Surg Endosc (1998) 12: 353354

Springer-Verlag New York Inc. 1998

Endoluminal stenting for benign colonic obstruction

R. Davidson, W. B. Sweeney
University of Massachusetts Medical Center, Department of Surgery, 55 Lake Avenue North, Worcester, MA 01655, USA Received: 10 June 1997/Accepted: 1 July 1997

Abstract. We report a case of complete descending colon obstruction due to diverticular disease that was initially managed by endoscopic stent placement followed by singlestage left colectomy with primary anastomosis. Traditional management of complete large bowel obstruction, whether due to benign or malignant disease, most often requires a temporary colostomy because of unprepared colon. In this case, preparation of the colon was accomplished by successful stenting of the benign colonic obstruction. We believe that endoscopic colonic stenting is an effective way of avoiding a temporary colostomy in patients with complete large bowel obstruction. Key words: Large bowel obstruction Colonic stent Endoprosthesis Diverticular stricture

The surgical management of colonic obstruction, regardless of its nature, is always a dilemma. Particular challenging are those lesions of the left colon and rectosigmoid region. Both diverticular disease and cancer affect this area of the colon frequently. Management has traditionally been surgical, including excision of the obstructing lesion and diversion of the fecal stream. A second procedure may eventually restore colonic continuity. This sequence of events, initiated by emergency exploration, often leads to increased morbidity and mortality [2]. Recently some investigators of malignant disease in the rectosigmoid colon have advocated the use of endoluminal stents to acutely decompress the bowel and avoid major surgery [7]. These stents have proven successful according to several published reports. We describe a case of colonic stenting for benign disease of the colon.

tion over a period of 12 months. She suffered from intermittent diffuse abdominal pain with bloating and occasional nausea and vomiting. She had also lost 15 lb in 1 month. Surgical consultation was obtained after an abdominal radiograph revealed a large bowel obstruction (Fig. 1). At the time of admission, her abdomen was notably distended but not tender. There was some hyerresonance on percussion but no evidence of a mass. She had no fever, and her white blood cell count and hematocrit were normal. She subsequently underwent a limited barium enema that demonstrated circumferential annular constriction of the descending colon with near total obstruction to retrograde flow (Fig. 2). After a frank discussion, the patient agreed to undergo colonic stenting in an attempt to achieve bowel decompression. The day following her admission, she underwent endoluminal stenting using both endoscopic and fluoroscopic guidance. The colonoscope was first passed to the distal point of the colonic stricture, which was at the descending to sigmoid colon junction. A guide wire was then placed through the stricture into the proximal dilated bowel under both fluoroscopic and endoscopic guidance. A catheter was placed over the guide wire, and a small amount of contrast was injected to accurately define the length of the stricture. The catheter was withdrawn and the guide wire left in place. Finally, two overlapping Wallstent enteral endoprostheses (Pfizer/Schneider, Minneapolis, MN 55442, USA) were introduced. A 22 60 mm stent was overlapped with a 22 90 mm stent in order to completely bridge the stricture (Fig. 3). After the deployment of both of these colonic stents in an overlapping fashion, immediate bowel decompression was achieved. Endoscopic evaluation of the stented colonic stricture was performed (Fig. 4). The patient tolerated the procedure well. Over the next few days, she underwent gradual colonic cleansing. On the 6th day following her stenting procedure, she underwent left colectomy with primary anastomosis (Fig. 5). The postoperative course was uneventful and she was discharged home on postoperative day 4, tolerating a regular diet and having formed bowel movements. The histopathology revealed diverticulitis with stricture of the left colon.

Discussion Endoscopic stenting of gastrointestinal strictures is not new. With advances in technology, the relatively limited role of endoscopically placed stents has expanded. Initially, endoscopic stenting was confined to the biliary tree, but recent innovations have expanded its role in both esophageal and colorectal disease. The initial colonic stents described were often crude, made of plastic, and associated with potential complications such as perforation and displacement of the stent [6]. Newer technology has expanded the role of stents in the colon and rectum. These self-expanding stents can now be passed through a typical colonoscopic channel. There are a considerable number of reports that describe stenting for malignant disease of the colon and rectum with

Case report
A 60-year-old woman with a history of hypertension and hypercholesterolemia was seen by her primary care physician for complaints of obstipa-

Correspondence to: W. B. Sweeney

354

Fig. 1. Abdominal radiograph demonstrating large bowel obstruction. Fig. 2. Barium enema examination revealing a circumferential anular stricture of the descending colon. Fig. 3. Abdominal radiograph demonstrating two overlapped metallic stents in good position with decompression of the colonic obstruction. Fig. 4. Endoscopic evaluation of the stented colonic stricture. Fig. 5. Left colectomy specimen opened, revealing the preoperatively placed endoluminal metallic stent.

good results [1, 35, 7]. We describe a case in which a benign colonic stricture was successfully stented. If stenting can be associated with low morbidity, then the ability to stent preoperatively allowing for decompression and bowel preparation would be preferable to proceeding with a twostage procedure. Certainly in patients who have debilitating comorbid factors, it would be beneficial to avoid two operations for benign disease.

References
1. Canon CL, Baron TH, Morgan DE, Dean PA, Koehler RE (1997) Treatment of colonic obstruction with expandable metal stents: radiologic features. Am J Roentgen 168: 199205 2. Leitman IM, Sullivan JD, Brams D, DeCosse JJ (1992) Multivariate analysis of morbidity and mortality from the initial surgical management of obstructing of the colon. Surg Gynecol Obstet 174: 513518 3. Mainar A, Tejero E, Maynar M, Ferral H, Kastaneda-Zuniga W (1996) Colorectal obstruction: treatment with metallic stents. Radiology 198: 761764 4. Raijman I, Siemans M, Marcon N (1995) Use of an expandable ultraflex stent in the treatment of malignant rectal stricture. Endoscopy 27: 273276 5. Rey JF, Romanczyk T, Greff M (1995) Metallic stents for palliation of rectal carcinoma: a preliminary report on twelve patients. Endoscopy 27: 501504 6. Rupp KD, Dohmoto M, Meffert R, Holzgreve A, Hohlbach G (1995) Cancer of the rectum-palliative endoscopic treatment. Eur J Surg Oncol 21: 644647 7. Tejero E, Fernandez-Lobato R, Mainar A, Montes C, Pinto I, Fernandez L, Jorge E, Lozano R (1997) Initial results of a new procedure for treatment of malignant obstruction of the left colon. Dis Colon Rectum 40: 432436

Conclusions We believe that endoscopic stenting is a relatively easy and effective method for decompressing the colon in both benign and malignant obstruction. Endoluminal stenting for malignant as well as benign colonic obstruction should lead to substantial decrease in the morbidity and mortality associated with this disease, as well as reducing the cost of treatment.

Surg Endosc (1998) 12: 348350

Springer-Verlag New York Inc. 1998

The effect of peritoneal air exposure on postoperative tumor growth

J. C. Southall, S. W. Lee, M. Bessler, J. D. Allendorf, R. L. Whelan
Columbia University College of Physicians and Surgeons, Columbia Presbyterian Medical Center, Department of Surgery, 161 Fort Washington Avenue, New York, NY 10032, USA Received: 14 May 1997/Accepted: 14 July 1997

Abstract Background: Previous work has demonstrated that cellmediated immune function is better preserved in rodents after laparoscopic than open surgery. The cause of this laparotomy-related immunosuppression is unclear. Some investigators have attributed it to the length of the incision; others, to peritoneal air exposure. It has also been shown that tumors in mice are more easily established and grow larger after sham laparotomy than after pneumoperitoneum. Lastly, the differences in tumor growth have been shown to be, at least in part, attributable to the immunosuppression that occurs after laparotomy. The purpose of this study was to determine if air pneumoperitoneum, presumably via immunosuppression related to peritoneal air exposure, is associated with increased tumor growth in the postoperative period. Methods: A total of 150 immunocompetent syngeneic mice received high-dose intradermal injections of mouse mammary carcinoma tumor cells. They were then randomized to undergo one of the following procedures: (a) anesthesia alone, (b) air insufflation (46 mm Hg), (c) CO2 insufflation, or (d) full laparotomy. No intraabdominal procedure was carried out. All procedures were 20 min long. After 12 days, the animals were killed and the mean tumor mass determined for each group. Results: All animals grew tumors. There was no significant difference in the mean tumor size of the anesthesia control, CO2 insufflation, and air insufflation groups (p > 0.85 by ANOVA). However, the laparotomy group tumors were 1.5 times as large as those of the other three groups (p < 0.05 by ANOVA). Conclusions: In this model, air insufflation did not significantly affect postoperative tumor growth, nor did CO2 pneumoperitoneum. However, full laparotomy was associated with increased tumor growth.

Key words: Laparoscopy CO2 pneumoperitoneum Pneumoperitoneum Tumor growth

Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 2021 March 1997 Correspondence to: R. L. Whelan

The introduction, development, and growth of laparoscopic techniques in general surgery has proceeded at an amazingly rapid rate. The purported clinical benefits of minimally invasive surgery include less postoperative pain, more rapid resumption of diet, and shorter hospitalization [58]. The physiologic basis of these benefits remains, for the most part, obscure. Few basic science studies had been done prior to the laparoscopic explosion that occurred in the early 1990s. Over the last 6 years, our laboratory and others have been attempting to elucidate the physiologic, immunologic, and oncologic impact of the laparoscopic and open methods of abdominal exposure. In both large and small animal models, it has been established that cell-mediated immune function, as measured by serial delayed-type hypersensitivity (DTH) testing, is significantly better preserved after pneumoperitoneum and laparoscopic bowel resection than after laparotomy and open resection [2, 10]. Because the diminished DTH response observed after laparotomy is not seen after minilaparotomy, it has been hypothesized that immune function differences are related to incision length [3]. Watson et al., who studied other indicators of immune function, also found that laparotomy resulted in immunosuppression [12]. However, they concluded that the exposure of the peritoneal cavity to lipopolysaccharides found in circulating air is the reason for the postoperative immunosuppression. In a separate series of murine studies, Allendorf et al. have shown that tumors are more easily established and grow more rapidly after laparotomy and open bowel resection than after pneumoperitoneum and laparoscopic bowel resection [1, 4]. Finally, in an experiment involving T-cell deficient mice, the tumor growth differences have been demonstrated to be attributable in large part to the immunosuppression that occurs after laparotomy [4]. The purpose of this study was to assess and compare tumor growth following air pneumoperitoneum, CO2 pneu-

349

moperitoneum, and laparotomy. If, in fact, it is exposure of the peritoneal cavity to air and not the incision that causes the immunosuppression seen after laparotomy, then tumor growth should be increased after air pneumoperitoneum.

Materials and methods

The protocol for this study was approved by the Columbia University Institutional Animal Care and Use Committee in accordance with FDA regulations. All animals were acclimated to a climate- and light cycle controlled environment for 24 h prior to investigations. Mice were fed standard laboratory rodent chow and tap water ad libitum. The mouse mammary carcinoma (MMC) tumor cell line was utilized for this study. This tumor line is derived from the MC2 cell line [11] obtained from Dr. J. Vaage at the Roswell Park Cancer Institute (Buffalo, NY, USA). MMC is an immunogenic cell line that shows a growth plateau 1214 days after tumor cell inoculation, after which 20% of tumors begin to spontaneously regress. MMC is syngeneic to the C3H/He mouse strain. This is an immunocompetent strain. It has been established in previous experiments that after an intradermal injection of one million tumor cells, >95% of control mice develop tumors [1]. On the day of operative intervention, tumor cells were prepared as a single-cell suspension for intradermal inoculation. MMC cells growing free-floating in RPMI 1640 medium supplemented with 10% fetal calf serum, 150 U/ml penicillin, and 150 mg/ml streptomycin were washed twice, counted, and resuspended in phosphate-buffered saline. A suspension of 107 cells per ml was prepared and mice were injected with 0.1 ml, for a total inoculum of 106 cells. Tumor cell viability throughout the experiment was determined to be >95% by trypan blue exclusion. A total of 150 56 week old female C3H/He mice (Charles River Laboratories, Wilmington, MA, USA) were used in this study. On the day of and prior to the intervention, the mice were restrained, shaved, and then given an intradermal injection of 0.1 ml of the tumor cell suspension (107 cells/ml) in the dorsal skin. Animals were then randomly assigned to one of the following four groups: (a) anesthesia control, (b) CO2 pneumoperitoneum, (c) air pneumoperitoneum, and (d) laparotomy. Immediately before the procedure, all animals were anesthetized by intraperitoneal injection of ketamine (50 mg/kg) and xylazine (5 mg/kg) in a total volume of 0.3 ml. Anesthesia control mice underwent no procedure and after 20 min were returned to their cages. In both air and CO2 pneumoperitoneum groups, insufflation was accomplished via an 18-gauge angiocath inserted into the abdomen. A pressure of 46 mm Hg was maintained for 20 min in each group. In the air pneumoperitoneum group, insufflation was carried out with ambient room air. Laparotomy group animals underwent a ziphoid to pubic symphisis midline incision that was stapled closed in one layer at 20 min. The animals were killed and the tumors excised and weighed on postoperative day 12. Representative tumor samples were taken from each group. The presence of tumor was confirmed via histologic sections evaluated by a pathologist. All data were collected in a blinded fashion. Differences among groups were analyzed for statistical significance by ANOVA.

Fig. 1. Tumor mass by postoperative day 12 after air insufflation, CO2 insufflation, and laparotomy. *p < 0.05 for CO2, air, and control groups versus laparotomy. **p value not significant for air versus control, air versus CO2, CO2 versus control.

Results Tumor nodules developed in all mice by postoperative day 12. The mean tumor size (62.2 52 mg) of the laparotomy group was 1.5 times larger than the control (40.3 39 mg), CO2 insufflation (39.5 43 mg), and air insufflation groups (37.9 34 mg) (p < 0.05 for all comparisons). There were no significant differences in tumor mass between the CO2 air, and anesthesia groups (p > 0.85 for all comparisons). The mortality of the study procedure was 2%.

Discussion Rat DTH studies from our laboratory have established that cell-mediated immune function is better preserved after lap-

aroscopic procedures than after the equivalent open procedures carried out via a full-length midline laparotomy incision [7]. Interestingly, significant immunosuppression was not found when a sham mini-laparotomy (one-half the ziphoid to pubis distance) was compared to pneumoperitoneum [3]. This finding suggested that the immunosuppression observed following a full laparotomy was related to the length of the incision. Watson et al., who looked at different indicators of immune function in a mouse model, have also shown that immunosuppression occurs after laparotomy [12]. They believe that this immunosuppression is related to factors in circulating air that cause lipopolysaccharide (LPS) translocation across the gut wall, which in turn has a systemic effect on the immune system. Therefore, they believe that peritoneal exposure to airnot the incision lengthis the cause of the documented differences in immune function that occur after laparotomy. Previous work from our laboratory, utilizing three different tumor cell lines, has also demonstrated that tumors in mice are more easily established and grow larger after laparotomy and open bowel resection than after pneumoperitoneum and laparoscopic bowel resection [1, 9]. Furthermore, in an experiment involving athymic mice, it has been established that these tumor growth differences are, at least in part, related to the immunosuppression that occurs after laparotomy [4]. Whereas tumors grew larger after laparotomy in immunocompetent mice, in athymic mice (without T cells) the open and pneumoperitoneum group tumors were of equal size. If peritoneal air exposure is the cause of the immunosuppression seen after laparotomy and if this immunosuppression results in more rapid tumor growth, then tumors that develop after air insufflation should be larger than those after CO2 pneumoperitoneum. The current study was designed to test this hypothesis. Utilizing the C3H/He mouse strain and the MMC tumor cell line, insufflation with room air was compared to pneumoperitoneum with CO2, laparotomy, and anesthesia alone. A high dose of tumor cells was injected intradermally immediately prior to the surgical intervention. As expected, tumors were found in all animals 12 days following the test procedures. There was no difference in the mean tumor size of the air, CO2, and anesthesia control groups (p < 0.85 for all comparisons) (Fig. 1). However, the

350

laparotomy group tumors were 1.5 times larger than those of the other three groups (p < 0.05). These findings are similar to those of an earlier study from our laboratory [1]. Our results suggest that while peritoneal air exposure does not significantly affect the rate of postoperative tumor growth, a full laparotomy incision is associated with larger tumors. This does not necessarily mean that peritoneal air exposure does not have immunomodulating effects. It is possible that peritoneal air exposure and a full laparotomy incision influence the immune system in different ways via separate pathways. If this is the case, then the immunosuppression related to a large incision is associated with increased postoperative tumor growth, whereas that related to air exposure is not. Our laboratory is presently studying the impact of large abdominal incisions and air exposure on lymphocyte proliferation rates in a murine model.
Acknowledgment: This investigation was made possible by generous support from the Ethicon division of Johnson and Johnson Corporation and the United States Surgical Corporation.

