as shown in Table 3 in the overall morbidity numbers, already surmounted the learning curve.

335

Conclusion The improvement in operative time is related to both experience and procedure type. Adhesions, due to prior surgery or inflammation making dissection tedious, is the most important technical factor which affects operative time. However, despite increased complexity, decreased operative time and morbidity rate reflect improved skills. Thus, the benefit from laparoscopic colorectal surgery in terms of operative time is best realized after 70 cases. References
1. Fowler DL, White SA, Anderson CA (1995) Laparoscopic colon resection: 60 cases. Surg Laparosc Endosc 6: 468471

2. Freys SM, Fuchs KH, Heimbucher J, Thiede A (1994) Laparoscopic adhesiolysis. Surg Endosc 8: 12021207 3. Hoffman GC, Baker JW, Fitchett CW, Vansant JH (1994) Laparoscopic-assisted colectomy, initial experience. Ann Surg 6: 732743 4. Monson JRT, Hill ADK, Darzi A (1995) Laparoscopic colonic surgery. Br J Surg 82: 150157 5. Nyhus LM (1992) Laparoscopic hernia repair: a point of view. Arch Surg 127: 137 6. Scott TR, Zucker KA, Bailey RW (1992) Laparoscopic cholecystectomy: a review of 12,397 patients. Surg Laparosc Endosc 2: 191198 7. Senagore AJ, Luchtefeld MA, Mackeigan JM (1995) What is the learning curve for laparoscopic colectomy? Am Surg 61: 681685 8. Wexner SD, Reissman P, Pfeifer J, Bernstein M, Geron N (1996) Laparoscopic colorectal surgery: analysis of 140 cases. Surg Endosc 10: 133136 9. Wishner JD, Baker JW, Hoffman GC, Hubbard II GW, Gould RJ, Wohlgemuth SD, Ruffin WK, Melick CF (1995) Laparoscopic-assisted colectomy. Surg Endosc 9: 11791183

Surg Endosc (1997) 11: 371375

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The role of endoscopic retrograde cholangiopancreatography in management of patients recovering from acute biliary pancreatitis in the laparoscopic era
G. Lin,1 A. Halevy,1 O. Girtler,1 R. Gold-Deutch,1 A. Zisman,1 E. Scapa2
1 2

Department of Surgery B, Assaf Harofeh Medical Center, Zerifin 70300, Affiliated to Sackler Faculty of Medicine, Tel Aviv University, Israel Institute of Gastroenterology, Assaf Harofeh Medical Center, Zerifin 70300, Affiliated to Sackler Faculty of Medicine, Tel Aviv University, Israel

Received: 9 February 1996/Accepted: 31 July 1996

Abstract Background: Traditionally an episode of acute biliary pancreatitis (ABP) is an indication for direct imaging of the biliary tree. The optimal approach may vary according to local expertise, and endoscopic retrograde cholangiopancreatography (ERCP) is the most common. The fact that the incidence of choledocholithiasis in patients recovering from ABP varies between 3 and 33% raises a question about the necessity of visualizing the biliary tree in all patients recovering from ABP. Methods: In order to evaluate this policy, we reviewed 48 ERCPs performed on patients recovering from ABP who were scheduled for laparoscopic cholecystectomy (LC). We checked the correlations between ERCP findings and the severity of pancreatitis, biochemistry values (which were sampled during the acute phase), and ultrasonographic (US) findings. Results: The ERCP demonstrated common bile duct (CBD) stones in 11 (22.9%) patients. US finding of a dilated CBD and maximal aspartate transaminase (AST) values higher than 90 units/l were significantly correlated with CBD stones (a relative risk [RR] of 2.95 with a 95% confidence interval [CI] for a dilated CBD and RR of 3.89 with a 95% CI of 1.1812.80 for an AST value higher than 90 units/l). No other parameters were significantly correlated with CBD stones. Conclusion: We, therefore, recommend performing a preoperative ERCP only on patients who present with an ultrasonographic finding of CBD dilatation. The correlation to high AST is still to be proven. Key words: Acute biliary pancreatitisEndoscopic retrograde cholangiopancreatographyCommon bile duct stonesCommon bile duct dilatation
Correspondence to: A. Halevy

Traditionally, an episode of acute biliary pancreatitis (ABP) is an indication for direct imaging of the biliary tree [2]. However, the approach to patients suspected of having common bile duct (CBD) stones has changed in the laparoscopic era [24]. As laparoscopic cholecystectomy (LC) has become the golden standard treatment for symptomatic cholelithiasis [3, 37, 39], most patients recovering from acute biliary pancreatitis (ABP) undergo LC [15, 29]. However, the optimal approach to the subgroup of patients suspected of having CBD stones remains uncertain [24] and may vary according to local expertise [8]; endoscopic retrograde cholangiopancreatography (ERCP) is the most common [5, 13, 19, 34, 44]. Some surgeons perform an ERCP after the operation on patients shown to have CBD stones by intraoperative cholangiography (IOC) [30, 43], but most surgeons advocate a preoperative ERCP, as failure of the endoscopist to remove CBD stones will influence further decisions [5]. The pathophysiology of ABP begins with an obstruction of the ampula of Vater by a biliary stone [12, 22]. This block is usually transient and the stone will find its way to the duodenum within a few hours or days [1, 22, 29]. Another point to consider is the incidence of choledocholithiasis in patients recovering from ABP, which varies between 3 and 33% [7, 27]. This creates speculation about the necessity of visualizing the biliary tree in all patients recovering from ABP. Some studies emphasize the significance of liver function tests (LFT) as predictors of choledocholithiasis [9, 11, 17, 19, 28], while others negate these findings [4, 21]. Dilatation of the CBD as demonstrated by ultrasonography (US) is a more accepted predictor of choledocholithiasis [9, 14, 22, 24]. Since the introduction of LC at the Assaf Harofeh Medical Center (September 1990), all patients over 40 years of age recovering from ABP have been referred for a preoperative ERCP, and an attempt at endoscopic sphincterotomy (ES) has been made when CBD stones are confirmed. Because we noticed, a low incidence of CBD stones in this

372

group of patients, we decided to reassess this policy. We tried specifically to define criteria for selecting the subgroup of patients who would be referred for ERCP prior to referral for LC. This article summarizes our experience and our conclusions regarding the role of ERCP in patients recovering from acute biliary pancreatitis. Methods Patient population
The study was based on 48 patients. The patients were scheduled for LC following recovery from ABP. Between September 1990 and March 1994, ERCP was performed on all patients by the same endoscopist (E.S.). All ERCPs were performed 322 days following the return of the serum and urinary amylase values to normal. The diagnosis of acute pancreatitis was determined by symptoms and signs, laboratory evidence of elevated serum and urinary amylase values, with or without demonstration of edema of the pancreas, by computerized tomography (CT). The diagnosis of cholelithiasis was confirmed by US scan performed within 24 h following arrival at the hospital. A biliary etiology for the pancreatitis was established after ruling out ethylism, hyperlipidemia, hypercalcemia, and, of course, the presence of cholelithiasis.

Table 1. Laboratory values Parameter Serum amylase Urinary amylase ALP AST ALT Bilirubin Highest value (units/l) (meanSD) 1,129 779 5,909 3,940 281 132 105 67 164 85 40 26(mol/l) No. of patients studied 48 48 43 40 41 48

Table 2. US findings No. of patients Small GB stones Big GB stones Normal CBD Dilated CBD Choledocholithiasis 42 6 40 8 6

sons score was 3.6 1.7 criteria present, range zero to seven criteria. CT findings Twenty-six patients underwent a CT scan. Fifteen scans were interpreted as normal. Four of the 15 patients with a normal CT scan were found to have CBD stones. Eleven scans revealed an edematous pancreas, but only one patient who presented with an abnormal CT scan was found to have CBD stones. Laboratory values Table 1 details the highest value of each laboratory parameter. ALT showed the most prominent rise to about four times the normal range. AST rose to about 2.5 times the normal, ALP 2.3 times, and bilirubin twice the normal values. US findings (Table 2) Forty-two patients had small (less than 1 cm) gallbladder stones and six patients large stones. There were 40 patients with normal CBDs and eight patients with dilated CBDs (>8 mm). Choledocholithiasis was diagnosed in six patients; only four of them were found to have CBD stones on ERCP. ERCP results (Table 3) The preoperative ERCP was technically successful in 43 (89.6%) out of the 48 patients. CBD stones were found in 11 (22.9%) patients. No additional biliopancreatic pathology was found. Eight patients underwent endoscopic sphincterotomy (ES). Three patients, under 40 years of age, did not undergo ES. Later, two of them had normal intraoperative cholangiographies (IOCs) and one was treated by open choledochotomy. Six patients presented with mild complica-

Data collection
All data were collected from hospital charts, laboratory charts, and endoscopic, US, and CT reports. Ransons score was calculated by totalling the number of criteria present [31]. Biochemistry results were considered abnormal in a patient with any of the following: bilirubin >22 mol/l, alkaline phosphatase (ALP) >120 units/l, aspartate transaminase (AST) >41 units/l, alanine transaminase (ALT) >39 units/l, serum amylase >160 units/l, and urinary amylase >700 units/l. The highest serum levels during hospitalization were recorded for each patient and used for the analysis. All ultrasound interpretations were made by the same radiologist. The CBD diameter (in millimeters) was measured at its midportion. The suspicion or presence of bile duct stones seen on US was recorded, as was the presence of numerous, small (less than 1 cm) gallbladder (GB) stones. CT studies were performed on 26 patients within 4 days after hospitalization and were interpreted by the same radiologist. ERCP findings with respect to the presence of bile duct stones and other biliopancreatic pathology were evaluated by one of the senior authors (A.H.).

Statistical analysis
For each of the variables listed below, a relative risk (RR) for finding CBD stones by an ERCP was calculated. The risk was considered significant when a confidence interval (CI) of 95% did not include the number 1. Continuous variables (age; laboratory values of bilirubin, ALP, AST, ALT, and amylases; duration of hospital stay; and the Ransons score) are described in the text as mean standard deviation (SD). For these variables, three to 20 cutoffs were done. The RR for finding CBD stones by ERCP was calculated for each cutoff value. CBD diameter at US was analyzed categorically with a cutoff value of >8 mm. CT findings were analyzed as a two-categorical variable: edematous or normal pancreas.

Results The mean age was 55.8 17.4 years, range 2385 years. The patient population comprised 36 women and 12 men. Hospital stay was 10.5 5.6 days, range 525 days. Ran-

373

Fig. 1. Correlation of AST values to ERCP findings.

Fig. 2. Correlation of CBD diameter to ERCP findings.

Table 3. ERCP findings No. of patients Normal ERCP Technical failure CBD stones Sphincterotomy Complications Additional biliopancreatic information 32 5 11 8 6 None

tions following the ERCP: Three patients had mild pancreatitis, two patients had chest pains with no ECG or cardiac enzyme abnormalities, and one patient had upper GI bleeding which resolved spontaneously. All patients with complications were hospitalized for a period of 1 to 4 days. Correlations There were no correlations between age, sex, hospital stay, Ransons score, or CT findings and the ERCP findings. Among the laboratory values, only AST values higher than 90 units/l could predict choledocholithiasis (Fig. 1). Seven out of 15 patients with AST values >90 units/l showed CBD stones on ERCP. Three of the 25 patients with AST values <90 units/l had CBD stones. The relative risk of CBD stones in patients with AST values >90 units/l was 3.89 (CI 1.1812.80) as compared to patients with AST values <90 units/l. There were no correlations between other laboratory parameters and the ERCP findings. Four out of seven patients with a US demonstration of dilated CBD stones (>8 mm) showed CBD stones on ERCP (Fig. 2). Seven out of 36 patients showed evidence of CBD stones on ERCP despite a normal US finding. The relative risk of CBD stones for patients with a dilated CBD was 2.94 (CI 1.177.40) when compared to patients with normal CBD. Discussion The advent of LC has revived the debate on the optimal management of patients with suspected CBD stones [5, 8,

13, 19, 24, 34, 44]. The intention is to maintain the benefits of minimally invasive surgery. As laparoscopic techniques to clear the CBD necessitate time, resources, and specialized equipment, ERCP has become the treatment of choice for CBD stones in many centers [5, 13, 19, 34, 44]. A close to 90% success rate in our study, as well as in other series [20, 36], together with a very low complication rate of 2.3 7% [9, 20, 36], shows the ERCP to be a safe and effective method of dealing with CBD stones. However, we have to remember that this procedure is not free of serious morbidity and morality [9, 20]. The optimal approach to CBD stones has not yet been established [40]. Moreover, there is no clear-cut consensus about the optimal time to perform exploration of the CBD endoscopically [8, 19]. A new therapeutic approach to ABP is an emergency ERCP. Preliminary reports suggest that an emergency ERCP may miminize the complications of ABP [6, 12, 26]. Most endoscopists still perform an ERCP following recovery from ABP [33] as ERCP alone can give rise to pancreatitis [7, 9]. Predictors of CBD stones have been studied extensively, but patients recovering from ABP represent a specific population. Traditionally, hyperamylasemia and history of pancreatitis are indications for CBD exploration [16, 32], but recent randomized controlled studies showed that they were not significant predictors of CBD stones [4]. Although the pathophysiology of ABP is induced by a CBD stone [22], clinical series suggest that unsustained hyperamylasemia reflects stone migration across the sphincter of Oddi [1, 22]. This phenomenon is demonstrated by two of our patients who showed CBD stones on US on admission and had a normal ERCP a few days later. CBD stones were found in 25.6% of our patients with ABP, which is in the higher portion of the range found in the literature (333%) [7, 15, 27]. This rate may be an exaggeration of the number of patients who require treatment, as two patients who were seen to have evidence of CBD stones on ERCP had a clear IOC. This was probably as a result of spontaneous passage of these stones. It seems that the incidence of CBD stones after ABP does not justify performing ERCP on a routine basis, especially when no other biliopancreatic information has been obtained. Investigation of all patients with IOC may still be considered.

374

Haver-Jensen et al. [18] and Barkun et al. [4] found that age was a significant predictor of CBD stones. We did not find either age or sex to be a predictor of CBD stones in the subpopulation of patients recovering from ABP. The severity of the pancreatitis was defined in our study by three parameters: Duration of hospital stay, Ransons score, and the shape of the pancreas determined by CT (which is considered to be correlated to the severity of the pancreatitis) [38]. None of these parameters was correlated to CBD stones. Many studies examine the LFT as a predictor of CBD stones. The chance of a patient with normal LFT having CBD stones is low [11, 17, 28], and most studies found that at least two of the LFTs have to be abnormal in order to become a significant predictor [9, 11, 17, 19, 35]. Many patients with ABP present with an abnormal LFT. In our study, as in another [35], ALT had the most prominent rise. Yet, as we examined each laboratory parameter individually, we found that only AST values above 90 units/l could predict CBD stones. We cannot find any theoretical explanation for this finding, and it may be a casual result. Our sample was too small to examine combinations of laboratory values or other parameters. Further studies are needed to clarify the role of AST values and combinations of parameters as predictors of CBD stones in this specific population. Many studies have found US criteria, especially dilatation of the CBD, to be useful predictors of CBD stones [9, 14, 18, 24], but others maintain that a dilated CBD alone may be an unreliable indicator [45], perhaps because of the ability of the CBD diameter to change its diameter over a short period of time [23, 25]. We found that a dilated CBD is a significant predictor of CBD stones. Gillams et al. [14] recommended performance of ERCP on every patient with US evidence of a dilated CBD, and this seems especially appropriate in patients recovering from ABP. Again, our numbers were too small to examine the predictive value of combinations of CBD dilatation with other abnormalities. Such combinations may increase the ability to predict CBD stones [9, 24]. The availability of modern US equipment enables precise measurement of the CBD diameter. For detecting CBD stones, however, US is not sensitive enough. Cronan [10] diagnosed 55% of CBD stones with US. In our study, four out of 11 patients with CBD stones as confirmed by ERCP were found by US. Accurate diagnosis of CBD stones mandates direct injection of contrast material into the biliary tree. As the availability of intraoperative criteria, which carry a strong predictive ability [17, 41, 42], have decreased in the era of LC, we will continue searching for noninvasive methods of predicting choledocholithiasis. For patients recovering from ABP, we conclude that an ERCP should be performed on those with a dilated CBD revealed on US.

References
1. Acosta JM, Ledesma CL (1974) Gallstone migration as a cause of acute pancreatitis. N Engl J Med 290: 484487 2. Armstrong CP, Taylor TV, Jeacook J, et al (1985) The biliary tract in patients with acute gallstone pancreatitis. Br J Surg 72: 551552 3. Barkun JS, Barkun AN, Sampolis JS, et al (1992) Laparoscopic versus mini-cholecystectomy: a randomized controlled trial. Lancet 2: 116 119

4. Barkun AN, Barkun JS, Fried G, et al (1994) Useful predictors of bile duct stones in patients undergoing laparoscopic cholecystectomy. Ann Surg 220: 3239 5. Boulay J, Schellenberg R, Brady PG (1991) Role of ERCP and therapeutic biliary endoscopy in association with laparoscopic cholecystectomy. Am J Gastroenterol 87: 837842 6. Brambs HJ, Ruckauer K, Scholmerich J, et al (1988) ERCP in acute pancreatitis. A preliminary report. Dig Surg 5: 156159 7. Conn M, Goldenberg A, Conception L, et al (1991) The effect of ERCP on circulating pancreatic enzymes and pancreatic protease inhibitors. Am J Gastroenterol 86: 10111014 8. Cotton PB, Baillie H, Pappas TN, et al (1991) Laparoscopic cholecystectomy and the biliary endoscopist. Gastrointest Endosc 37: 9497 9. Cotton PB, Lehaman G, Vennes J, et al (1991) Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 37: 383393 10. Cronan JJ (1986) US diagnosis of choledocholithiasis: a repraisal. Radiology 161: 133134 11. DelSanto P, Kazarian KK, Rogers JF, et al (1985) Prediction of operative cholangiography in patients undergoing elective cholecystectomy with routine liver chemistries. Surgery 98: 711 12. Fan ST, Lai EC, Mok FP, et al (1993) Early treatment of acute biliary pancreatitis by endoscopic papilotomy. N Engl J Med 328: 228232 13. Frazee RC, Roberts J, Symmonds R, et al (1993) Combined laparoscopic and endoscopic management of cholelithiasis and choledocholithiasis. Am J Surg 166: 702706 14. Gillams A, Cheslyn-Curtis S, Russell RCG, et al (1992) Can cholangiography be safely abandoned in laparoscopic cholecystectomy? Ann R Coll Surg Engl 74: 248251 15. Graham LD, Burrus RG, Burns RP, et al (1994) Laparoscopic cholecystectomy in biliary pancreatitis. Am Surg 60: 4043 16. Hashmonai M, Arrison R, Schramek A (1980) Indications for exploration of the bile ducts. Int Surg 65: 239245 17. Haver-Jensen M, Karesen R, Nygard K, et al (1985) Predictive ability of choledocholithiasis indicators. Ann Surg 202: 6468 18. Haver-Jensen M, Larsen R, Nygard K, et al (1993) Prospective randomized study of routine intraoperative cholangiography during open cholecystectomy: long-term follow-up and multivariable analysis of choledocholithiasis. Surgery 113: 318323 19. Hawasli A, Lloyd L, Pozios V, et al (1993) The role of endoscopic retrograde cholangio-pancreaticogram in laparoscopic cholecystectomy. Am Surg 59: 285289 20. Helwig A, Dinkel E, Mundinger A, et al (1990) The place of ERCP in pancreatic diagnosis. The change caused by sonography and computed tomography. Radiologe 30: 413419 21. Hunter JG (1992) Laparoscopic transcytic common bile duct exploration. Am J Surg 163: 5358 22. Kelly TR (1976) Gallstone pancreatitis: pathophysiology. Surgery 80: 488492 23. Lygidakis NJ (1984) The incidence and significance of common bile duct dilatation in biliary calculous disease. World J Surg 8: 327334 24. McEntee G, Grace PA, Bouchier-Hayes D (1991) Laparoscopic cholecystectomy and the common bile duct. Br J Surg 78: 385386 25. Mueller PR, Ferucci JT, Simeone JF, et al (1982) Observations on the distensibility of the common bile duct. Radiology 142: 467472 26. Neoptolemos JP, Hall AW, Finlay DF, et al (1984) The urgent diagnosis of gallstone in acute pancreatitis: a prospective study of three methods. Br J Surg 71: 230233 27. Paloyan D, Simarowitz D, Skinner DB (1975) The timing of biliary tract operations in patients with pancreatitis associated with gallstones. Surg Gynecol Obstet 141: 737739 28. Parnthaler P, Sandbichler P, Schmid TN, et al (1990) Operative cholangiography in elective cholecystectomy. Br J Surg 77: 399400 29. Pellegrini CA (1993) Surgery for gallstone pancreatitis. Am J Surg 165: 515518 30. Perissat J, Huibregtse K, Keane FBV, et al (1994) Management of bile duct stones in the era of laparoscopic cholecystectomy. Br J Surg 81: 799810 31. Ranson JHC, Rifkind DM, Rases DF, et al (1974) Prognostic signs and the role of operative management in acute pancreatitis. Surg Gynecol Obstet 139: 6981 32. Reiss R, Deutsch AA, Nudelman I, et al (1984) Statistical value of

375 various clinical parameters in predicting the presence of choledochal stones. Surg Gynecol Obstet 159: 273276 Rosseland AR, Solhavg JH (1984) Early or delayed endoscopic papilotomy (EPT) in gallstone pancreatitis. Ann Surg 199: 165167 Roy A, McAlister V, Passi RB (1993) ERCP in management of choledocholithiasis with laparoscopic cholecystectomy. Can J Surg 36: 8184 Saltztein EC, Peacock JB, Thomas MD (1982) Preoperative bilirubin, alkaline phosphatase and amylase levels as predictors of common duct stones. Surg Gynecol Obstet 154: 381384 Scheeres DE, Simon I, Ponsky JL (1990) Endoscopic retrograde cholangiopancreatography in general surgery practice. Am Surg 56: 185 191 Schirmer BD, Edge SB, Dix J, et al (1991) Laparoscopic cholecystectomy: treatment of choice for symptomatic cholelithiasis. Ann Surg 213: 665676 Scholmerich J, Gross V, Johannsesson T, et al (1989) Detection of biliary origin of acute pancreatitis. Comparison of laboratory tests, ultrasound, computed tomography and ERCP. Dig Dis Sci 34: 830 833 39. Soper NJ, Stockmann PT, Dunnegan DL, et al (1992) Laparoscopic cholecystectomy the new gold standard? Arch Surg 127: 917923 40. Stiegmann GV, Goff JS, Mansour A, et al (1992) Precholecystectomy endoscopic cholangiography and stone removal is not superior to cholecystectomy, cholangiography and common duct exploration. Am J Surg 163: 227230 41. Taylor TV, Torrance B, Rimmer S, et al (1983) Operative cholangiography: is there a statistical alternative? Am J Surg 145: 640643 42. Taylor TV, Armstrong CP, Rimmer S, et al (1988) Prediction of choledocholithiasis using a pocket microcomputer. Br J Surg 75: 138140 43. Traverso LW, Kozarek RA, Ball TJ, et al (1993) Endoscopic retrograde cholangiopancreatography after laparoscopic cholecystectomy. Am J Surg 165: 581586 44. Vitale GC, Larson GM, Wieman TJ, et al (1993) The use of ERCP in the management of common bile duct stones in patients undergoing laparoscopic cholecystectomy. Surg Endosc 7: 911 45. Wilson TG, Hall JC, Watts JMCK (1986) Is operative cholangiography always necessary? Br J Surg 73: 637640

33. 34. 35. 36. 37. 38.

Surg Endosc (1997) 11: 397401

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Optimal port locations for endoscopic intracorporeal knotting

G. B. Hanna,1,2 S. Shimi,1 A. Cuschieri1
1 2

Department of Surgery, Ninewells Hospital and Medical School, University of Dundee, Dundee Tayside DD1 9SY, Scotland Surgical Skills Unit, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, Scotland

Received: 23 July 1996/Accepted: 4 October 1996

Abstract. Port location is crucial for endoscopic manipulations. The aim of the study was to investigate the influence of manipulation, azimuth, and elevation angles of instruments on endoscopic intracorporeal knotting. The standard task was tying a surgeons knot. Manipulation angles of 30, 45, 60, 75, and 90 with equal and unequal azimuth angles and elevation angles of 0, 30, and 60 were investigated. The endpoints were the execution time and parameters of knot analysis. The execution time was shorter with 60 than with either 90 or 30 manipulation angles (p < 0.0001 and p < 0.01). Equal azimuth angles resulted in a shorter execution time than wide unequal angles (p < 0.001). A combination of 60 manipulation angle with 60 elevation angle had the shortest execution time (p < 0.001) and highest performance quality score (p < 0.02). A range of 4575 manipulation angles with equal azimuth angles is recommended. As the manipulation angle increases, the elevation angle has to increase accordingly. Key words: Port sites Knot Instruments Needle drivers

The aim of the study was to investigate the effect of manipulation, azimuth, and elevation angles on endoscopic task performance during intracorporeal knot tying since these angles determine optimal port placement. Materials and methods
Experiment I concerned the influence of varying manipulation and azimuth angles. Experiment II was designed to investigate the effect of different elevation angles on task performance (Fig. 1).