9. Southall JC, Lee SW, Allendorf JDF, Bessler M, Whelan RL (1997) Colon adenocarcinoma and B-16 melanoma grow larger after laparotomy versus laparoscopy in a murine model [Abstract]. Dis Colon Rectum 40: A20 10. Trokel MJ, Bessler M, Treat MR, Whelan RL, Nowygrod R (1994) Preservation of immune response after laparoscopy. Surg Endosc 8: 13851387 11. Vaage J, Pepin K (1985) Morphological observations during developing concomitant immunity against a C3H/He mammary tumor. Cancer Res 45: 659666 12. Watson RW, Redmond HP, McCarthy J, Bouchier-Hayes D (1995) Exposure of the peritoneal cavity to air regulates early inflammatory responses to surgery in a murine model. Br J Surg 82: 10601065

Discussion Dr. Talamini: In Baltimore, as Im sure in many communities across the country, there is the belief that when a patient with cancer has air touch the tumor that it suddenly explodes in growth, and I guess this, perhaps, can finally put that theory to rest. Dr. Marcus: How did you come up with the time of 12 days? Do you think that some of the effect of not having a difference between your air and CO2 might just be the amount of time for your cell-mediated response? Dr. Southall: Actually, the 12 days is a time point where tumor growth is at its maximum, and where differences will be maximized. Beyond 12 days, there is a growth plateau, at which point differences will be minimized, essentially. Dr. Nduka: Id like to congratulate you and your group for your excellent work in this field, this complicated field of oncological laparoscopy. My question is concerning your model. Do you think it is appropriate to use an intradermal model for studying what, in effect, is an intraperitoneal problem? Dr. Southall: Well, we were studying the effects of carbon dioxide gas and air gas. Insufflation is associated with two processes: (1) a direct process of insufflation gas and pressure on a tumor, intraperitoneal tumor, and (2) on systemic effect of carbon dioxide or air gas. This model focuses on the systemic effect, and for that reason the tumor is placed in a position where it is not directly exposed to the gases.

References
1. Allendorf JDF, Bessler M, Kayton ML, Oesterling SD, Trear MR, Nowygrod R, Whelan RL (1995) Increased tumor establishment and growth after laparotomy versus laparoscopy in a murine model. Arch Surg 130: 649653 2. Allendorf JDF, Bessler M, Whelan RL, Trokel M, Laird DA, Terry MB, Treat MR (1996) Better preservation of immune function after laparoscopic-assisted vs. open bowel resection in a murine model. Dis Colon Rectum 39s: S6772 3. Allendorf JDF, Bessler M, Whelan RL, Trokel M, Laird DA, Terry MB, Treat MR (1996) Postoperative immune function varies inversely with the degree of surgical trauma. Surg Endosc 11: 427430 4. Allendorf JDF, Marvin MR, Bessler M, Whelan RL (1996) Tumors grow larger after laparotomy vs laparoscopy in immunocompetent mice but not in athymic mice. Surg Forum 47: 150152 5. Barkun JS, Barkun AN, Sampalis JS, Fried G, Taylor B, Wexler MJ, Goresky CA, Meakins JL (1992) Randomised controlled trial of laparoscopic versus mini cholecystectomy. Lancet 340: 11161119 6. Gadacz TR, Talamini MA (1991) Traditional versus laparoscopic cholecystectomy. Am J Surg 161: 336338 7. Grace PA, Quereshi A, Coleman J, Keane R, McEntee G, Broe P, Osborne H, Bouchier-Hayes D (1991) Reduced postoperative hospitalization after laparoscopic cholecytectomy. Br J Surg 78: 160162 8. Reddick EJ, Olsen DO (1989) Laparoscopic laser cholecystectomy. Surg Endosc 3: 131133

Case reports
Surg Endosc (1998) 12: 351352 Springer-Verlag New York Inc. 1998

Operative technique for thoracoscopic transmyocardial laser revascularization

A. Pietrabissa,1 A. Milano,2 U. Bortolotti,2 F. Mosca1
1 2

Istituto di Chirurgia Generale e Sperimentale, Universita ` di Pisa, Ospedale di Cisanello, via Paradisa 2, 56124 Pisa, Italy Cardiochirurgia, Universita ` di Pisa, Pisa, Italy

Abstract. We describe herein the operative steps used to perform a transmyocardial laser revascularization by thoracoscopy. A special technique and specific equipment are required for the efficacy and safety of the procedure. Our preliminary results with this novel approach suggest that it could be a valid alternative to the thoracotomic procedure. Key words: Myocardial revascularization Laser surgery Thoracoscopy

Transmyocardial laser revascularization (TMR) is a new approach in the treatment of coronary artery disease for patients deemed unsuitable for either angioplasty or bypass grafting [1]. In this procedure, which is usually performed through a left anterior thoracotomy without cardiopulmonary bypass, a series of small transmural channels are created in reperfusable areas of the left ventricular wall of the beating heart by repeated laser applications. The clinical benefit of TMR seems to be a consequence of the direct perfusion of oxygenated blood from the left ventricle into the myocardial vascular network. However, the exact mechanism by which cardiac perfusion is improved and angina relief is provided is not yet fully understood [2]. This report is based on our preliminary experience with TMR via the thoracoscopic approach in two patients. It details the operative steps that we used to perform the procedure.

Technique Under general anesthesia with nonselective endotracheal ventilation, the patient is placed in the right lateral decubitus position with a 30 break of the operating table at the level of the mid-thorax. A 25 angled telescope is advanced through a 10-mm port placed in the sixth intercostal space at
Correspondence to: A. Pietrabissa

the posterior axillary line. A positive intrapleuric pressure of 6 mm Hg is maintained by CO2 insufflation. Two operative trocars are subsequently inserted: a 10-mm trocar in the seventh space at the anterior axillary line and another one in the fourth intercostal space at the mid-axillary line. The second trocar has a 5-mm flexible cannula to allow the introduction of coaxially curved instruments (Storz, Tuttlingen, Germany) that will facilitate the safe opening of the pericardium. A long pericardial incision is created anterior to the left phrenic nerve with the use of graspers and coaxially curved scissors. In cases where the patient has undergone previous cardiac surgery (usually through a mid-sternotomy), the adhesions between the left ventricle and the internal aspect of the pericardial sac are carefully divided. The superior edge of the pericardial window is then suspended by two interrupted stitches to the anterior thoracic wall. Sutures mounted on 5-cm straight needles are used for this purpose. The straight needle is hand-driven percutaneously through the anterior thoracic wall until a sufficient length emerges from the pleural surface. The needle is then grasped internally by a pair of needle holders. It is passed through the superior edge of the incised pericardium and then reversed and guided inside-out through the thoracic wall in the vicinity of the entrance point. The two ends of the suture are grasped with a Kelly clamp and placed in gentle traction to maximize the surgeons access to the left ventricle. An average of 30 channels, 1 mm in diameter each, are then created in the left ventricle by the tip of a holmium laser probe (Eclipse Surgical Technologies, Inc., Sunnyvale, CA, USA). This probe is advanced inside an Olsens cholangiograsper (Storz) to reach the anterior and lateral epicardial surface of the left ventricle; it is then activated (Fig. 1). The base and the posterior aspect of the left ventricle are approached with the aid of a specially designed right-angle applicator, through which the flexible optical fiber of the holmium laser is easily advanced (Fig. 2). Bleeding from the epicardial holes created by laser application usually stops spontaneously within a few minutes; occasionally, transient local compression is required. Compression is exerted by a peanut pledget holder inserted and retracted in-

352

side a reducer tube. Finally, the pericardial incision is partially closed by approximating the cut edges with a few interrupted stitches to allow intrapleural drainage of pericardial fluid. A pleural aspiration drain is left and removed 48 h later after pulmonary reexpansion. Comment A wide range of procedures can now be performed via the thoracoscopic approach, with a subsequent reduction in the requirement for postoperative analgesia and a shorter recovery period [3]. TMR is a simple operation currently indicated for patients in whom conventional invasive methods of revascularization are no longer an option. Our early experience with thoracoscopic TMR in two patients indicates that this new approach is feasible and safe. This technique provides a good alternative to the conventional thoracoscopic procedure, where the trauma of access constitutes the largest component of total operative insult [4]. The fact that these patients tend to be elderly and chronically ill reinforces the potential benefits of the minimally invasive approach to TMR in terms of recovery, short-term disability, and reduced costs. References
1. Horvath KA, Mannting F, Cummings N, Shernan SK, Cohn LH (1996) Transmyocardial laser revascularization: operative techniques and clinical results at two years. J Thorac Cardiovasc Surg 111: 10471053 2. Kohmoto T, Fisher PE, Gu A, Zhu SM, Yano OJ, Spotnitz HM, Smith CR, Burkhoff D (1996) Does blood flow through holmium: YAG transmyocardial laser channels? Ann Thorac Surg 61: 861868 3. Mulder DS (1993) Pain management principles and anesthesia techniques for thoracoscopy. Ann Thorac Surg 56: 630632 4. Cushieri A (1995) Whither minimal access surgery: tribulations and expectations. Am J Surg 169: 919

Fig. 1. Creation of transmyocardial channels on the lateral aspect of the left ventricle by the holmium laser probe introduced inside the Olsens cholangiograsper. Fig. 2. The base of the left ventricle is reached with a specially designed right-angle applicator for the flexible fiber of the holmium laser.

Surg Endosc (1998) 12: 305309

Springer-Verlag New York Inc. 1998

Spilled gallstones after laparoscopic cholecystectomy

A relevant problem? A retrospective analysis of 10,174 laparoscopic cholecystectomies
M. Scha fer,1 C. Suter,1 Ch. Klaiber,2 H. Wehrli,2 E. Frei,2 L. Kra henbu hl1
1 2

Department of Visceral and Transplantation Surgery, Inselspital University of Berne, CH-3010 Berne, Switzerland Swiss Association for Laparoscopic and Thoracoscopic Surgery (SALTS)

Received: 4 April 1997/Accepted: 9 July 1997

Abstract Background: Spilled gallstones after laparoscopic cholecystectomy may cause abscess formation, but the exact extent of this problem remains unclear. Method: The data (collected by the Swiss Association of Laparoscopic and Thoracoscopic Surgery) on 10,174 patients undergoing laparoscopic cholecystectomy at 82 surgical institutions in Switzerland between January 1992 and April 1995 were retrospectively analyzed with special interest in spilled gallstones and their complications. Results: In 581 cases (5.7%) spillage of gallstones occurred; 34 of these cases were primarily converted to an open procedure for stone retrieval. Of the remaining 547 cases only eight patients (0.08%) developed postoperatively abscess formation requiring reoperation. Conclusions: Spillage of gallstones after laparoscopic cholecystectomy is fairly common and occurs in about 6% of patients. However, abscess formation with subsequent surgical therapy remains a minor problem. Removal of spilled gallstones is therefore not recommended for all patients, but an attempt at removal should be performed whenever possible. Key words: Laparoscopy Cholelithiasis Spilled gallstones Complications after laparoscopic cholecystectomy Abdominal abscess

Laparoscopic cholecystectomy has rapidly become the standard treatment for symptomatic cholelithiasis and acute cholecystitis in western countries [7, 12]. Today, laparoscopic cholecystectomy has shown to be a safe and effective procedure in experienced hands and carries therefore a low morbidity and mortality rate [2, 3, 9]. However, introduction of laparoscopic cholecystectomy was associated with some new complications, which have been rare with tradi-

tional open cholecystectomy. An increased incidence of major bile duct injuries, especially during the learning curve, has been reported by many authors [8, 10]. Iatrogenic bile duct injuries probably represent the most serious complication of laparoscopic cholecystectomy and have therefore attracted the most attention. Perforation of the gallbladder with intraabdominal spillage of gallstones is a common problem after laparoscopic cholecystectomy although it is not considered to be as serious. The reported rate of gallbladder perforation after laparoscopic cholecystectomy varies from 10 to 32% [1, 4, 11, 14]. Mostly, inadvertent opening of the gallbladder occurs during the dissection from the liver bed. Furthermore, rupture of the gallbladder can either occur by tearing the gallbladder with grasping forceps or during the extraction through the abdominal wall. With open cholecystectomy intraperitoneal spillage of gallstones is easily recognized, and the lost stones can be retrieved without problems. Thus, retained gallstones were uncommon, and only one patient has been reported with intraabdominal abscess formation several years after open cholecystectomy [5]. The exact morbidity and complication rate of spilled gallstones after laparoscopic cholecystectomy are not well investigated. It is known that spilled gallstones may cause abscess formation [6]. But the incidence of complications due to spilled gallstones seems to be low or even very low [11]. Thus, it remains unclear if stone spillage should be considered an indication for conversion to an open procedure for stone retrieval. The aims of this retrospective study were to investigate the frequency, the complications, and therapy of spilled gallstones after laparoscopic cholecystectomy. Patients and methods
The Swiss Association for Laparoscopic and Thoracoscopic Surgery (SALTS) prospectively collects the data of patients undergoing cholecystectomy at 82 surgical institutions (universities, county and district hospitals, and surgeons in private practice) in Switzerland. More than 350 single items of data including personal records, ASA classification, indication for surgery, preoperative investigations, intraoperative findings, operative

Correspondence to: L. Kra henbu hl

306 Table 1. Characteristics of patients, indication for LC and operating time Overall CE group 10,174 3,103:7,071 (31:69) 52.6 (296) 7,579 (74.5) 1,049 (10.3) 1,546 (15.2) 26.8 55.5 17.7 LC with spilled gallstones, no conversion 547 249:298 (46:54) 55.4 (1992) 369 (67.5) 82 (15.0) 96 (17.5) 9.2 54.8 36.0 LC with spilled gallstones and conversion 34 15:19 (44:56) 59.1 (3082) 24 (70.6) 8 (23.5) 2 (5.9) 0 47.1 52.9 LC with spilled gallstones and complications 8 5:3 (62:38) 62.1 (3979) 5 (62.5) 3 (37.5) 0 (0.0) 0 62.5 37.5

Characteristic No. of patients M:F (%) Mean age, range (years) Indication (%) CCL AC Other Operating time (%) <60 min 60120 min >120 min

AC, acute cholecystitis; CE, cholecystectomy; CCL, cholecystolithiasis; LC, laparoscopic cholecystectomy; Other: common bile-duct stones, gallbladder polyps, biliary pancreatitis.

complications, conversion rate, postoperative morbidity, reoperation rate, and mortality were recorded for every patient on a specially designed computerized worksheet [13]. The data of 10,174 laparoscopic cholecystectomies performed between January 1992 and April 1995 were retrospectively analyzed with special regard for intraabdominally spilled gallstones. All patients with lost gallstones were identified. Their medical records and operative reports were collected from the referring surgical institutions or the general practitioner; they were then asked to answer an additional questionnaire concerning further details and follow-up. If necessary, the institutions were visited or telephoned by one of the investigators to collect data which were missing. The follow-up of all patients was guaranteed by contacting either the general practitioner or the patient. Data collected from chart review, telephone call, and questionnaire were then carefully reviewed, and the patients who had had complications following spilled gallstones were identified. These few cases were then further investigated using a previously created protocol with selected criteria from the literature. Results are expressed as mean, standard deviation, and range values, respectively.

followed by acute cholecystitis (10.3%) and some less frequent indications (15.2%, i.e., choledocholithiasis, biliary pancreatitis, polyps of the gallbladder). Patients undergoing laparoscopic cholecystectomy for acute cholecystitis had a higher incidence of intraabdominal gallstone spillage. Therefore, in the group with conversion to an open procedure for stone retrieval, as well as in the group with serious postoperative complications, 23.5% and 37.5% of the patients underwent laparoscopic cholecystectomy for acute cholecystitis, respectively (Table 1).

Operating time The French procedure, with the surgeons position between the patients legs, is the most common surgical technique in Switzerland. As shown in Table 1, more than 80% of the laparoscopic cholecystectomies in the overall laparoscopic group were performed within 120 min. However, spillage of gallstones markedly increased the operating time, and at least one-third of these operative procedures lasted more than 120 min. Of course, conversion to the open procedure caused an even longer operating time. Nevertheless, patients with serious postoperative complications due to spilled gallstones did not have longer operating times than patients with simple spillage of gallstones.

Results Characteristics of patients There were 3,103 male (30.5%) and 7,071 female patients (69.5%). The overall mean age at the time of operation was 52.6 years (range 296 years). Among the 10,174 patients, there were 581 patients (5.7%) with intraoperative gallstone spillage into the peritoneal cavity. Their mean age was 55.6 years (range 1992 years); 547 of these 581 operations were finished laparoscopically, whereas in 34 cases the operation was converted to an open procedure during which all the spilled gallstones were removed. Only eight patients (0.08%, five men and three women) with a mean age of 62.1 years (range 3979 years), could be identified as having had a serious postoperative complication due to intraabdominally lost gallstones. The baseline data are shown in Table 1. Indications of laparoscopic cholecystectomy Symptomatic cholecystolithiasis was the main indication for patients undergoing laparoscopic cholecystectomy (62.5%),

Intraabdominally spilled gallstones: postoperative complications, postoperative reoperation, and intervention rate As previously mentioned, there were 581 patients (5.7%) with intraabdominally spilled gallstones. Nearly all the procedures were finished laparoscopically; only 34 patients (5.8%) had a conversion to an open procedure for stone retrieval. Comparing these two patient groups, we found a considerably lower rate (5.9 vs 10.8%) of local postoperative complications in the group with conversion. In particular, no complications due to common bile-duct stones, bile leakage, bleeding, or spilled gallstones occurred. However, the systemic complication rate, i.e., cardiac failure, pulmo-

307 Table 2. Intraabdominally spilled gallstones: complications, reoperation, and intervention rate LC with spilled gallstones no conversion (n 547) Postoperative complications (%) Local: CBD stones Bile leakage Intraabd. abscess formation Intraabd. bleeding Ileus Trocar herniation Trocar hematoma Wound infection Systemic: Pulmonary embolism Cardiopulmonary insuffic. Sepsis Urinary tract infection Other Reoperation rate (%) 59 (10.8) 8 18 8 7 1 1 10 6 42 (7.7) 1 15 2 5 19 5.5 LC with spilled gallstones and conversion (n 34) 2 (5.9) 0 0 0 0 0 0 0 2 4 (11.8) 2 1 0 0 1 0

CBD, common bile duct; LC, laparoscopic cholecystectomy.

nary complications, and thromboembolism, was higher (11.8 vs 7.7%) in the group with conversion. The reoperation rate was 0% and 5.5% for the groups with conversion and without conversion, respectively. Bile leakage, intraabdominal abscess formation, and bleeding were the main reasons for reoperation. None of the 34 patients whose operations were converted to an open procedure for stone retrieval developed any further problems from spilled gallstones (Table 2). Serious postoperative complications due to spilled gallstones requiring reoperation Among the 10,174 patients undergoing laparoscopic cholecystectomy we identified eight patients (0.08% for the overall group or 1.4% of those with spilled gallstones, respectively) with serious postoperative complications following intraabdominally spilled gallstones. There were five men and three women. Mean age at the time of operation was 62.1 years (range 3979 years). Five patients underwent laparoscopic cholecystectomy for cholecystolithiasis, while the remaining three were operated on for acute cholecystitis. Six operations were elective procedures, whereas two were performed as emergencies. All the patients, bar one, developed intraabdominal abscess formation requiring reoperation. Three of these seven patients not only developed an intraabdominal abscess but also fistulas and abscess formation into the abdominal wall. In one patient who complained postoperatively of upper abdominal pain, gallstones had become sandwiched between the liver and the diaphragm intraoperatively. Reoperation was always performed by open access. Four patients were reoperated in the early postoperative course (221 days); the remaining four patients were reoperated on after 2.3, 4.5, 5.0, and 18.4 months. Gallstones were found in seven patients. In four cases, bacterial growth was detected; in the remaining four cases, bacterial culture was either negative or not performed. Unfortunately, chemical analysis of the retrieved gallstones was only per-

formed in one case. Also, the number and the size of the gallstones were only poorly documented (Table 3).

Follow-up Follow-up time was 1656 months. None of the eight patients with abscess formation who had required reoperation had further complications. They all recovered fully and their further postoperative course was uneventful. The same uneventful postoperative course was found in the patient group with conversion to an open procedure.

Discussion The purpose of our current study was to investigate the clinical relevance of spilled gallstones after laparoscopic cholecystectomy. To this end, the data of 10,174 laparoscopic cholecystectomies performed at 82 surgical institutions in Switzerland were retrospectively analyzed. Intraoperative spillage of gallstones occurred in 5.7% of our cases. Since Fitzgibbons has suggested that spillage of gallstones occurs in about two-thirds of gallbladder perforation [4], the effective perforation and rupture rate of the gallbladder is probably even higher, although it did not occur in our study. The perforation rate of the gallbladder according to the literature varies considerably, from 10 to 32% [1, 4, 10, 11]. The exact number of lost gallstones was not recorded by SALTS and thus could not be evaluated. However, in the small group of patients with postoperative complications, more than one lost gallstone was found in the peritoneal cavity. Symptomatic cholelithiasis was the most frequent indication for laparoscopic cholecystectomy, followed by acute cholecystitis. Patients with acute cholecystitis had an increased rate of spilled gallstones as well as a higher complication rate. On the other hand, patients who underwent

308 Table 3. Patients with serious postoperative complication due to intraabdominally spilled gallstones Gallstones Patient 1 2 3 4 5 6 7 8 Sex M M M M M F F F Age (years) 39 53 62 75 79 41 69 79 ASA 1 1 2 2 2 1 2 3 Ind. CCL AC AC AC CCL CCL CCL CCL Reop (days) 150 2 21 3 552 10 70 132 Complication Fistula of the umbilicus, intraabdominal abscess Irritation of the diaphragm Intraabdominal abscess Intraabdominal abscess Intraabdominal abscess Abscess formation intraabdominally and abdominal wall Fistula of the abdominal wall Intraabdominal abscess Bacterial culture Not performed Negative Enterococcus faecalis Negative Streptococcus milleri Pseudomonas aeruginosa Not performed Escherichia coli + + + + + + <5 mm 510 mm >10 mm + + size unknown no GS found

AC, acute cholecystitis; CCL, cholecystolithiasis; GS, gallstone; Ind., indication; Reop., reoperation (days after the first operation).

laparoscopic cholecystectomy for cholelithiasis had a lower conversion rate and also a lower complication rate due to spilled gallstones. Spillage of gallstones extended the operating time, but there was no difference between the patient group with simple spillage of gallstones and those with postoperative complications. Although we found 581 patients (5.7%) with intraabdominally lost gallstones, spillage of gallstones only led to conversion to an open procedure for stone retrieval in 34 cases. Therefore, it can be suggested that either the majority of the spilled stones were completely retrieved or they were so utterly lost in the peritoneal cavity that the surgeons decided not to search for them. The exact reasons why the spilled gallstones led to conversion are unknown. Different complication rates were found for local and systemic complications. Local complications were markedly more frequent in the group with simple spillage of gallstones compared to the group with conversion for stone retrieval. In particular, no complications due to common bile-duct stones, bile leakage bleeding, or spilled gallstones occurred. Additionally, there were no complications related to laparoscopic access such as trocar hematoma and herniation. However, systemic complications were more frequent after conversion to an open procedure. Since these patients were older and their operating time was longer, they possibly had an increased risk of systemic complications. Although intraabdominal spillage of gallstones was not an uncommon problem, complications following such spillage were very rare, and only eight patients were identified among these 10,174 cases. All these patients (five men and three women) had a simple spillage of gallstones, and the laparoscopic cholecystectomies were uneventful. The indication for laparoscopic cholecystectomy was symptomatic cholelithiasis in five patients and cholecystitis in three patients. Thus, this small patient group showed a threefold increase in the rate of acute cholecystitis compared to the overall cholecystectomy group. Leakage of infected bile and gallstones are probably responsible for this increased complication rate. Furthermore, the inflamed wall of the gallbladder is vulnerable and the local inflammatory reaction makes the dissection more difficult. Intraabdominal abscess formation, which is also the most frequent complica-

tion in the literature, occurred in seven patients [6]. One patient was reoperated on 2 days after the laparoscopic cholecystectomy and so in all likelihood had not yet developed abscess formation in this short postoperative period. But all of these eight patients were reoperated on with open access. Four patients were reoperated on early after laparoscopic cholecystectomy (within the first 3 postoperative weeks), among whom were all three cases with acute cholecystitis. The remaining four patients developed abscess formation and cutaneous fistula in the late postoperative course. Gallstones of different size and number were found in seven cases. Chemical analysis was only performed in one case, which revealed bilirubin stones. In four cases, bacterial growth was detected with four different types of bacteria. In two cases, the bacterial cultures were sterile, and in the remaining two cases no bacterial culture was performed. During the follow-up time, all patients who underwent reoperation due to septic complications or with conversion to an open procedure for stone retrieval developed further complications, but no mortality occurred. Therefore, all these additional operative procedures were successful. The final question is whether spillage of gallstones is an indication for conversion to an open procedure for stone retrieval. Although spillage of gallstones may lead to severe postoperative complications, the incidence and the mortality are low, even very low. Thus, obligatory conversion to an open procedure for stone retrieval is inappropriate. However, an attempt should always be made to remove spilled gallstones and to irrigate the abdominal cavity. In conclusion, spillage of gallstones during laparoscopic cholecystectomy is a common problem (5.7%), but serious postoperative complications are fortunately very rare (0.08%). Elderly patients with acute cholecystitis with infected bile and spilled stones may have an increased risk of intraabdominal abscess formation. Therefore, perforation and rupture of the gallbladder should be prevented whenever possible. In cases of spilled gallstones, the surgeon must try to retrieve these lost gallstones and the abdominal cavity should be irrigated to dilute the infected bile and spilled gallstones. But there is no need for obligatory conversion to an open procedure for stone retrieval because the incidence and the mortality rate of serious complications are so low.

309 Acknowledgments. Thanks are due to all the following surgeons and surgical institutions who contributed to this national study: Public Hospital Kantonsspital Aarau Spital Aarberg Bezirksspital Affoltern, Zu rich Kantonsspital Baden Kantonsspital Basel Claraspital Basel Klinik f. Viszerale u. Transplantationschirurgie, Inselspital Bern Tiefenauspital Bern Zieglerspital Bern Regionalspital Biel Bezirksspital Thierstein, Breitenbach Oberwalliser Kreisspital, Brig-Glis Kantonsspital Bruderholz, Basel-Land Kreisspital Bu lach Ho pital La Chaux-de-Fonds Kantonsspital Chur Spital Davos, Davos-Platz Thurgauisches Kantonsspital, Frauenfeld Kantonsspital Glarus Spital Grenchen Bezirksspital Grossho chstetten Kantonales Spital, Heiden Kantonales Spital, Herisau Regionalspital Horgen Bezirksspital Fraubrunnen, Jegenstorf Regionalspital Surselva, Ilanz Regionalspital Langenthal CHUV, Lausanne Ospedale Civico, Lugano Kreisspital Ma nnedorf Ospedale Mendrisio Spital Menziken Bezirksspital Meyriez, Murten Kantonsspital Mu nsterlingen Surgeons in pivate practice PD Dr. J. Baltensweiler, Zu rich Dr. D. Baumgartner, Luzem Dr. U. Brand, Genf Dr. J. Burri, Freiburg Dr. J. Dbaly, Bern Dr. C. Ghielmetti, Thun Dr. M. Herrmann, Onex

Kreisspital Muri, Aargau Ho pital de la Providence, Neuenburg Bezirksspital Niederbipp Bezirksspital Oberdiessbach Kantonsspital Olten Ho pital de Zone, Payerne Kreisspital Pfa ffikon Paracelsus-Spital Richterswil Spital Richterswil Gemeindespital Riehen Kreisspital Ru ti, Ru ti, Zu rich Kantonsspital Obwalden, Sarnen Kantonsspital Schaffhausen Spital Limmattal, Schlieren Zu rich Bu rgerspital Solothurn Kantonsspital St. Gallen Bezirksspital Sumiswald Regionalspital Thun Spital Thusis Spital Uster Kantonales Spital, Uznach Ospedale Italiano di Lugano, Viganello Kantonales Spital, Walenstadt Spital Wattwil Kreisspital Wetzikon Spital Wil, St. Gallen Kantonsspital Winterthur Bezirksspital Zofingen Spital Neumu nster, Zollikerberg Stadtspital Triemli, Zu rich Unispital Zu rich Stadtspital Waid, Zu rich Kantonsspital Zug Bezirksspital Zweisimmen

Dr. Dr. Dr. Dr. Dr. Dr.

Ch. U. Krayenbu hl, Zu rich K. H. Leemann, Zug A. Osterwalder, Lugano P. Petignat, Biel Th. P. Ricklin, Zu rich R. A. Schori, Bern

References
1. Catarci M, Zaraca F, Scaccia M, Carboni M (1993) Lost intraperitoneal stones after laparoscopic cholecystectomy: harmless sequela or reason for reoperation? Surg Laparosc Endosc 3: 318322 2. Cuschieri A, Dubois F, Mouiel J, Mouret P, Becker H, Buess G, Trede M, Troidl H (1991) The European experience with laparoscopic cholecystectomy. Am J Surg 161: 385387 3. Deziel DJ, Millikan KW, Economou SG, Doolas A (1993) Complications of laparoscopic cholecystectomy: a national survey of 4,292 hospitals and an analysis of 77,604 cases. Am J Surg 165: 914 4. Fitzgibbons RJ, Annibali R, Litke BS (1993) Gallbladder perforation and gallstone removal: open versus closed laparoscopy and pneumoperitoneum. Am J Surg 165: 497504 5. Jacob H, Rubin KP, Cohen MC, Kahn IJ, Kan P (1979) Gallstones in a retroperitoneal abscess: a late complication of perforation of the gallbladder. Dig Dis Sci 24: 964966 6. La uffer JM, Kra henbu hl L, Baer HU, Mettler M, Bu chler MW (1997) Clinical manifestations of lost gallstones after laparoscopic cholecystectomy. A case report with review of the literature. Surg Laparosc Endosc 7: 103112 7. NIH consensus conference statement on gallstones and laparoscopic cholecystectomy. (1993) Am J Surg 165: 390398

8. Russell JC, Walsh SJ, Mattie AS, Lynch JT (1996) Bile duct injuries, 19891993. A statewide experience. Arch Surg 131: 382388 9. Schlumpf R, Klotz HP, Wehrli H, Herzog U (1993) Laparoskopische Cholezystektomie in der Schweiz: Kritischer Ru ckblick auf die ersten 3,722 Fa lle. Chirurg 64: 307313 10. Shea JA, Healey MJ, Berlin JA, Clarke JR, Malet PF, Staroscik RN, Schwartz JS (1996) Mortality and complications associated with laparoscopic cholecystectomy. Ann Surg 224: 609620 11. Soper NJ, Dunnegan DL (1991) Does intraoperative gallbladder perforation influence the early outcome of laparoscopic cholecystectomy? Laparosc Endosc 1: 156161 12. Soper NJ, Stockmann PT, Dunnegan DL, Ashley SW (1992) Laparoscopic cholecystectomy: the new gold standard. Arch Surg 127: 917 923 13. Wehrli H, Klaiber Ch, Frei E, Metzger A, Bu hler M (1995) The Swiss experience with laparoscopic cholecystectomy. In: Bu chler MW, Frei E, Klaiber Ch, Kra henbu hl L (eds) Five years of laparoscopic cholecystectomy: a reappraisal. Progress in Surgery, vol 22. Karger, Basel, pp 4655 14. Wetscher G, Schwab G, Fend F, Glaser K, Ladurner D, Bodner E (1994) Subcutaneous abscess due to gallstones lost during laparoscopic cholecystectomy. Endoscopy 26:324325

History
Surg Endosc (1998) 12: 359360 Springer-Verlag New York Inc. 1998

J. B. Murphy, M.D.
Of buttons and blows
L. Morgenstern
Division of General Surgery, Department of Surgery, Cedars-Sinai Research Institute, Cedars-Sinai Medical Center, and Department of Surgery, UCLA School of Medicine, 444 South San Vicente Boulevard, Los Angeles, CA 90048-1869, USA Received: 19 September 1995/Accepted: 27 September 1995

John Benjamin Murphy (18571916) (Fig. 1), fabled surgeon extraordinary and stormy petrel of American surgery [1] at the turn of the century, deserves a note of recognition in the annals of endoscopy and biliary surgery. In 1912, in the Surgical Clinics of North America [2], Murphy described the insertion of a cystoscope into a cholecystostomy tract for removal of a residual stone. This was accomplished successfully by passing a hook through the cystoscope. The maneuver was an early forerunner of a similar endoscopic maneuver in use today for residual biliary calculi. Cholecystostomy, however, was J.B. Murphys second choice for the operative treatment of cholecystic disease.