Task
The standardized task consisted of tying an intracorporeal surgeons knot from a 200-mm length of 2/0 silk inserted through a block of light yellow foam. The foam block was 25 75 155 mm. A longitudinal groove in the middle of the back of the foam housed a rubber tube. The foam was mounted with Velcro on the oblique surface of a right-angled block of wood (Fig. 2). Two silk threads were passed round the rubber tube from the front; one at each end of the block, at a distance of 30 mm from the middle of the foam. At each site, the thread was passed around the rubber tube fixed behind the foam with a 10mm distance between the entry and exit points on the foam. Equal lengths of thread were left protruding for the surgeon to form the knot. The assembled task rig was stabilized in place inside a box (450 mm 350 mm 250 mm). The base, sides, and back were made of wood while the front and the top were made of brown cardboard attached to the wood with Velcro. Strips of neoprene were sutured to the cardboard. Into those strips, trocars were inserted to allow maneuverability of the instruments while retaining the port positions. The trainer was placed on a table of adjustable height so that the surgeon could hold the needle drivers with his elbow at a right angle and the arm in the horizontal plane.

The position of the instrument ports in relation to each other and to the optical port is an important determinant of the ease of performance of an endoscopic procedure and of its execution time. For bimanual tasks, three angles (manipulation, azimuth, and elevation) govern optimal port sites. The manipulation angle is the angle between the two instruments (active and assisting). The azimuth angle describes the angle between either instrument and the optical axis of the endoscope. The elevation angle of the instrument is defined as the angle between the instrument and the horizontal plane.

Equipment

Endoscopes. 10-mm 0, 30, 45 endoscopes (Karl Storz, Tuttlingen, Germany) were used.1

Correspondence to: A. Cuschieri

Light sources. Storz cold-light fountain 450-V light sources were connected to endoscopes using 495NL fiber-optic light cables of 3.5-mm diameter and 180-cm length (Karl Storz, Tuttlingen, Germany).

398

Fig. 1. Diagrammatic representation of the azimuth, manipulation, and elevation angle.

target surface at the knot location with the end of the endoscopes at distances of 125 mm and 150 mm from the knot, respectively. Needle drivers were inserted to make 0 or 30 and 60 elevation angles and 30 and 60 manipulation angles (comprising equal azimuth angles). Twelve instrument positions were investigated.

Fig. 2. Task rig.

Subjects and procedure. Ten surgeons with differing clinical experience who received training in knot tying in the advanced laparoscopic courses in Dundee participated in each experiment. The surgeons had corrected eyesight and each performed six knots in each instrument position dictated by the experimental design. Each participating surgeon performed the required intracorporeal knots in three sessions, tying two knots in each of the instrument positions per session. The order of instrument position during each session was allocated in a random sequence. After tying the intracorporeal surgeons knot, the thread loop around the tube was divided at the opposite pole to the knot. All the knots performed by each operator were numbered and coded for the operator and position. The time taken to execute each knot was measured from the moment the operator grasped the handles of the needle drivers until the instruments were released on completion of each intracorporeal knot.

Camera and monitor. The Storz Endovision 9050-PB single-chip camera (Karl Storz, Tuttlingen, Germany) and Sony PVM-1443MD highresolution (>600 TV lines) monitors (Sony, Tokyo, Japan) were used for the experiments. The monitor was placed directly facing the surgeon at a distance of 1 m with the center of the monitor at eye level.

Knot testing apparatus

The knot testing apparatus was set up using an Instron tensiometer (model 1026, Instron Ltd., UK). Knots were strained between the clamps of the tensiometer and signals from the load cell (max. 20 kg) were fed to a signal conditioning unit which provided a gain of 100 and a filter to remove the high-frequency noise. The modified signal was recorded by an analogueto-digital conversion card with a sampling frequency of 25 Hz (Advantech PCL-812PG) inserted in the computer recording system. A snapshot program (HEM Data Corp., Michigan, USA) controlled the conversion card and displayed the load cell waveform. A data analysis program (written in Matlab, The Math Work Inc., Natick, Massachusetts, USA) was used to smooth the data (by digital filtering), remove zero offsets, apply the load cell calibration factor, synchronize each record using a digital trigger, derive extension correlates to the force data, and calculate knot-quality parameters. The extension values were obtained from the tensiometer jaw separation speed (50 mm/min) and the sampling frequency. The two ends of the divided loop of the knot were inserted between the jaws of the clamps and strained by the tensiometer.

Needle drivers and trocar cannulae. The intracorporeal knot tying was performed using a pair of 26173SK Cuschieri needle drivers (Karl Storz, Tuttlingen, Germany). Endoscopes were introduced though reusable 11mm ports (Karl Storz, Tuttlingen, Germany) while needle drivers were inserted though reusable 5.5-mm ports (Wisap, Germany).

Experimental procedures

Experiment I. The forward-oblique 30 endoscope was introduced such that the optical axis of the endoscope was perpendicular to the target surface at the knot location with the end of the endoscope at a distance of 125-mm from this point. Needle drivers were inserted perpendicular to the plane of the oblique surface of the foam block, making a 60 elevation angle. Manipulation angles of 30, 45, 60, 75, and 90 comprising equal azimuth angles were investigated. The 60 manipulation angle was also investigated with the following unequal azimuth angles: 37.5/22.5, 22.5/ 37.5, 45/15, and 15/45 with the first angle on the right side of the surgeon. Nine instrument positions were investigated.

Endpoints of knot analysis

A number of parameters were obtained on analysis of the force-extension curves for systems of untied ligatures and knots. The maximum force necessary to break the knot was defined as the breaking force, while the slipping force (required to undo the knot without breakage of the thread) was defined as the average force of the plateau of the curve. Measurements were obtained from the initial slope of the curves in the form of work done and integrated force. The slope measurements were obtained over a section

Experiment II. Endoscopes with 0 and 45 directions of view were introduced such that the optical axis of the endoscope was perpendicular to the

399 Table 1. Effect of manipulation angles of the instruments on endoscopic knot tying (median and interquartile range) Angle 30 Time (seconds) Force (Newtons) Work (Newtons m) KQS (%) PQS
a

45 62 (27.5) 24.5291 (3.9503) 0.0276 (0.0112) 34.1395 (14.7283) 35.0828 (23.0856)

60 61 (21.5) 24.6279 (3.4943) 0.0240 (0.0107) 29.4109 (15.1830) 33.7411 (21.2977)

75 63 (17.75) 24.2747 (4.5969) 0.0248 (0.0124) 30.2085 (17.7260) 33.0296 (25.9378)

90 67 (23.5) 24.5108 (3.4976) 0.0232 (0.0091) 28.0498 (14.7697) 27.9039 (21.0339)

p valuea <0.0001 0.1 0.07 0.1 <0.01

76 (31.75) 23.4502 (3.0817) 0.0254 (0.0107) 28.5575 (14.3918) 26.1115 (17.4001)

Kruskal-Wallis one-way ANOVA

that was short enough to avoid the breaking or slipping events but long enough to permit sufficient integration and thus reduce the noise contribution. The breaking or slippage force reflects the strength of the knot while the measurements of the slope of the curve reflect the degree of tightening. A single parameter of knot quality which reflects the strength of the knot and degree of tightening, a knot quality score (KQS), is derived as follows: knot breaking or slipping force integrated force for the knot KQS = 100% thread breaking force integrated force for the thread The time of execution of the task was used as a measure of task efficiency. To relate the quality of the knot to task efficiency, a performance quality score (PQS) was obtained. Performance quality score knot quality score/execution time score. The execution time score relates the execution time for each knot to a norm-referenced execution time which, in the present study, was defined as the mean execution time for the entire population.

differ between manipulation angles studied. However, both 60 and 45 manipulation angles yielded a significantly higher PQS than 30 angles (p < 0.01). Experiment II. Three hundred sixty knots were tied at each manipulation angle. Thirty and 38 knots slipped on using 60 and 30 manipulation angles, respectively, and these knot slippage rates were not significantly different (p 0.3). The 60 manipulation angle had a lower execution time and a higher force (median 80 s, 24.76 Newton) than 30 angle (median 90 s, 24.2 Newton) with p value <0.0001 and <0.05, respectively (Mann-Whitney U-test). There was no significant difference in the work done or KQS. In addition, the 60 manipulation angle yielded a better PQS (median 34.37) than 30 manipulation angle (median 29.06) with p < 0.0001.

Statistical analysis
The data on the execution time, force, work done, knot quality, and performance quality scores were not normally distributed. Nonparametric tests such as Kruskal-Wallis one-way ANOVA and Mann-Whitney test were used as appropriate. Significance was set at the 5% level.

Azimuth angles In experiment I, 60 knots were tied at each angle. The number of slipped knots was three, five, 12, four, and six with 30/30, 15/45, 45/15, 22.5/37.5, and 37.5/22.5 azimuth angles, respectively. Logistic regression showed no significant difference in the number of slipped knots (p 0.26). Table 2 shows the median and interquartile range of the output parameters when using different azimuth angles. Equal azimuth angles had the shortest execution time. Equal (30/30) or narrow unequal azimuth angles (22.5/37.5) resulted in a shorter execution time than wide unequal angles (45/15) (p < 0.01Mann-Whitney U-test). There was no significant difference in the force, work done, or KQS between the different azimuth angles. The disposition of the wider azimuth angles, on the right or left side, did not influence the execution time, the force, work done, KQS, or PQS (p > 0.1Mann-Whitney U-test).

Results Manipulation angles Experiment I. Sixty knots were tied at each angle. The number of slipped knots were 6, 2, 3, 6, and 4 when using 30, 45, 60, 75, and 90 manipulation angles, respectively. Logistic regression showed no significant difference in the number of slipped knots (p 0.26). Table 1 shows the median and interquartile range of the output parameters when using different manipulation angles; the 60 manipulation angle had the shortest execution time. The execution time was significantly shorter when using the 60 manipulation angle than for the 90 or 30 angles (p < 0.0001, p < 0.01, respectivelyMannWhitney U-test). There was no significant difference in the execution time between manipulation angles of 45, 60, and 75 (p 0.1). The force, work done, and KQS did not

Elevation angle Table 3 shows the median (interquartile range) of parameters of knot analysis for each elevation angle in experiment

400 Table 2. Effect of azimuth angles of the instruments on endoscopic knot tying (median and interquartile range) Azimuth angles 30/30 Time (seconds) Force (Newtons) Work (Newtons m) KQS (%) PQSy
a

15/45 66 (17.5) 24.3685 (4.1717) 0.0260 (0.0101) 31.6392 (16.4124) 29.7095 (19.3980)

45/15 70 (22.5) 24.3668 (5.9574) 0.0242 (0.0100) 26.4081 (19.7535) 27.0929 (19.3273)

22.5/37.5 65 (22) 24.3235 (4.5446) 0.0296 (0.0106) 33.2258 (16.8082) 34.3745 (28.3848)

37.5/22.5 60 (22) 24.6448 (4.3111) 0.0258 (0.0109) 32.2351 (18.7304) 34.1124 (22.2828)

p valuea <0.005 0.7 0.1 0.3 0.08

61 (21.5) 24.6279 (3.4943) 0.0240 (0.0107) 29.4109 (15.1830) 33.7411 (21.2977)

Kruskal-Wallis one-way ANOVA

Table 3. Effect of elevation angles on endoscopic knot tying at each manipulation angle; median (interquartile range) Manipulation angle 30 Elevation angles Time (seconds) Force (Newtons) Work (Newtons m) KQS (%) PQS
a

60 0 100 (39.75) 25.1321 (4.0905) 0.0269 (0.0122) 32.9994 (17.6808) 28.6283 (21.3405) p <0.0001 <0.001 <0.001 <0.01 0.6
a

60 90 (36.5) 23.1983 (4.1609) 0.0273 (0.0113) 32.4999 (18.6140) 30.8835 (26.8366)

30 84 (28.5) 24.3230 (4.3134) 0.0228 (0.0118) 26.6288 (16.6542) 29.0648 (19.7476)

60 71.5 (25) 24.2581 (3.7509) 0.0248 (0.0130) 30.3694 (19.1540) 40.6000 (24.1492)

30 71.5 (30) 24.8011 (4.3629) 0.0227 (0.0113) 27.9867 (17.5059) 33.0647 (24.8939)

0 92 (30) 24.9952 (3.6118) 0.0265 (0.0118) 33.6971 (17.0235) 32.3494 (20.9361)

pa <0.001 0.2 <0.01 <0.01 <0.02

Kruskal-Wallis one-way ANOVA

II. At 30 manipulation angle; 18, ten and ten knots slipped out of 120 knots tied at 0, 30, and 60 elevation angles (p 0.09). Logistic regression showed that the different elevation angles studied did not affect the number of knots that slipped when stressed. The 30 elevation angle yielded a shorter execution time than 0 and 60 angles (p < 0.0001, p < 0.005, respectivelyMann-Whitney U-test). On the other hand, knots performed in the 30 elevation angle had a lower work done than either 0 or 60 angles (p < 0.005, p < 0.001, respectively) when stressed. Knots performed with 60 elevation angle had a lower force than 0 or 30 angles (p < 0.001, p < 0.05 respectively) but no significant difference was found between 30 and 0 angles (p 0.1). Knots performed with 30 elevation angle had a lower quality score than either 0 or 60 angles (p < 0.005, p < 0.05). At 60 manipulation angle; nine, fourteen, and seven knots were slipped out of 120 knots tied at 0, 30, and 60 elevation angles. Logistic regression showed that the number of slipped knots was not significantly different between elevation angles (p 0.64). The 60 and 30 elevation angles had a shorter execution time than the 0 angle (p < 0.0001). The work done by knots formed with either 60 or 0 elevation angles was higher than that with the 30 angle (p < 0.05, p < 0.005, respectively). The 60 elevation angle was attended with a higher PQS than 30 and 0 angles (p < 0.05, p < 0.005, respectively).

Surgeon and practice In experiment I the number of slipped knots ranged between one and 11 per 54 knots tied by each surgeon with a significant difference between surgeons (p < 0.01 on logistic regression). Table 4a shows the median and interquartile range of execution time and quality scores for each surgeon. There was a significant difference between surgeons in the execution time and all knot parameters (p < 0.0001). No significant difference in the execution time and knot quality parameters was observed between the first and subsequent knots during each session (p > 0.3Mann-Whitney U-test). Practice improved the execution time (67, 66, 62 s, p < 0.005) and the PQS (29.26, 27.8, 34, p < 0.05KruskalWallis one-way ANOVA) between the three sessions. There was no significant difference in the force, work done, or KQS with practice. Similar findings relating to surgeons performance were encountered during experiment II. The number of knots that slipped when strained in this experiment ranged from one to 18 per 108 knots tied by each surgeon (p 0.12 on logistic regression). Table 4b shows the median and interquartile range of the execution time and quality scores for each surgeon. A significant difference between surgeons in the execution time and all knot parameters was observed (p < 0.0001). However, in experiment II, a significant improve-

401 Table 4a. Effect of the surgeon on endoscopic knot tying (median and interquartile range) in experiment I Surgeon 1 Time (second) KQS (%) PQS 65 (16.25) 16.1093 (11.0706) 16.6155 (14.1655) 2 52.5 (15) 26.8058 (12.5977) 34.6975 (17.5051) 3 72 (21.5) 25.4027 (11.1662) 20.6908 (17.1533) 4 43 (17.5) 29.2318 (13.8631) 48.1868 (32.0202) 5 63 (13.25) 44.8225 (10.4072) 48.6740 (12.7610) 6 57.5 (12.25) 31.8741 (13.4712) 36.3133 (22.3668) 7 75 (20.25) 37.3667 (11.2442) 32.2867 (13.0417) 8 88.5 (26.75) 31.7495 (15.0629) 24.3057 (13.6546) 9 77 (23.25) 31.1324 (11.1587) 24.0704 (15.2151) 10 59.5 (12) 30.6639 (15.1197) 34.9352 (18.5570)

Table 4b. Effect of the surgeon on endoscopic knot tying (median and interquartile range) in experiment IIa Surgeon 1 Time (seconds) KQS (%) PQS
a

2 62.5 (26.25) 23.5513 (16.0325) 31.2210 (20.5610)

3 95.5 (32) 19.7805 (15.9982) 17.6486 (14.2923)

4 65 (28.75) 29.9100 (17.2255) 44.2222 (26.8107)

5 84 (30) 43.2332 (14.5613) 47.4857 (22.1383)

6 75 (22.5) 37.2222 (16.6852) 45.0438 (27.6953)

7 90 (34.25) 36.5053 (11.6522) 35.6612 (17.6889)

8 95 (31.75) 27.4430 (19.9349) 25.6022 (20.8270)

9 110 (45.75) 34.0277 (15.2545) 27.8654 (13.7791)

10 86 (27.5) 23.4244 (10.2866) 23.1928 (15.1270)

83 (34.5) 30.8989 (10.7545) 33.6695 (20.0114)

Surgeons 2, 4, 5, 6, 7, and 10 participated in both experiments

ment in the execution time was observed between the first and subsequent knots (median 88.5, 83.5 s, respectively, p < 0.05) but no significant difference was found in knot quality parameters (p > 0.1Mann-Whitney U-test). Furthermore, practice improved the execution time (median 92, 85, and 82 s, p < 0.001) and the PQS (median 28.21, 32.36, and 34.75, p < 0.05Kruskal-Wallis one-way ANOVA). Discussion The placement of ports for advanced endoscopic operations is currently dictated by surgeon preference based on individual experience since there have never been any reported studies on optimal placement based on objective data. The results of the present investigation, which was designed to address this issue, indicate that as far as intracorporeal knotting is concerned, maximal efficiency and quality performance are obtained with a manipulation angle ranging between 45 and 75; the ideal manipulation angle is 60. Manipulation angles below 45 or above 75 are accompanied by increased difficulty and degraded performance. The operating ports for the active and assisting instruments, e.g., needle drivers, should be placed subsequent to the insertion of the endoscope such that they subtend a manipulation angle within this rage and as close to 60 as is possible. The results of this study have also shown that task efficiency was better with equal than with unequal azimuth angles on either side of the optic. In practice this may be difficult to achieve but wide azimuth inequality should be avoided since this degrades task efficiency (for intracorporeal knotting) irrespective of the side, right or left, of the azimuth angle predominance.

When a 30 manipulation angle is imposed by the anatomy or build of the patient, the elevation angle should be also 30 as this combination carries the shortest execution and an acceptable knot quality score. Likewise, with a 60 manipulation angle, the corresponding optimal elevation angle which yields the shortest execution time and optimal quality performance is 60. When manipulations are carried out without an elevation angle, intracorporeal knot tying is slower although the knot quality is good. The objective data on the ideal angles for efficient endoscopic manipulations can be incorporated into a software system which works out the ideal port sites commensurate with the build of the patient and the operative task in hand. This aspect of the work, which represents an example of computer-assisted operating planning, is currently being addressed by the group. In the meantime, the best ergonomic layout for endoscopic surgery consists of a manipulation angle ranging from 45 to 75 with equal azimuth angles on either side of the optic. There is a direct correlation between the manipulation and the elevation angle. Thus, wide manipulation angles necessitate wide elevation angles for optimal performance and task efficiency. This study has also demonstrated the positive effect of practice and the variable performance by different surgeons which reflects different levels of skills that the individual brings to the endoscopic task. There will always be individuals who are more gifted than others. The majority improve with practice but reach different levels of performance.
Acknowledgment. This work was supported by a grant from the Scottish Office, Home and Health Department, St. Andrews House, Edinburgh. We are grateful to the surgeons who participated in the study.

Surg Endosc (1997) 11: 403

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The author replies

We would like to thank Dr. Halpern for his supplementary remarks concerning our case report on a false aneurysm of a hepatic artery branch after laparoscopic cholecystectomy. We have read the paper by Dr. Genyk, Dr. Keller, and Dr. Halpern with great interest and admit that we were not earlier aware of this publication. Dr. Genyk and co-authors stress a possible relationship between the use of laser-assisted dissection and injury of the hepatic artery with pseudoaneurysm formation. Since we only use conventional electrocautery during laproscopic surgery, our patient demonstrates that this complication may also occur without the use of laser-assisted dissection. R. J. Porte
Department of Surgery Leiden University Hospital P.O. Box 9600 9600 R C Leiden The Netherlands

Surg Endosc (1997) 11: 383384

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Small-bowel obstruction and perforation

A rare complication of an esophageal stent
T. H. Henne, B. Schaeff, V. Paolucci
Department of Surgery, University of Frankfurt am Main, Theodor Stern Kai 7, D-60590 Frankfurt, Germany Received: 23 August 1995/Accepted: 23 March 1996

Abstract. To bridge a malignant stenosis after esophageal cancer recurrence two silicon-coated wall stents were inplanted in a 52-year-old patient within 6 months. Two weeks after the second stent was implanted, clinical examination showed dislocation of the prosthesis. Intraoperatively the two stents were found sticking in the side-to-side jejunostomy of a former Billroth II operation, leading to perforation there. Key words: Wire mesh stent Wall stent Esophageal carcinoma

The implantation of silicon-coated stents is an effective and safe procedure for avoiding dysphagia caused by inoperable esophageal stenosis [1, 2]. We present a rare complication: a small-bowel obstruction followed by perforation due to migration of two esophageal stents.

Materials and methods

R.T. is a white female 52-year-old patient. After a Billroth II resection in 1979 due to benign ulcers she developed a squamous cell carcinoma of the esophagus in 1994 (pT3, pN 2, M0). Esophagectomy was performed and passage was restored with colon interposition. Six months later she complained of dysphagia. Endoscopy revealed a stenosis in the cervical anastomosis, and in accordance with the macroscopic and microscopic aspect as well as the tumor staging, a recurrent cancer was assumed. A silicon-coated wall stent (25-mm diameter, 10-cm length) was implanted. The patient was discharged symptom-free after 4 days. Six months later she suffered again from dysphagia. Endoscopy showed a stenosis at the oral end of the stent. A second stent was implanted, lengthening in telescope fashion the first stent (Fig. 1). She left the hospital after 6 days, symptom-free. Two weeks later at a control checkup she referred to abdominal pain over 10 days. Oral food intake was without problem and she had no weight loss. Clinical examination revealed a tender mass in the left lower abdominal quadrant.

Fig. 1. Two stents in the cervical anastomosis.

Correspondence to: V. Paolucci

Sonography showed a circular hyperechoic figure (Fig. 2) and X-ray (Fig. 3) proved the diagnosis of stent migration. She was operated the same day. Intraoperatively the stents were found sticking in the side-to-side jejunostomy, performed during the former Billroth II operation, with a small perforation. The anastomosis was resected and reconstructed. The histologic examination of the resected specimen revealed a lymphatic carcinomatosis. Figure 4 shows one of the removed stents. During the post-

384

Fig. 4. Removed stent with obstruction.

effective procedure. Coating of the stents with silicon avoids tumor growth through the wall of the stent but seems to increase the risk of stent dislocation [3]. In our case, lengthening a 25-mm-diameter siliconcoated wall stent with a second one led to dislocation of both stents into the jejunum, followed by obstruction and finally perforation of the small bowel. While dislocations during implantation or in the first 2 days are well known, late dislocationsas in our caseare rare. We believe that, due to the pressure of two stents, local necrosis developed, thus increasing the diameter of the former stenosis. The future will show whether dislocation of stents can be avoided by further improvement of stent technology. The arrest of the stents with following perforation was caused by the right angular side-to-side anastomosis of the jejunum performed during the former Billroth II operation. Passage through colojejunostomy and the following perforation of the small bowel were uneventful. We suggest waiting for normal passage in patients with dislocated stents in the small bowel and operating if the patients become symptomatic. Conclusion
Fig. 2. Sonographic aspect of the stent in the small bowel. Fig. 3. Abdominal X-ray, showing the stents in the abdomen.

We recommend clinical checks of patients with stents routinely in individually adapted intervals. References

operative course again a stent was implanted due to dysphagia. The patient was discharged after 3 weeks.

Discussion Palliation of dysphagia in patients with inoperable esophageal stenosis by implantation of self-expanding stents is an

1. Knyrim K, Wagner HJ, Bethge N, Keymling M, Vakil NA (1993) A controlled trial of expansile metal stent for palliation of esophageal obstruction due to inoperable cancer. N Engl J Med 18: 13021307 2. Neuhaus H, Hoffmann W, Dittler HJ, Nierdermeyer HP, Classen M (1992) Palliation of malignant dysphagia. Endoscopy 24: 405410 3. May A, Hochberger J, Fleig WE, Hahn EG, Ell C (1994) Silikonummantelte und nicht-ummantelte Metallendoprothesen (Wallstent) zur sophagus und Cardia. Endoskopie Palliation maligner Stenosen von O heute 2: 151156

Technique
Surg Endosc (1997) 11: 390392

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Thoracoscopic transdiaphragmatic left adrenalectomy

An experimental study
E. Pompeo, W. Coosemans, P. De Leyn, G. Deneffe, D. Van Raemdonck, T. Lerut
Department of General Thoracic Surgery, Katholiek University of Leuven, Gasthuisberg Hospital, Herestraat 49, 3000 Leuven, Belgium Received: 6 November 1995/Accepted: 28 May 1996

Abstract. The endoscopic approach to adrenal glands has been limited to laparoscopic and retroperitoneal access due to the relative inaccessibility of the retroperitoneal space in the former case and to the limited working space in the latter. We undertook this study to investigate the possibility of performing a left adrenalectomy through a thoracoscopic transdiaphragmatic approach in a swine model. Five pigs were anesthetized, intubated, and ventilated. Four laparoscopic trocars were inserted and a left pneumothorax was accomplished by CO2 insufflation. A peripheral posterior phrenotomy was made starting from the aortic hiatus and extended laterally for about 6 cm. The resulting pneumoretroperitoneum facilitated the subsequent identification of anatomic structures, allowing an easy exposure of the left adrenal gland. The gland was progressively dissected downward, interrupting the tributary vessels with endoscopic clips, and it was finally extracted through one of the trocar ports. Adrenalectomy was accomplished in all the animals without intraoperative mortality. Complications included splenic injury with prolonged bleeding and difficulty in performing the diaphragmatic suture (one case each). Through this approach direct and rapid exposure of the left adrenal gland was allowed, and adrenalectomy was accomplished in all the animals. These results suggest further investigations of the clinical application of this procedure. Key words: Adrenal gland Endoscopic adrenalectomy Thoracoscopy

of videoscopy in the management of many surgical pathologies, allowing adrenalectomy to be performed either through a laparoscopic [4, 5, 8, 9, 11] or a retroperitoneal endoscopic approach [1]. Mack et al. [7] recently reported on a clinical thoracoscopic transdiaphragmatic adrenal biopsy for metastatic carcinoma. Despite this, thoracoscopic adrenalectomy has never been described. Here we report on an investigation into the possibility of performing a left adrenalectomy through a thoracoscopic transdiaphragmatic approach in a swine model.