His preference was for cholecystoenterostomy, over either cholecystostomy or cholecystectomy. To facilitate the joining of the gallbladder to the intestine, Murphy devised a hollow-cored spherical device divisible into two hemispheres, over which the purse-stringed segments to be anastomosed could be affixed. The two halves of the brass button, when then joined and locked together, achieved a rapid apposition of the two segments. The metal sphere was then passed within weeks after insertion, leaving the anastomosis intact. He called this device the Anastomosis Button. It is better known as the Murphy Button (Fig. 2). Six days after using his newly devised Anastomosis

Fig. 1. John Benjamin Murphy (18571916). Fig. 2. The Murphy button as originally described by J. B. Murphy.

360

Fig. 3. The original Murphy maneuver for eliciting Murphys Sign. The middle finger of the left hand is used as the anvil. Fig. 4. The original Murphy maneuver for eliciting Murphys Sign. The hammer-stroke is delivered sharply by the edge of the right hand over the gallbladder area.

Button [3] for the first time on a dog, he performed a cholecystoenterostomy on a 35-year-old, severely jaundiced woman, joining the gallbladder with the jejeunum by means of the button. The gallstones were not removed. Time from opening of the peritoneum until the closing of same, eleven minutes. In Murphys hands, use of the button reduced operating time for this anastomosis tenfold. After visiting Murphys clinic, the Mayo brothers, William and Charles, adopted the Murphy Button for use in the Mayo Clinic, where the button was in common usage until 1935. But neither the Murphy Button nor cholecystoenterostomy for gallstone disease were durable contributions of this remarkable surgeon. He is probably best remembered eponymically for his Murphys sign in acute cholecystitis (Figs. 3, 4). As originally described by Murphy, Murphys sign entailed percussion of the right midsubcostal region with the bent middle finger of the left hand, using the right hand to strike the dorsum of the left hand with hammer-like blows [4]. As commonly elicited today by legions of medical students, interns, residents, and their elders, it is a fist per-

cussion, which Murphy described for demonstrating CVA tenderness in patients with ureteral obstruction rather than for acute cholecystitis. Murphys niche in the history of American surgery is secure. A small part of that niche was carved by his exploits in the biliary tract, by endoscope, anastomotic button, and the ubiquitously used but now transmuted Murphys sign, the diagnostic blow signaling acute cholecystitis.

References
1. Davis L (1938) Surgeon extraordinary; the life of J. B. Murphy. Foreword by A. J. Cronin. George G. Harrap London & Co. 2. Murphy JB (1912) Surgical clinics of John B. Murphy, M.D. at Mercy Hospital, Chicago. Cholelithiasis 1: 417 3. Murphy JB (1892) Cholecysto-intestinal, gastrointestinal, enterintestinal anastomosis, and approximation without sutures. Med Rec 42: 665 4. Schmitz RL, Oh TT (eds) (1993) The remarkable surgical practice of John Benjamin Murphy. University of Illinois Press, Urbana-Champaign, p 29

Surg Endosc (1998) 12: 301304

Springer-Verlag New York Inc. 1998

Laparoscopic endobiliary stenting as an adjunct to common bile duct exploration

K. S. Gersin, R. D. Fanelli
Department of Surgery, Berkshire Medical Center, 725 North Street, Pittsfield, MA 01201, USA Received: 28 March 1992/Accepted: 3 August 1997

Abstract Background: The management of common bile duct stones (CBDS) in the era of operative laparoscopy is evolving. Several minimally invasive techniques to remove CBDS have been described, including preoperative endoscopic retrograde cholangiopancreatography (ERCP), postoperative ERCP, lithotripsy, laparoscopic transcystic common bile duct exploration, and laparoscopic choledochotomy with common bile duct exploration (CBDE). Because of the risks and limitations of these procedures, we utilize laparoscopically placed endobiliary stents as an adjunct to CBDE. Methods: Sixteen patients underwent laparoscopic common bile duct exploration (LCBDE) by either choledochotomy or the transcystic technique with placement of endobiliary stents. These patients were identified during laparoscopic cholecystectomy as having occult choledocholithiasis, using routine dynamic intraoperative cholangiography. Results: CBDS were successfully removed in all patients as demonstrated by completion cholangiography and intraoperative choledochoscopy. Eighty percent of patients were discharged the following day; the first three patients in this series were observed for 48 h prior to discharge. No patient required T-tube placement and closed suction drains were removed the morning after surgery. Stents were removed endoscopically at 1 month. Six- to 30-month follow-up demonstrates no complications to date. Conclusions: Laparoscopic endobiliary stenting reduces operative morbidity, eliminates the complications of T-tubes, and allows patients to return to unrestricted activity quickly. We recommend laparoscopically placed endobiliary stents in patients undergoing LCBDE.

Key words: Laparoscopic endobiliary stenting ERCP Laparoscopic common bile duct exploration Laparoscopic cholecystectomy Common bile duct stones

Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 2021 March 1997 Correspondence to: R. D. Fanelli, Surgical Specialists of Western New England, P.C., 510 North Street, Pittsfield, MA 01201, USA

The management of common bile duct stones (CBDS) in the era of laparoscopic cholecystectomy is evolving. Several techniques for treatment of CBDS have become popular, including preoperative endoscopic retrograde cholangiopancreatography (ERCP), postoperative ERCP, endobiliary lithotripsy, laparoscopic common bile duct exploration (LCBDE), and conversion to laparotomy with bile duct exploration [1, 2, 7, 14, 1921, 23]. Several novel techniques such as intraoperative ERCP and antegrade endoscopically guided sphincterotomy have been advocated by some but remain largely unavailable to the majority of surgeons [8, 9]. LCBDE has been performed by the transcystic route as well as by choledochotomy with stone extraction. Although each of the above techniques addresses the problem of CBDS, none is ideal. Preoperative ERCP is appropriate for some patients, but those with occult CBDS will not benefit from selective application of this technique. The routine use of postoperative ERCP may require that some patients undergo a second operative procedure for stone extraction if endoscopic clearance is unsuccessful. Contact lithotripsy requires expensive specialty equipment which is unavailable in many hospitals. Thus, the majority of laparoscopic surgeons faced with the problem of CBDS must rely on LCBDE or conversion to laparotomy for open CBDE with placement of drains and T-tubes. Procedures which rely on the placement of T-tubes require that patients remain hospitalized for several days postoperatively. Even after discharge, indwelling T-tubes are uncomfortable, require continuous management, and significantly limit patient activity because of the risk of dislodgment. Conversion to open surgery increases hospital

302

Fig. 1. Choledocholithiasis discovered during routine intraoperative dynamic cholangiography. Note the stone in the proximal ductal system. Fig. 2. After creating a choledochotomy a soft wire basket is used to extract the CBDS. Baskets are also used via the transcystic route when feasible. Fig. 3. An alternative to basket retrieval is forced irrigation of the distal CBD after placement of a biliary balloon catheter. Small stones and debris may be flushed from the duct. Fig. 4. Once clearance of all CBDS is confirmed, a 450 cm by 0.035 Zebra guidewire is introduced through the cholangiogram catheter and placed across the ampulla where it gently curls within the duodenum. Fig. 5. A 5-cm-long 7-French Microvasive Solopass Percuflex endobiliary stent is advanced over the guidewire until its distal tip lies within the duodenum. Once the stent is positioned and deployed, the delivery mechanism and guidewire are removed. Fig. 6. After successful placement of the stent, the choledochotomy is closed with sutures and the cystic duct is divided. Laparoscopic cholecystectomy is then completed, and a closed suction drain is placed.

lengths of stay, prolongs convalescence, and is associated with greater morbidity than laparoscopic surgery [6, 12, 16, 22]. We describe a method of LCBDE with placement of an endobiliary stent which obviates the need for T-tube placement, eliminates the morbidity of open CBDE, and allows patients to return to activity as quickly as those undergoing laparoscopic cholecystectomy without CBDE [17]. The laparoscopic placement of endobiliary stents also guarantees access to the common bile duct, ensuring success in cases

where postoperative ERCP may still be necessary, such as multiple stones or stones high in the proximal biliary tree. Materials and methods
Sixteen patients undergoing laparoscopic cholecystectomy for symptomatic cholelithiasis during a 3-year period were found on routine intraoperative dynamic cholangiography to have CBDS (Fig. 1). Each patient underwent laparoscopic transcystic CBDE or laparoscopic choledochotomy with CBDE. Under dynamic, fluoroscopic imaging, soft wire

303 baskets and balloon irrigation catheters were used to clear the CBDS (Figs. 2 and 3). Completion cholangiography was routinely performed to ensure the adequacy of stone clearance after LCBDE. Intraoperative choledochoscopy was performed in all cases where choledochotomy was utilized and in cases of transcystic LCBDE where cystic duct size was sufficient for introduction of the choledochoscope (Karl Storz 40 flexible fiberoptic ureteroscope, model #11274AA). Once successful CBDS clearance was confirmed, a 0.035-inch-wide by 450-cm-long Zebra guidewire (Microvasive catalog #5168) was placed across the ampulla into the duodenum either through the cystic duct or the choledochotomy (Fig. 4). Under digital fluoroscopic guidance, a 7-French, 5-cm SoloPass Percuflex endobiliary stent (Microvasive catalog #3405) was advanced over the guidewire until the distal tip entered the duodenum (Fig. 5). The injection of contrast through the stent delivery mechanism allows excellent visualization of the stents location, ensuring appropriate positioning across the ampulla. After successful deployment of the stent, cystic duct stump ligation was accomplished (Fig. 6) using either endoclips (Ethicon endoscopic rotating multiple clip applier ER320) or 2-0 Vicryl ligatures (Ethicon catalog #J286G). When choledochotomy had been utilized, closure was accomplished with interrupted 4-0 Vicryl sutures (Ethicon catalog #J214). A closed suction drain was placed after completion of the laparoscopic cholecystectomy and patients were admitted for overnight observation.

Results All 16 patients had successful laparoscopic clearance of CBDS as demonstrated by completion cholangiography. Choledochoscopy was an effective adjunct when fluoroscopically guided soft wire basket stone extraction was unsuccessful or when multiple stones were present. The first three patients in this series were observed for 48 h prior to discharge; all subsequent patients have been discharged the morning after surgery. None of these patients required Ttube placement and all patients with occult CBDS during the study period were managed by this method. Closed suction drains had been placed routinely and were removed prior to discharge after inspection revealed no evidence of bile leak. Stents were removed endoscopically at 1 month, and retrograde cholangiograms obtained during ERCP demonstrated no retained stones, bile duct strictures, or bile leaks. Two patients did not undergo postoperative ERCP because of spontaneous passage of their endobiliary stents noted on plain radiographs taken for evaluation of nonbiliary pathology. Six- to 30-month follow-up demonstrates no complications to date.

the risk that patients with retained CBDS will require a second surgical procedure. Although we were able to clear CBDS in all cases, others have reported a CBDS clearance rate ranging from 81 to 100% after LCBDE [4, 9, 10]. In cases of incomplete stone removal, the transampullary endobiliary stent will provide effective biliary decompression, eliminating concern for postoperative gallstone pancreatitis, cholangitis, and obstructive jaundice. Since the use of endobiliary stents has been shown to be effective in the treatment of postoperative bile leaks, their use in protecting the primary closure of the choledochotomy described herein limits the risk of bile leak [13]. Opponents may argue that a postoperative ERCP required for stent removal may not be cost effective. However, since completion cholangiography and choledochoscopy are highly effective in limiting the incidence of retained CBDS, with a 0% retained stone rate in this series, postoperative ERCP will not be necessary in many cases in which there was complete removal of CBDS. Stent removal may be accomplished easily with esophagogastroduodenoscopy to avoid the risks and expense of ERCP. Additional hospital costs are reduced with the realization that with the exception of the first three patients in our series, all patients were discharged the morning following surgery. These results compare favorably with a recent review by Ferzli in which the lengths of stay following LCBDE ranged from 1.7 to 12.0 days [11]. In conclusion, the discovery of occult CBDS demands a rational treatment approach that doesnt increase morbidity, hospital expense, or negate the beneficial effects of laparoscopic cholecystectomy. Many laparoscopic surgeons have mastered LCBDE, but their patients must still endure the limitations of indwelling T-tubes. The above technique allows CBDS to be removed expeditiously without the problems of T-tubes and preserves the benefits of laparoscopic cholecystectomynamely, quick recovery and early return to full activity. Our experience demonstrates this to be a safe and effective procedure.
Acknowledgment. The authors acknowledge medical illustrator Robin Brickman from Williamstown, Massachusetts, USA.

References Discussion Laparoscopic endobiliary stent placement eliminates reliance on postoperative T-tubes. Patient comfort is improved and there are no complications related to indwelling T-tubes including biliary leaks, common bile duct obstruction, duodenal erosions, and retained portions of T-tubes [3, 5, 15, 18]. Our results demonstrate that transductal laparoscopic placement of endobiliary stents following LCBDE allows early return to activity, provides decompression of the common bile duct, and guarantees successful cannulation of the CBD during postoperative ERCP when necessary. There are no complications from stent placement to date and all stents were successfully removed postoperatively. Surgeons unable to clear CBDS entirely during LCBDE because of multiple proximal stones or low insertion of the cystic duct may employ this technique since it eliminates
1. Abu-Khalaf A (1995) Endoscopic removal of retained common bile duct stones in patients with T tube in situ. Surg Laparosc Endosc 5: 1720 2. Arregui ME, Davis CJ, Arkush AM, Nagan RF (1992) Laparoscopic cholecystectomy combined with endoscopic sphincterotomy and stone extraction or laparoscopic choledochoscopy and electrohydraulic lithotripsy for management of cholelithiasis with choledocholithiasis. Surg Endosc 6: 1015 3. Benakis P, Nicolakis D, Triantafillidis JK (1994) Successful endoscopic removal of part of a T-tube from the common bile duct. Surg Endosc 8: 11681174 4. Berci G, Morganstern L (1994) Laparoscopic management of common bile duct stones. A multi-institutional SAGES study. Surg Endosc 8: 11681174 5. Bernstein DE, Goldberg RI, Unger SW (1994) Common bile duct obstruction following T-tube placement at laparoscopic cholecystectomy. Gastrointest Endosc 40: 362365 6. Cagir B, Rangraj M, Maffuci L, Ostrander LE, Herz BL (1994) A retrospective analysis of laparoscopic and open cholecystectomies. J Laparendosc Surg 4: 89100

304 7. Carroll B, Chandra M, Papaioannou T, Daykhovsky L, Grundfest W, Phillips E (1993) Biliary lithotripsy as an adjunct to laparoscopic common bile duct stone extraction. Surg Endosc 7: 356359 8. Curet MJ, Pitcher DE, Martin DE, Zucker KA (1995) Laparoscopic antegrade sphincterotomy. A new technique for the management of complex choledocholithiasis. Ann Surg 221: 149155 9. DePaula AL, Hashiba K, Bafutto M (1994) Laparoscopic management of choledocholithiasis. Surg Endosc 8: 13991403 10. Ferzli GS, Massaad A, Ozuner G, Worth MH (1991) Laparoscopic exploration of the common bile duct. Surg Gynecol Obstet 4: 419421 11. Ferzli GS, Hurwitz JB, Massaad AA, Piperno B (1996) Laparoscopic common bile duct exploration: a review. J Laparendosc Surg 6: 413 419 12. Gadacz TR (1993) U.S. experience with laparoscopic cholecystectomy. Am J Surg 165: 450454 13. Jenkins MA, Ponsky JL, Lehman GA, Fanelli RD, Bianchi T (1994) Treatment of bile leaks from the cystohepatic ducts after laparoscopic cholecystectomy. Surg Endosc 8: 193196 14. Jones DB, Soper NJ (1996) The current management of common bile duct stones. Adv Surg 29: 227233 15. Kacker LK, Mittal BR, Sikora SS, Ali W, Kapoor VK, Sacena R, Das BK, Kaushik SP (1995) Bile leak after T-tube removala scintigraphic study. Hepatogastroenterology 42: 975978 16. Kelley JE, Burrus RG, Burns RP, Graham LD, Chandler KE (1993) Safety, efficacy, cost, and morbidity of laparoscopic versus open cholecystectomy: a prospective analysis of 228 consecutive patients. Am Surg 59: 2327 17. Lange V, Rau HG, Schardey HM, Meyer G (1993) Laparoscopic stenting for protection of common bile duct sutures. Surg Laparosc Endosc 3: 466469 18. Mosimann F, Schneider R, Mir A, Gillet M (1994) Erosion of the duodenum by a biliary T-tube: an unusual complication of liver transplantation. Transplant Proc 26: 35503551 19. Petelin JB (1993) Laparoscopic approach to common duct pathology. Am J Surg 165: 487491 20. Stoker ME (1995) Common bile duct exploration in the era of laparoscopic surgery. Arch Surg 130: 265268 21. Swanstrom LL, Marcus DR, Kenyon T (1996) Laparoscopic treatment of known choledocholithiasis. Surg Endosc 10: 526528 22. Unger SW, Rosenbaum G, Unger HM, Edelman DS (1993) A comparison of laparoscopic and open treatment of acute cholecystitis. Surg Endosc 7: 377379 23. Waters GS, Crist DW, Davoudi M, Gadacz TR (1996) Management of choledocholithiasis encountered during laparoscopic cholecystectomy. Am Surg 62: 256258

but often this is just performed with a gastroscope, just to snare the stent and remove it, so we do feel it to be cost effective. Dr. Petelin: Congratulations, again, on a nice paper. I would echo the comments of the previous questioner, and I would suggest that youre not preserving all the benefits of laparoscopic cholecystectomy by leaving the stent. If its a dissolvable stent or one that will pass on its own that probably is the case, but the need to perform a second procedure does increase the risk. We have used primary closure in a number of our patients without an indwelling stent, and it has worked, so I guess I would ask you to consider the subsequent procedure and bring us back some details on how much that does cost. Dr. Gersin: Thank you, Dr. Petelin. Again, we did stent all of our patients, even if we explored them transcystically. In fact, it may prove to be unnecessary. However, due to our concerns of distal spasm and edema, we felt that it was safe just to provide a stent in those patients. I should also add that the stent is an excellent adjunct if youre unable to completely clear those stones, because it does guarantee successful cannulation if postoperative ERCP is necessary. Dr. Minasi: If you are taking your patients back and endoscoping them at one month, and not restudying them, then youre accepting a 13% chance of missing retained stones that you didnt see at the time of your initial clearance. Do you have any comment on that? Dr. Gersin: We havent had any problems in the 30-month follow-up to date on retained stones, so that if, in fact, there are retained stones, it would be, again, a very low percentage, and no patients have had any complications from that. We do perform routine intraoperative choledochoscopy and cholangiography to try to minimize the incidence of there being retained stones. Dr. Cosgrove: I enjoyed your presentation. My colleague, Marie Chen, has looked at the role of intraoperative ERCP for the common duct stones. We have about 20 cases. She has reported this in abstract form. They have been successful removing the stones in all but two cases. Can you comment on what you feel the role of intraop ERCP is doing to laparoscopic cholecystectomy? Dr. Gersin: Yes, I think thats a technique that certainly, currently is evolving. I think its a good technique. We have tried it on several of our patients. With the first two patients there was difficulty, due to positioning. We are more used to having our patients in the left lateral position, and the patients who have it intraoperatively are obviously supine. I do think its good; however, you must keep in mind that there are several surgical colleagues who at this time are unable to perform ERCP, so if youre a surgeon who can perform intraoperative ERCP I think its an excellent idea.

Discussion Dr. Schirmer: Thank you very much. I appreciate your study. It was very nice. But I do have one concern, and that is what youre basically saying by your argument against the T-tube is that it prolongs recovery and its, therefore, more costly. However, have you compared that, the difference in the hospitalization, with having to have a T-tube versus the cost to go back and pull that stent out, because thats another endoscopy, and thats a fairly costly procedure as well. Dr. Gersin: Right. We feel that the cost was offset by the fact that our patients were discharged within 24 hours, whereas in most of the literature the patients who did require T-tubes and common bile duct explorations were in the hospital for a period of time much longer than 24 hours. Second of all, the patients do require a repeat endoscopy,

Surg Endosc (1998) 12: 327330

Springer-Verlag New York Inc. 1998

Biliary stenting is more effective than sphincterotomy in the resolution of biliary leaks
J. M. Marks,1,3 J. L. Ponsky,2 R. B. Shillingstad,1 J. Singh3
1 2

Department of Surgery, Case Western Reserve University School of Medicine, 11100 Euclid Avenue, Cleveland, OH 44106, USA Department of Endoscopic Surgery, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA 3 Mount Sinai Medical Center, One Mount Sinai Drive, Cleveland, OH 44106, USA Received: 19 March 1997/Accepted: 14 July 1997

Abstract Background: Biliary fistulae may occur following surgical injury, abdominal trauma, or inadequate closure of a cystic duct stump. These leaks are most often managed by drainage of the associate biloma and either endoscopic sphincterotomy or placement of a biliary endoprosthesis to decrease the pressure gradient between the bile duct and the duodenum created by the muscular contraction of the ampullary sphincter. In a previous study, we demonstrated a statistically significant reduction in ductal pressures following stent placement as compared to sphincterotomy. The goal of this present study was to determine if reduction in ductal pressures correlates clinically with the resolution of biliary leaks in an animal model. Methods: Fourteen mongrel dogs underwent laparotomy, cholecystectomy without closure of the cystic stump, and a lateral duodenotomy to identify the major papilla. The dogs were then randomized into three groups. Group I (n 5) was a control group undergoing closure of the duodenotomy only. Group II (n 4) underwent sphincterotomy. Group III (n 5) underwent placement of a 7 Fr 5 cm biliary endoprosthesis prior to duodenotomy closure. A drain was placed adjacent to the cystic duct stump in all groups. Drain output was recorded daily. The biliary leak was considered resolved when the output was <10 cc/day. Regardless of suspected fistula closure, the drains were not removed until 2 weeks postprocedure. Necropsy was performed to identify undrained intraperitoneal bile. Statistical analysis was performed using Students paired t test. Results: All dogs had bile leaks identified on postoperative day 1. The number of days required for resolution of bile leak in group I (mean SEM) was 7.60 0.87 days, as compared to 6.75 0.80 days for group II and 2.60 0.24 days for group III. There was no significant difference in the duration of bile leak between groups I and II (p 0.445).
Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 1922 March 1997 Correspondence to: J. M. Marks

Group III, however, had a significant reduction in the duration of biliary fistulae as compared to both groups I and II (p < 0.005). At autopsy, persistent bilomas were identified in 80% of group I, 25% of group II, and 0% of group III. None of the dogs showed evidence of dehisence of the duodenotomy closure site as a source of bile leak. Conclusions: Biliary stenting significantly reduces the time to resolution of cystic duct leaks as compared to sphincterotomy in a canine model. The results obtained in this study support the use of biliary endoprostheses in the management of biliary leaks and fistulae. Key words: Biliary fistulae Endoscopic stent placement Endoscopic sphincterotomy

Biliary fistulae may occur following surgical injury, penetrating or blunt abdominal trauma, or secondary to inadequate closure of a cystic duct stump or accessory duct [12, 21]. The majority of biliary tract leaks occur from the extrahepatic biliary tree. The management of these leaks includes drainage of the intraperitoneal bile collection and decompression of the bile duct. This requires the integrated use of multiple modalities, including computed tomography, scintigraphy scanning, and endoscopic retrograde cholangiography [3, 20, 25]. The preferred endoscopic treatment, however, has yet to be determined. Endoscopic sphincterotomy by electrosurgical division of the sphincter of Oddi is thought by some to provide adequate drainage of the common bile duct [13]. Others, however have proposed the placement of biliary endoprostheses or nasobiliary tubes. The goal of both treatments is to decrease the pressure gradient between the bile duct and the duodenum created by the muscular contraction of the ampullary sphincter. Previous work by our lab showed that biliary endoprosthesis placement led to a significant reduction in bile duct pressures, as compared to sphincterotomy. The goal of the present study was to determine if this reduction in ductal pressures correlates clinically with the resolution of biliary fistulae.

328

Materials and methods

Fourteen mongrel dogs (mean weight, 20 kg) were fasted for 12 h preoperatively. After being anesthetized with intravenous thiopental (1020 mg/ kg), acepromazine (2.5 mg), and atropine (0.08 mg/kg), they were endotracheally intubated with maintenance on a Harvard Pump respirator. Additional doses of thiopental were given as necessary to maintain adequate anesthesia. Following midline laparotomy, the gallbladder in all animals was removed without closure of the cystic duct stump. A lateral duodenotomy was then created 6 cm distal to the pylorus and the major papilla identified. The dogs were then randomized into three groupsa control group (group I), a sphincterotomy group (group II), and a biliary stent group (group III). Randomization was performed by the selection of a sealed envelope with a 3 5 card detailing the specific group. Group I, the control dogs, underwent a sham procedure with immediate closure of the duodenotomy. In group II, sphincterotomies were performed following the duodenotomy with a standard endoscopic 5-Fr 20-mm double-channel papillotome (Wilson-Cook Medical Inc., Winston-Salem, NC, USA). A 10-mm sphincterotomy was made using blended cutting current delivered from a standard electrocautery generator (Valley Lab, Boulder, CO, USA). The dogs of group III, which was randomized to receive a biliary stent, underwent cannulation of the major papilla with a 0.035 in guide wire (Wilson-Cook Medical Inc.) and placement of a 7-Fr 5-cm Cotton-Leung biliary endoprosthesis (Wilson-Cook Medical Inc.). The duodenotomy was closed in all cases with interrupted 3-0 silk sutures. A closed drain system (Baxter Healthcare Corp., Deerfield, IL, USA) was placed adjacent to the cystic duct stump and tunneled subcutaneously to exit posteriorly between the scapulae. Protective wraps were placed around the dogs abdomen to prevent the dog from removing the drains. All dogs received clear liquids the night following surgery and were advanced to a regular diet the next day. Drain output was recorded daily. The biliary fistula was considered resolved when the output dropped to <10 cc per day, but the drains were not removed until 2 weeks after surgery, when the dogs were euthanized with a lethal dose of pentobarbitol. A necropsy was performed to identify undrained intraperitoneal bile or a persistent biloma. All protocols were approved by the Animal Research Committee prior to this study. Statistical analysis was performed using Students paired t test. Statistical significance was assigned to p < 0.05. Results are expressed as mean SEM.

Fig. 1. Daily outputs from the drains were reported for the control (solid triangles), sphincterotomy (open circles), and the biliary stent groups (solid squares).

Results In this study, 14 dogs were randomized to group I (n 5), group II (n 4), or group III (n 5). All dogs had biliary leaks identified on the 1st postoperative day. The number of days required for resolution of the biliary leak in group I was 7.60 0.87 days, as compared to 6.75 0.80 days for group II and 2.60 0.24 days for group III (Fig. 1). There was no significant difference in the duration of biliary leaks between group I and group II (p 0.445). The group that underwent biliary stent placement, however, had a significant reduction in the duration of biliary leak as compared to both the control and sphincterotomy groups (p < 0.005). At autopsy, persistent bilomas were identified in 80% of group I, 25% of group II, and 0% of group III. None of the animals had duodenotomy dehisences identified as a source for bile leakage at autopsy. Discussion We designed this study to compare in a prospective randomized fashion the two endoscopic techniques currently available for the treatment of bile duct leaks. In previous animal experiments, we found that biliary stent placement was superior to endoscopic sphincterotomy in overcoming the outflow resistance of the ampullary sphincter. There-

fore, we set out to determine the clinical significance of this reduction in biliary pressures and ascertain if it correlated with the actual resolution of bile duct leaks. The number of days required for resolution of bile leaks in the control dogs (group I) was 7.60 0.87 days; it was 6.75 0.80 days in the dogs undergoing sphincterotomy (group II) and 2.60 0.24 days in the biliary stent group (group III). There was no significant difference between the control and sphincterotomy groups in the duration of the bile leaks (p 0.445). The biliary stent group, however, had a significant reduction in the duration of bile leaks as compared to both the control and sphincterotomy groups (p < 0.005). At autopsy, no bilomas were identified in any of the dogs who had undergone biliary stent placement, whereas four of the five control dogs and one of the four sphincterotomy dogs did have persistent intraperitoneal bile at autopsy. The incidence of common bile duct injuries following open cholecystectomy has been reported to be 0.070.1% [1, 16]. In contrast, the incidence of common bile duct injuries following laparoscopic cholecystectomy appears to be at least twice that of open cholecystectomy; recent reports have estimated the incidence to be 0.20.3% [17, 31, 32, 38]. The rate of bile duct injury varies according to the experience of each institution performing laparoscopic cholecystectomy. In one series, the incidence of bile duct injury in institutions having performed <100 cases was 0.65%, as compared to 0.42% in those that had performed >100 [11]. Bile leaks were identified in 0.29% of patients in this series and represented the most common postoperative complication. In most series, cystic duct leaks comprise 20% of all biliary tract injuries following laparoscopic cholecystectomy [36]. The management of biliary fistulae has evolved with the advancement of endoscopic techniques. In a review of 77 cases of endoscopic management of postoperative biliary leaks, Binmoeller et al. found the treatment to be successful in 95% of cases [4]. This led to resolution of the biliary leaks in 82% of the patients. The practice of intraperitoneal bile drainage in combination with biliary duct decompression is the standard therapeutic regimen. Percutaneous drains have been placed under sonographic or CT guidance, and decompression of the biliary tree has been performed

329

utilizing endoscopic sphincterotomy and/or biliary stent placement. Whether sphincterotomy or stent placement results in shorter time to closure of biliary leaks has yet to be established. Most of the series investigating the results of these two endoscopic techniques comprised only small numbers of patients, and there have been no prospective randomized trials comparing the two methods. Several authors have reported small series of patients with successful resolution of biliary fistulae following electrosurgical division of the ampullary sphincter [7, 8, 18, 27, 28]. In addition, Geenen et al. found a significant reduction in the common bile duct to duodenal pressure gradient at 12 months following endoscopic sphincterotomy [14]. No studies, however, have investigated the immediate pressure changes following endoscopic sphincterotomy. It is our belief that the results of the Geenen study do not reflect the early changes following endoscopic sphincterotomy. The local inflammatory process that occurs following electrosurgical division of the sphincter may prevent immediate biliary decompression. This hypothesis is supported by our previous work, which identified no significant acute reduction in common bile duct pressures following sphincterotomy as compared to baseline in a canine model [37]. The use of biliary endoprostheses has been recommended by other investigators as a more effective method for equilibrating the pressures between the biliary tree and the duodenum than sphincterotomy alone [4, 15, 16, 19, 30, 34, 35, 37]. In addition, Mortensen and Kruse argued that endoscopic stent placement was superior to endoscopic sphincterotomy because it preserves the sphincter of Oddi. The complication rate from stent placement was very low in their report; there was only one case of cholangitis [23]. Smith et al. documented the successful management of several high-volume biliary cutaneous fistulae, which healed promptly after placement of biliary stents [30]. Traverso et al. also prefer endoscopic stent placement for the management of cystic duct leaks following cholecystectomy. They reported that it allowed for successful decompression of the biliary tree with almost immediate amelioration of the biliary fistula [33]. Moulton et al. presented three cases of biliary leaks following pediatric liver trauma that were all managed successfully with ERCP and transampullary biliary decompression [24]. The major disadvantage to biliary stent placement is that it requires a second endoscopic procedure to remove the stent. One alternative to stent placement is the use of a nasobiliary tube [2, 5, 9, 22]. These tubes can be removed without a repeat endoscopy, although a follow-up cholangiogram is usually performed. The disadvantages of nasobiliary tubes include patient discomfort, fluid losses, electrolyte disturbances, and possibly prolonged hospitalization. Complications of endoprosthesis placement include biliary obstruction, cholangitis, stent migration, and pancreatitis. The major complications of endoscopic sphincterotomy can be divided into immediate and long-term risks. The immediate complications of endoscopic sphincterotomy include bleeding, duodenal perforation, and pancreatitis. The long-term complications of endoscopic sphincterotomy are as yet unknown. However, the postsphincterotomy stenosis rate may be as high as 613% in patients followed for >10 years [6]. In addition, the use of endoscopic sphincterotomy

in patients with small- or normal-caliber bile ducts can lead to higher complication rates, as has been reported by Sherman et al. [29]. The placement of 10-Fr biliary stents, however, does not routinely require a sphincterotomy or papillotomy. In conclusion, biliary stenting significantly reduces the time to resolution of cystic duct leaks as compared to sphincterotomy in this canine model. Although the sample numbers are small and variations in the severity of fistulae leading to statistical bias may exist due to different baseline intraductal pressures, cystic duct length, and CBD diameter, the results obtained in this study support the use of biliary endoprostheses rather than endoscopic sphincterotomy in the management of simple biliary fistulae.
Acknowledgments. We are grateful to Wilson-Cook Medical Inc., WinstonSalem, NC, USA, for their support of this work. We also thank Jane Dostal and Brian Dunkin, M.D., for their assistance in the preparation of this manuscript.