Materials and methods

Five pigs (mean weight 20 kg) were anesthetized with 0.5 ml/kg of a sodium penthobarbital solution intravenously. All animals received human care in compliance with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85-23, 1985). The study was approved by the Institutional Animal Care and Use Committee. After single-lumen tracheal intubation, ventilation was carried out at a tidal volume of 15 ml/kg and a rate of 15 cycles/min with a model 608 Harvard ventilator (South Natick, MA). Four 10-mm laparoscopic trocars were inserted through the chest into the left pleural cavity as illustrated in Fig. 1. A left pneumothorax was accomplished to collapse the left lung using CO2 insufflation at a mean pressure of 10 mmHg via one of the predisposed trocars. A 0 telescope was then inserted, followed by other instrumentation. A peripheral posterior phrenotomy was made starting from the lateral side of the aorta at the hiatus level and extended for about 6 cm. Once the diaphragmatic fibers divided, the retroperitoneal space was disclosed. Interestingly this maneuver was facilitated by the presence of positive intrapleural pressure due to CO2 insufflation. In fact, this created a pneumoretroperitoneum which resulted in a wider exposure of the surgical area. Through this approach Gerotas fascia was then opened. The left adrenal gland was easily recognizable due to its yellowish color, which contrasted with the smooth, dark red surface of the kidney (Fig. 2). The spleen, which in the pig is longer and more medial than in humans, needed to be gently retracted medially. Thus the pancreatic tail and the splenic artery were disclosed and made accessible on the medial side of the field. Afterward the gland was progressively dissected downward, following the correct plane between the gland and the kidney; the small tributary vessels were interrupted with endoscopic clips and divided. Finally, the inferior pole of the adrenal gland was gradually freed from the renal vein by blunt dissection, which was continued until the central adrenal vein was visualized and isolated. The

Numerous surgical approaches to the adrenal gland have been proposed, including anterior, posterior, lumbar, and thoracoabdominal access [6]. However, recent advances in endoscopic surgical instrumentation have expanded the role

Correspondence to: T. Lerut

391

Fig. 1. Position of all four endoscopic ports.

Fig. 3. Adrenal vessels are interrupted with endoscopic clips and divided.

Fig. 4. Diaphragmatic suture is performed by running endosuture technique.

Fig. 2. Transdiaphragmatic access to Gerotas fascia, which is incised. The spleen is retracted medially.

Discussion The importance of and indications for adrenal surgery have increased during this century as a better understanding of adrenal pathophysiology has evolved. The recent development of endoscopic laparoscopic surgery made it possible to perform, via this approach, adrenalectomy [4, 5, 8, 9, 11] and even nephrectomy [2, 3]. Laparoscopic adrenalectomy was reported on the first time in 1992 by Suzuki and Kawabe [11] and by Gagner et al. [4], and subsequently by others [5, 9, 11]. Brunt et al. [1] described their method of retroperitoneal endoscopic adrenalectomy in an animal model. Mack et al. [7] described in 1993 the first clinical thoracoscopic adrenal biopsy performed for suspicion of nonsmall-cell-lung-cancer (NSCLC) adrenal metastasis. Stimulated by this report, we investigated the possibility of performing a transdiaphragmatic thoracoscopic adrenalectomy in a swine model. Indeed, the use of a posterior transpleural approach for adrenal open surgery is not new, as it was proposed by Young [13] more than 50 years ago while searching to find a useful application in selected patients undergoing adrenalectomy. Endoscopic laparoscopic adrenalectomy has been performed in cases of primary aldosteronism, Cushings syndrome, pheochromocytoma, and even NSCLC metastasis [4, 8, 9]. An additional indication could be small nonfunctioning adrenal adenoma [11]. The same indications have been postulated for the Young approach

adrenal vein was then interrupted with endoscopic clips and divided (Fig. 3). Finally, adrenalectomy was completed by dividing the remaining peritoneal adhesions, and the gland was pulled out of the chest through one of the trocars. After accurate revision of the hemostasis, the diaphragmatic defect was sutured with a running nonabsorbable 0/0 suture utilizing endosuture technique (Fig. 4). At this point the left lung was reexpanded by cessation of CO2 inflation and the operative time was calculated.

Results Adrenalectomy could be performed in all the pigs. The operative time ranged from 150 to 215 min (mean 165 min). The blood loss ranged from 30 to 200 ml (mean 76 ml). Prolonged operative time in two cases was, respectively, due to splenic injury in one case and difficult diaphragmatic suturing in the other. In particular, in the former case the spleen was damaged by a lung endorectractor used to retract the spleen medially, and the resulting bleeding was only partially controlled. In the latter case the difficulty was related to a too-peripheral phrenotomy which resulted in an insufficient muscular layer at the external edge of the suture line. Despite this, all the pigs survived the surgical procedure, remaining alive until the operation was interrupted. The mean weight of the excised adrenal glands was 1.5 g; the average size of 23 10 4 mm.

392

and could be applied to this thoracoscopic approach. In this respect, it has indeed been reported that removal of NSCLC metastatic adrenal gland can be indicated in selected cases, resulting in acceptable long-term survival [10, 12]. This experimental study demonstrated that left adrenalectomy was feasible in all the cases through the transdiaphragmatic thoracoscopic approach. Some difficulties relate to the necessity of an adequate and atraumatic retraction of the spleen and to suturing the diaphragm using endosuture technique. For this reason, adequate training in performing endoscopic sutures is mandatory. Interestingly, the use of CO2 insufflation of the pleural cavity to about 10 mmHg resulted in a double advantage. It was possible to perform the entire procedure without double-lumen intubation, and once the diaphragm opened, intrapleural positive pressure created a useful pneumoretroperitoneum which facilitated the subsequent dissection. As emerged in a previous report, through the laparoscopic approach adrenalectomy is more demanding to the left side than to the right one [3]; moreover, difficulties can be related to the need for extensive mobilization and retraction of overlying intraabdominal organs to reach the retroperitoneum from the peritoneal cavity. The retroperitoneal endoscopic approach allows a more direct access to the adrenals but presents some disadvantages: (1) The working space in the retroperitoneum is small and limits the number of ports one can insert as well as the dissection angle. (2) There is the potential for injury to the kidney during initial insertion of the Veress needle and the first trocar [1]. Thoracoscopy, too, allows direct and rapid exposure of the left gland and results in a wider operative field related to the presence of the nonvirtual pleural cavity. In addition, left adrenalectomy could be performed by this approach even in patients who previously had abdominal surgery, which generally contraindicates laparoscopic surgery. Disadvantages relate to (1) the unilateral access to the left adrenal gland; (2) the use of a transpleural, transdiaphragmatic route, and (3) necessity of suturing the diaphragmatic defect at the end of the procedure. The theoretical

disadvantage of double-lumen intubation necessity could be avoided by using CO2 insufflation of the pleural cavity. In conclusion, transdiaphragmatic thoracoscopic left adrenalectomy was easily accomplished in this experimental study, suggesting further evaluation of the clinical application of the procedure. However, further investigation will be required to elucidate advantages and disadvantages over the laparoscopic approach and to provide incidental indications. References
1. Brunt LM, Molmenti EP, Kerbl K, Soper NJ, Stone AM, Clayman RV (1993) Surg Laparosc Endosc 3: 300306 2. Clayman RV et al. (1991) Laparoscopic nephrectomy: initial case report. J Urol 146: 278282 3. Coptcoat M et al. (1992) Laparoscopic nephrectomy: the Kinks clinical experience. J Urol 6: 127133 4. Gagner M, Lacroix A, Bolte E (1992) Laparoscopic adrenalectomy in Cushings syndrome and pheochromocytoma. N Engl J Med 327: 1033 5. Go H, Takeda M, Takahashi H, Imai T, Tsutsui T, Mizusawa T, Nishiyama T, Morishita H, Nakajima Y, Sato S (1993) Laparoscopic adrenalectomy for primary aldosteronism: a new operative method. J Laparoendosc Surg 3: 455459 6. Guz BV, Straffon RA, Novick AC (1989) Operative approaches to the adrenal gland. Urol Clin North Am 16: 527534 7. Mack MJ, Aronoff RJ, Acuff TE, Ryan WH (1993) Thoracoscopic transdiaphragmatic approach for adrenal biopsy. Ann Thorac Surg 55: 772773 8. Meurisse M, Joris J, Hamoire E, Hubert B, Charlier C (1995) Laparoscopic removal of pheochromocytoma. Why? When? and Who? Surg Endosc 9: 431436 9. Miccoli P, Iacconi P, Conte M, Goletti, Buccianti P (1995) Laparoscopic adrenalectomy. J Laparoendosc Surg 5: 221225 10. Reyes L, Parvez Z, Nemoto T, Regal A-M, Takita H (1990) Adrenalectomy for adrenal metastasis from lung carcinoma. J Surg Oncol 44: 3234 11. Suzuki K, Kawabe K (1992) Laparoscopic adrenalectomy. Urol Surg 5: 753758 12. Twomey P, Montgomery C, Clark O (1982) Successful treatment of adrenal metastases from large-cell carcinoma of the lung. JAMA 248: 581583 13. Young HH (1936) A technique for simultaneous exposure and operation on the adrenals. Surg Gynecol Obstet 54: 179

Surg Endosc (1997) 11: 385386

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Management of obliterating stricture after coloanal anastomosis

P. Reissman, J. J. Nogueras, S. D. Wexner
Department of Colorectal Surgery, Cleveland Clinic Florida, 3000 W. Cypress Creek Road, Fort Lauderdale, FL 33309, USA Received: 20 June 1995/Accepted: 11 October 1995

Abstract. We present an unconventional approach to the management of a severe stricture with complete luminal obliteration after a coloanal anastomosis which was protected with a diverting loop ileostomy. The colonoscope was inserted in an antegrade fashion into the defunctionalized limb of the loop ileostomy and advanced to the level of the stricture. Under colonoscopic vision, a Kelly clamp was carefully introduced transanally through the stricture into the proximal lumen. The strictured anastomosis was then dilated with calibrated Hegar dilators. Periodic dilatations followed by closure of the ileostomy completed the management. The technique obviated the need for a more extensive surgical procedure. Key words: Coloanal anastomosis Anastomotic stricture

Anastomotic strictures after coloanal anastomosis, stapled or handsewn, are not uncommon [7, 9]. However, in the majority of cases, serial dilatation is sufficient and surgical revision is rarely required. We present a case of anastomotic stricture with complete obliteration of the lumen after a handsewn coloanal anastomosis for a recurrent villous tumor of the rectum. Due to the complete obliteration, blind manual dilation was unsafe. Therefore, antegrade colonoscopy with dilation under direct vision was performed. The case and technique are presented.

a complete obliteration of the anastomosis was noted. Gentle manual manipulation was unsuccessful, as well as an attempt to insert a 21-gauge soft catheter to obtain a contrast study. More forceful attempts with blind insertion of an instrument or dilator seemed unsafe. The patient underwent an evaluation under anesthesia in the lithotomy position which confirmed the finding of a complete obliteration of the anastomosis. Subsequently, while the patient was under general anesthesia, an antegrade colonoscopy through the distal limb of the loop ileostomy to the level of the stricture was performed (Fig. 1). The colonoscope light source was identified by anoscopy and provided a landmark of the patent lumen proximal to the site of obliteration. At this point, under direct visualization of the stricture by the colonoscope on a video monitor, a blunt-tip Kelly clamp was gently introduced transanally and directed toward the stricture by the perineal surgeon who was observing the video monitor. The Kelly clamp was passed through the stricture into the proximal lumen (Fig. 1). The stricture was dilated using the clamp, followed by digital examination (Fig. 2) and then by serial Hegar dilators. Close inspection of the stricture site revealed an intact rectal wall. Subsequent management consisted of periodic dilatations wtih Hegar dilators, a gastrograffin enema which excluded extravasation, and closure of the ileostomy. The patient had an uneventful recovery and 6 months after the closure of the ileostomy he continues to have spontaneous bowel evacuations.

Comment Anastomotic strictures are associated with postoperative morbidity and are observed in 25% of colonic anastomosis [8]. The etiology of anastomotic stricture after coloanal anatomosis is multifactorial. The case of severely strictured anastomosis with luminal obliteration is unusual and likely occurred as a result of delayed postoperative examination which would routinely take place 46 weeks after the initial procedure. At the postoperative visit, a gentle digital rectal examination is performed, thereby dilating the anastomosis prior to the closure of the ileostomy. Subsequently, stool passage through the anastomosis provides repeated dilation. However, in the case presented here, due to cardiac disorders, the patients follow up was delayed for 4 months and rectal examination was not performed during this period. The management of anastomotic strictures requires knowledge of the length of the strictures and the luminal direction. Water-soluble contrast enemas are a safe approach to providing an image of the strictures. In this case, direct colonoscopic evaluation of a stricture provides lim-

Technique
A 67-year-old male patient underwent a restorative proctectomy with a handsewn coloanal anastomosis and diverting loop ileostomy for a recurrent villous tumor. The anastomosis was performed transanally at the dentate line using a single layer of interrupted 3.0 Vicryl sutures. The immediate postoperative course was unremarkable. However, after discharge, the patient developed bradyarrhythmias which necessitated insertion of a permanent pacemaker. Therefore, the first postoperative office visit and rectal examination were 4 months after the initial procedure. At that time,

Correspondence to: J. J. Nogueras

386

Fig. 1. Schematic illustration of the technique. The colonoscope is introduced through the distal loop of the ileostomy or colostomy and is advanced to the stricture. Under the colonoscopic visualization and control, a blunt-tip Kelly is introduced transanally through the stricture into the proximal colon. Dilation of the stricture is subsequently performed. Fig. 2. Colonoscopic view of the digital examination performed immediately after the initial instrumental dilation.

ited information, since an adult colonoscope is unable to transverse colonic strictures in up to 54% of cases [5]. In such cases, pediatric endoscopes with smaller luminal diameters may be more successful in passing through the stricture [1]. There are a variety of techniques available for the dilation of anastomotic strictures. These include balloon dilation utilizing transendoscopic and nontransendoscopic placement or over a wire for combined techniques [3]. Balloon dilation can be successful in up to 75% of patients. This technique requires placement of the balloon intraluminally within the stricture [6]. Transendoscopic placement of the balloon dilators may also be performed over a guidewire and under fluoroscopic control. DeLange and Shaffer describe a technique of wireguided balloon dilatation without the use of the colonoscope [4]. A guidewire is advanced to a point proximal to the stricture under fluoroscopic control, allowing for placement of a balloon for dilation. This technique is useful for rectal and low anastomotic strictures. Bedogni et al. described a double endoscopy technique for colonic strictures distal to a colostomy [2]. These authors describe a 96% success rate with this technique. Other options in the management of anastomotic strictures include rigid dilations, electrocautery incisions, laser photoablation, and surgical revision. Severe anastomotic stricture with complete obliteration of the lumen after coloanal procedures will commonly require surgical intervention for reconstruction of the anastomosis. Blind insertion of dilators is unsafe in such cases as a false root with perforation of the rectal wall may occur. However, as an alternative to surgical intervention, we employed the technique of antegrade colonoscopy through the defunctionalized limb of a loose ileostomy in order to visualize the stricture through the proximal rectal lumen. This enabled

controlled dilation of the stricture by transanal insertion of a blunt instrument followed by dilators. This procedure is feasible since the majority of patients who undergo a coloanal anastomosis will also have a temporary protecting ileostomy or colostomy which may be used for the antegrade insertion of the colonoscope. This simple technique may obviate the need for a more extensive surgical procedure.

References
1. Bat Z, Williams CB (1989) Usefulness of pediatric colonoscopes in adult colonoscopy. Gastrointest Endosc 35: 329332 2. Bedogni G, Ricci E, Pedrazzoli C, Conigliaro R, Barbieri I, Bertoni G, Contini S, Serafini G. (1987) Endoscopic dilation of anastomotic colonic stenosis by different techniques: an alternative to surgery. Gastrointest Endosc 33: 2123 3. Bernstein D, Marten HD (1994) Endoscopic management of colonic strictures: technology and follow up. In: Barkin JS, OPhelan CA (eds) Advanced Therapeutic Endoscopy (2nd ed). Raven Press, New York, pp 231241 4. Delange EE, Shaffer HA (1991) Rectal strictures: treatment with fluorscopically guided balloon dilation. Radiology 178: 475479 5. Forde KA, Treat MR (1985) Colonoscopy in the evaluation of strictures. Dis Colon Rectum 28: 699701 6. Kozarek RA (1986) Hydrostatic balloon dilation of gastrointestinal stenosis: a rational survey. Gastrointest Endosc 32: 1520 7. Luchtefeld MA, Milsom JW, Senagore A, Surrell JA, Mazier WP (1989) Colorectal anastomotic stenosis: results of a survey of the ASCRS membership. Dis Colon Rectum 32: 7336 8. Oz MC, Forde KA (1992) Endoscopic treatment of colonic stricture. In: Cameron JL (ed) Current surgical therapy, 4th ed. Mosby-Yearbook, St Louis, pp 493496 9. Venkatesh KS, Ramanujam PS, McGee S (1996) Hydrostatic balloon dilation of benign colonic anastomotic strictures. Dis Colon Rectum 35: 78991

Surg Endosc (1997) 11: 366370

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The efficacy of laparoscopic surgery in the diagnosis and treatment of peritonitis

Experience with 107 cases in Mexico City
J. Cueto,1,2 O. D az,1 D. Garteiz,1 M. Rodr guez,1 A. Weber1
1 2

Department of Surgery, American British Cowdray Hospital, Sur 138 esq. Av. Observatorio, Col Ame ricas, 01120 Me xico Department of Surgery, Louisiana State University, 1542 Tulane Avenue, New Orleans, LA 70112, USA

Received: 15 March 1996/Accepted: 29 August 1996

Abstract Background: Peritonitis continues to be an important cause of morbidity and mortality and often an etiologic diagnosis is unclear. To evaluate the efficacy and safety of laparoscopy the authors analyzed their 5-year experience with this modality of treatment. Methods: A review was made of 107 consecutive nonselected laparoscopic procedures performed between October 1990 and November 1995. The diagnosis was established by clinical, laboratory, and imaging findings and confirmed by laparoscopy and/or laparotomy. Results: An etiologic diagnosis was unclear in 35% of the cases and was established in all by laparoscopy; 94 patients (87.9%) were successfully treated by laparoscopy while 13 (12.1%) required conversion. Mortality was 4.6%; 14% had postoperative complications and 7.4% had reoperations. Conclusions: Laparoscopic surgery is safe and very efficient in the diagnosis and treatment of patients with peritonitis. In most instances a definitive treatment can be carried out without conversion and has the additional and wellknown advantages of minimally invasive surgery. Key words: Peritonitis Morbidity Mortality

since the 1950s [18, 19], the use of this method for therapeutic purposes is very recent. Surgeons are now performing diagnostic and therapeutic laparoscopic procedures in patients with acute cholecystitis, appendicitis, perforated peptic ulcer, intestinal obstruction, perforated diverticuli, and pelvic disorders with similar good results [58, 16]. The following report evaluates our experience in 107 consecutive laparoscopic procedures in patients with acute peritonitis.

Materials and methods

Between October 1990 and November 1995, the authors performed 107 consecutive, nonselected laparoscopic procedures in patients with acute peritonitis (excluding acute noncomplicated appendicitis and cholecystitis). Patients were treated at the American British Cowdray Hospital, Spanish Hospital, National Institute of Perinatology, and other hospitals in Mexico City and elsewhere. Seventy female and 37 male patients aged 8 to 92 (mean 50) are reported. In all cases, the acute abdominal syndrome was established by clinical, laboratory, and imaging findings and confirmed by laparoscopy (Table 1). During the preoperative period patients were evaluated and treated aggressively to correct electrolyte, acidbase, and metabolic disturbances. All received appropriate intravenous antibiotic therapy. Patients 40 years or older were evaluated by a cardiopulmonary specialist before surgery. All patients were hemodynamically stable before surgery and were operated under endotracheal general anesthesia with pulse oximetry and capnography monitoring. In some instances, additional invasive measures such as Swan-Ganz and direct arterial catheters were required. Pneumoperitoneum was established by transumbilical puncture with a Veress needle in 80% of the patients. Five percent required extraumbilical needle placement (mainly the left upper quadrant) due to previous periumbilical surgical scars and suspected adhesions. The remaining 15% had severe abdominal distension and required an open technique for the induction of pneumoperitoneum [5, 16, 32]. Digital lysis of adhesions through the first port was necessary in several instances. In all patients, pneumoperitoneum pressure was kept at an average of 810 mmHg throughout the procedure. Some cases required placement of additional ports to allow introduction of irrigation and aspiration cannulas, atraumatic retractors, etc.

Peritonitis continues to be an important cause of morbidity and mortality and is commonly associated with acute disease of the abdominal and pelvic viscera [1, 5, 14, 16, 19]. Although the diagnostic value of laparoscopy in patients with an acute abdominal syndrome has been recognized

Correspondence to: J. Cueto, Teapa 4, Col. Lomas de Chapultepec, 11000 Me xico, D.F., Me xico

367 Table 1. Causes of acute peritonitis Diagnosis Ruptured appendix Intestinal obstruction Perforated cholecystitis Biliary pancreatitis Pelvic disorders Blunt abdominal trauma Perforated Diverticulitis Perforated peptic ulcer Penetrating abdominal trauma Residual abscess Total
a

Table 2. Mortality Number of cases 20a 20b 15c 13 13d 12 6 4e 2 2 107 Sex Female Male Female Male Female Age 91 75 72 71 28 Diagnosis Intestinal obstruction Acute necrotizing pancreatitis Intestinal obstruction Cholangitis Pelvic inflammatory disease Cause of death Myocardial infarction (11th postoperative day) Multiple organ failure Sepsis and respiratory failure (14th postoperative day) Multiple organ failure Endotoxic shock (1 month after surgery)

Four cases with periappendiceal abscess. Acute noncomplicated appendicitis excluded b Including one case of sigmoid volvulus c One case with abscess. Acute noncomplicated cholecystitis excluded d Including ruptured ectopic pregnancies, ovarian disorders, pelvic inflammatory disease, and one ruptured uterus due to carcinoma e One with subphrenic abscess

sepsis, one for a pancreatic debridement and drainage (laparoscopic), and one laparotomy which turned out to be negative. Discussion The results reported by Llanio and Sarle, among others [18, 19], in the diagnosis of acute peritonitis in an emergency room setting are well known. It has been demonstrated that laparoscopy under local anesthesia in some instances allows an adequate visualization of the entire abdominal and pelvic cavities, providing a diagnosis that would not be possible with other methods such as radiology or conventional laboratory tests. In our experience, this minimally invasive technique established a precise diagnosis in 100% of the cases. Laparoscopy seems to be a safe technique for evaluation of patients with acute abdominal syndrome regardless of the affected organ, the presence of adhesions, or the extent of the diffuse peritonitis. Given the excellent exposure and other benefits of laparoscopy [12, 20, 24] it is not surprising to find that many groups are beginning to report their results in the treatment of acute peritonitis. In our series, definitive treatment was possible by laparoscopy in 87.9% of patients with excellent postoperative results, such as those recently reported by Geis [16]. The well known advantages of this method, such as reduced postoperative pain, reduced ileus, reduced hospital stay, and faster return to normal life activities, are also observed in cases of acute peritonitis. In addition, there is a decreased incidence of incisional infections and pulmonary complications which are common in patients with septic abdominal disease that require long surgical incisions. In these patients, the reduced traumatic effect of the surgical procedure and the possible conservation of a better immune response are critical for their recovery [30]. The role of laparoscopy has also been evaluated in critical-care patients for the diagnosis of abdominal conditions in which conventional surgery represents an elevated risk [25, 27]. For the evaluation and treatment of all patients we followed a simple algorithm, shown in Fig. 1. In some instances of severe bowel loop dilation, it was impossible to obtain adequate retraction and exposure even after the introduction of 10-mm fan-type atraumatic retractors. In two patients with perforated colonic diverticulae, early in this experience it was very difficult to perform the dissection and achieve adequate hemostasis with a very thickened edematous mesocolon, and rather than incur in an undue risk, a laparotomy was done.