References
1. Andren-Sandberg A, Johannson S, Bengmark S (1985) Accidental lesions of the common bile duct II. Results and treatment. Ann Surg 20: 452455 2. Barthel JS, Sastri SV, Landsbaum C (1989) Closure of cystic duct stump leak by nasobiliary tube drainage. J Clin Gastroenterol 11: 574577 3. Bezzi M, Silecchia G, Orsi F, Materia A, Salvatori FM, Fiocca F, Fantini A, Basso N, Rossi P (1995) Complications after laparoscopic cholecystectomy: coordinated radiologic, endoscopic, and surgical treatment. Surg Endosc 9: 2936 4. Binmoeller KF, Katon RM, Shneidman R (1991) Endoscopic management of postoperative biliary leaks: review of 77 cases and report of two cases with biloma formation. Am J Gastroenterol 86: 227231 5. Burmeister W, Koppen MO, Wurbs D (1985) Treatment of a biliocutaneous fistula by endoscopic insertion of a nasobiliary tube. Gastrointest Endosc 31: 279281 6. Cotton PB, Lehman G, Vennes J, Geenen JE, Russell RCG, Meyers WC, Liguory C, Nickl N (1991) Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 37: 383393 7. Davids PHP, Rauws EAJ, Tytgat GNJ, Huibregtse K (1992) Postoperative bile leakage: endoscopic management. Gut 33: 11181122 8. Del Olmo L, Merono E, Moreira VF, Garcia T, Garcia-Plaza A (1988) Successful treatment of postoperative external biliary fistulas by endoscopic sphincterotomy. Gastrointest Endosc 34: 307309 9. Deviere J, van Gansbeke D, Ansay J, deToeuf J, Cremer M (1987) Endoscopic management of a post-traumatic biliary fistula. Endoscopy 19: 136139 10. Deziel DJ (1994) Complications of cholecystectomy: incidence, clinical manifestations, and diagnosis. Surg Clin North Am 74: 809823 11. Deziel DJ, Millikan KW, Economou SG, Doolas A, Ko S-T, Airan MC (1993) Complications of laparoscopic cholecystectomy: a national survey of 4,292 hospitals and an analysis of 77,604 cases. Am J Surg 165: 914 12. Edelman DS (1994) Bile leak from the liver bed following laparoscopic cholecystectomy. Surg Endosc 8: 205207 13. Geenen JE, Hogan WJ, Shaffer RD, Stewart ET, Dodds WJ, Arndorfer RC (1977) Endoscopic electrosurgical papillotomy and manometry in biliary tract disease. JAMA 237: 20752078 14. Geenen JE, Toouli J, Hogan WJ, Dodds WJ, Stewart ET, Mavrelis P, Riedel D, Venu R (1984) Endoscopic sphincterotomy: follow-up evaluation of effects on the sphincter of Oddi. Gastroenterology 87: 754758 15. Goldin E, Katz E, Wengrower D, Kluger Y, Haskel L, Shiloni E, Libson E (1990) Treatment of fistulas of the biliary tract by endoscopic insertion of endoprostheses. Surg Gynecol Obstet 170: 418423 16. Goldin E, Libson E, Rachmilewitz D (1987) Endoscopic insertion of endoprostheses in the treatment of postoperative cutaneous biliary fistula. Surgery 102: 8890

330 17. Gilliland TM, Traverso LW (1990) Modern standards for comparison of cholecystectomy with alternative treatments for symptomatic cholecystectomies with emphasis on long-term relief of symptoms. Surg Gynecol Obstet 170: 3944 18. Hsu D (1987) Endoscopic sphincterotomy in the management of biliary-cutaneous fistula. Am J Gastroenterol 82: 10781080 19. Janardhanan R, Brodmerkel GJ Jr, Turowski P, Gregory DH, Agrawal RM (1986) Endoscopic retrograde cholangiopancreatography in the diagnosis and management of postcholecystectomy cystic duct leaks. Am J Gastroenterol 81: 474476 20. Kaufmann SL, Kadir S, Mitchell SE, Chang R, Kinnison ML, Cameron JL, White RI Jr (1985) Percutaneous transhepatic biliary drainage for bile leaks and fistulas. AJR 144: 10551058 21. Klotz HP, Schlumpf R, Largiader F (1992) Injury to an accessory bile duct during laparoscopic cholecystectomy. Surg Laparosc Endosc 2: 317320 22. Leung JWC, Chung SCS, Sung JY, Metreweli C (1988) Endoscopic management of postoperative biliary fistula. Surg Endosc 2: 190193 23. Mortensen J, Kruse A (1992) Endoscopic management of postoperative bile leaks. Br J Surg 79: 13391341 24. Moulton SL, Downey EC, Anderson DS, Lynch FP (1993) Blunt bile duct injuries in children. J Pediatr Surg 28: 795797 25. Mueller PR, Ferrucci JT Jr, Simeone JF, Cronan JJ, Wittenberg J, Neff CC, van Sonnenberg E (1983) Detection and drainage of bilomas: special considerations. AJR 140: 715720 26. Mullen JP, Carr RE, Rupnik EJ, Knapp RW (1976) 1000 cholecystectomies, extraductal palpation and operative cholangiography. Am J Surg 131: 672675 27. ORahilly S, Duignan JP, Lennon JR, OMalley E (1983) Successful treatment of a post-operative external biliary fistula by endoscopic papillotomy. Endoscopy 15: 6869 28. Ponchon T, Gallez JF, Valette PJ, Chavaillon A, Bory R (1989) Endoscopic treatment of biliary tract fistulas. Gastrointest Endosc 35: 490498 Sherman S, Ruffolo TA, Hawes RH, Lehman GA (1991) Complications of endoscopic sphincterotomy: a prospective series with emphasis on the increased risk associated with sphincter of Oddi dysfunction in nondilated bile ducts. Gastroenterology 101: 10681075 Smith AC, Schapiro RH, Kelsey PB, Warshaw AL (1986) Successful treatment of nonhealing biliary-cutaneous fistulas with biliary stents. Gastroenterology 90: 764769 Soper NJ (1991) Laparoscopic cholecystectomy. Curr Probl Surg 28: 587655 Southern Surgeons Club (1991) A prospective analysis of 1518 laparoscopic cholecystectomies. N Engl J Med 324: 10731078 Traverso LW, Kozarek RA, Ball TJ, Brandabur JJ, Hunter JA, Jolly PC, Patterson DJ, Ryan JR, Thirlby RC, Wechter DG (1993) Endoscopic retrograde cholangiopancreatography after laparoscopic cholecystectomy. Am J Surg 165: 581586 Vitale GC, Stephens G, Wieman TJ, Larson GM (1993) Use of endoscopic retrograde cholangiopancreatography in the management of biliary complications after laparoscopic cholecystectomy. Surgery 114: 806814 Woods MS, Shellito JL, Santoscoy GS, Hagan RC, Kilgore WR, Traverso LW, Kozarek RA, Brandabur JJ (1994) Cystic duct leaks in laparoscopic cholecystectomy. Am J Surg 168: 560565 Woods MS, Traverso LW, Kozarek RA, Tsao J, Rossi RL, Gough D, Donohue JH (1994) Characteristics of biliary tract complications during laparoscopic cholecystectomy. Am J Surg 167: 2732 Youngelman DF, Marks JM, Ponsky T, Ponsky JL (1997) Comparison of bile duct pressures following sphincterotomy and endobiliary stenting in a canine model. Surg Endosc 11: 126128 Zucker KA, Bailey RW, Gadacz TR, Imbembo AL (1991) Laparoscopic guided cholecystectomy. Am J Surg 161: 3644

29.

30. 31. 32. 33.

34.

35. 36. 37. 38.

EndoScope: world literature reviews

Surg Endosc (1998) 12: 355358 Springer-Verlag New York Inc. 1998

Original articles from a wide range of international surgical journals are selected by our editors and presented here as a structured summary and critical review. EndoScope serves as a quick and comprehensive survey of the expansive endoscopic literature from all the corners of the globe.

Common bile duct exploration and laparoscopic cholecystectomy: role of intraoperative ultrasonography

Randomized, prospective, single-blind comparison of laparoscopic versus small-incision cholecystectomy

Majeed, AW, et al Lancet (1996) 347: 989994 A prospective randomized comparison of elective laparoscopic and small-incision cholecystectomy in 200 patients involving four surgeons was conducted in Sheffield, UK. All subjects had symptomatic gallstones confirmed by ultrasonography and randomized for elective cholecystectomy by either method. The four participating surgeons had a minimum of 40 laparoscopic cholecystectomies each as either the principal operator or assistant, and no comment is made about their small-incision experiences. The smallincision method requires a high transverse sub-xiphoid incision over the junction of the cystic duct and common bile duct. The length of the incision is tailored to individual cases (median of 7 cm). Standard instruments are used for the fundus-last dissection and, finally, as in laparoscopy, all cases are accompanied by intraoperative cholangiogram. The authors report that more time was required for laparoscopic cholecystectomies when compared to the smallincision procedure (median 65 min vs 40 min, respectively, and including cholangiogram time). They found no difference between the groups for hospital stay, time back to work, and time to full activity. Of the 205 meeting the criteria for elective cholecystectomy between 1992 to 1995, only five cases were eliminatedbecause the patients refused to be randomized, hepatic metastasis, or cirrhosis was present. Although patient selection is randomized, the four surgeons participating in the study were still able to affect the outcome of the studya design flaw that is always an inherent problem in this type of study. The longer time reported for the laparoscopic procedure may not be a significant finding, as it can certainly be shortened. Finally, the complication reported is notable for one bile duct injury only in the laparoscopic group, which was fortunately identified after an intraoperative cholangiogram. An interesting observation is that the overall conversion rate in the laparoscopic group turned out to be 20%a figure that is much higher than reported in the literature. Certainly, the authors have shown that for the surgeons well adapted to the small-incision technique, similar post-

Santambrogio, R, et al J Am Coll Surg (1997) 185: 17 The role of laparoscopic ultrasound (LUS) in determining whether a patient has choledocholithiasis during a laparoscopic cholecystectomy (LC) was prospectively studied. Two hundred sixteen consecutive patients with symptomatic gallstones were enrolled in the study. All underwent a standard preoperative algorithm including ultrasonography and measurements of liver function tests on the working day before operation. Patients at high risk for common bile duct (CBD) stones underwent ERCP before LC. Laparoscopic ultrasonography examination during cholecystectomy was routinely performed to identify stones unsuspected preoperatively. One hundred seventy-seven (82%) of the patients were determined to be low risk for choledocholithiasis and 39 patients (18%) were at high risk and had preoperative ERCP. In 17 patients (43.5%) CBD stones were found, and in 16 patients (41%) they were removed by endoscopic sphincterotomy. In all patients LUS documented the intra- and extrahepatic ducts, but the distal tract of the CBD was not well visualized in eight cases. Small stones were found in the CBD of eight patients. A subsequent CBD exploration or intraoperative cholangiogram confirmed the diagnosis in all but one patient. Retained CBD stones were found in two patients during the follow-up period. An endoscopic sphincterotomy was performed in those patients successfully. Based on these data the authors suggest that laparoscopic ultrasonography may be a real alternative to cholangiography during laparoscopic cholecystectomy. The authors also point out that considerable ultrasonographic experience is required for LUS to be performed successfully. Comments: This paper is an interesting evaluation of laparoscopic ultrasonography to determine choledocholithiasis. It would have been beneficial if the study had looked at the role of laparoscopic ultrasound in delineating the anatomy to prevent CBD injuries. The paper did not allude to any advantages of laparoscopic ultrasonography vs cholangiography to justify learning this new skill for the purpose of detecting choledocholithiasis.

356

operative recovery results can be achieved as with the laparoscopic method. However, the bias for or against the procedure is impossible to remove from the surgeons in the study, and therefore true randomization remains questionable.

Endoscopic monitoring of operative procedures during cardiac surgery

Miyagawa, H, et al Surg Today Jpn J Surg (1997) 27: 434438

Laparoscopic or open splenectomy for hematologic disease: which approach is superior?

Friedman, RL, et al J Am Coll Surg (1997) 185: 4954 A retrospective review of 137 patients who underwent laparoscopic (LS) and open splenectomy (OS) at the CedarsSinai Medical Center in Los Angeles, California, was performed. The study was performed to compare safety, outcome, and costs of laparoscopic and open splenectomy for a variety of hematologic diseases. Diagnosis, age, gender, operative time, blood loss, splenic weight, time to resumption of oral diet, postoperative hospital stay, morbidity, mortality, and costs were analyzed by multivariate statistical analysis. Patients who underwent laparoscopic splenectomy had significantly shorter hospital stay and time to resumption of an oral diet (p < 0.01). Operative costs were higher, although the total direct costs were not. Idiopathic thrombocytopenic purpura patients had earlier resumption of an oral diet after LS, shorter postoperative stay, and comparable OR time. Five patients (7%) were converted, with outcomes similar to OS except for greater operative time and cost. Grade II complications occurred in three LS and four OS patients; grade III in three OS patients; and grade IV in two OS patients. There were two major complications of LS and eight of OS, with two deaths. Multivariate analysis revealed that operative time and time to resumption of oral intake were significantly related to age, diagnosis, operative technique, and splenic weight. The duration of postoperative hospitalization was related to operative technique, splenic weight, and major complications. Costs (direct and operative) were related to age, splenic weight, and major complications, but not to operative technique. The authors concluded that LS results are influenced by splenic weight, disease, and age. Splenic weight appears to be the crucial determinant of operative time and length of hospitalization. LS is a superior treatment for patients with idiopathic thrombocytopenic purpura and patients with small spleens. Comments: This is a very well organized review comparing laparoscopic and open splenectomy. However, we must keep in mind that this is a retrospective review and patient selection for open or laparoscopic splenectomy was based on surgeon preference as well as time of resumption of oral intake and length of hospitalization. Perhaps a large prospective randomized trial comparing laparoscopic and open splenectomy might draw the same conclusions.

This is a case report of five patients who underwent intraoperative monitoring using an endoscope video system during open heart surgery. The scope was used to visualize (1) the inside of the aorta to identify tears during hemiarch replacement for an acute aortic dissection, (2) an anastomosis and debris via a left ventriculotomy for a left ventricle aneurysm, (3) the right ventricular outlet tract and location of an atrial septal defect (ASD) in a patient with an ASD and a right ventricular outlet tract resection, (4) the anastomosis during a coronary artery bypass graft (CABG) in a 14-year-old with Kawasakis disease, and (5) any leakage during a mitral valve plasty and aortic valve replacement. The authors concluded that an endoscopic video monitoring system is useful not only for thoracic surgery but also for cardiac surgery. Comments: Using an endoscope for magnification during CABG seems much more cumbersome than directly viewing the anastomosis using magnifying glasses. Perhaps an endoscope may be useful in viewing lesions that are intracardiac and otherwise difficult to view directly. The new field of minimally invasive coronary artery surgery will doubtless engender many such novel developments.

A video-assisted thoracoscopic surgical technique for interruption of patent ductus arteriosus

Tsuboi, H, et al Surg Today Jpn J Surg (1997) 27: 439442 This study describes a technique for closure of a patent ductus arteriosus (PDA) using video-assisted thoracoscopic surgery (VATS). Five patients with a mean age of 3 years and mean weight of 13.7 kg were operated. Under general anesthesia, two 10-mm trocars and two or three 5-mm trocars were inserted through the left thoracic wall. A video camera and specially designed surgical tools including scissors, dissectors, and a clip applicator were introduced. The ductus was dissected, and two titanium clips were applied to interrupt the ductus completely. Closure of the PDA was successful in all patients. The only complication was in one patient who developed hoarseness for 2 weeks postoperatively. The hospital stay ranged from 7 to 12 days. The authors concluded that VATS is a safe and effective technique for achieving closure of PDA. Comments: The paper describes a technique which seems to be an excellent alternative to a classical left thoracotomy.

357

Further experience with this procedure and larger studies in the literature might make this procedure the preferred technique for closure of a PDA.

A comparison of open and laparoscopic cholecystectomy for patients with cirrhosis

Saeki, H, et al Surg Today Jpn J Surg (1997) 27:411413 This is a retrospective study of 13 patients who underwent cholecystectomy for gallstones over a 6-year period: Seven had laparoscopic cholecystectomy (LC) and six had open cholecystectomy (OC). No statistical differences were observed in the duration of surgery or the intraoperative blood loss between the two groups; however, the C-reactive protein level in the serum was significantly higher in the OC group than in the LC group. LC was followed by a significantly shorter hospital stay (p < 0.05) and earlier resumption of a normal diet (p < 0.05) in comparison to OC. All of the patients who underwent OC had an uneventful clinical course; however, 9 of the patients who underwent LC suffered from intractable ascites postoperatively. The difference in the cost of hospitalization between the two groups was not statistically significant. The authors conclude that the therapeutic significance of performing LC in patients with cirrhosis should be assessed after carefully evaluating all factors including mortality, morbidity, and cost effectiveness. Comments: This is a retrospective study with a very small number of patients. No conclusions can be drawn from this paper. The authors stated that caution should be exercised when selecting laparoscopic cholecystectomy for patients with cirrhosis because one patient in the group developed severe ascites leading to liver failure and death 5 months after operation. But one may not generalize because no statistical significance is possible with such a small number of patients. The authors did not propose any hypothesis as to why they think that there would be any difference between comparing LC and OC in patients with and without cirrhosis.

water-soluble contrast enema, computed tomography, and cystoscopy were used to establish the diagnosis. All patients were managed by a one-stage excision of the colovesical fistula combined with simultaneous colon resection using a linear stapler cutter. The mean operative time was 3 h 50 min. There were no reports of any complications and full oral feeding was initiated on the 3rd postoperative day. The mean total hospital stay was 13 days. The authors concluded that a one-stage excision of colovesical fistulas combined with simultaneous colon resection is feasible laparoscopically, but the training and experience of the surgeon are essential for the success of this method. Comments: This article was a very well written case report. It would be interesting to know how low down the sigmoid colon was resected. It may be difficult to use the staplers to resect the sigmoid colon down to the peritoneal reflection to prevent recurrence of the diverticular disease.

Emergency minilaparotomy cholecystectomy for acute cholecystitis: prospective randomized trialimplications for the laparoscopic era
Assalia, A, et al World J Surg (1997) 21: 534539 This article is a prospective controlled trial comparing minicholecystectomy (MC) in cases of acute cholecystitis to conventional cholecystectomy (CC). Sixty consecutive patients with acute cholecystitis were prospectively randomized to undergo MC (30 cases) or CC (30 cases). The two groups were well matched with regard to age, sex, weight/ height index, previous upper abdominal surgery, and APACHE II scores. The mean length of incision was 5.5 cm in the MC group compared to 13.5 cm in the cc group. There was no signicant difference with regard to operative time, operative difficulty on a 1 to 10 scale, and complication rate. Significantly lower analgesia requirements were noted in the MC group: 27.5 14.6 mg of morphine sulfate compared to 44.5 13.2 mg in the CC group (p < 0.001). Twenty-two patients (73%) in the MC group were reported to return to normal daily activities 2 weeks after the operation, compared to only 12 (40%) in the CC group (p 0.0028). The authors concluded that MC is safe and applicable as an emergency procedure for acute cholecystitis. It is superior to CC in terms of convalescence and cosmesis. The results of MC in the setting of acute cholecystitis compare favorably with the published results of laparoscopic cholecystectomy (LC). Comments: This study is a well-designed prospective controlled study. One thing which was not controlled for in the study was the number of days it took to go to the operating room from the time of onset of symptoms, which has previously be shown by Garber et al (Surg Endosc [1997] 11: 347350) to affect the conversion rate and morbidity for

Treatment of colovesical fistulas by laparoscopic surgery: report of five cases

Nassiopoulos, K, et al Dig Surg (1997) 14: 5660 This study is a case report on five cases of colovesical fistulas that were treated by laparoscopic surgery. In all cases the clinical picture was characterized by pneumaturia and signs of diverticular disease of the sigmoid. Barium or

358

laparoscopic cholecystectomy. The conclusion that MC compares favorably with LC is not completely accurate. The average length of stay of 3.1 1.0 days is higher than other studies on laparoscopic cholecystectomy for acute cholecystitis. Perhaps a prospective study comparing MC to LC would determine if MC is comparable to LC.

Laparoscopic cholecystectomy for acute cholecystitis: prospective trial

Eldar, S, et al World J Surg (1997) 21: 540545 This is a prospective study to determine the indications for and the optimal timing of laparoscopic cholecystectomy (LC) following the onset of acute cholecystitis. It also evaluates preoperative and operative factors associated with conversion from LC to open cholecystectomy (OC) in the presence of acute cholecystitis. The study consisted of 130 patients, of which 93 (72%) underwent successful LC and 37 (28%) needed conversion to OC. The conversion rate of acute gangrenous cholecystitis (49%) was significantly higher than that for uncomplicated acute cholecystitis (4.5%) and for hydrops (28.5%) and empyema of the gallbladder (28.5%). Patients with an operative delay of 96 h or less from the onset of acute cholecystitis had a conversion rate of 23%, whereas a delay of more than 96 h was associated with a conversion rate of 47% (p 0.022). The complication rate was 8.5% in the LC group and 27% in the converted group. Patients over 65 years of age, with a history of biliary disease, a nonpalpable gallbladder, WBC count over 13,000/cc, and acute gangrenous cholecystitis were independently associated with a high LC conversion rate; male patients, finding large bile stones, serum bilirubin over 0.8 mg/dl, and WBC count over 13,000/cc were independently associated with a high complication rate following laparoscopic surgery with or without conversion. The authors concluded that LC can be performed safely for acute cholecystitis, with acceptably low conversion and complication rates. Different forms of cholecystitis carry various conversion and complication rates in selected cases. LC for acute cholecystitis should be performed within 96 h of the onset of disease. Predictors of conversion and complications may be helpful when planning the laparoscopic approach to acute cholecystitis. Comments: This is a large prospective study on acute cholecystitis and has conclusions similar to those previously reported by Garber et al. (Surg Endosc [1997] 11: 347350) in that the conversion rate is dramatically different within 4 days of onset of symptoms. The conversion rate for less than 96 h of symptoms was much higher in this paper, 23%, as compared to 3.6% reported by Garber et al. The experience of the surgeons was not well defined, and perhaps this played a role in the difference. Concerning the higher morbidity rate after conversion in this study, a posible bias is the association of several comorbid factors and the fact that conversion has been performed in the most difficult cases.

Early postoperative results of a prospective series of laparoscopic vs. open anterior resections for rectosigmoid cancers
Goh, YC, et al Dis Colon Rectum (1997) 40: 776780 This study prospectively compared postoperatively laparoscopic (LAR) with open (OAR) anterior resection in patients with rectosigmoid cancer. Forty consecutive patients were divided into two groups. Twenty patients without a palpable abdominal mass under anesthesia were subjected to LAR, and 20 patients with a palpable mass underwent OAR. Median lengths of distal margin of clearance beyond the tumor were 4 cm and 4.5 cm in the LAR and OAR groups, respectively. Median numbers of lymph nodes harvested were 20 and 19 for the LAR and OAR groups, respectively. Median operating times were 90 min and 73 min in the LAR and OAR groups, respectively. Blood loss was 50 ml in both groups. There was no intraoperative complication in either group, and no laparoscopic patient was converted to an open procedure. Median length of extraction site incision in the LAR group was 5.5 cm and the length of incision in the OAR group was 18 cm (p < 0.002). Median length of follow-up was 8 months for the LAR group and 10 months for the OAR group. The mean postoperative stay for the LAR group was 5 days and 5.5 days for the OAR group. The authors concluded that there were no significant differences between the two groups with regard to duration of parenteral analgesia, starting of fluid and solid diet after surgery, time to first bowel movement, or time to discharge from the hospital. Comments: This article brings attention to a very controversial area in laparoscopic surgery: colon cancer. This study reported no evidence of port site recurrence with a follow-up of only 8 months; however, the incidence of port site recurrence ranges from 0 to 4.5% in the literature. Given the results of this paper indicating no improvement in postoperative course for patients with LAR, one should think carefully about affecting the outcome of a potentially curable disease. Currently there are large multicenter studies looking at this issue. The authors certainly seem to be adept, as they are able to complete the procedure in an expeditious fashion.

Reviewers for this issue: S. M. Garber, J. M. Sackier, F. Chae.

Surg Endosc (1998) 12: 310314

Springer-Verlag New York Inc. 1998

Common bile duct injuries during laparoscopic cholecystectomy that result in litigation
B. J. Carroll, M. Birth, E. H. Phillips
Department of Surgery, Cedars-Sinai Medical Center, 8700 Beverly Boulevard, Los Angeles, CA 90048, USA Received: 3 April 1997/Accepted: 5 July 1997

Abstract Background: Iatrogenic common bile duct injury is the worst complication of laparoscopic cholecystectomy. The goal of this study is to increase awareness of the problem and educate surgeons about the consequences of these injuries. Methods: A total of 46 bile duct injuries were analyzed by review of medical records, cholangiograms, videotapes, and surgeon statements. All cases were involved in malpractice litigation. Results: All types of injuries were represented. There were 15 transections, 11 excisions, 6 lacerations, 8 clip impingements, 3 burns, 2 bile leaks, and 1 cystic duct leak. In all, 72% of these injuries occurred in elective cases in which there was no acute inflammation. Cholangiograms were performed in 16 cases, but they were misinterpreted in 11 of them. Injury type and severity was similar in patients with and without cholangiography. A total of 80% of these injuries were not detected at the initial surgery. The average delay in diagnosis was 10 days. Complications were worse in patients with delayed diagnosis. Primary surgeons had less successful outcomes from repairs than referral surgeons (27% versus 79%). In 86% of cases, litigation was resolved in favor of plaintiffs by settlement or verdicts. The average award was $214,000. Conclusions: Factors that predispose to lawsuits include treatment failures in immediately recognized injuries, complications that result from delays in diagnosis, and misinterpretation of abnormal cholangiograms. Injury prevention can be improved by increased awareness of common mistakes. Improved cholangiographic technique and interpretation should decrease injury severity, delays in diagnosis, and subsequent morbidity. Key words: Laparoscopic cholecystectomy Injuries Common bile duct Litigation
Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 20-21 March 1997 Correspondence to: B. J. Carroll, 8635 West 3rd Street, Suite 795 West, Los Angeles, CA 90048, USA

The benefits of laparoscopic cholecystectomy have been achieved against the backdrop of an increased incidence of bile duct injuries. A total of 98 studies have examined the outcome of 78,747 laparoscopic cholecystectomies. The incidence of bile duct injuries ranged from 0.36% to 0.47% [6]. However, selection bias, nonreporting of injuries, and a lack of long-term follow-up have led to an underestimation of the true incidence of injuries. This study was undertaken to increase the awareness of the types, consequences, and financial impact of laparoscopic bile duct injuries that result in litigation. Materials and methods
We reviewed 46 cases involving bile duct injury sustained during laparoscopic cholecystectomy, which were supplied by 41 attorneys. Each case was already involved in malpractice litigation. Complete medical records, intraoperative cholangiograms (obtained in 16 cases), and postoperative cholangiograms were studied. Videotape documentation of the operation was available in six cases.

Results The injuries occurred between September 1990 and March 1996. There were 41 women and five men. Their average age was 38 years. The average for women was 37 years; for men, it was 50 years. Bile duct injuries occurred whether the operations were performed by inexperienced or experienced laparoscopic surgeons. However, data regarding the surgeons prior experience were available in only 27 cases. Seven injuries occurred in the surgeons first 10 cases, 11 within the surgeons first 50 cases, four between 50 and 100 cases, and five injuries occurred after experience with >100 cases. All types of bile duct injuries were represented in these 46 cases. There were 15 transections, 11 excisions, 6 lacerations, 8 cases of clip impingement, three electrocautery burns, two bile leaks from the liver bed, and one cystic duct leak. The Bismuth level and injury type is shown in Table 1. Intraoperative factors and etiologies for injuries were determined by a review of records, cholangiograms, video documentation, and reoperative findings. Thirteen injuries

311 Table 1. Classification of injuries Injury type Bismuth level 1 2 3 4

a

Transection 1 10 1 3

Excision 0 7 3 1

Laceration 1 3 1 1

Impingement 2 6 0 0

Burn 1 0 1 1

Totala 5 26 6 0

Plus (one) cystic leak and two Luschka leaks not classified by Bismuth level.

occurred in the presence of acute inflammation (severe in seven cases, moderate in six cases) and 33 in conjunction with chronic scarring (severe in one case, moderate in 14 cases, minimal in 18 cases). In 25 of 46 cases (54%), no problem was recorded in the operative report or patient chart following laparoscopic cholecystectomy. In nine cases (20%) there was chronic scarring, adhesions, and thickening of the hilar tissue that was retrospectively thought to have contributed to injury. A short cystic duct was implicated by the operating surgeon causally in eight cases (17%), but this fact was noted only after recognition of the injury. Excessive bleeding was reported in three cases. In two cases, surgeons confused the common bile duct for a large cystic duct. In two additional cases, the common hepatic duct was confused for an accessory duct draining directly into the gallbladder after the common bile duct had been mistaken for a cystic duct and had been transected below the junction with the cystic duct. The mechanisms of injury were as follows: Inadequate Calot dissection/confusion of normal anatomy misidentification of CBD as cystic duct (17) cystic clips impinging on CBD (3) misidentification of CHD (2) Misinterpretation of cholangiography Blind clipping/cauterization near hilum Unknown/unrecognized Failure to recognize Luschka duct 22

11 5 6 2
Fig. 1. Intraoperative cholangiogram demonstrating nonvisualization of hepatic duct performed prior to common bile duct transection.

In 37 patients (80%), bile duct injuries were not immediately recognized during the initial surgery. In 16 patients who had intraoperative cholangiography, only two (13%) had immediate recognition of injury. Also, cholangiography was not associated with decreased injury severity in this study. However, 11 of 16 patients (69%) had abnormal cholangiograms that were misinterpreted intraoperatively. If these 11 cholangiograms had been interpreted properly, the injuries would have been discovered immediately, and more severe excisional injuries could have been prevented (Figs. 1 and 2). Therefore, in 13 cases (81%), the surgeon could have identified the injury at the time of LC. In addition, attempts at cholangiogram were aborted in two cases. Conversely, in 23 of 30 cases (77%) where cholangiography was not performed, surgeons failed to recognize injuries intraoperatively. The cholangiography errors were as follows: Nonvisualized hepatic duct Extravasation of uncertain origin Aborted/incomplete procedure 6 4 1

In the 37 patients with delayed recognition of bile duct injury, the average delay was 10 days (range, 331 days). The presenting symptoms were as follows: Elevated liver function tests Pain in abdomen/shoulder Nausea/vomiting Fever Ileus Leukocytosis 34 (92%) 32 (87%) 17 (46%) 9 (24%) 6 (16%) 5 (14%)

The average postoperative bilirubin in patients with bile duct injuries was 4.5 mg% (range, 119). In 27 of 37 patients (73%) with a diagnosis of bile duct injury that was made postoperatively, additional delays resulted from diagnostic and treatment failures. Endoscopic retrograde cholangiograms were unsuccessful or misinterpreted in six cases

312 Table 2. Treatment of injuries Type of repair Choledocho-/hepaticojejunostomy Primary end-to-end repair Closure with T-tube Drain only Remove clip on CBD Stent only Ligation of CBD (in error) N/A, not applicable. No. of cases 29 7 4 3 1 1 1 Success (%) 11 (44) 3 (43) 3 (75) N/A 1 (100) 1 (100) N/A

Fig. 2. Intraoperative cholangiogram showing extravasated contrast performed after a common hepatic duct injury.