Results As mentioned before, the precise cause of acute peritonitis was unclear before surgery in 35% of the cases. Differential diagnosis was especially difficult in female patients with right lower quadrant pain and in the elderly. In all of these cases, diagnosis was finally established through laparoscopy, and our findings were usually acute appendicitis, pelvic inflammatory disease, intestinal obstruction, and one case of pancreatitis. In spite of multiple adhesions and generalized peritonitis in many cases, a complete and adequate exploration of the entire abdominal and pelvic cavity was possible in every instance. Ninety-four patients (87.9%) were successfully treated by laparoscopy while 13 (12.1%) required conversion to laparotomy. Of the cases that required conversion, eight were due to technical difficulties and equipment failure (mostly in the first 20 procedures of our series), two to perforated colonic diverticuli, one to an injury of an intestinal loop, and one to uncontrollable bleeding. The hysterectomy also required a formal laparotomy. There were two transoperative lacerations of epigastric vessels which were controlled without the need for conversion. The mortality rate was 4.6% (5 patients) (Table 2). A 72-year old female patient with severe respiratory insufficiency (basal pO2 of 39 mmHg) developed two distal ileal well-circumscribed, round, 1-cm perforations on the 3rd postoperative day following laparoscopic lysis of adhesions with an internal volvulus. A 91-year-old female died of myocardial infarction 11 days after surgery, two patients died of multiple organ failure secondary to sepsis, and one died of endotoxic shock 1 month later. There were postoperative complications in 14% (15 patients); one documented pulmonary embolism, seven postoperative prolonged ileus, two cases of pneumonia, four with umbilical cellulitis, and one acute umbilical herniation (epiplocele) corrected with local anesthesia in the emergency room. Eight cases (7.4%) required reoperation. Three were residual abscesses (all originated from diffuse peritonitis due to perforated appendix)three for persistent abdominal

368

Fig. 1. Algorithm for the evaluation and treatment of patients with acute abdominal syndrome.

In cases that required conversion to laparotomy, the incision performed was well planned and smaller than usual. When a stoma was required, such as in cases of bowel resection, one of the trocar ports was slightly extended and used for formation of the stoma [6, 23]. Of the five patients who died, the 28-year-old female was admitted from another hospital 22 days after what was called a normal vaginal delivery with septicemia from a pelvic suppuration, and a gynecological team performed a total abdominal panhysterectomy. She was taken to the intensive care unit; her condition improved gradually and the septic process appeared to be under control. Twelve days later she developed upper abdominal pain, nausea, and vomiting, and an ultrasound confirmed the clinical diagnosis of acute cholecystitis. She was taken to the operating room where a laparoscopic cholecystectomy was done in 40 minutes, uneventfully. There was no leakage of the pneumoperitoneum (810 mmHg) in the infraumbilical vertical incision. Her clinical condition improved again and she was transferred 3 days later to the surgical floor. Unfortunately, 4 days later she again developed signs and symptoms of pelvic suppuration, and after several days of different schemes of intravenous antibiotic therapy based on blood and vaginal cultures for Staphylococcus aureus, she was reoperated by the gynecological team and an infected pelvic hematoma was drained, but she continued to deteriorate and died 1 month after her hysterectomy. A postmortem study could not be done. The patient with acute necrotizing pancreatitis tolerated the cholecystectomy, biliary drainage, and peritoneal irrigation well but developed signs of sepsis 10 days later. At the

reoperation (laparoscopy) a debridement and drainage procedure was done which the patient tolerated well. After 7 days of intensive care that included total parenteral nutrition, intravenous antibiotic therapy, etc., the patient was transferred to the surgical floor afebrile, and gradually was able to tolerate liquids, had normal bowel function, etc., but unexpectedly had a cardiorespiratory arrest and died in his room. A postmortem study could not be done and the probable causes of death were pulmonary thromboembolism and/or subclinical sepsis. A 91-year-old female had undergone lysis of adhesions causing acute small-bowel obstruction and was discharged on the 5th postoperative day with normal bowel movements, taking a bland diet. She was readmitted with severe chest pain to the emergency room on the 11th postoperative day and expired in the coronary care unit the same day with a proven acute myocardial infarction. A 71-year-old male was operated 2 days after his admission to the hospital with acute cholecystitis, cholangitis, and jaundice (of 25 mg %), choluria, and acholia of 3 weeks duration. A retrograde cholangiography could not be done since he had had a B-11 gastrectomy 15 years before. Once the coagulation parameters were corrected, a transhepatic percutaneous cholangiography was performed which revealed a lithiasic obstruction with massive dilation of the intra- and extrahepatic biliary ductal system. He was taken to the operating room and the laparoscopic procedure was converted to laparotomy because of technical difficulties. A cholecystectomy and common bile duct exploration were done and a T tube was inserted, but his condition deteriorated steadily and he died with hepatorenal and respiratory insufficiency 7 days later. A 72-year-old female was referred by a gastroenterologist with the diagnosis of small-bowel obstruction and a laparoscopic adherecniolysis was done uneventfully. On the 3rd postoperative day she developed signs of an acute abdomen and severe respiratory insufficiency. A formal laparotomy was done and two round, well-circumscribed distal ileal perforations were found. Both were sutured (instead of performing a resection) since the bowel appeared in good condition and there was no evidence of obstruction. She was taking to the intensive care unit where she remained intubated. Her condition deteriorated and she died on the 14th postoperative day. In analyzing the demise of these five patients, it appears that the laparoscopic procedure could be directly related to a lethal complication only in the patient with postoperative distal ileal perforations, although a careful review of the videotape doesnt show any instance during which the bipolar electrocautery and/or any other surgical instrument could have produced the injury to the bowel wall. Nevertheless this is the most likely explanation for this known complication [24] that occurs in conventional surgery, too. Her chronic, severe respiratory insufficiency due to heavy smoking for over 50 years very likely was another contributing factor for her death. Although our experience in acute peritonitis by the laparascopic approach is limited, there are some aspects of this surgical approach that require special mention: 1. Induction of pneumoperitoneum and insertion of the first trocar is a blind procedure that entails the risk of injury

369

shown, to this date there is no evidence that HIV can be transmitted in this way [2, 3]. We recommend evacuating the smoke and pneumoperitoneum with a suctionirrigation system. Based on these results and those of other groups, it must be stated that therapeutic procedures in patients with acute peritonitis should only be carried out by surgeons with extensive laparoscopic experience, for they require dexterity and the ability to perform two-handed dissection, suturing, and other techniques. A good example of this is the thorough exploration of the small and large intestines required in patients with bowel obstruction. The bowel must be explored, in its full length and manipulation must be done gently with adequate instruments to avoid lacerations. Whatever the procedure being carried out, it will always be the surgeons expertise and judgment that will determine up to what point laparoscopy should, or should not, be continued. References
1. Berci G (1993) Elective and emergency laparoscopy. World J Surg 17: 815 2. Bouillot J, Delmi N, Kazatchkine M, Fernandez F, Pikketti C, Alexandre JH (1995) Role of laparoscopic surgery in the management of acute abdomen in the HIV positive patients. J. Laparoendosc Surg 5: 101104 3. Carroll BJ, Rosenthal RJ, Phillips EH, Bonet H (1995) Complications of laparoscopic cholecystectomy in HIV and AIDS patients. Surg Endosc 9: 874878 4. Conlon KC, Dougherty E, Klunstra D, Coit DG, Turnbull Ad, Brennan MF (1996) The value of minimal access surgery in the staging of patients with potentially resectable peripancreatic malignancy. Ann Surg 223: 134140 5. Cueto GJ, Rojas DO, Weber SA (1994) Laparoscopia como me todo diagno stico y terape utico de la peritonitis. In: Cueto GJ, Weber SA (eds) Cirug a laparosco pica. Interamericana-McGraw Hill, Mexico City, 181183 6. Cueto GJ, Melgoza OC, Rojas DO, Serrano BF, Weber SA (1994) Laparoscopic colectomy for cancer. Report of two cases. J Exp Clin Cancer Res 13: 409413 7. Cueto GJ, Rojas DO, Weber SA (1993) Es u til la laparoscopia en el diagno stico y terape utica del s ndrome abdominal agudo? Rev Gastroenterol Mex 58: 360 8. Cueto GJ, Serrano BF, Weber SA (1993) The value of laparoscopy in the diagnosis and treatment of acute peritonitis. Poster session. First European Congress. E.A.E.S. June 35, 1993, Cologne, Germany 9. Cueto GJ, Weber SA, Serrano BF (1993) Laparoscopic treatment of perforated ulcer. Surg Laparosc Endosc 3: 216218 10. Cuschieri A, Jakimowicz, Spreenwel J (1996) Laparoscopic distal 70% pancreatectomy and splenectomy for chronic pancreatitis. Ann Surg 223: 280285 11. Cuttat JF (1994) Laser surgery versus electrosurgery: point and counterpoint. Chapter 16. In: Steichen FM, Welter R (eds) Minimally invasive surgery and new technology. Quality Medical, St Louis, MO, pp 94102 12. Easter D, Cuschieri A (1992) The utility of diagnostic laparoscopy for abdominal disorders. Arch Surg 127: 379383 13. Eubanks S, Newman L, Lucas G (1993) Reduction of HIV transmission during laparoscopic procedures. Surg Laparosc Endosc 3: 25 14. Forde K, Treat M (1992) The role of peritoneoscopy (laparoscopy) in the evaluation of the acute abdomen in critically ill patients. Surg Endosc 6: 219221 15. Fry D (1993) Reduction of HIV transmission during laparoscopic procedures. Surg Laparosc Endosc 3: 1 16. Geis WP, Kim HC (1995) Use of laparoscopy in the diagnosis and treatment of patients with surgical abdominal sepsis. Surg Endosc 9: 178182 17. Kim HB, Gregor M, Boley S, Cleinhaus S (1993) Digitally assisted

Fig. 2. Visualization, aspiration, and irrigation of purulent material and fibrinous adhesions.

2.

3.

4.

5.

6.

7. 8.

of intraabdominal organs, especially in patients with distended intestinal loops. A safe alternative is the use of an open technique and the placement of a Hassans cannula [5, 16]. When the open approach has been selected, a digital exam of the periumbilical area and careful lysis of surrounding adhesions should be done. In patients with periumbilical scars due to previous surgery we recommend placement of the initial puncture at an extraumbilical site, preferably the left upper quadrant, to avoid accidental organ injury. The abdominal and pelvic cavities can be explored in such detail with the laparoscopic approach that it can be said that this method is better when compared to the one performed through small incisions like the ones used for appendectomy, etc. It is important to note that although initially retroperitoneal exploration was contraindicated by some groups [25, 27] it is routinely employed by others now [4, 10, 31]. In cases of unclear preoperative diagnosis, laparoscopy (even with local anesthesia and in an emergency room setting) can shorten the observation period and avoid the need for expensive laboratory and imaging tests. When using electrocautery, it must be emphasized that bipolar instruments are preferable when compared to monopolar because of the risk of distant injury and perforation with the latter [11, 21, 22, 29]. Visualization and aspiration of purulent material, blood, bile, or intestinal fluids is much more precise with laparoscopy. Irrigation under pressure allows better debridement of abscesses and fibrinous formations (Fig. 2). This method is superior to the customary flooding of the surgical field with solution in conventional surgery [5, 9, 16, 28]. When conversion to laparotomy is required, the surgical incision can be accurately placed. Finally, two possible additional advantages of the laparoscopic approach are the reduced immunosuppression observed in response to the surgical trauma, and recently, Fry [15] reported a reduced risk of transmission of viral infections such as HIV in surgeons and medical personnel [13], a point also made by Bouillot et al. [2]. On the other hand, although the expulsion of live particles during the evacuation of pneumoperitoneum has been

370 laparoscopic drainage of multiple intraabdominal abscesses. J Laparoendosc Surg 5: 477479 Llanio R, Soto A, Ferret O, Gimenez G, Nordase O (1976) Diagnostic de labdomen aigu par laparoscopie. Experience portant 6400 cas. 3 European Congress of Gastrointestinal Endoscopy. Abstract. June July 1976; 11. Budapest O (ed) Llanio R, Sarle H (1956) Interet de la peritoneoscope chez politraumatises. Marseille Chirurg 8: 8286 Milhriro A, Sousa FC, Manso S, Leitao F (1994) Metabolic response to cholecystectomy: open versus laparoscopic approach. J Laparoendosc Surg 4: 311318 Nduka CC, Super PA, Monson JRT, Darzi AW (1994) Cause and prevention of electrosurgical injuries in laparoscopy. J Am Coll Surg 179: 161170 Reindenbach HF, Buess G (1992) Ancillary technology: electrocautery, thermocoagulation and laser. Chapter 4. In: Cushieri A, Buess G, Perissat J (eds) Operative manual of endoscopic surgery. SpringerVerlag Berlin, Heidelberg, pp 4460 Roe AM, Barlow AP, Durdey P, Elteingham WK, Spiner HJ (1994) Indications for laparoscopic formation of intestinal stomas. Surg Laparosc Endosc 4: 345347 Schrenk P, Woisetschlager R, Wayand W, Rieger R, Sulzbacher H (1994) Diagnostic laparoscopy: a survey of 92 patients. Am J Surg 168: 348351 Serrano BF, Weber SA, Cueto GJ (1994) Evaluacio n laparosco pica del paciente con traumatismo contuso del abdomen. In: Cueto GJ, Weber SA (eds) Cirug a laparosco pica. Interamericana-McGraw Hill, Mexico City, pp 185187 Silvester KG, Paskin DL, Schuricht AI (1994) Combinated laparoscopic-endoscopic gastrostomy. Surg Endosc 8: 10721075 Sosa JL, Markley M, Sleeman D, Puente V, Carrillo E (1995) Laparoscopy in abdominal gunshot wound. Surg Laparosc Endosc 5: 417 419 Sullivan GC, Murphy D, OBrien MG, Ireland A (1996) Laparoscopic management of generalized peritonitis due to perforated colonic diverticula. Am J Surg 171: 432434 Testas P (1994) The dangers of and correct procedures in laparoscopic electrosurgery. Chapter 17. In: Steichen FM, Welter R (eds) Minimally invasive surgery and new technology. Quality Medical, St. Louis, MO, pp 102104 Trokel M, Bessler J, Treat MR, Wekan RL, Nowygrog R (1994) Preservation of immune response after laparoscopy. Surg Endosc 8: 13851388 Watanable Y, Sato M, Kimura S (1993) A new endoscopic drainage laparoscopic technique of pancreatic abscesses. J Laparoendosc Surg 5: 489494 Weber SA, Serrano BF, Cueto GJ (1993) Te cnicas de neumoperitoneo. In: Sepu lveda A (ed) Cirag a laparosco pica. Ediciones Video Cirug a, Santiago de Chile, 1: 121126

18.

26. 27.

19. 20. 21. 22.

28.

29.

30.

23. 24. 25.

31.

32.

Erratum
Surg Endosc (1997) 11: 409410

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Due to an error in printing, the following two pages of abstracts were omitted from Surgical Endoscopy, Volume 11, Number 2. They are reprinted here in their entirety.

410

Surg Endosc (1997) 11: 336340

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Laparoscopic vs open appendectomy

A randomized clinical trial
G. Kazemier,1 G. R. de Zeeuw,1 J. F. Lange,2 W. C. J. Hop,3 H. J. Bonjer1
1 2

Department of General Surgery, University Hospital Dijkzigt, Dr Molewaterplein 40, 3015 GD Rotterdam, The Netherlands Department of General Surgery, St. Clara Hospital, Olympiaweg 350, 3078 HT Rotterdam, The Netherlands 3 Department of Biostatistics, Erasmus University, Dr Molewaterplein 50, 3015 GE Rotterdam, The Netherlands

Abstract Background: A randomized clinical trial was performed to compare open appendectomy (OA) and laparoscopic appendectomy (LA). Methods: 201 patients with similar characteristics of appendicitis were randomized to either OA or LA. Operative time and technique, reintroduction of diet, postoperative pain, use of analgesia, hospital stay, and complications were documented. Results: 104 patients were allocated to the OA group and 97 to the LA group. Postoperative pain was significantly less in the LA group on the 1st (p < 0.001) and 2nd (p < 0.001) postoperative day, resulting in less use of analgesics on both days (p < 0.001). Restoration of diet was similar in both groups. Mean operative time was longer in the LA group: 61 vs 41 min (p < 0.001). Postoperative complications did not differ in either group, except for wound infections (six OA group vs zero LA group, p < 0.05). Mean hospital stay was similar in both groups. Conclusions: LA results in less postoperative pain and fewer wound infections. The laparoscopic procedure is technically more demanding to perform, resulting in longer operative time. Key words: Appendicitis Appendectomy Laparoscopy Randomized

pneumoperitoneum in patients with peritonitis, the use of small incisions for open appendectomy (OA) and performance of appendectomy at nightly hours appear to have held back most surgeons from employing laparoscopic techniques in patients with appendiceal disease. However OA has several disadvantages. Exploration of the entire abdominal cavity is impossible through a McBurney musclesplitting incision, which increases the chance to miss, e.g., pelvic inflammatory disease, sigmoiditis, cholecystitis, and gastric perforation in patients with normal appendices. Due to the highly variable clinical picture of appendicitis, the rate of negative appendectomies is 2030% in adults [2]. Wound infection is a common complication following OA, occurring in 8.520% of all patients [12]. Laparoscopy allows thorough exploration of the peritoneal cavity and appears to be associated with less wound infections than open surgery, as LA is associated with minimal contact between the appendix and the skin. To answer questions on superiority and safety of LA compared to OA we performed a randomized clinical trial in our teaching hospitals focusing on operative time, postoperative pain, use of analgesics, restoration of diet, complications, and hospital stay.

Patients and methods

Preoperative course and trial entry Appendectomy is done annually in almost 16,000 patients in The Netherlands, accounting for 6% of all surgical procedures performed yearly [8]. Laparoscopic techniques to remove the appendix were described prior to laparoscopic cholecystectomy [5, 9, 16]. In spite of the high incidences of appendicitis, widespread employment of laparoscopic appendectomy (LA) did not follow. Reluctance to create a
Patients were asked to participate in this trial if a clinical diagnosis of acute appendicitis was most likely, if a McBurney muscle-splitting incision in the right lower quadrant was considered a suitable approach, and if they were suitable for laparoscopy and laparotomy. Trial entry started November 1, 1993, and ended in March 1996. Patients were excluded if they were pregnant, were under 11 years of age, or were thought unable to understand fully the aspects of randomization or were otherwise incapable of providing informed consent. Patients requiring elective appendectomy were also excluded. The study protocol was reviewed and approved by the ethics committees of both participating institutions. After the decision to operate had been taken and after fully informed consent was obtained, the type of procedure was determined for each

Correspondence to: G. Kazemier

337 patient by drawing a cardon which the type of surgery to be performed was writtenfrom an opaque envelope. Computer-generated blocked random numbers were used to assign LAP or OPEN to either card. Age, gender, and duration of symptoms in hours of every patient were recorded. Table 1. Clinical characteristics of patients in both groups LA (n97) Mean age in years (SD) Gender: male female Mean duration of symptoms in hours (SD) 30.8 (14.5) 52 (53.6%) 45 (46.4%) 39.8 (36.4) OA (n104) 33.7 (17.6) 59 (56.7%) 45 (43.3%) 36.3 (27.6)

Preoperative care and antibiotics

Surgery was done under general anesthesia. All patients received 1 g cefotaxime and 500 mg metronidazole intravenously at the time of induction. In the case of nonperforated appendicitis, antibiotics were not continued; in the case of perforated appendicitis, patients received 750 mg cefuroxime per 6 h and 500 mg metronidazole per 8 h intravenously until temperature remained below 37.5 C for 48 h, with a maximum of 5 days. All patients had a nasogastric tube during surgery. Patients did not receive a urinary catheter routinely.

Postoperative care and follow-up

Analgesia was recorded and prescribed on demand and consisted of 1 mg/kg pethidine, maximally every 6 h on the 1st day and 1 g of paracetamol, maximally every 6 h on the 2nd postoperative day. Postoperative pain was scored by each patient every 6 h postoperatively on a visual analogue scale (VAS) consisting of a 10-cm-long horizontal line without graduations varying from no pain at all on the left side to unbearable pain on the right side. Afterward, the VAS was scored by measuring the length in millimeters left of the patients mark. Reintroduction of a liquid diet was recorded and was defined as the first day on which over 1 l of fluid oral intake was tolerated. Reintroduction of a solid diet was defined as the 1st day on which a normal solid meal was tolerated. Hospital stay was recorded and defined as the number of postoperative days spent in hospital, including days spent in hospital after possible readmission because of causes related to the initial operation. Day 1 was defined as the day of operation. Postoperative complications were recorded both in the hospital and at follow-up. Wound infection was defined as edema and redness around any wound or purulent discharge. Ileus was defined as the inability to tolerate over 1 l of fluid on the 3 postoperative day. Pulmonary infection was defined as consolidation on a chest X-ray combined with positive cultures of sputum. Data were collected by two departmental secretaries and by surgical trainees not directly involved in the operations.

Operative care
All open or laparoscopic surgery was performed or supervised by surgeons or surgical trainees with experience of more than 15 open and laparoscopic appendectomies. Normal training practices were continued during the trial.

Open appendectomy. Open surgery was done through a 6-cm McBurney muscle-splitting incision in the right lower quadrant. The appendix was removed with ligation of the stump with an absorbable suture; the appendiceal stump was not buried routinely. The incision was extended if necessary. A normal appendix was always removed at open surgery. In that case, an attempt was made to visualize the right ovary and right fallopian tube in women and the distal 100 cm of ileum to detect a possible Meckels diverticula. Saline lavage was not performed routinely. Drainage tubes were not left in the abdominal cavity. The skin incision was closed with 3-0 Nylon (Ethilon; Ethicon, Sommerville, NJ, USA) unless a perforated appendicitis was found, in which case the skin wound was left open.

Statistical analysis
Means and standard deviations (SD) of variables were calculated and analyzed for both groups using Wilcoxon rank sum test and the Fisher exact test for proportions. A p value of <0.05 was considered significant. All analyses were performed on an intention-to-treat basis; this included analyses of data of patients whose laparoscopic operation had to be converted to an open operation.

Laparoscopic appendectomy. For LA the patient was in a supine position, with both surgeon and assistant on the left side and video monitor on the right side of the patient. The CO2 pneumoperitoneum was established by use of a Veress needle in the St. Clara Hospital and by use of an open technique and a Hassons trocar in the University Hospital Dijkzigt. The Veress needle and/or first trocar were placed below the umbilicus or in the left lower quadrant if an operative scar was present at the umbilicus. A 0 laparoscope was inserted at the umbilicus and two reusable canulas were introduced under direct vision: one 1012 mm trocar in the left lower quadrant laterally to the rectus muscle and one 5-mm trocar in the midline just above the pubic bone. The operation was performed with the operating table in Trendelenburg position, tilted 1020 to the left. The abdominal cavity was explored, and after the diagnosis of acute appendicitis had been confirmed or other diagnoses had been excluded, appendectomy was begun by dissection and division of the appendicular artery between clips or by electrocautery. The appendix stump was secured after division of the mesentery and divided between Chromic catgut loops (Ethibinder; Ethicon, Sommerville, NJ, USA). If the base of the appendix was heavily inflamed, an endoscopic linear stapling device (Endo-GIA-30; US Surgical Corp, Norwalk, CT, USA) was applied over the base of the caecum to resect the appendix safely. The stump of the appendix was never buried. The appendix was retrieved through the canula in the left lower quadrant or by use of a plastic bag (Endocatch; US Surgical Corp, Norwalk, CT, USA). A normal appendix was always removed, unless a definite other diagnosis responsible for the patients clinical course was found on laparoscopic exploration of the abdominal cavity. Lavage was performed routinely using 1 l of 0.9% saline solution if blood or purulent material was left after appendectomy or if blood obscured adequate vision. Drainage tubes were not left in the abdominal cavity. The skin incisions were closed in every case using 3-0 Nylon (Ethilon; Ethicon, Sommerville, NJ, USA). Operative time was taken as the time between the first incision and application of dressings to the wounds. Extension of the incision in open surgery or conversion from laparoscopic to open surgery was done at the surgeons discretion. All removed appendices were sent for histological examination.

Results Patients and observations during operation A total of 201 patients were enrolled in the trial. Ninetyseven patients (48.3%) were allocated to LA; 104 patients (51.7%) were allocated to an open operation. One patient who was allocated to the LA underwent primarily an open operation due to miscommunication; this patient stayed in the LA group during analysis. There were no significant differences between the two groups with respect to mean age, gender, and mean duration of symptoms at the time of enrollment (Table 1). The numbers of other diagnoses found at exploration, and the number of histologically normal, inflamed, and perforated appendices removed, were comparable in both groups (Table 2). One appendix was not removed in the LA group because bilateral pelvic inflammatory disease was found. In all other cases, in both groups, the appendix was removed, because no other definite diagnosis could be found or because of involvement of the appendix in the process. Retrocecally located appendices were found in 24 (24.7%) patients in the LA group and in 32 (30.8%) patients in the OA group.

338 Table 2. Clinical and histological diagnoses in both groups LA (n97) Acute appendicitis Nonperforated Perforated Normal appendix No other diagnosis Meckels diverticulum Pelvic inflammatory diseasea Hyperplasia of lymphoid tissue in ileocecal corner Carcinoma of appendix Carcinoid of appendix Carcinoma of right ovary Cecal diverticulitisb Sigmoid diverticulitisb
a b b

Table 3. Postoperative course in both groups OA (n104) 90 (86.5%) 72 (82.7%) 18 (17.3%) 11 (10.8%) 9 (8.7%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) Table 4. Pre- and postoperative complications in both groups LA (n97) Wound infection Ileal perforation Sigmoid perforation overlooked Ileus Urinary tract infection Pulmonary infection Esophagitis Intraabdominal abscess * p < 0.05 1 (1%) 5 (5.2%) 1 (1%) 1 (1%) OA (n104) 6 (5.8%)* 1 (1%) 3 (2.9%) 1 (1%) 1 (1%) Postoperative pain VAS day 1 VAS day 2 Analgesia day 1 Analgesia day 2 Reintroduction of diet Liquid diet in days Solid diet in days Hospital stay Stay in days * p < 0.001 ** p: NS LA (n97) mean (SD) 35.3 (23.7) 18.7 (17.6) 1.3 (0.7) 1.0 (0.8) 1.3 (0.2) 2.1 (0.4) 3.7 (2.5) OA (n104) mean (SD) 58.7 (25.0)* 34.0 (24.5)* 2.2 (0.9)* 1.8 (0.9)* 1.4 (0.3)** 2.2 (0.3)** 4.4 (3.9)**

85 (87.6%) 69 (83.5%) 16 (16.5%) 8 (8.3%) 6 (6.2%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) 2 (2.1%)

appendix not removed appendix involved in process

Conversion of laparoscopic to open operation was necessary in 12 patients (12%). In this group, the appendix was located retrocaecally in nine cases and was surrounded by dense infiltrate in ten cases. The laparoscopic operation was converted to a normal McBurney muscle-splitting incision in nine cases and to a transverse, transrectal incision in two cases (both because of cecal diverticulitis); in one case a lower midline laparotomy was necessary to perform a resection because of carcinoma of the right ovary. Extension of the McBurney muscle-splitting incision in the OA group was necessary in six patients (5.8%). In five cases, lateral extension of the incision was sufficient. Once a lower midline laparotomy was necessary to remove a perforated Meckels diverticula. During LA, one ileal perforation caused by electrocautery was oversewn laparoscopically during the initial operation. The skin was left open in 19.4% of patients in the OA group, while the skin was closed in all patients in the LA group. The mean operative time was significantly longer for LA than for OA61 24 vs 42 18 min (p < 0.001; mean SD). Postoperative course Postoperative pain was significantly less in the LA group on both the 1st and 2nd postoperative days, resulting in lower scores on the VAS and in less use of analgesics (Table 3). There were no statistically significant differences for reintroduction of diet, or hospital stay (Table 3). Postoperative complications did not differ except for the number of wound infections, which were significantly lower in the LA group compared to the OA group (0 vs 6 p < 0.05) (Table 4). Postoperative ileus was seen in five cases in the LA group and three cases in the OA group (p NS). Ileus followed operation because of nonperforated appendicitis in four out of five in the LA group and in two out of three in the OA group. One patient in each group developed ileus after operation for perforated appendicitis. One intraabdominal abscess was seen at the base of the cecum in the OA group in a patient operated on because of perforated appendicitis. This abscess was treated successfully with percutaneous

drainage. Esophagitis, successfully treated with omeprazole, was seen in one case following OA because of acute appendicitis in a patient with a history of alcohol abuse. Urinary tract infection occurred in one female patient who had an LA for acute appendicitis; she was treated successfully with antibiotics. One reoperation was necessary during the study. A 41-year-old male in the OA group had to be reoperated 2 days after removal of an inflamed appendix because of high fever and persisting abdominal pain. At reoperation through a midline laparotomy, a long sigmoid colon was found with perforated diverticular disease and abscess in the right and middle abdomen. A Hartmanns procedure was performed. The patient recovered uneventfully. Colon continuity was restored after 6 months without negative sequelae.

Discussion Laparoscopic techniques have revolutionalized gallbladder surgery without any randomized clinical trial supporting the change from open cholecystectomy to laparoscopic cholecystectomy. LA, on the other hand, has been shown in several randomized, controlled trials to be superior where postoperative pain or use of analgesia [1, 6, 7, 10, 15], number of postoperative complications [1, 10, 14, 15], hospital stay [1, 13, 14, 15] and return to normal activities [1, 6, 7, 10, 15] are concerned. Despite this evidence LA has not become the gold standard in treating acute appendicitis. This may be partly because of inconsistency in the literature, because other studies show no differences at all between OA and LA

339

or only disadvantages for the laparoscopic technique [17], or because appendectomy through a muscle-splitting incision is already considered minimally invasive surgery. Possibly the often-acute aspect of appendectomy, hampering surgical training and motivation of anaesthesiologists or even surgeons at night, might also contribute to the reluctance to introduce LA as therapy of first choice to treat acute appendicitis in all cases. In our study we showed that even with normal training practices continued during the study, important advantages can be achieved with reduced postoperative pain, less use of analgesics, and less wound infections. The only disadvantage for LA in our study was the considerably longer mean operative time for LA: 61 vs 42 min and possibly related higher operative costs. Earlier reports showed or suggested lower rates of ileus and quicker restoration of diet following laparoscopic surgery. We did not find lower numbers of ileus in the LA group, nor did we find a quicker restoration of diet following LA. The presence of postoperative ileus was not related to perforated appendicitis, because four out of five patients in the LA group and two out of three patients in the OA group who developed ileus had only contained appendicitis. Only one patient in each group developed ileus after operation for perforated appendicitis. Insufflation of CO2 in the peritoneal cavity has been theorized to spread pus intraperitoneally when a purulent intraabdominal infection is present. Therefore LA would be expected to result in a higher rate of postoperative intraabdominal abscesses and conversion to open operation is advocated by many authors if there is evidence of complicated appendicitis [1, 3]. In our study only one postoperative abscess was observed after OA and none after LA. We noticed that laparoscopic irrigation of the peritoneal cavity could be performed effectively. In our study, evidence is lacking that laparoscopy is deleterious in generalized peritonitis. Laparoscopy has been advocated as a diagnostic tool to decrease the rate of negative appendectomies. This could not be analyzed in this trial because all appendices were removed in both groups unless a definite other diagnosis was found. Yet this rate could have been reduced by 9.5%, if all normal appendices had been left in situ. However, detecting all appendeceal pathology on the serosal side of the appendix can be difficult [11] and laparoscopic examination of an appendix is affected by laparoscopic experience and quality of the video imaging system. Therefore we advocate removal of the appendix if no definite other diagnosis can be found at explorative laparoscopy. We found that a rigid, inflexible appendix is a consistent feature of appendiceal pathology which dictates removal. Although the number of patients with a definite other diagnosis for whom a laparoscopic exploration could be beneficial was low in this trial, the patient who was found to have a perforated sigmoiditis at relaparotomy might have benefited from inspection of the entire abdominal cavity instead of only limited inspection, such as was performed through a McBurney incision at the initial operation. Our hypothesis is that in this patient, thickening of the appendix was due to generalized peritonitis, caused by sigmoiditis, and mistaken as acute appendicitis. At laparoscopy a perforated sigmoiditis would probably not have been overlooked. To determine whether the higher efforts and costs are worthwhile to perform LA instead of OA, one should take

into account not only direct advantages like the use of the laparoscope to increase diagnostic power, less postoperative pain, and less wound infections, but also long-term effects such as possibly decreased number of postoperative adhesions following LA [4]. De Wilde performed laparoscopy 3 months after OA and LA. Eighty percent of patients who had OA developed adhesions, and only 10% of the patients in the LA group. Especially if one wants to perform cost benefit analyses, this aspect could be of utmost importance to determine the long-term costs of open appendectomy. In this trial it was shown that even in a teaching hospital setting, LA can be performed safely and effectively, although both trainer and trainee should be aware of specific risks of minimally invasive surgery. One of our intraoperative complications consisted of ileal perforation due to electrocautery. One of the clamps used to coagulate the mesentery was held against the ileum unnoticed because only part of the tip of clamp was shown on the screen before electrocautery was started. If only limited experience exists with laparoscopic procedures this can happen easily. We consider, on the other hand, LA to be a perfect teaching model, following laparoscopic cholecystectomy, in which to teach more advanced minimally invasive techniques. We think that LA is superior over OA regarding postoperative pain and postoperative complications. Long-term follow-up studies are necessary to determine a possible decrease of late bowel obstruction. Because of the increased operative time and possibly related higher direct costs, LA might not be the best way to treat acute appendicitis for every doctor and every patient at every hospital.
Acknowledgment. The authors are greatly indebted to Manon Hemmes and Rita Stege for culling and processing all data. The authors would like to thank all surgeons and residents from St. Clara Hospital and University Hospital Dijkzigt who participated in recruitment, treatment, and follow-up of the patients involved in this trial.

References
1. Attwood SEA, Hill ADK, Murphy PG, Thornton J, Stephens RB (1992) A prospective randomized trial of laparoscopic versus open appendectomy. Surgery 112: 497501 2. Berry J, Malt R (1984) Appendicitis near its centenary. Ann Surg 200: 567575 3. Bonanni F, Reed J, Hartzell G, Trostle D, Boorse R, Gittleman M, Cole A (1994) Laparoscopic versus conventional appendectomy. J Am Coll Surg 179: 273278 4. De Wilde RL (1991) Goodbye to late bowel obstruction after appendicectomy. Lancet 338: 1012 5. Dubois F, Icard P, Berthelot G, Levard H (1990) Coelioscopic cholecystectomy. Preliminary report of 36 cases. Ann Surg 211: 6062 6. Frazee RC, Roberts JW, Symmonds RE, Snyder SK, Hendricks JC, Smith RW, Custer MD, Harrison JB (1994) A prospective randomized trial comparing laparoscopic versus open appendectomy. Ann Surg 219: 725731 7. Hansen JB, Smithers BM, Schache D, Wall DR, Miller BJ, Menzies BL (1996) Laparoscopic versus open appendectomy: prospective randomized trial. World J Surg 20: 1721 8. Hesselink EJ (1995) In Trends in surgery. SIG health care information. p 14 9. Kok de A (1983) The laparoscopic mini appendectomy. Act Endosc 13: 56 10. Kum CK, Ngoi SS, Goh PMY, Tekant Y, Isaac JR (1993) Randomized controlled trial comparing laparoscopic and open appendicectomy. Br J Surg 80: 15991600 11. Lau WY, Fan ST, Yiu TF, Chu KW, Suen HC, Wong KK (1986) The

340 clinical significance of routine histopathologic study of the resected appendix and safety of appendiceal inversion. Surg Gynecol Obstet 162: 256258 12. Lewis FR, Holcroft JW, Boey J, Dunphy JE (1975) Appendicitis: a critical review of diagnosis and treatment in 1,000 cases. Arch Surg 110: 677684 13. Martin LC, Puente I, Sosa JL, Bassin A, Breslaw R, McKenney MG, Ginzburg E, Sleeman D (1995) Open versus laparoscopic appendectomy; a prospective randomized comparison. Ann Surg 222: 256262 14. Mompean JAL, Campos RR, Paricio PP, Aledo VS, Ayllon JG (1994) Laparoscopic versus open appendicectomy: a prospective assessment. Br J Surg 81: 133135 15. Ortega AE, Hunter JG, Peters JH, Swanstrom LL, Schirmer B, The Laparoscopic Appendectomy Group (1995) A prospective randomized comparison of laparoscopic appendectomy with open appendectomy. Am J Surg 169: 208213 16. Semm K (1983) Endoscopic appendectomy. Endoscopy 15: 5964 17. Tate JJT, Dawson JW, Chung SCS, Lau WY, Li AKC (1993) Laparoscopic versus open appendectomy: prospective randomised trial. Lancet 342: 633637

Review article
Surg Endosc (1997) 11: 321325

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Endoluminal surgery
H. Spivak, J. G. Hunter
Department of Surgery, H124C Emory University Hospital, 1364 Clifton Road NE, Atlanta, GA 30322, USA

Abstract Background: Progress in laparoscopic surgery and the employment of minimally invasive techniques have led to the emergence of a new branch in this field, laparoscopic endoluminal surgery (LES). Methods: LES encompasses all surgical procedures that involve intentional trocar or instrument penetration into the lumen of the gut in order to perform a surgical procedure. The integrity of the organ is preserved except for the small enterotomy sites and the operated area. The procedure is technically demanding and requires sound endoscopic skills. Results: The experience in LES has primarily involved the stomach and right colon for two reasons. First, these organs have a larger volume/surface ratio in the gut; therefore, they are accessible for the endoluminal instruments. Second, even limited resection of the stomach or colon can potentially carry a significant morbidity. LES minimizes the extent of this resection. Conclusions: In this paper we review recent developments in LES, discuss the technical aspects of the procedure, and recommend its applicability. Key words: Endoluminal surgery Polyp Pancreatic pseudocyst Early gastric cancer

Gastric endoluminal surgery The short history of LES of the stomach begins with a related but somewhat controversial procedure, flexible endoscopic gastric surgery. Early experience using this technique was presented in 1970 by Tsuneoka and Uchida [18], who performed polypectomy for the treatment of polypoidtype early gastric cancer. In 1979, Mizushima et al. [9] introduced laser thermocoagulation, and other new methods have followed. Tissue destruction techniques, such as laser, local coagulation by ethanol injection, electrocoagulation, and microwave, have led to unsatisfactory results since the

Correspondence to: J. G. Hunter

most accurate assessment of the infiltration depth of a cancer is the histologic examination of an intact specimen. Tissue resection, including strip biopsy and polypectomy, is the preferred endoscopic method for treating gastric polyps. The strip biopsy technique requires injection of 35 ml of saline through a needle into the submucosal layer of the target lesion (Fig. 1). The infusion pressure causes the lamina propria to strip off at the submucosal layer. The elevated tissue is then grasped with forceps, tied with a snare, and resected using high-frequency current [17]. Polypectomy without submucosal injection using the snare or grasper has also been employed [6]. Endoscopic resection is an option for early gastric cancer, especially in high-risk patients [2, 6, 7, 17]. In Japan, a large series of endoscopic resections of early gastric cancer reported survival results similar to those associated with traditional operative procedures [6, 7, 15]. In the United States, experience has been limited, with indications for this procedure generally being limited to patients who are not candidates for gastric resection [2]. However, data from the Japanese series have established that early gastric cancer can be treated using endoscopic techniques with minimal morbidity and organ function preservation. In the early 1990s, minimally invasive surgery had a significant effect on gastrointestinal surgery. Improvements in laparoscopic instruments and skills resulted in an abundance of upper abdominal procedures, and laparoscopic gastric surgery gained increased popularity. The recent introduction of medical therapy for Helicobacter pylori infection has offered effective healing of most peptic ulcers, and elective antiulcer surgery has become a rare phenomenon. Unfortunately, such progress has not been made in the case of neoplasm, and surgical therapy is still the only effective treatment for cancerous and noncancerous lesions. Prior to the laparoscopic era, gastric lesions that required excision or hemostasis that could not be accomplished by peroral endoscope routinely required open laparotomy. Furthermore, functional procedures that utilized the stomach for internal drainage of pancreatic pseudocyst required generous opening of the anterior stomach wall in addition to the traditional laparotomy. These large gastrotomies or partial gastrectomies carried substantial morbidity, which was especially frustrating when treating benign or small lesions. Current laparoscopic techniques have minimized operative trauma and permitted resection of anterior wall lesions

322

Fig. 1. The strip biopsy is performed by injection (A) of 3 to 5 ml of saline into the submucosal layer followed by (B) snare resection.

Table 1. Reported experience of LES Author and year Way et al. [19] 1994 Gagner [5] 1994 Hunter 1994-1996 No. of patients 9 1 4 Diagnosis 9 pancreatic pseudocysts Pancreatic pseudocyst 1 leomyoma gastroesophageal junction 2 antral polyp 1 giant cecal polyp 6 early gastric cancer 1 giant polyp 1 leomyoma 14 early gastric ca. 7 early gastric ca. 2 pancreatic pseudocyst 1 leomyoma 1 bleeding polyp 1 chronic uncler 1 trichobezoar Procedure Cystgastrostomy Cystgastrostomy Submucosal excision Snare excision Snare excision Mucosal/submucosal excision Mucosal/submucosal excision Mucosal excision (local anesthesia) Cystgastrostomy Full thickness Partial thickness Fragment excision Unpublished data Retrograde cecal colonoscopy Comments 1 failure

Ohashi [11, 12] 1994/5 Ohgami et al. [13] 1996 Murai et al. [10] 1996 Filipi et al. [3] 1996

8 14 6 6

1 postoperative bleeding

Wound infection Converted to open

of the stomach using endostapler instruments. However, not all lesions are accessible to laparoscopic wedge resection (posterior wall, cardia). In addition, wedge resection may be too aggressive for small benign lesions or for some early gastric cancers in high-risk patients. Nevertheless, the laparoscopic formula is applicable to the stomach to facilitate dissection of certain lesions or to control localized bleeding. Thus far, the reported experience in LES of the stomach has been limited (Table 1). The procedure is still in its infancy and carries many technical difficulties. In 1994, Way et al. [19] presented a series of nine patients who underwent cystogastrostomy for pancreatic pseudocyst. The technique involves insertion of an 18G needle wrapped in an elastic sheet with a balloon on the tip. Following insertion into the gastric lumen, the balloon is inflated and the obturator and sleeve are passed down the sheet, creating a 5-mm port that snugly grips the gastric wall following removal of the obturator. A total of three ports are made in this manner. Following deflation of the intraperitoneal gas, a 5-mm scope is introduced through one of the ports. Using electrocautery, the cyst is entered, widely opened, and drained. Biopsies are then taken, the trocars are removed, and the stomach incisions are closed laparoscopically. The authors reported a complete success in eight of their nine

patients. The one complication involved inadequate drainage due to a small incision. Hospital stay for all patients was 2 days, and average blood loss was less than 30 ml. Also in 1994, Gagner [5] described a similar approach for drainage of a large pancreatic pseudocyst. At 3-month follow-up, CT scan showed almost complete resolution in that case. Ohashi [11] in 1994 was the first to describe endoluminal resections of gastric mucosal or submucosal lesions. Then in 1995 [12], he presented a series of eight laparoscopic intragastric procedures, in which two patients underwent enucleation or polypectomy and six had mucosal resections. Preoperatively, a gastroscopic examination of the stomach is made to visualize the lesion and to determine adequate insertion sites for three trocars. The points for trocar placement should meet the anterior wall of the stomach. The author suggested placement of a nasogastric tube with an inflated balloon in the duodenum to prevent airflow from the stomach to the intestine. Three abdominal wall incisions are made at selected spots and two stay sutures are applied for each of the planned gastric trocar sites. Hasson cannulae are inserted through these sites into the gastric lumen under gastroscopic monitoring and the balloons are inflated. The resection is accomplished by dissection of the mucosal margins with forceps, electrocautery, and/or laser. In 1996, Ohgami et al. [13] presented a laparoscopic

323

approach to the treatment of early gastric cancer. Using an endoluminal technique similar to that described by Ohashi [12], 24 patients underwent laparoscopic wedge resection of the stomach and 14 underwent intragastric mucosal resection. The histology reports showed that all of the resections carried adequate margins. Based on these results, the authors recommend laparoscopic wedge resection for early gastric lesions of the anterior wall and lesser and greater curvatures and LES for posterior, cardiac, and prepyloric lesions. Murai et al. [10] in 1996 presented a new approach to resecting early gastric cancer. They describe insertion under local anesthesia of a percutaneous endoscopic gastrostomy (PEG), containing a metal rod, placed in the lumen of the anterior wall of the stomach. The lesion is grasped using an endoluminal rigid grasper and resected by snare forceps that are inserted through a gastrofiberscope channel. The investigators reported excellent results in six out of seven patients undergoing resections for early gastric cancer, with one patient requiring endoscopic management of a postoperative bleeding complication. Filipi et al. [3] described a similar technique for trocar insertion using a PEG. Their series contained a diversity of cases. There were two complications among six patients. One case of chronic fibrotic gastric ulcer required conversion to an open technique, and a postoperative wound infection was encountered in another patient. Our gastric endoluminal experience has included successful extraction of an 6-cm submucosal leiomyoma of the gastroesophageal junction in a 55-year-old male with epigastric fullness and difficulty belching. Following induction of general anesthesia, the lesion was manipulated and excised using the endoluminal instruments. Then it was cut to small pieces and removed via the oral route. The mucosal defect was repaired with a running Vicryl suture. Postoperatively, a nasogastric tube was left in place for 24 h. The patient was tolerating a clear diet on postoperative day 1 and was discharged home on postoperative day 2. The pathology report showed a benign tumor, adequately extracted, and at 2-year follow-up there were no recurrences or late complications. In other cases, we performed LES in a similar fashion and extracted 1.52 cm antral polyps using a dissection or snare. The patients had an uneventful postoperative course and were discharged home after 1 day. These cases demonstrate the value of endoluminal surgery. Without this technique, such lesions would have required extensive surgery with potential serious morbidity.

Fig. 2. Gastric endoluminal procedure utilizing endogastric scope, grasper, and cutting devices.

Technique
As a rule, anterior wall lesions are best treated by a limited wedge resection, while posterior, cardia, or gastroesophageal junction lesions can be dealt with using the endoluminal technique. We use two basic approaches. The first is performed under general anesthesia and involves intraperitoneal placement of the laparoscope followed by insertion of 510-mm trocars into the stomach. This method is preferred for longer cases, such as large leiomyoma of the gastroesophageal junction. The other approach is performed under local anesthesia (with intravenous sedation) and is carried out by placing the trocars directly into the stomach using the PEG technique [14]. The latter is preferred for small lesions and high-risk patients. The PEG method for trocar insertion also can be applied to the first approach. Details of the two approaches follow. 1. The patient is placed in a supine position on the operating table, with

monitors mounted above the patients head as close as possible to the 10 and 12 oclock positions. After induction of general anesthesia, the patients legs are spread on a modified fractured table using leg boards to avoid the problem of Allen stirrups. Pneumatic compression boots and a Foley catheter are placed. An upper endoscopy is performed to establish the exact location of the planned procedure. A nasogastric tube with a balloon is placed in the duodenum, and the balloon is inflated to occlude the pylorus. The stomach is inflated with air, and according to the anatomic location of the intended procedure, positions for the endoluminal trocars and camera are chosen and marked on the abdominal wall. Transillumination of the stomach wall can facilitate locating an appropriate trocar site. The stomach is deflated while the endoscope remains in place. After a pneumoperitoneum has been established, a right paramedian 10-mm trocar is introduced into the peritoneum between the two planned endoluminal trocar positions (at least 3 cm apart). In case of tumor resection, a survey of the abdomen is performed, searching for unexpected suspicious lymph nodes or extension of the gastric tumor beyond the serosa. Under laparoscopic and endoscopic guidance, two 5-mm trocars are inserted into the stomach at the chosen positions (Fig. 2). The pneumoperitoneum is released, the camera and 10-mm trocar are withdrawn, and a third trocar is introduced for passage of the camera into the stomach. The circumflex margin of the incision is marked using Bovie cautery on the gastric mucosa. Mucosal lesions can be elevated by injection of 5 ml of saline to the submucosa. Submucosal lesions are best treated with traction and sharp dissection by applying the electrocautery or laser. The tumor is placed in a bag and extracted through either one of the ports or the peroral route. 2. The patient and monitor positions are the same as those used for the first procedure. Local anesthesia with sedation can be applied. Following endoscopic surveillance of the stomach, a PEG containing a metal rod as described earlier is placed and secured to the abdominal wall in the usual fashion [14]. The tube is cut about 1 cm from the abdominal wall, and a laparoscopic grasper is introduced via the tube into the stomach. The laparoscopic dissector and camera are inserted into the lumen through a second and third PEG. The intraluminal procedure is carried out in the same fashion as described above.

324

Colonic endoluminal surgery The polyp theory of cancer development has led to enormous efforts to survey and biopsy all colonic lesions. Colonoscopy is the best tool for the diagnosis and management of almost all mucosal growths. However, there are a few limitations in its employment. First, even in experienced hands, the proximal segments of the colon, namely, the right colon and cecum, cannot be reached by colonoscopy in about 5% of cases. Second, some of the polyps, especially those in the right colon, are large sessile adenomas that can be very difficult to remove by colonoscopy alone. It has been suggested that 3550% of colonic neoplasms are found in the right colon, especially in the elderly [1, 16]. When a polyp cannot be removed by colonoscope due to difficulties in access or anatomic features of the polyp, the alternative has been to explore the abdomen and resect the involved segment. This can sometimes lead to resection of the entire right side of the colon for benign disease. Laparoscopic surgery for such a lesion is a safe option but still requires resection of the segment that carries the polyp. Endoluminal surgery is an alternative for removal of some difficult or inaccessible polyps in the colon. The lesions that are most suitable for such a resection are those that are located at the posterior wall and at some parts of the lateral walls of the right colon and cecum. We employed the endoluminal technique in a 56-year-old patient who was found to have a 5-cm benign-appearing pedunculate cecal polyp. The polyp could not be reached by colonoscopy, and the patient returned for elective endoluminal surgery. The procedure was performed by transabdominal endocecal placement of a pediatric endoscope. The polyp was identified at the posterior wall of the cecum and excised using a snare that was introduced via the endoscopic channel (see Technique section below). The patient had an uneventful postoperative course and was discharged home the next day. To our knowledge, there are no similar cases in the literature, but the technique is relatively simple and should be included in the armamentarium of the endoscopic surgeon. Technique
The patient is placed in a supine position on the operating table. Induction of general anesthesia, monitor placement, and patient leg position are the same as those described for gastric endoluminal surgery. A 5-mm laparoscope and one or two additional 5-mm trocars are introduced in the usual fashion into the peritoneal cavity. Two stay sutures are placed intracorporeally 3 cm apart, and a small cut is made in between at the anterior aspect of the cecum. A trocar is inserted to the cecum under laparoscopic visualization. A pediatric endoscope follows through this trocar into the cecum, and the polyp is identified. A snare is introduced by the endoscopic channel, and the polyp is encircled and resected using electrocautery (Fig. 3). The specimen can be left in the cecum and retrieved later using bowel catharsis. At completion of the resection, the cecal trocar is removed and the cecotomy closed with two intracorporeal sutures. Finally, the polyp dissection and trocar sites are checked for perforation or leakage, and the instruments are retrieved from the abdomen. An alternative approach that we have not yet tried would be to combine an intraoperative colonoscopy to assist with visualization, manipulation of the polyp, and retrieval of the specimen.

Fig. 3. Colonic endoluminal procedure. A pediatric scope is introduced to the cecum, and the polyp is resected by the snare.

Discussion Laparoscopic endoluminal surgery has thus far been utilized in a limited number of cases. It is best suited for gastric

lesions but can be employed for selective proximal colonic polyps or bleeding sources, as well. A large-volume elastic organ, the stomach is easily accessible by the oral route, and during inflation its anterior boundaries reach the abdominal wall, facilitating placement of endoluminal trocars. These unique features, in contrast to most other organs, make the stomach an excellent organ for LES. Laparotomy combined with gastric resection or large gastrotomy carries unproportionally greater trauma than is necessary to treat small, benign lesions or localized bleeding vessels. The guiding principles for the application of minimal access surgery apply here perfectly. Early gastric cancer (mucosal and some submucosal lesions) is a more complicated issue in the application of endoluminal surgery. The main obstacle to treating early gastric cancer by a limited peroral endoscopic excision is the concern that it is a suboptimal treatment for lesions that have already metastasized to lymph nodes [15, 17]. Indeed, lymph node metastasis in early gastric cancer is found in 3% and 20% of mucosal and submucosal lesions, respectively [7, 15]. The vast majority of experience in limited peroral endoscopic resection for early gastric cancer comes from Japan. The data that are presented in Table 1 point out that while LES performed in Japan was primarily for early gastric cancer, the procedure in the West has been utilized only for benign lesions. Nevertheless, the advantage of LES over peroral resection is the ability to perform a thorough laparoscopic exploration of the abdomen and stomach prior to dissection. Suspicious lymph nodes or lymph node sampling can be taken for frozen sections and the stomach can be surveyed for any extension of the tumor beyond what was preoperatively presumed. In this respect, LES may be a better option for early gastric cancer. After a precise preoperative diagnosis and verification of anatomic location have been made, selective benign lesions or bleeding can be handled by LES. Careful consideration should be given to the potential extension of submucosal benign lesions through the serosa. Such an extension would require a different approach and probably would

325

be best managed by laparoscopic extraluminal dissection of the seromuscular layers, keeping the mucosa intact [4, 8], or by laparoscopic wedge resection. In summary, lesions of the posterior wall of the stomach, cardia, gastroesophageal junction, and prepyloric aria can be technically managed by LES. Excision of small benign lesions or localized procedures for symptomatic lesions in high-risk patients may be performed under local anesthesia. Larger tumors or early gastric cancer excision (when indicated) should be done under general anesthesia, with exploratory laparoscopy as part of the procedure. Colonic LES can be utilized to avoid extensive resection for difficult polyps or for a bleeding source in the proximal right colon. Care should be taken, as with gastric endoluminal surgery, to tightly close the trocar incisions and to carefully observe the operated site for bleeding or perforation. In conclusion, LES is a promising alternative to traditional gastric surgery in selected cases. Although the rationale for its application may appear persuasive, it should be kept in mind that these procedures are technically demanding and often time-consuming. Nevertheless, the characteristics of the stomach and right colon make them uniquely suitable for LES, especially with the current techniques and instrument sizes. Future research and development may extend LES application to the entire large and small intestine, or to common bile duct stenting procedures.
Acknowledgment. We thank Ms. Cathy Alden for her editorial assistance in the preparation of this manuscript.

References
1. Chapman I (1963) Adenomatous polyps of the large intestine: incidence and distribution. Ann Surg 157: 223226 2. Farley DR, Donohue JH (1992) Early gastric cancer. Surg Clinic North Am 72: 401421

3. Filipi CJ, Katada N, Hinder RA, Lowham A, Cornet DA (1996) Percutaneous endoscopic gastrostomy assisted endo-organ surgery [abstract]. SAGES meeting, 1317 March 1996, Philadelphia 4. Fowler DL, White S (1991) Laparoscopic resection of a submucosal gastric lipoma: a case report. J Laparoendosc Surg 1: 303306 5. Gagner M (1994) Laparoscopic transgastric cystogastrostomy for pancreatic pseudocyst [abstract]. Surg Endosc 8: 239 6. Haruma K, Sumii K, Inoue K, Teshima H, Kajiyama G (1990) Endoscopic therapy in patients with inoperable early gastric cancer. Am Coll Gastroenterol 85: 522526 7. Hioki K, Nakane Y, Yamamoto M (1990) Surgical strategy for early gastric cancer. Br J Surg 77: 13301334 8. Lacy AM, Tabet J, Grande L, Garcia-Valdecasas JC, Fuster J, Delgado S, Visa J (1995) Laparoscopic-assisted resection of gastric lipoma. Surg Endosc 9: 995997 9. Mizushima K, Harada K, Okamura T (1979) Clinical application of YAG laserThe second report. Gastroenterol Endosc 21(11): 1289 1296 10. Murai R, Ando H, Mitsumori N, Hada T, Fuijioka S, Nagayama A, Itsubo K (1996) Our technique of percutaneous transgastric wall endoscopic mucosal resection, PTEMR, for early gastric cancer [abstract]. SAGES meeting, 1317 March 1996, Philadelphia 11. Ohashi S (1994) Laparoscopic intragastric surgery: is it a new concept in lap surgery [abstract]? Surg Endosc 8: 256 12. Ohashi S (1995) Laparoscopic intraluminal (intragastric) surgery for early gastric cancer. A new concept in laparoscopic surgery. Surg Endosc 9: 169171 13. Ohgami M, Otani Y, Kumai K, Kubota T, Kitajima M (1996) Laparoscopic curative surgery for early gastric cancer. (Poster) SAGES meeting, 1317 March 1996, Philadelphia 14. Ponsky JI, Gauderer MWL (1981) Percutaneous endoscopic gastrostomya nonoperative technique for feeding gastrostomy. Gastrointest Endosc 27: 911 15. Sano T, Kobori O, Muto T (1992) Lymph node metastasis from early gastric cancer. Br J Surg 79: 241244 16. Spivak H, Valdemalle D, Friedman I, Nussbaum M (1996) Colorectal surgery in the octogenarian. J Am Coll Surg 183: 4650 17. Tada M, Murakami A, Karita M, Yanai H, Okita K (1993) Endoscopic resection of early gastric cancer. Endoscopy 25: 445450 18. Tsuneoka K, Uchida T (1970) Endoscopic polypectomy of the stomach. Second World Congress of Gastrointestinal Endoscopy, Copenhagen 19. Way LW, Legha P, Mori T (1994) Laparoscopic pancreatic cystgastrostomy: the first operation in the new field of intraluminal laparoscopy surgery [abstract]. Surg Endosc 8: 235

Surg Endosc (1997) 11: 354358

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Fully thoracoscopic pulmonary lobectomy and specimen extraction through rib segment resection
Preliminary report
E. C. Poulin, R. Labbe
Department of Surgery, Ho pital du Saint-Sacrement, Universite Laval, 1050 Chemin Ste-Foy, Que bec, QC, Canada G1S 2L6 Received: 20 August 1996/Accepted: 19 September 1996

Abstract Background: A technique of fully thoracoscopic pulmonary lobectomy with rib-segment resection for specimen extraction is described, and preliminary results in 18 patients are presented. Methods: Surgery is performed through four 15-mm ports. For all lobes except one, the surgeon operates in front of the patient, where the rib spaces are widest and rib-space trauma is less. When lobar dissection is complete, specimen extraction is performed after resection of a rib segment proportional to tumor size. Muscle section is kept to a minimum. There is no rib retraction. Results: There were no deaths, three conversions to open surgery, and three major complications. Average postoperative stay was 5.4 days for patients without complications and 9.6 days for patients with complications. In total six patients presented with some degree of air leaks, and two had post-thoracotomy pain (>2 months duration). The literature is reviewed to analyze current techniques and to define parameters of a truly minimally invasive pulmonary lobectomy. Conclusions: This technique is safe and promising; however, thoracoscopic lobectomy still needs refining. Before valid randomized studies comparing thoracoscopic lobectomy and muscle-sparing thoracotomy or posterolateral thoracotomy can be credible, technical issues related to the production of a truly minimally invasive procedure should be resolved.

Key words: Thoracoscopy Pulmonary lobectomy VATS lobectomy

Despite the development of laparoscopic and thoracoscopic equivalents for most current surgical procedures in recent years, thoracoscopic pulmonary lobectomy has received limited interest. The precise identification of bronchial and vascular anatomy is perceived as difficult, and the possibility of severe hemorrhage has led many thoracic surgeons to express reservations about thoracoscopic pulmonary lobectomy, especially for malignant disease, where precise dissection planes must be respected for adequate oncologic resection. Many authors have reported various techniques of pulmonary lobectomy aided by thoracoscopy (e.g., videoassisted, or VATS lobectomy), and the use of a utility or access incision [7, 1214, 1618], but few have reported pulmonary lobectomy done completely thoracoscopically [15]. We present a technique of fully thoracoscopic pulmonary lobectomy (FTPL) with extraction of the specimen through a small rib-segment resection site made after the lung resection is completed and report the preliminary clinical results in the first 18 patients. Furthermore, because a number of prospective studies report equivocal results on the value of VATS lobectomy, we review currently reported techniques to help define a truly minimally invasive procedure for pulmonary lobectomy. Patients and methods
For anesthesia during thoracoscopic pulmonary lobectomy, a doublelumen endotracheal tube is used to deflate the lung completely on the operated side [9], which reduces the blood flow in the ipsilateral lung vessels and makes thoracoscopic ligature much less menacing. The patient is placed in a lateral decubitus position with flank bolster and flexion of the operative table to increase the size of the rib space. The surgeon stands anterior to the patient, and the video monitor is placed posteriorly, except for resection of the right middle lobe. The anatomical structures are ap-

Presented at the annual meeting of the Society of American Gastroenterological and Endoscopic Surgeons, Philadelphia, Pennsylvania, USA, April 1996 Correspondence to: E. C. Poulin, The Wellesley Central Hospital, University of Toronto, 217B Jones Building, 160 Wellesley Street, Toronto, Ontario, Canada, M4Y 1J3

355

Fig. 1. Position of surgical instruments for thoracoscopic pulmonary lobectomy. A, B For all lobes except one, the surgeon uses an anterior approach where the rib space is wider to minimize rib-space trauma. C The right middle lobe, because of its anterior location, is better approached from behind. D Diagram of rib-segment excision for specimen extraction (camera port site).

proached directly in the fissure, where most of the dissection is done, except for resection of the right middle lobe, an anteriorly situated structure, for which the surgeon stands behind the patient with the video monitor in front (Fig. 1). The operation is performed through four ports. Two 15-mm-long ports are situated along the usual incision line of a posterolateral thoracotomy. One port, made near the posterior axillary line, serves as the principal camera port; the choice of this site places the camera directly over the plane of the greater fissure, where most of the dissection will occur. The second port site on the imaginary line of a standard posterolateral thoracotomy incision is situated below the tip of the scapula posteriorly and serves as the main retraction site. Should the surgery be converted to an open procedure, two of the incisions would therefore be included in the open thoracotomy incision. The third and fourth ports (15-mm-long) are chosen to allow two-handed surgery in a parallel axis with the robot-held camera. These last two ports are situated on each side of the plane of the usual thoracotomy incision, one or two rib spaces above and below the camera port site. They are placed slightly lower than the camera port, at the level of the anterior axillary line, and at the end of the procedure, the chest tube will be placed through one of these incisions (Fig. 1). In obese patients, longer port-site incisions are required. To minimize the possibility of intercostal nerve trauma from torquing an instrument in the rib space during the procedure, no port sleeves are used. This setup allows the surgeon to operate freely with both hands and with good working angles. The camera remains out of the way but in a parallel axis with the working instruments, which facilitates triangulation. To optimize exposure at times, more than one instrument can be introduced at any port site. The camera can also be moved accordingly at any port site to permit better identification of anatomical structures before a decision is reached regarding ligature and section. All structures are either stapled with the linear stapler or ligated with the aid of a knot pusher. The lobar bronchus and larger vessels are usually stapled, and for cost concerns all smaller vessels are doubly ligated on the patient side. Clips are used on the specimen side when it is deemed safe to do so, but they are not used on the patient side for fear of slippage as a result of the movement of the mediastinum. Interlobar and mediastinal nodes are routinely sampled separately to optimize staging [14]. Minithoracotomy is not used during the dissection. Once the lobe is completely mobilized, a deperiosted rib segment of a size equivalent to the tumor diameter is excised at the main camera port site in the posterior axillary

line. This step avoids all forms of rib retraction to extract the specimen. The specimen is retrieved in a sturdy sterile plastic bag to prevent any wound contamination or seeding. Between November 1993 and August 1995, thoracoscopic pulmonary lobectomy was offered to selected patients with bronchiectasis or a histologically diagnosed and staged peripheral small-diameter tumor with the understanding that the procedure would be converted to open surgery if adequate oncologic resection could not be provided or if there was danger of a technical mishap. All patients with a malignancy underwent mediastinoscopy. All had normal arterial blood gases and pulmonary function (forced expiratory volume in 1 s > 1.5 l) adequate to tolerate pulmonary lobectomy. All malignancies were T1\T2N0, except one, in which a microscopic foci of carcinoma was found in one interlobar node in the final pathology report. Follow-up ranged between 12 and 28 months (average 20.3 months).

Results This descriptive cohort consisted of 12 selected men and six women ranging in age from 19 to 78 years (average 55.5 years). Their mean weight was 68.7 kg (range 51100 kg). For three patients (16%), the procedure was converted to open surgery because of an inability to assess thoracoscopically the extent of surgery required (two cases) or the extent of mediastinal involvement (one case). Two underwent open pneumonectomy and one could not be resected. There were no conversions for technical reasons or intraoperative complications. All pulmonary lobes were resected: left lower lobe (n 5), right lower lobe (n 1), right middle lobe (n 3), right upper lobe (n 5), left upper lobe (n 1) (Table 1). The average blood loss was 242 ml (range, 90700 ml), and no blood was given during surgery. The average extraction incision measured 5.1 cm (range, 3.56.5 cm). Mean operating time was 3 h 29 min (range, 2 h 15

356 Table 1. Distribution of fully thoracoscopic pulmonary lobectomy by site (n 15) Lobe Right Upper Middle Middle with segment of upper Lower Upper Lower No. of cases 5 2 1 1 1 5

Left

Table 2. Patients undergoing fully thoracoscopic pulmonary lobectomy (n 18) classified according to pathology Pathology Benign Malignant Histoplasmosis Bronchiectasis Bronchogenic carcinoma Adenocarcinoma No. of cases 1 3 6 8

min4 h 40 min). Of the 15 patients who successfully underwent thoracoscopic lobectomy, three had recurrent infection or bleeding from bronchiectasis, one had histoplasmosis, and 11 had lung cancer (three bronchogenic carcinoma, eight adenocarcinoma) (Table 2). The average size of excised tumors was 2.3 cm (range, 0.55 cm). After resection, all bronchial and specimen margins were free of disease. For the first six patients, postoperative pain was treated with epidural morphine analgesia for the first 2448 h and acetaminophen thereafter. Supplemental codeine or meperidine was required in three of these patients. The pain severity did not justify morphine epidural analgesia, however; so most of the remaining patients received a 100-mg indomethacin suppository immediately before surgery. Their postoperative regimen consisted of indomethacin suppositories every 12 h for 2 or 3 days with acetaminophen. Codeine or meperidine was used sparingly when required. Patients tended to complain more frequently of pain at the chest tube site than at the site of incisions. Normal diet was resumed in an average of 1.7 days (range, 14 days). In the postoperative period, one patient bled from an intercostal artery and required one transfusion, one required bronchoscopy for lobar atelectasis, and one patient (6%) presented an air leak that lasted 10 days (20% morbidity). Major air leaks are usually defined as lasting more than 10 days. Five other patients (33%) presented air leaks which, although lasting less than 10 days, were a more important factor in delaying early discharge than postoperative pain. The average postoperative stay was 6.2 days (range, 411 days). The three patients with postoperative complications had an average postoperative stay of 9.6 days (range, 711 days), whereas the 12 patients without complications stayed an average of 5.4 days (range, 38 days). All patients, except one returned to usual activities within 3 weeks. Two patients presented with residual post-thoracotomy pain (>2 months duration). Discussion A number of publications have demonstrated the safety of thoracoscopic pulmonary lobectomy despite occasional re-

ports of linear stapler failure producing hemorrhage [3]. These few incidents point out the necessity of one-lung ventilation and complete deflation on the affected side as blood flow and vessel size are thus reduced in the operated lung, making vessel control less arduous. It also has been suggested that to maximize the safety of stapling large vessels, the knife can be removed from the staple cartridge before stapling, which will result in the application of six rows of staples without cutting. A small, sharp scissors is then required to cut between the staple lines once hemostasis is ascertained. Alternatively, it is simpler to place a thoracoscopic vascular clamp on the afferent side of the vessel before releasing the stapler [1, 14, 17]. We routinely take one of these precautions, and, like many authors, have not witnessed stapler failure. Despite demonstration of its safety, it has been difficult so far to demonstrate that there are undeniable advantages to thoracoscopic pulmonary lobectomy. It may be that there remain conceptual problems in designing a truly minimally invasive pulmonary lobectomy, which has been borne out in studies comparing various techniques of video-assisted thoracoscopic pulmonary lobectomies (VATS) vs modified open thoracotomies with muscle sparing. Two frequently quoted studies compare VATS with lobectomy through muscle-sparing thoracotomy (MST). Guidicelli et al., reporting on 67 patients, showed that VATS lobectomy patients (44 patients) had less pain and a longer operation than patients who underwent MST (23 patients). There were no differences in hospital stay or any definite benefit to VATS lobectomy. Their VATS technique, however, included a 1015-cm incision in the intercostal muscle and 4-cm rib retraction. In their MST, the intercostal muscle incision ranged from 15 to 20 cm [5]. It would seem that the difference in the severity of surgical trauma between the VATS lobectomy and the MST performed by these authors may not be enough to show significant statistical advantage of one technique over the other. Furthermore, their study was not randomized, and the choice of the surgical technique was left to the discretion of the participating surgeons, with one of the participants performing 75% of the VATS procedures. Furthermore, the fact that one group was twice as large as the other precludes valid comparison. It is also unfair to compare the work of surgeons experienced with the technique of MST lobectomy with that of a surgeon still in the learning curve of VATS lobectomy. Kirby et al. carried out a randomized study comparing MST lobectomy (30 patients) and VATS lobectomy (25 patients) and demonstrated essentially no significant differences between the two groups in operating time, intraoperative blood loss, duration of chest tube drainage, or length of hospital stay [8]. Although this study is better designed, it nevertheless produced only 25 VATS patients from three different centers in 2 years, a fact that probably introduces a learning-curve bias. Furthermore, the authors admit to using a 68-cm access thoracotomy incision and limiting rib spreading until the end of the procedure and then only if it was necessary to remove the specimen. Before valid randomized comparison is made between thoracoscopic lobectomy and lobectomy by MST or standard thoracotomy, five important areas must be addressed. First, the fact that VATS lobectomy means different things

357

to different people demonstrates that this technique has not yet produced consensus among surgeons and that the technical elements contributing to the goal of minimal operative trauma are still being defined. Ideally, this procedure should be performed without cutting chest-wall muscles, with minimal intercostal muscle cutting, and no rib spreading. Chest-wall and intercostal muscle incisions along with the effects of rib spreading are generally viewed as the principal causes of acute and chronic pain after thoracic procedures, and they are also largely responsible for postoperative shoulder joint dysfunction [4, 6, 1012]. Second, intercostal rib-space management should be flawless. Trauma to the intercostal nerves should be minimized by using small (5 mm) instruments whenever possible in the intercostal space. Large port sleeves should be avoided, and 5-mm instruments should be placed directly into the chest. There will remain exceptions to this, as there is presently no good alternative to the 12-mm linear stapler; however, the stapler does not remain in the intercostal space for any prolonged time. Thinner alternatives to the 10-mm thoracoscope that produce adequate thoracic cavity illumination should be found. The long-standing presence of the 10-mm thoracoscope torquing in a small intercostal space could well be responsible for the chronic pain syndrome that we have observed in two of our patients. In this regard, the use of standard thoracic surgery instruments through the port sites, as advocated by some authors, may cause too much trauma and may not be ideal [5, 14, 17, 18]. Surgeons have also found that with this technique, perfect hemostasis at the port sites is required to minimize nuisance soiling to the thoracoscope and the operating instruments. Use of the cautery deep in the intercostal space for this purpose needs to be done carefully under direct vision and minimum wattage (<30 W) to decrease the possibility of intercostal nerve injury from the cautery and possible related pain. Whenever possible, thoracoscopic procedures should be performed with the surgeon standing in front of the patient because rib spaces are wider and afford better odds against causing rib-space injury. Whenever access is needed posteriorly, small (5-mm) instruments should be used and torquing should be avoided (Fig. 1). Third, issues regarding specimen extraction need to be resolved. Although there is a consensus for putting the specimen in a sturdy plastic bag to avoid contamination or seeding during extraction, the optimal extraction technique must be determined. We believe that extracting a tumor mass 5 cm through a 1-cm rib space cannot be done without some rib spreading, even if the intercostal muscles are sectioned up to 8 cm or more as suggested by various authors [5, 7, 8]. We also postulate that most rib-space injuries occur the first time the ribs are spread, whether this is done by a retractor or a plastic bag containing a pulmonary lobe and nodule. We present an alternative extraction technique in which a rib segment is resected before extraction. Preliminary clinical observation seems to indicate that this is a relatively innocuous procedure that could well stand comparison with other extraction techniques. We have used costal cartilage resection for many years to improve staging in open anterior mediastinotomy and have always been impressed by the fact that little pain is produced by this maneuver. McKenna, in his series of VATS lobectomy, reports a 7-cm rib-segment excision for a 6-cm tumor in which the

Table 3. Desirable characteristics of minimally invasive thoracoscopic lobectomy Small skin incisions Minimal muscle section Chest wall Intercostal No rib spreading No intercostal nerve damage Appropriate size of instruments vs size of rib space Minimal torquing of instruments Avoidance of thermal injury to nerve from cautery Avoidance of port sleeves Prevention of air leaks Staples Abrasion Talc Pleurectomy Minimally traumatic specimen extraction Rib-segment excision Specimen morcellation

patient experienced no more pain than his other patients [14]. One should note that rib-segment resection is performed after completion of the pulmonary resection rather than at the beginning of the procedure to avoid complete rib resection if conversion to an open procedure occurs; alternatively, full posterolateral intercostal muscle section after rib segment resection should be avoided in the event of a conversion. Rossi and Litwin described their experience with FTPL and also described specimen morcellation in the plastic bag to reduce the trauma caused by the other extraction techniques [15]. This technique may well be the most minimally traumatic extraction technique of all, but we were not comfortable with this concept for several reasons. Although it may or may not affect patient outcome, specimen morcellation in cancer cases makes evaluation of bronchial margin or other specimen margins more complicated to assess with precision. Because so little pain was observed clinically with small rib-segment resection, we did not want to add another contentious issue to this procedure. Fourth, before randomized comparison is made between thoracoscopic lobectomy and lobectomy by MST or standard thoracotomy, technical improvements to the thoracoscopy technique are still needed in a few areas to decrease the rate of postoperative pulmonary leaks. The visualization of anatomic elements is enhanced in thoracoscopic lobectomy so that it becomes relatively easy to identify the interlobar plane and dissect it with cautery when required. Most likely, to do so produces microscopic lesions in the visceral pleura and results in air leaks. Air leaks are an important factor in determining the length of postoperative stay after thoracoscopic pulmonary lobectomy; therefore, all efforts should be made to minimize them [2]. When sufficient areolar tissue is found in the interlobar plane, cautery dissection should be replaced with sharp scissors dissection, and when the lung parenchyma of adjacent lobes is closely juxtaposed, linear staples should be used liberally. When air leaks are observed after the lung resection is complete, all efforts should be made to correct the problem before closure by using staplers, pleural abrasion, talcum poudrage, pleural flap, or pleurectomy (Table 3). Finally, comparison of various techniques for pulmo-

358 Table 4. Principal causes of chronic pain after thoracotomy Rib fracture or periosteal damage Neurovascular bundle injury Neuroma Anterior and posterior costal articulation injury Costochondritis Costochondral dislocation Tumor recurrence Others Local infection Pleurisy Frozen shoulder

truly minimally invasive technique for pulmonary lobectomy. References

1. Acuff TE, Mack MJ, Landreneau RJ, Hazelrigg SR (1993) Role of mechanical stapling devices in thoracoscopic pulmonary resection. Ann Thorac Surg 56: 749751 2. Bhatnagar NK, Berndt S (1993) A solution to prolonged air leak after video-assisted thoracoscopic lobectomy. Ann Thorac Surg 59: 260 261 3. Craig SR, Walker WS (1995) Potential complications of vascular stapling in thoracoscopic pulmonary resection. Ann Thorac Surg 60: 11571159 4. Dajczman E, Gordon A, Kreisman H, Wolkove N (1991) Long term postthoracotomy pain. Chest 99: 270274 5. Giudicelli R, Thomas P, Lonjon T, Ragni J, Morati N, Ottomani R, Fuentes PA, Shennib H, Noiclerc M (1994) Video-assisted minithoracotomy versus muscle sparing thoracotomy for performing lobectomy. Ann Thorac Surg 58: 712718 6. Heitmiller RF (1988) Thoracic incisions. Ann Thorac Surg 22: 601 7. Kirby TJ, Rice TW (1993) Thoracoscopic lobectomy. Ann Thorac Surg 56: 784786 8. Kirby TJ, Mack MJ, Landreneau RJ, Rice TW (1995) Lobectomy video-assisted thoracic surgery versus muscle-sparing thoracotomy. A randomized trial. J Thorac Cardiovasc Surg 109: 9971001 9. Lamb JD (1993) Anesthesia for thoracoscopic pulmonary lobectomy. Can J Anesth 40: 10731075 10. Landreneau RJ, Hazelrigg SR, Mack MJ, Dowling RD, Burke D, Gavlick J, Perrino MK, Ritter PS, Bowers CM, DeFino J (1993) Postoperative pain-related morbidity: video-assisted thoracic surgery versus thoracotomy. Ann Thorac Surg 56: 12851289 11. Landreneau RJ, Mack MJ, Hazelrigg SR, Naunheim K, Dowling RD, Ritter P, Magee MJ, Nunchuck S, Keenan RJ, Ferson PF (1994) Prevalence of chronic pain after pulmonary resection by thoracotomy or video-assisted thoracic surgery. J Thorac Cardiovasc Surg 107: 1079 1085 12. Landreneau RJ, Mack MJ, Hazelrigg SR, Dowling RD, Acuff TE, Magee MJ, Ferson PF (1992) Video-assisted thoracic surgery: basic technical concepts and intercostal approach strategies. Ann Thorac Surg 54: 800807 13. Liu HP, Chang CH, Lin PJ, Chang JP, Hsieh MJ (1995) Thoracoscopic-assisted lobectomy. Preliminary experience and results. Chest 107: 853855 14. McKenna RJ Jr (1994) Lobectomy by video-assisted thoracic surgery with mediastinal node sampling for lung cancer. J Thorac Cardiovasc Surg 107: 879881 15. Rossi L, Litwin DEM, Gowda K (1996) Anatomic thoracoscopic lobectomy (ATL) without minithoracotomy: preliminary experience. Surg Laparosc Endosc 6: 4955 16. Roviaro G, Rebuffat C, Varoli F, Vergani C, Mariani C, Maciocco M (1992) Videoendoscopic pulmonary lobectomy for cancer. Surg Laparosc Endosc 2: 244247 17. Walker WS, Carnochan FM, Tin M (1993) Thoracoscopy assisted pulmonary lobectomy. Thorax 48: 921924 18. Yim APC, Ko K, Chau WC, Ma C, Ho JKS, Kyaw K (1996) Videoassisted thoracoscopic anatomic lung resections. The initial HongKong experience. Chest 109: 1317

nary lobectomy will require assessment of long-term postthoracotomy pain (>2 months) by impartial observers. A pulmonary medicine group reported that after posterolateral thoracotomy, when patients were stratified in 1-year intervals up to 5 years, from 30 to 70% of patients reported pain when interviewed. Although the syndrome was not severe in most instances, 44% reported that pain interfered with their lives, and 9% required one or more forms of ongoing therapy [4]. Therefore, a truly minimally invasive thoracoscopic pulmonary lobectomy must be designed to avoid the type of surgical trauma that produces chronic pain syndromes (Table 4). Conclusions

Fully thoracoscopic pulmonary lobectomy with ribsegment resection for specimen extraction is a safe and promising technique that results in short postoperative stays. Thoracoscopic lobectomy still requires refinement. Elements of the procedure that correlate with acute or chronic pain and surgical trauma in general can be improved. Rib-space management, prophylaxis of air leaks, and extraction techniques with this technology require more study, as they may have direct impact on outcome. Before valid randomized studies comparing thoracoscopic lobectomy and MST or posterolateral thoracotomy can be credible, technical issues related to the production of a truly minimally invasive procedure should be resolved. Surgeons conducting these studies should also have eliminated the learning-curve bias of thoracoscopic lobectomy of all five pulmonary lobes. Most of the currently described VATS pulmonary lobectomy techniques probably represent an intermediary step between conventional posterolateral thoracotomy and a

Surg Endosc (1997) 11: 341346

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Cardiopulmonary responses to intravenous infusion of soluble and relatively insoluble gases

M. W. Roberts, K. A. Mathiesen, H. S. Ho, B. M. Wolfe
Department of Surgery, University of California Davis Medical Center, 4301 X Street, Room 2310, Sacramento, CA 95817, USA Received: 4 April 1996/Accepted: 21 August 1996

Abstract Background: Carbon dioxide is the current gas of choice for pneumoperitoneum, but hemodynamic and acidbase effects secondary to its systemic absorption have been reported. Various studies have suggested inert gases as alternatives. Methods: We studied the cardiopulmonary responses to intravenous infusion of carbon dioxide, nitrous oxide, argon, helium, and nitrogen in anesthetized swine. The gas was infused into the femoral vein at a rate of 0.1 ml kg1 min1 for 30 min. The changes in end-tidal CO2, mean arterial pressure, hemodynamics, and arterial blood gases were compared to baseline values. Results: No animals died during infusion of the soluble gases (CO2 and N2O). Three of the five pigs infused with nitrogen died suddenly at 20 and 30 min of infusion. The animals in the insoluble gas groups (Ar, He, N2) experienced clinical pulmonary gas embolism and severe acidemia, hypercapnea and tachycardia. Conclusions: Venous gas embolism is poorly tolerated when the gas is relatively insoluble. Insoluble gases should not be used for pneumoperitoneum when there is any risk of venous gas embolism. Key words: Laparoscopy Venous gas embolism Inert gas

system during abdominal insufflation [6, 7, 23] coupled with the trend toward longer, more complex procedures during which these systemic effects may become profound. Some of the alternatives to carbon dioxide pneumoperitoneum which are currently under consideration include abdominal-wall lifting devices and different insufflating gases. Nitrous oxide has long been proposed as an alternate gas, but it has been generally avoided in laparoscopic surgery due to concern over its potential combustibility [9]. Insoluble or chemically inert gases such as nitrogen, argon, and helium have also been suggested as alternative insufflating agents. However, potential problems with venous gas embolism are expected to be more severe with these relatively insoluble gases because of their low solubilities in the blood [19, 22]. Therefore, any evaluation of an alternative gas for pneumoperitoneum should include a critical assessment of the cardiopulmonary responses to the potential venous gas embolism. This study compares the cardiopulmonary responses to venous infusion of soluble gases vs relatively insoluble gases with the hypothesis that the more insoluble gases have a higher risk of significant venous embolism and a more profound impact on acidbase balance and hemodynamics. Materials and methods Animal preparation

Investigators are actively searching for alternatives to carbon dioxide pneumoperitoneum during laparoscopic surgery [4, 9, 10, 13, 18]. The impetus for such a search is the accumulation of clinical and experimental evidence of the negative impact of CO2 absorption on the cardiopulmonary
Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), Orlando, Florida, USA, 1214 March 1995 Correspondence to: B. M. Wolfe

Twenty-four pigs of either sex (mean weight: 47.7 15.5 kg) were used for this study, with protocols for care and use approved by our Institutional Animal Care and Use Committee. The animals were fasted with free access to water for 24 h prior to experiments. For induction of general anesthesia, pigs were premedicated with an intramuscular injection of atropine sulfate (0.044 mg kg1) and acepromazine maleate (0.5 mg kg1). Ten minutes later, they were anesthetized with intramuscular ketamine hydrochloride (20 mg kg1). They were then endotracheally intubated, placed supine, and ventilated with 100% oxygen using a veterinary anesthesia ventilator (model 2000, Hallowell Engineering and Manufacturing) and a halothane dispenser (Fluotec 3, Cyprane North America, Inc.) which administered 12% halothane for maintenance of anesthesia. Next, animals were instrumented for cardiopulmonary monitoring. Si-

342

Fig. 1. The changes in MPAP in response to venous gas embolism. There was a profound and statistically significant increase in MPAP with infusion of each of the relatively insoluble gases (argon, helium, and nitrogen). There were only slight changes in MPAP with the soluble gases at this rate of infusion, but these changes were statistically significant in the carbon dioxide group.

lastic femoral venous and arterial lines (0.062 inch ID) were inserted via a left groin cutdown. The arterial line was advanced up to the abdominal aorta for measurement of the mean arterial pressure and collection of arterial blood gas samples, while the venous line was advanced into the inferior vena cava for infusion of gas. A pulmonary artery thermodilution catheter (Edwards Critical Care Division, Irvine, CA) was placed via a right internal jugular vein cutdown and was connected to a multichannel monitor and recorder (7754 System, Hewlett Packard, Palo Alto, CA) with pressure transducers positioned at the level of the right atrium and calibrated with a mercury manometer. End-tidal CO2 levels (ET-CO2) were continuously monitored by capnometry (model 1260, Novametrix Medical Systems, Wallingford, CN).

Table 1. Solubilities of the experimental gases in water and plasma, expressed as Ostwald coefficients (ml gas @ 37C per ml fluid per ATA) [19, 22] MW CO2 N2O Ar He N2 44 44 40 4 28 Water 0.62300.6420 0.44000.4780 0.02700.0303 0.00960.0101 0.01370.0145 Plasma 0.68600.6990 0.4540 0.02410.0281 0.0086 0.01310.0137

Experimental protocol
Following surgical preparation, all animals were observed and supported until hemodynamics were stable. Minute ventilation was adjusted to obtain normal arterial blood gases and was then fixed throughout the rest of the experiment. Baseline data were collected over a 1-h period. The animals were divided into five experimental groups to receive intravenous infusion of one of the five gases (CO2, N2O, Ar, He, N2) at a rate of 0.1 ml kg1 min1 for 30 min via the femoral venous catheter. Venous gas infusion rate was controlled by using high-pressure medical-grade gas with a regulator connected to a 120-mm Manostat flowmeter (Fisher Scientific, Santa Clara, CA). Hemodynamics and ET-CO2 measurements were obtained every 5 min during the infusion period. Arterial blood gas analysis was performed every 15 min. The infusion was followed by a 30-min recovery period with all experimental data collected at 15 min after the end of infusion. The experiments were concluded when all animals that survived were euthanatized.

insoluble gases. All analyses were done using computer statistical software (Instat Instant Biostatistics, GraphPad Software, San Diego, CA).

Results Evidence of pulmonary arterial embolism Table 1 lists the Ostwald coefficients of the experimental gases in water and plasma for comparison purposes [19, 22]. Carbon dioxide is 1.5 times more soluble than nitrous oxide in plasma, but both gases are relatively soluble. Carbon dioxide is 2480 times more soluble in plasma than the three insoluble gases studied (argon, helium, nitrogen), while nitrous oxide is 1653 times more soluble in plasma than these gases. No deaths occurred as a result of intravenous infusion of either of the soluble gases CO2 and N2O at our experimental rate. In the relatively insoluble groups, three out of the four pigs infused with N2 died during the experiment, two at 20 min and one at 30 min of infusion. Although all groups received intravenous gas, only the Ar, He, and N2 groups showed clinical signs of pulmonary arterial gas embolism, as evidenced by the changes in mean pulmonary arterial pressure (MPAP) and ET-CO2 (Figs. 1 and 2). The maximum changes in the soluble gas groups were mild, with MPAP increased from 13.6 0.4 mmHg to 15.5 0.7 mmHg in the CO2 group (p 0.0017), and from

Analysis
All data are reported as mean SEM. Analysis of variances for repeated measurements was performed for each group. A p value less than 0.05 was considered significant interaction, and when this occurred, Dunnetts multiple range post hoc test was performed to identify the source of differences from baseline. For the N2 group, the data were analyzed for the first 15 min of infusion due to death of the pigs and a paired t-test was performed for blood-gas variables since only baseline and 15-min values were available. The following null hypotheses were tested: (1) Intravenous infusion of gases at the rate of this experiment (0.1 ml kg1 min1) does not cause acidbase imbalance or hemodynamic change. (2) There is no difference in the degree of pulmonary gas embolism created by the soluble vs relatively

343

Fig. 2. End-tidal CO2 levels closely reflected the changes in MPAP. Intravenous infusion of insoluble gases led to significant but reversible reduction in ET-CO2 levels. The slight changes in the carbon dioxide group were statistically significant.

14.4 0.8 mmHg to 16.2 2.2 mmHg in the N2O group (p 0.8106). The decrease in ET-CO2, from 30.8 2 mmHg to 29.8 1.6 mmHg in the CO2 animals was statistically significant (p 0.0176). For the relatively insoluble gas groups, there were more marked increases in MPAP and decreases in ET-CO2. Animals infused with helium or nitrogen experienced similar rates of change. Mean PAP increased by 109% in the helium group (from 13.3 0.3 mmHg to 27.8 2.6 mmHg) and by 147% in the nitrogen group (from 13.1 1.2 mmHg to 32.3 1.4 mmHg) after 15 min of IV infusion. The rates of increase in MPAP were 0.96 mmHg/min and 1.28 mmHg/min for helium and nitrogen groups, respectively. At the same time, ET-CO2 dropped at a rate of 1.74 mmHg/min and 1.82 mmHg/min for helium and nitrogen groups, respectively. Animals infused with argon also experienced changes in MPAP and ET-CO2, suggesting pulmonary arterial gas embolism, with a slower initial rate of 1.3 mmHg/min for MPAP changes but a similar magnitude of change (23.25 4.48 mmHg) after 20 min (Fig. 2). The slight but statistically significant changes in MPAP and ET-CO2 seen in the CO2 group were still present after 15 min of recovery. The Ar and He groups had returned to baseline MPAP values at 15 min. ET-CO2 levels had returned to baseline values in the He group at that time but remained low in the argon group. Only one nitrogen animal survived to the 15-min recovery point.

with ET-CO2 level during venous gas embolism caused by the relatively insoluble gases (Fig. 3). The argon group experienced the most severe hypercapnea, with a significant change in PaCO2 from 41.9 2.5 mmHg to 74.6 10.0 mmHg at 30 min (p 0.0044), associated with a severe acidemia of 7.22 0.05 (p 0.0007) and tachycardia of 115 13 beats/min ( p 0.0467). The increase in MAP was not statistically significant (Table 2). Helium induced similar changes in arterial blood gas results, with PaCO2 as high as 63.5 6.2 mmHg (p 0.0012) and pH as low as 7.25 0.04 (p 0.0003) at 30 min. Animals infused with helium quickly developed tachycardia in addition to a significant drop in mean arterial pressure to 52 8 mmHg (p 0.0056). Although HR and MAP values returned to baseline within 15 min after infusion for the Ar and He groups, the changes in acidbase balance were not corrected at that time (Table 2). Nitrogen infusion resulted in a 75% mortality rate during the infusion period. The animals quickly developed significant hypercapnea (PaCO2 61.2 4.7 mmHg, p 0.0159) and acidemia (pH 7.29 0.03, p 0.0040), in addition to tachycardia (HR 117 8 beats/min, p 0.0110). Deaths were sudden, most likely secondary to cardiac events, but we did not perform necropsy on these animals.

Discussion Cardiopulmonary changes associated with pulmonary arterial embolism In addition to the changes in MPAP and ET-CO2, animals infused with the relatively insoluble gases also experienced severe acidemia, hypercapnea, and tachycardia (Table 2). With infusion of the more soluble gases, CO2 and N2O, arterial blood gases remained essentially unchanged, and there were no significant changes in heart rate or mean arterial pressure. Arterial CO2 tension did not correlate well Carbon dioxide has been the traditional gas for pneumoperitoneum during laparoscopic surgery due to its rapid excretion, good patient tolerance, nonflammability, low cost, and availability. However, numerous clinical and laboratory studies have demonstrated the potential problems associated with the transperitoneal absorption of CO2 during laparoscopy. In one study, healthy patients undergoing 20 mmHg pneumoperitoneum with CO2 in Trendelenburg position experienced a statistically significant rise in PaCO2, with a fall in pH [1]. A study by Liu and colleagues showed that in-

344

Fig. 3. Evidence of the discrepancy between ET-CO2 levels and systemic carbon dioxide tension during intravenous infusion of insoluble gases. ET-CO2 level is not a reliable monitor of arterial CO2 tension when there is a problem with gas exchange at the alveoli.

Table 2. Acidbase balance and hemodynamics during intravenous gas infusion (mean SEM) Infusion Gases pH CO2 N2O Ar He N2 CO2 N2O Ar He N2 CO2 N2O Ar He N2 CO2 N2O Ar He N2 n 6 6 4 4 4 6 6 4 4 4 6 5 4 4 4 6 6 4 4 4 Baseline 7.49 0.01 7.46 0.02 7.42 0.03 7.43 0.02 7.47 0.01 34.8 0.9 38.1 2.9 41.9 2.5 39.9 1.5 36.9 1.4 99 8 94 6 98 9 98 3 104 3 53 3 52 3 47 4 60 7 56 7 15 min 7.48 0.01 7.45 0.03 7.33 0.02 7.30 0.02a 7.29 0.03 36.7 1.4 39.9 4.6 54.5 4.0 57.7 2.3a 61.2 4.7 94 7 90 6 96 8 127 17 118 8a 52 3 50 3 41 1 54 8 39 2 30 min 7.48 0.02 7.42 0.05 7.22 0.05a 7.25 0.04a N/A 36.3 1.2 45.1 7.3 74.6 10.0a 63.5 6.2a N/A 94 7 96 9 115 13 141 24a N/A 51 2 52 3 44 2 52 8a N/A Recovery 7.48 0.01 7.42 0.04 7.24 0.05a 7.28 0.05a N/A 36.8 1.5 42.8 5.8 66.3 7.8a 56.3 5.9a N/A 94 6 94 7 110 7 130 21 N/A 50 2 52 3 46 1 55 7 N/A ANOVA (p value) NS NS 0.0007 0.0003 0.0040b NS NS 0.0044 0.0012 0.0159b NS NS 0.0467 0.0142 0.0110 NS NS NS 0.0056 NS

PaCO2 (mmHg)

HR (beats/min)

MAP (mmHg)

a b

Significant vs baseline, Dunnetts post hoc test. Paired t-test results.

creased minute ventilation was required to correct rising PaCO2 in patients undergoing laparoscopic cholecystectomy [14]. Intraoperative acidemia as low as 7.33 may occur during laparoscopic cholecystectomy in up to 29% of patients [24]. In a porcine model of laparoscopic cholecystectomy, pulmonary excretion of CO2 increased up to 70% above baseline after 1 h of CO2 insufflation, without any concomitant changes in oxygen consumption. At the same time, the ratio of minute ventilation to total pulmonary CO2 excretion decreased and the respiratory quotient increased steadily, suggesting no evidence of either respiratory compromise or increased metabolism [7]. In a randomized cross-over comparison between CO2 and nitrogen pneumoperitoneum without cholecystectomy, the same findings were confirmed in the CO2 group, but not in the nitrogen group [8]. These data demonstrated conclusively that there

was substantial systemic CO2 absorption during pneumoperitoneum and that the pharmacological effects of the absorbed CO2, not the mechanical effects of the intraabdominal pressure, are responsible for the problems associated with CO2 insufflation. In addition, all of these studies documented progressive hypercapnea and acidemia during CO2 pneumoperitoneum with a fixed minute ventilation. In view of these potential deleterious effects of carbon dioxide absorption during laparoscopy, an active search for alternatives to CO2 pneumoperitoneum has taken place. The current contenders for replacement of CO2 as the insufflating agent for laparoscopic surgery are nitrous oxide, argon, helium and nitrogen. Nitrous oxide has been used extensively in laparoscopic gynecology. This gas has no known metabolic consequences from absorption during pneumoperitoneum, yet it also has a relatively high solubility in the

345

blood, similar to that of CO2 (Table 2). However, it has not been used during general laparoscopic surgical procedures due to the great concern regarding the lack of suppression of combustion [9]. Inert gases such as argon and helium have been proposed as attractive alternatives because they would theoretically avoid the metabolic effects of CO2 and the problems with combustion [3, 13]. Argon is of particular interest because of the recent development of the argon beam coagulator. It has been shown, however, that argon gas may not be physiologically inert since it caused significant changes in systemic vascular resistance and stroke volume during pneumoperitoneum in a pig model [4]. Furthermore, a porcine study revealed that venous gas embolism may occur during argon-enhanced coagulation of the liver [17]. Nitrogen is also of interest because its low blood solubility might prevent its transperitoneal absorption during pneumoperitoneum. In our previous study of the mechanical vs pharmacological effects of CO2 pneumoperitoneum, it was shown that nitrogen pneumoperitoneum does not create the metabolic changes seen with the use of CO2 [8]. On the other hand, the low solubilities of gases such as argon, helium, and nitrogen may significantly increase the risks associated with the potential problem of venous gas embolism during pneumoperitoneum. Venous gas embolism is a rare but potentially lethal condition in man. Gas bubbles in the venous system have been associated with neurosurgical procedures, traumatic venotomies, and diving decompression. This complication has also been reported to occur during laparoscopic surgery [2, 3, 12, 16]. The severity of venous gas embolism depends on the amount and type of gas embolized and the site and rate of administration. Significant bubbles in the venous circulation may occlude blood vessels, impairing flow. Upon entering the pulmonary circulation, embolized gas bubbles are dispersed into the pulmonary capillary bed. There, they alter vascular resistance and blood flow distribution, leading to physiologic shunting [6]. Physiologic dead space increases, gas exchange is impaired, and endtidal CO2 falls. When massive, venous gas embolism may obstruct the pulmonary outflow tract, leading to acute right ventricular failure. Paradoxical arterial embolism may occur, due to intracardiac shunts or transpulmonary passage of the bubbles [15, 20]. Subsequent compromise of the coronary blood flow is presumed to be the cause of death, either from intracardiac gas or from pulmonary vascular obstruction leading to reduced blood flow from the right to the left heart [5, 21]. In addition to their mechanical effects, gas bubbles can also cause biophysical effects at the bloodbubble interface. They incite an inflammatory reaction by activation of the Hageman factor, which in turn activates the complement cascade, leading to leakage of intravascular fluid and to microvascular sludging, resulting in decreased tissue perfusion and ischemia [11]. The main objective of this study is to determine the relative risk of venous gas embolism of soluble vs nonsoluble gases; therefore, each gas was infused at the same rate and site of administration. A porcine model was chosen for this study since young pigs have pulmonary circulation anatomically similar to that of adult humans [20]. We used an infusion model because venous gas embolism occurring during laparoscopic procedures is best simulated using continuous infusion as opposed to bolus injection. Gases were

infused through a 0.062-inch internal diameter femoral venous catheter. Bubble size was not measured since any bubbles which remained undissolved upon reaching the heart would coalesce and then be fractured into bubbles of different sizes by the pumping of the ventricle [21]. Mean pulmonary arterial pressure (MPAP) was used as a clinical measure of the degree of pulmonary embolism. Although Doppler is more sensitive for the detection of intravascular microbubbles, MPAP remains a reliable indicator of clinically significant pulmonary gas embolism. Furthermore, while halothane may act as a dilator of the pulmonary circulation and therefore may lower the threshold for bubble breakthrough into the arterial system [25], all five groups received halothane anesthesia, so this study effectively shows the relative responses to venous infusion with the different gases. Our data revealed that venous gas embolism with relatively insoluble gases (argon, helium, nitrogen) is tolerated very poorly. The animals had significant elevations in mean PAP and PaCO2, along with decreases in the end-tidal CO2 and arterial pH. These findings are consistent with venous gas embolization. Soluble gases (CO2 and N2O), on the other hand, are relatively well tolerated. Although there were statistically significant changes in mean PAP and ETCO2 in the CO2 group, these changes were slight and were not indicative of clinically significant embolization. There were no significant changes in the other parameters that were measured in these two soluble-gas groups. It is possible that at the rate of infusion in the current model, N2O and CO2 molecules may have been partially or completely dissolved by the time they arrived at the right ventricle. The metabolic consequences of gas infusion were more severe in animals infused with relatively insoluble gases. In fact, 75% (3/4) of animals infused with nitrogen gas died during the infusion period. Whether the cause of death in these animals was paradoxical arterial embolism or acute right ventricular failure could not be determined with our methodology. A 33% patent foramen ovale rate has been reported in pigs [20], but we did not have necropsy data to confirm or rule out this possibility. On the other hand, it is known that nitrogen is not a physiologically inert gas. Pulmonary emboli of nitrogen gas have been shown to produce acute changes in pulmonary mechanics similar to those following autologous thrombo-emboli [11], presumably via activation of thrombin, platelet aggregation, and acute airway constriction. In the same study, neither argon nor helium caused such effects on pulmonary mechanics, presumably due to their noble gas characteristics. However, in surviving animals infused with these inert gases in our study, the profound hypercapnea and acidemia suggest significant respiratory compromise, most likely secondary to intrapulmonary shunting. The responses to intravenous infusion of relatively insoluble gases observed in our experimental model clearly demonstrated the inherent risk of these gases if they are used to create pneumoperitoneum for laparoscopic surgery. Given the same amount and site of inadvertent venous gas embolism, the less soluble gases result in more significant clinical sequellae. They induced higher and faster pulmonary hypertension and more profound systemic hypercapnea and acidemia. We caution against the use of these gases for pneumoperitoneum during laparoscopic surgery. Any po-

346

tential gaseous alternative to CO2 for pneumoperitoneum must be subjected to thorough study of its potential pharmacological effects as well as its risk of pulmonary gas embolism.
Acknowledgment. This work was supported in part by an educational grant from Ethicon Endo-Surgery, Inc.

References
1. Alexander GD, Brown EM (1969) Physiologic alterations during pelvic laparoscopy. Am J Obstet Gynecol 105: 10781081 2. Beck DH, McQuillan PJ (1994) Fatal carbon dioxide embolism and severe haemorrhage during laparoscopic salpingectomy. Br J Anaesth 72: 243245 3. Cottin V, Delafosse B, Viale JP (1996) Gas embolism during laparoscopy. A report of seven cases in patients with previous abdominal surgical history. Surg Endosc 10: 166169 4. Eisenhauer DM, Saunders CJ, Ho HS, Wolfe BM (1994) Hemodynamic effects of argon pneumoperitoneum. Surg Endosc 8: 315320 5. Geoghegan T, Lam CR (1953) The mechanism of death from intracardiac air and its reversibility. Ann Surg 138: 351359 6. Gildenberg PL, OBrien RP, Britt WJ, Frost EA (1981) The efficacy of Doppler monitoring for the detection of venous air embolism. J Neurosurg 54: 7578 7. Ho HS, Gunther RA, Wolfe BM (1992) Intraperitoneal carbon dioxide insufflation and cardiopulmonary functions. Laparoscopic cholecystectomy in pigs. Arch Surg 127: 928933 8. Ho HS, Saunders CJ, Gunther RA, Wolfe BM (1995) Effector of hemodynamics during laparoscopy. CO2 absorption or intraabdominal pressure? J Surg Res 59: 497503 9. Hunter JG, Staheli J, Oddsdottir M, Trus T (1995) Nitrous oxide pneumoperitoneum revisited. Is there a risk of combustion? Surg Endosc 9: 501504 10. Ivankovich AD, Miletich DJ, Albrecht RF, Heyman HJ, Bonnet RF (1975) Cardiovascular effects of intraperitoneal insufflation with carbon dioxide and nitrous oxide in the dog. Anesthesiology 42: 281287

11. Khan MA, Alkalay I, Suetsugu S, Stein M (1972) Acute changes in lung mechanics following pulmonary emboli of various gases in dogs. J Appl Physiol 33: 774777 12. Lantz PE, Smith JD (1994) Fatal carbon dioxide embolism complicating attempted laparoscopic cholecystectomy. Case report and literature review. J Forensic Sci 39: 14681480 13. Leighton TA, Liu SY, Bongard FS (1993) Comparative cardiopulmonary effects of carbon dioxide versus helium pneumoperitoneum. Surgery 113: 527531 14. Liu SY, Leighton T, Davis I, Klein S, Lippmann M, Bongard F (1991) Prospective analysis of cardiopulmonary responses to laparoscopic cholecystectomy. J Laparoendosc Surg 1: 241246 15. Moon RE, Camporesi EM, Kisslo JA (1989) Patent foramen ovale and decompression sickness in divers. Lancet 1: 513514 16. Moskop RJ Jr, Lubarsky DA (1994) Carbon dioxide embolism during laparoscopic cholecystectomy. South Med J 87: 414415 17. Palmer M, Miller CW, Van Way III CW, Orton EC (1993) Venous gas embolism associated with argon-enhanced coagulation of the liver. J Invest Surg 6: 391399 18. Paolucci V, Gutt CN, Schaeff B, Encke A (1995) Gasless laparoscopy in abdominal surgery. Surg Endosc 9: 497500 19. Severinghaus JW, Stupfel M, Bradley AF (1956) Accuracy of blood pH and PCO2 determination. J Appl Physiol 9: 189196 20. Vik A, Brubakk AO, Hennessy TR, Jenssen BM, Ekker M, Slordahl SA (1990) Venous air embolism in swine. Transport of gas bubbles through the pulmonary circulation. J Appl Physiol 69: 237244 21. Vik A, Jenssen BM, Brubakk AO (1994) Comparison of haemodynamic effects during venous air infusion and after decompression in pigs. Eur J Appl Physiol 68: 127133 22. Weathersby PK, Homer LD (1980) Solubility of inert gases in biological fluids and tissues. A review. Undersea Biomed Res 7: 277296 23. Wittgen CM, Andrus CH, Fitzgerald SD, Baudendistel LJ, Dahms TE, Kaminski DL (1991) Analysis of the hemodynamic and ventilatory effects of laparoscopic cholecystectomy. Arch Surg 126: 9971001 24. Wittgen CM, Naunheim KS, Andrus CH, Kaminski DL (1993) Preoperative pulmonary function evaluation for laparoscopic cholecystectomy. Arch Surg 128: 880885 25. Yahagi N, Furuya H, Sai Y, Amakata Y (1992) Effect of halothane, fentanyl, and ketamine on the threshold for transpulmonary passage of venous air emboli in dogs. Anesth Analg 75: 720723

Surg Endosc (1997) 11: 351353

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Preemptive analgesia in the laparoscopic patient

C. G. Mixter III, T. R. Hackett
Exeter Hospital, Exeter, NH 03833, USA Received: 5 April 1996/Accepted 2 November 1996

Abstract Background: One hundred consecutive laparoscopic patients were prospectively followed in the Post-anesthesia Recovery Unit (PAR) in a community hospital. Methods: Data was collected regarding (1) intraoperative administration of ketorolac, (2) instillation of local anesthesia into the wound, and (3) requirements for analgesic administration in the PAR. Those patients receiving both forms of preemptive analgesia required less narcotic administration in the PAR. The results were highly significant. Results: Based on these data and the observance of markedly reduced pain in patients during the early postoperative period, an institutional plan of care was developed which has resulted in the virtual elimination of the need to administer narcotics to patients undergoing routine laparoscopic surgical procedures. Conclusions: The resultant plan of care, which includes preemptive analgesia, rapid ambulation, early feeding, and routine timed administration of non-narcotic pain medications, is presented. Data is also presented which demonstrates a more rapid discharge of patients from the hospital. Key words: Laparoscopic Pain Preemptive Analgesia

Study population and data collection

One hundred consecutive patients who underwent laparoscopic procedures were studied (Table 1). All patients undergoing laparoscopy, whether for gynecological or general surgical purposes, were included as long as the patient did not have any additional procedures performed. Unknown to the physicians, patients of all surgeons (ten) performing laparoscopy were studied. All anesthesia was given by a group of five anesthesiologists and no attempt was made to standardize either the amount of medication administered or the type of anesthetic agents used. Additionally, no attempt was made to standardize the care of this group of patients in the Postanesthesia Recovery Unit (PAR). Four data elements were recorded by the PAR staff: (1) whether or not a patient had received bupivacaine in the OR, (2) whether or not a patient had received ketorolac in the OR, (3) whether the patient had received neither drug in the OR, and (4) whether or not the patient received narcotic medication in the PAR. Data was not evaluated until 100 patients had been accrued to the study.

For many years various authors have advocated the administration of medication before pain stimulation occurs [3]. The success of such preemptive analgesia is variable [1, 2, 4, 6]. In early 1992, one of the authors (C.G.M.) began to use intravenous ketorolac (60 mg) preemptively during laparoscopic surgical procedures. Additionally, 30 cc of 0.5% bupivacaine was injected into each trocar site. These patients appeared to require less analgesia during the postoperative period. As a result, a simple study was devised to test the validity of this observation.

Correspondence to: C. G. Mixter III, Perry Medical Office Building, Suite 301, 3 Alumni Drive, Exeter, NH 03833, USA

Results From 30 to 60-mg of ketorolac was administered to patients receiving the drug. In general, 10 to 30 cc of 0.5% bupivacaine was used for subcutaneous injection of the trocar sites. The data relating the intraoperative use of the study medications and the administration of narcotics in the PAR are presented in Tables 2 and 3. A multivariate analysis was conducted utilizing the Log-Linear Model. Using this model the data was felt to be highly significant and the probability that a patient will receive narcotic medication in the PAR is 25% if the patient receives both bupivacaine and ketorolac intraoperatively, 45% if the patient received only ketorolac, 59% if the patient received only bupivacaine, and 81% if the patient received neither drug. The data is presented in graphical format in Fig. 1. Earlier discharge of patients has resulted; 44% of all elective cholecystectomy patients have been discharged on the day of surgery in comparison to 23% previously. Most patients who do not return to their homes on the day of surgery have had surgery performed late in the day and are discharged the next morning. Over 98% of laparoscopic inguinal hernia repairs have been discharged on the day of surgery. Most importantly, no patients have required narcotic medications following discharge and there have been no readmissions in this group. No renal problems were noted on any patients. It was also noted that those patients receiving both

352 Table 1. Type of laparoscopic procedure Cholecystectomy Herniorraphy Tubal ligation Appendectomy Lysis Bilateral Pelvic nodes Oopherectomy Hernia Excision of cyst of tube Cystectomy Colostomy Cholecystectomy & liver biopsy Cholecystectomy & EGD Aspiration ovarian cyst Aspiration of cyst Appendectomy Grand total 26 24 23 8 5 3 1 1 1 1 1 1 1 1 1 1 1 100 Table 2. Medication administered in the OR to the study group Analgesia study 10-92 to 6-93 100 pts receiving medication in the OR OR medication Totals Ketorolac + bupivacaine 63 Ketorolac only 19 Bupivacaine only 11 Neither drug 7

Table 3. PAR narcotic requirements of patient groups Patients receiving medication in PAR Recovery room medication given Operating room medication Totals Ketorolac + bupivacaine 16 Ketorolac only 8 Bupivacaine only 6 Neither drug 6 No recovery room medication given Ketorolac + bupivacaine 47 Ketorolac only 11 Bupivacaine only 5 Neither drug 1

Fig. 1. Graph of percentage of patients in each group receiving narcotics in the PAR.

medications intraoperatively complained of virtually no right-sided shoulder pain during the postoperative period and none required narcotic medication after discharge from the PAR.

Discussion Preemptive analgesia has been advocated in patients undergoing surgical procedures. Laparoscopic procedures offer the opportunity to maximally utilize these techniques. The incisions are small and localized and can easily be infiltrated with an long-acting local anesthetic. Blocking of the afferent nerve fibers in combination with the use of a sys-

temic nonsteroidal analgesic agent has been demonstrated to reduce the need for narcotic medication in the PAR. The advantages of reducing narcotic usage are evident. Patients are more alert and cooperative. They can ambulate more rapidly. They have less nausea and able to take fluids and food earlier during recovery period [5]. After the completion of this study a program was instituted to maximize the benefits of preemptive analgesia in our patients. Unless contraindicated all patients receive ketorolac and bupivacaine intraoperatively. Since the completion of our study, concern has been raised regarding the safety of ketorolac as a prophylactic agent. We have not observed increased bleeding in our patients and a recently published study confirms the safety of administration of ketorolac [7]. We now employ several techniques to accomplish this aim. We employ a multidisciplinary approach [8]. Nursing has high expectations of our laparoscopic patients, and rapid discharge is thereby encouraged. A preoperative antiemetic is used which reduces the painful stimulation of postanesthetic retching and vomiting. Early ambulation of the patients reduces muscle spasm and pain. At discharge these laparoscopic patients are instructed to take 400 mg of ibuprofen every 4 h (not PRN) for the first 24 h, providing continuing analgesic coverage. Preemptive analgesia in laparoscopic patients coupled with the avoidance of postoperative narcotics leads to early discharge, excellent patient acceptance of surgery, potential reduction of complications, and more cost-effective care. Further studies are needed to provide additional documentation of these observations.
Acknowledgment. The authors express appreciation to Jessica Utts at the University of California at Davis for the statistical analysis of the data.

353

References
5. 1. Higgins MS, Givogre JL, Marco AP, Blumenthal PD, Furman WR (1994) Recovery from outpatient laparoscopic tubal ligation is not improved by preoperative administration of ketorolac or ibuprofen. Anesth Analg 1994: 274280 2. Huffnagle HJ, Norris M, Leighton G (1996) Ilioinguinal iliohypogastric nerve blocksbefore or after cesarean delivery under spinal anesthesia? Anesth Analg 82: 812 3. Kehlet H (1994) Postoperative pain reliefa look from the other side. Reg Anesth 19: 369377 4. Michaloiakou C, Chung F, Sharma S (1996) Preoperative multimodal

6.

7.

8.

analgesia facilitates recovery after ambulatory laparoscopic cholecystectomy. Anesth Analg 82: 4451 Petros JG, Realica R, Ahmad S, Rimm EB, Robillard RJ (1995) Patientcontrolled analgesia and prolonged ileus after uncompleted colectomy. Am J Surg 70: 371374 Ready LB, Brown CR, Stahlgren LH, Egan KJ, Ross B, Wild L, Moodie JE, Jones SF, Tommerassen M, Trierwieler M (1994) Evaluation of intravenous ketorolac administered by bolus of infusion for treatment of postoperative pain. Anesthesiology 80: 12771286 Strom BL, Berlin JA, Kinman JL, Splitz PW, Hennessy S, Feidman H, Kimmel S, Carson JL (1988) Parenteral ketorolac and risk of gastrointestinal and operative site bleeding. JAMA 275: 376382 Wheatly RG, Samaaan (1995) Postoperative pain relief. Br J Surg 82: 292294

Case reports
Surg Endosc (1997) 11: 381382

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Metastatic melanoma of gastrocolic ligament treated with minimally invasive approach

V. I. Sreenivas,1,2,* M. Damen,1 S. Agarwal,1 V. Pothula1
1 2

Hospital of St. Raphael, 1450 Chapel Street, New Haven, CT 06511 USA Department of Surgery, Yale University School of Medicine, New Haven, CT 06511 USA

Received: 19 April 1996/Accepted: 19 May 1996

Abstract. A patient with metastatic melanoma of the gastrocolic ligament was rendered clinically tumor free using a minimally invasive procedure. The technique and the rationale for the procedure are reviewed. Key words: Metastatic melanoma Laparoscopy and melanoma

Melanoma metastasizes both by lymphatic and hematogenous routes. Surgical excision of nodal metastasis is the only effective method for cure and local disease control. Elective regional lymph node dissection (ERLD) has been shown to improve survival in patients whose tumor is between 1.0 and 4.0 mm thick. ERLD is not indicated in patients with thin (<1 mm) melanoma as these lesions are generally localized, and the cure rate following wide excision is high. The procedure is not indicated in patients with thick (>4 mm) lesions as the risk of distant metastasis is high [4]. The role for surgical intervention in patients with hematogenous spread is less clear and the scope limited since the metastic involvement may be multiple and widespread. In the subset of patients who present with solitary metastasis or successive localized metastases amenable for excision, survival may be prolonged by metastasectomy. These patients should be rendered clinically tumor free with minimal discomfort and a short hospital stay. We report on a patient who had metastatic melanoma of the gastrocolic ligament successfully removed using a laparoscopic technique. Case report
A 59-year-old woman 6 years earlier had been diagnosed as having a 0.8-mm melanoma of the left lower eyelid, which was widely excised. Five

months later she was found to have metastasis to the left parotid lymph node. She underwent left parotidectomy and radical neck dissection followed by chemotherapy. Four years later she underwent splenectomy for splenic metastasis found on a routine follow-up computed tomographic (CT) scan of the abdomen. No other metastasis was seen in the abdomen. She received chemo- and immunotherapy. Two years later on a routine follow-up CT scan she was found to have a 2 1 cm mass in gastrocolic ligament suspected to be a metastastic lesion (Fig. 1). No other metastastic lesions were identified. Laparoscopic excision of the mass was undertaken. At operation few adhesions were found in the left upper quadrant of the abdomen from previous splenectomy. No metastatic lesions were found other than the one in the gastrocolic ligament. The lesser sac was entered after opening the gastrocolic ligament and the tumor was free posteriorly. The gastrocolic ligament was detached from stomach and colon and the contained tumor mass was excised with surrounding healthy tissue all around it. Hemostasis was obtained using an endovascular stapler. The specimen was placed in a retrieval bag and removed from the abdominal cavity. Postoperative course was uncomplicated and the patient was discharged on the following day. Pathologic examination confirmed the diagnosis of metastatic melanoma and the margins of resection were clear.

Correspondence to: V. Sreenivas, 1423 Chapel St., New Haven, CT 06511, USA

Discussion In their series of patients who had surgical resection of metastatic melanoma Fuen et al. [2] report a median survival of 18 months. However, in the subgroup of patients who were rendered clinically tumor free through successive resections the median survival was 36 months. They observed that the disease-free interval between the diagnosis of melanoma and the appearance of metastasis was an important prognostic factor; the longer the interval the better the prognosis. Thirteen months was the time interval between the favorable and the unfavorable group. According to Karakousis et al. [3] the number of metastases is also an important prognostic factor. In their series, patients with a solitary metastasis had a median survival of 26 months and estimated 5-year survival rate of 33%. The respective figures for those with two to four lesions were 19 months and 11% and for more than four lesions, 9 months and 13%. Laparoscopy offers several advantages in the management of patients with metastatic melanoma. As a diagnostic tool for staging the disease it is superior to other noninvasive procedure (CT scan, sonography, and liver scintigraphy). Caldironi et al. [1] in a series of 297 patients found that in approximately 3% of the patients the stage was

382

the fear that laparoscopy may predispose for disseminated metastasis and/or trocar site recurrence, as has been observed with patients with intraabdominal malignancy [5, 6]. The exact mechanism and the magnitude of this complication are as yet unknown. It appears that the benefit of laparoscopy in terms of decreased pain and disability outweighs the possible disadvantage of tumor dissemination and trocar site recurrence in this group of patients with poor prognosis.

References
1. Caldironi MW, Nitti D, Schiavon M, Rossi CR, Aldino MT, Azzena B (1989) Laparoscopy in the abdominal staging of melanoma. Eur J Cancer Clin Oncol 25: 223226 2. Feun LG, Gutterman J, Burgess MA, Hersh EM, Mavligit G, McBride CM, Benjamin RS, Richman SP, Murphy WK, Gerald PB, Brown BW, Mountain CF, Leavens ME, Freirch EJ (1982) The natural history of resectable metastatic melanoma (Stage IV A melanoma). Cancer 50: 16561683 3. Karakousis CP, Velez A, Driscoll DL Takita H (1994) Metastasectomy in malignant melanoma. Surgery 115: 295302 4. Mansfield PF, Lee JE, Balch CM (1994) Cutaneous melanoma: current practice and surgical controversies. Curr Probl Surg 31: 306313 5. Sailer M, Debus S, Fuchs KH, Thiede A (1995) Peritoneal seeding of gallbladder cancer after laparoscopic cholecystectomy. Surg Endosc 9: 12981300 6. Savalgi RS (1995) Mechanism of abdominal wall recurrence after laparoscopic resection of colon cancers. Semin Laparosc Surg 2: 158162

Fig. 1. CT scan of the abdomen showing metastatic melanoma in gastrocolic ligament (arrow).

changed from I to III and in 8% from II to III following diagnostic laparoscopy. In 9% of the patients the stage was downgraded. As a therapeutic tool it can render the patient clinically tumor free with less pain and disability compared to laparotomy. It also could help in identifying patients who would not benefit from resection because of multiple and or widespread metastasis in the abdomen. A possible disadvantage for laparoscopic removal of metastatic melanoma is

Surg Endosc (1997) 11: 404

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Gastrointestinal recovery of laporoscopic vs open colon surgery

I look forward to every issue of Surgical Endoscopy and eagerly read each article. Surgical Endoscopy is indeed a leader in endoscopic as well as laparoscopic surgery. I was somewhat appalled, however, to see the Gastrointestinal recovery of laparoscopic versus open colon surgery article in an issue of Surgical Endoscopy [1]. This was a poorly conceived, poorly executed, and misleading article. In this article an attempt was made to compare GI tract recovery in open surgery to results for laparoscopic surgery on a series of unrelated colonic procedures. The open group of procedures required an average of 200 min or less and the laparoscopic procedures required more than 400 and in some instances almost 500 min to complete. The authors wrote off this 8-h laparoscopic procedure as the result of resident training, but nevertheless tried to correlate the findings in a logical manner with the shorter, albeit resident-completed, open procedures. This is close to irresponsible and can only lead to errors in interpretation of the data presented. The data showed that the recovery of ileus from each of the procedures was approximately the same, but that earlier hospital discharge with laparoscopic surgery occurred despite taking twice to three times as long per given operation. This was apparently a tribute to individual patients resilience or perhaps indeed to lesser insult from the laparoscopic technique. However, the conclusion drawnthat there is no benefit from laparoscopic surgery in resolution of the ileusis quite opposite that seen in many other series, including those of Jacobs and Plasencia, Kim, Petelin, Ed Phillips, and my series. In these reasonably timed procedures the return of GI function with resolution of ileus (although not measured scientifically) was at 2 to 3 days and many times much before this. The basic difference between these latter surgeons and the work presented in the Virginia article is that these surgeons were experienced laparoscopic surgeons whose operating times were reasonable and not nearly as long as those presented in this article. It is interesting that the remainder of the data for the most part substantiates data seen by other investigatorsthat is, the blood loss from each laparoscopic surgery was much less despite the longer duration of the procedure and the amount of analgesia postop was less as well. I would hope that the investigators would rethink their technique and strongly consider learning to do the procedure in a reasonable period of time (such as that recently demonstrated by Dr. Ed Phillips, being approximately the same as open time or at least not more than, one-and-a-half times the amount that of open procedure) and repeat their study. I think then and only then could we get worthwhile data regardless of what the outcome might be. It is interesting that another article which was perhaps a little better controlled (I know the investigators very well) will be presented in the EAES meeting by Tittle, Shippers, and Schumpelick shows opposite findings. I am very interested in this subject and intrigued by the exact opposite findings in these two studies. Perhaps the doctors from Charlottesville, Virginia, would be interested in reading the article by doctors Shippers and Tittle. Reference
1. Hotoyezaya M, Dix J, Meutis EP, Minasi JS, Schirmer BD (1996) Gastrointestinal recovery following laparoscopic vs open colon surgery. Surg Endosc 10: 485489

M. E. Franklin, Jr.
4242 East Southcross Suite 1 San Antonio, TX 78222 USA

Editorial
Surg Endosc (1997) 11: 319320

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Appendectomylaparoscopic or open?
In this issue of Surgical Endoscopy, the Rotterdam Group report the results of a prospective randomized controlled trial (RCT) comparing open appendectomy (OA) with laparoscopic appendectomy (LA) in patients with a clinical diagnosis of acute appendicitis. This is the ninth reported RCT on this subject and the present study is one of the largest to date, with 201 randomized patients. Collectively, the verdict gained from the majority of these studies is clearly in favor of LA on two countssignificant reduction is postoperative pain and significant reduction in the postoperative wound infection rate. By contrast only four such RCTs have documented a shorter hospital stay. This is not surprising since hospital stay must reflect the severity of the septic insult to the patient at the time of the emergency admission to hospital. There is no logic to back the opponents of LA that it does not confer any benefit simply because there is no significant reduction in the postoperative hospital stay. Although important, hospital stay is one of many variables which must be taken into account in the final adjudication of the best approach, open or laparoscopic for any surgical disorder including acute appendictis. Latent benefits of laparoscopic appendectomy There are undoubted latent benefits of the laparoscopic approach, some of which have been documented by the Rotterdam trial. These merit consideration since they risk being overlooked in the cost-effective argument. Laparoscopic appendectomy gives a better evaluation of the peritoneal cavity than is abtained by the standard gridiron OA. The procedure allows rapid inspection of the paracolic gutters and the pelvic cavity which is not possible with the open grid-iron approach. The case reported in the Dutch RCT of a patient with missed perforated diverticular disease in the OA arn is a case in point. This advantage is of particular relevance to females in the child bearing age who develop right iliac fossa pain as laparoscopy provides the best means for diagnosis of pelvic inflammatory and pelvic adnexal disease. Reduced postoperative pain after LA should translate itself in better pulmonary function and reduced incidence of patchy collapse and consolidation. This will probably be of clinical significance in patients at risk for these complications such as individuals with chronic obstructive airway disease and chronic smokers. Although no difference in the postoperative chest infection rate was observed in the Rotterdam study, the randomized cohorts studied consisted of young adults (mean ages 30.8y33.7y) and information on the pre-operative respiratory status and smoking habits were not recorded. The incidence of postoperative chest infection (1/201) reported in this RCT is unusually low and may indicate a low detection rate. Reduced adhesion formation. This has been largely overlooked in the continuing debate but it is, nonetheless a substantive gain for LA in the long-term. One study reported an adhesion rate of 80% after OA compared to 10% after LA when patients were laparoscoped three months after surgery [1]. Adhesion formation is now one of the most common causes of intestinal obstruction. Its role in the development of chronic abdominal pain, or though less certain, cannot be ignored. Downside to laparoscopic appendectomy There are problems which have to be addressed in relation to LA. Perhaps the most contentious is laparoscopic removal of a normal appendix. The Rotterdam group advocate routine removal of the appendix in these cases. I would take issue with such a policy. The argument that detection of appendiceal pathology cannot be made from inspection and probe palpation during laparoscopy is not tenable. Even if the serosa does not appear to be inflamed, an obstructed appendix and its mesoappendix always feels rigid. Irrespective of its safety, laparoscopic removal of a normal appendix is unnecessary, uses operating time, hospital beds and may be complicated. However infrequent, one major complication after such a procedure is one too many. Routine removal of a laparoscopically normal appendix would be counterproductive, if with the more widespread recourse to LA, the unnecessary appendectomy rate were to escalate significantly. Surely the sensible policy should be to take the appendix out when in doubt, but not always. Bacterial translocation from peritoneal cavity into the systemic circulation. This remains a matter of some concern and controversy. According to one experimental study in the rat positive pressure insufflation of the peritoneal cavity increases the translocation of E. coli from the peritoneal cavity into the blood stream [2] and one non-randomized clinical study reported two septic deaths in 14 patients with perforated gastric ulcer treated by laparoscopic repair [3]. However, the overwhelming majority of reported clinical studies have not detected any increase risk and the Rotterdam group confirms this. Nonetheless the possibility of a deleterious effect cannot be ruled out and case selection

320

with avoidance of the laparoscopic approach in the seriously ill and immuno-compromised patient seems prudent until more information becomes available from experimental studies and randomized clinical trials. References
1. De Wilde RL (1991) Goodbye to late bowel obstruction after appendicectomy. Lancet 338: 1012 2. Evasovitch MR, Clark TC, Horattas MC, Holda S, Treen L (1996) Does pneumoperitoneum during laparoscopy increase bacterial translocation? Surg Endosc 10: 117679

3. Eypasch E, Spangenberger W, Ure B et al. (1994) Laparoskopische und konventionelle ubernahungen perforierter epetischer ulzera-eine gegenuberstellung. Chirurg 65: 44550

A. Cuschieri
Department of Surgery Ninewells Hospital and Medical School Dundee DD1 GSY UK