(16%). Biloma was dismissed as a duct of Luschka leak or minor cystic duct leak in six cases. Symptoms were misdiagnosed as common duct stone or ulcer in five patients. Subhepatic fluid collection was dismissed as irrigant or blood in four cases. The patient initially refused testing in four instances, and shoulder or abdominal pain was not evaluated in two patients. The injuries were treated as shown in Table 2. Successful repair of the injury by the primary surgeon occurred in only six of 22 cases (27%). Repair by a more experienced surgeon at a tertiary center with special expertise in bile duct repair occurred in 19 of 24 cases (79%). All primary end-to-end repairs (seven cases) were done by the primary surgeons; the success rate was 43%. Three end-to-end repairs were performed during the initial surgery, and all of these strictured within 6 months. Three end-to-end repairs were performed in the postoperative period (days 2, 4, and 51) and have been successful to date. One end-to-end repair performed 59 days after injury strictured. The reason for the failure of repairs performed by primary surgeons was stricture or anastomotic leak in 14 of 16 cases (88%). Immediate repair was successful in only one of eight repairs, whereas repairs performed for injuries diagnosed postoperatively were successful in 21 of 25 cases (84%). These data probably reflect the type of repair and the experience of the operator rather than the timing of the repair. Complications other than the initial bile duct injury occurred in most patients. Eight of the nine patients (89%) who had immediate recognition and attempted repair of the duct injury had postoperative complications of either leak or stricture resulting in cholangitis. Twenty-three of the 37 patients (62%) who had delayed recognition of injury suf-

fered complications including 12 leaks/strictures, three abscesses, three small bowel obstructions, one wound infection, one dehiscence, one ventral hernia, one gastrointestinal bleed, one acute respiratory distress syndrome (ARDS), and one death. There was one death due to bile peritonitis and sepsis that resulted from a cystic duct leak. This 63-year-old patient had undergone reexploration with drainage on the 7th postoperative day. To date, litigation has been resolved in 30 cases. Twenty-one cases were settled out of court, with an average settlement of $221,000 (range, $30,0001,300,000). In five cases, the plaintiffs prevailed at trial, with an average award of $214,000 (range, $125,000240,000). In four cases, there was a defense jury verdict. In three of four defense verdicts, the injury was caused by clip impingement. The fourth defense verdict involved an excision of a segment of the common hepatic duct in an inflamed acute case without cholangiography. There was a delayed diagnosis, but repair at a tertiary center proved successful. Discussion The spectrum of iatrogenic bile duct injuries ranges from minimal to life-threatening. Since the cases reviewed represent only patients who were involved in litigation, they are not necessarily representative of all bile duct injuries nor the frequencies of those injuries. Nevertheless, all major injury types, ranging from clip impingements to excision of the entire extrahepatic biliary tree, are represented. Injuries diagnosed and treated at the time of injury as well as injuries identified and treated later by the treating surgeon and experts are represented. Some patterns seem to predispose to lawsuits. Most of these patients had failures of immediate repairs or complications following delays in diagnosis. From a medical-legal standpoint, 86% of cases in this study were resolved in favor of plaintiffs through settlements or verdicts. It has been previously suggested that the high rate of biliary injury associated with laparoscopic cholecystectomy is the result of the learning curve [2, 3]. However, other investigators have noted an ongoing problem well past the learning period [4, 7]. Nine injuries in this study occurred after the surgeons 50th case, and five of these were after the 100th case (including ductal transections and excisions). Clearly, no surgeon is immune from the risk of bile duct injury, and no case is simply routine. Technical errors were the primary cause of these injuries. Injuries resulted from misidentification of normal

313

anatomy in 48% of cases. Seventy-two percent of injuries occurred in nonacute cases. The majority of injuries resulted when the common bile duct was mistaken for the cystic duct due to inadequate dissection in the triangle of Calot, failure to identify the underside of the gallbladder, and possibly in part to excessive upward tension placed on the duct. Other technical errors, such as cautery injury and blind or close clipping to the common duct, occurred in 11%. We recommend using blunt dissection starting posterolaterally at the presumed cystic ductgallbladder junction, avoiding cautery, sharp dissection, and clipping until the triangle of Calot and the underbelly of the gallbladder are fully exposed both laterally and medially. All injury types (lacerations, excisions, burns, and clip impingements) were seen even when intraoperative cholangiography was done. Despite the aid of cholangiography, misinterpretation of radiographic evidence of bile duct injury occurred frequently. The most common cholangiographic abnormalities associated with bile duct injuries were nonvisualized intrahepatic ducts and contrast extravasation. A cholangiogram must show contrast in the right and left hepatic ducts and the duodenum. Extravasation of contrast can mean that an injury is present. Intraoperative review of cholangiograms by an experienced radiologist can be helpful. Although the surgeons retrospectively indicted short cystic ducts, acute inflammation, chronic scarring, and excessive bleeding, no obvious problems were encountered during the initial surgery in 54% of these cases. For this reason, use of selective cholangiography in cases only when there is a recognized problem, confusion of anatomy, or possible common duct stones will not affect the incidence or severity of bile duct injuries. Cholangiography can prevent the most devastating of these injuriesa misidentification of the common for the cystic duct, with excision of the extrahepatic bile ducts. Additionally, liberal use of cholangiography should increase the surgeons awareness of normal, abnormal, and incomplete studies. It can also improve the surgeons ability to perform cholangiography so that fewer studies will be aborted. The injuries that were not recognized during surgery presented with postoperative symptoms. The average delay in diagnosis was 10 dayseven though almost all patients were seen within 7 days of surgery. Delays resulted from misinterpretation of the typical signs and symptoms of bile leak or duct obstruction, including protracted complaints of shoulder and abdominal pain, nausea, vomiting, and elevation of liver function tests. The absence of fever, leukocytosis, and ileus did not preclude the presence of bile duct injury, especially duct obstruction. Noninvasive studies, starting with ultrasonography or HIDA scan, are highly sensitive. Endoscopic retrograde cholangiography is usually definitive. It allows for both precise diagnosis and treatment, even though several endoscopic retrograde cholangiograms (ERCP) failed or were misinterpreted in this review. Delays in diagnosis were associated with complications that increased damages in 62% of cases. Long-term morbidity of biliary injury is also related to the location and extent of injury. Higher ductal injuries (especially Bismuth 4) are associated with an increased risk of postreconstructive failure even after appropriate repair [2]. There is evidence that early recognition and repair of a

bile duct injury by an experienced surgeon leads to improved outcome [5]. In this series, 20% of repairs were performed immediately, but outcomes were poor anyway. Of nine immediate repairs, eight ultimately failed, requiring reoperation. However, in this selected group of cases, failure of the initial attempt at repair may have predisposed to litigation. It seems likely that patients who had successful immediate repairs were not included in this study because they were less likely to sue. Many surgeons attempted repairs that were unlikely to succeed due to their own lack of experience with bile duct injuries. Primary surgeons tried to avoid major duct reconstruction in the hope that repair over a T-tube would suffice and would be easier to explain than hepaticojejunostomies. Patients with a failed initial attempt at end-to-end repair underwent an average of two subsequent percutaneous balloon dilations and two reoperations prior to successful recovery. One of these patients still has ongoing problems. Several patients injuries were only worsened by inappropriate attempts at primary repair. For example, a 3-0 silk suture end-to-end repair under tension was performed. Surgeons with expertise in bile duct repairs had better results than primary surgeons (79% versus 27% success) in spite of the fact that they were dealing with more challenging patients who were often in poor condition. When a surgeon diagnoses a bile duct injury, consideration should be given to the resources and circumstances of the surgeon and the patient. In some situations, it is entirely appropriate to place a catheter in the injured duct, drain the area, and transfer the patient for definitive care to a tertiary center. If an injury is suspected postoperatively, thorough evaluation prior to reoperation is advised. Specific identification of the site of injury can usually be accomplished before reoperation. Stabilization of the injury with stents placed by ERCP or percutaneous transhepatic radiographic procedures and percutaneous drainage can be achieved in most cases. If bile duct reconstruction is required, appropriate referral can be made electively. Bile duct injuries will always be the worst complication of cholecystectomy. However, their incidence and severity can be reduced. Knowledge of how and why they occur, coupled with more liberal use of cholangiography, can help to prevent debilitating and life-threatening sequelae. References
1. Chapman WC, Halevy A, Blumgart L, Benjamin I (1995) Post cholecystectomy bile duct strictures. Arch Surg 130: 597604 2. McMahon AJ, Fullarton G, Baxter JN, ODwyer PJ (1995) Bile duct injury and bile leakage in laparoscopic cholecystectomy. Br J Surg 82: 307313 3. Meyers WC, Club TS (1991) A prospective analysis of 1518 laparoscopic cholecystectomies. N Engl J Med 324: 10731078 4. Morgenstern L, McGrath MF, Carroll BJ, Paz-Partlow M, Berci G (1995) Continuing hazards of the learning curve in laparoscopic cholecystectomy. Am Surg 61: 914918 5. Pitt HA, Miyamoto T, Parapatis SK, Tompkins RK, Longmire WP (1982) Factors influencing outcome in patients with postoperative biliary strictures. Am J Surg 144: 1421 6. Shea JA, Healey MJ, Berlin JA, Clarke JR, Malet PF, Staroscik RN, Schwartz JS, Williams SV (1996) Mortality and complications associated with laparoscopic cholecystectomy: a metaanalysis. Ann Surg 224: 609620 7. Strasberg SM, Hertl M, Soper NJ (1995) An analysis of the problem of biliary injury during laparoscopic cholecystectomy. JACS 180: 101 125

314

Discussion Dr. Rattner: Ive had the unfortunate experience of repairing about 30 bile duct injuries, and subsequently serving as an expert witness for our defense insurer. In reading your abstract, I was dismayed to find how many of the verdicts had gone in the plaintiffs favor. In our state, and in Connecticut, where Ken Kern did a very nice study, it follows a typical medical malpractice situation of a third for the plaintiff, a third settled, a third for the defendant. I think a factor that has led to plaintiffs verdicts, and I must say, theres only one case in Massachusetts thats been litigated and was won by the defendant surgeon, which is when the surgeon attempts to repair an injury. Its important to recognize that if youve got a problem and you dont know how to fix it or youre not experienced, youre best off transferring it to somebody who does. The first repair is critical, and if it succeeds, the risk of a lawsuit is much less than if you do a primary common duct repair and it fails. This may lead to six months in the hospital and reoperations. My question for you is whether the presence of videos helped or hurt the surgeons in their defense. Dr. Carroll: Four out of five videotapes that were available did not help the surgeon. Theres only one case of a duct of Luschka injury that has an absolutely normal videotape, and I think it will help the surgeon. In general, the videotapes do not help the surgeon to settle a case. In my series, 86% were settled or had plaintiff verdicts; 14% were defense verdicts. Most of defense verdicts were clip impingements, but there was one excision of the bile duct where two cuts were made in the bile duct and a piece taken out of the middle that was defended in North Dakota.

Dr. Soper: We, too, unfortunately, have had a large experience. The videotapes, similar to this morning, do cause you to lose your breakfast some of the time. I think some of the points youve made are key; the critical view of the dissection in the triangle of Calot, or the hepatocystic triangle; making sure that in the early postoperative period anything that deviates from the normal course should be suspected as a common duct injury. How did you arrive at the denominator of cases in the study that youve done? Are these just ones that were referred to you for management, or ones that you were asked to help with the litigation? Dr. Carroll: They were referred by defense lawyers and plaintiff lawyers and hospital peer reviews for an opinion as to whether the cases met the standard of practice throughout the United States. I think your point about patients who are off the bell-shaped curve in the postoperative period and need a thorough evaluation is crucial. Dr. Soper: Do you know what the ultimate denominator in the United States for this last year was, in terms of number of malpractice suits specifically in regard to bile duct injuriesdo we know that number? Dr. Carroll: No, but the important fact is that in the published literature there was just a meta-analysis of 98 papers on bile duct injuries since 1990. In 78,000 laparoscopic cholecystectomies, the incidence was approximately 0.3 to 0.4 percent. There is probably under-reporting of this injury and the injury rate is probably higher, probably in the range of one in 200 cases, in my opinion.

News and notices

Surg Endosc (1998) 12: 383385 Springer-Verlag New York Inc. 1998

New Address for the European Association for Endoscopic Surgery (E.A.E.S.)
Effective January 1, 1997, the 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

of endosurgical Education and Research Center. Active participation in clinical and basic science research projects is also encouraged. For further information, please contact: Debbie Moser Tel: 202-994-8425 or, send curriculum vitae to: Dr. Jonathan M. Sackier Director, Washington Institute of Surgical Endoscopy George Washington University Medical Center Department of Surgery 2150 Pennsylvania Avenue, N.W. 6B-417 Washington, DC 20037, USA

Volunteer Surgeons Needed Northwestern Nicaragua Laparoscopic Surgery Teaching Program, Leon, Nicaragua
Volunteer surgeons are needed to tutor laparoscopic cholecystectomy for this non-profit collaboration between the Nicaraguan Ministry of Health, the National Autonomous University of Nicaragua, and Medical Training Worldwide. The program consists of tutoring general surgeons who have already undergone a basic laparoscopic cholecystectomy course. Medical Training Worldwide will provide donated equipment and supplies when needed. For further information, please contact: Medical Training Worldwide Ramon Berguer, MD, Chairman Tel: 707-423-5192 Fax: 707-423-7578 e-mail: berguer.r@martinez.va.gov

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

Under the direction of Sir 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 purposebuilt 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 Sir Cuschieri, and the latest books on endoscopic surgery are on display in our Resource area. Course fee including lunch and course materials is $860. 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

Courses at the Washington Institute of Surgical Endoscopy

We are delighted to offer a variety of regular courses in laparoscopic techniques year round. In addition, special arrangements can be made for individual tuition in all minimally invasive disciplines. CME credit is available and course fees depend on the instruction offered. For further information, please contact: Debbie Moser Washington Institute of Surgical Endoscopy 2150 Pennsylvania Avenue, N.W. Suite 6-B Washington, DC 20037 USA Tel: 202-994-8425, or 1-888-8WISEDOC Fax: 202-994-0567

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

Each month Sir A. 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 Sir Cuschieri and his team. The course is endorsed by SAGES. Course fee including lunch and course materials is $1850.

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

384 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

Cours Europeen de Chirurgie Laparoscopique (European Course on Laparoscopic Surgery) under the auspices of E.A.E.S. University Hospital Saint-Pierre (U.L.B.) May 12May 15, 1998 November 17November 20, 1998 Brussels, Belgium
Course Director: G.B. Cadiere, MD Universite Libre de Bruxelles (U.L.B.) Department of G.I. Surgery University Hospital Saint-Pierre The course will include live demonstrations and interactive dialogue with the operating surgeons and a video forum with discussions of videotapes, technical details, and pitfalls. Topics include functional gastric surgery (Nissen-Toupet-gastroplasty), colon (colectomy-rectopexy), hernia (trans-/preperitoneal approach, balloon), retroperitoneoscopy, splenectomy, needle surgery, biliary surgery, and new technologies. Surgeons include J. Bruyns, G.B. Cadiere, J. Himpens, J. Leroy, and M. Vertruyen. The official language for the May course is French with simultaneous translation provided into English. The official language for the November course is English with no simultaneous translation. Internet site: http://www.LAP-SURGERY.com For further information, contact: Scientific Information Mrs. Solange Izizaw C.H.U. Saint-Pierre Service de Chirurgie Digestive Rue Haute 322 B-1000 Bruxelles, Belgium email: coelio@resulb.ulb.ac.be or Administrative Secretariat Conference Services S.A. Avenue de lObservatoire 3, bte 17 B-1180 Bruxelles, Belgium email: conference.services@skynet.be

The Practical Aspects of Laparoscopic Fundoplication Surgical Skills Unit University of Dundee Scotland, UK
A three-day course, led by Sir A. 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

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

Advanced Laparoscopic Suturing and Surgical Skills Courses MOET Institute San Francisco, CA, USA
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

6th World Congress of Endoscopic Surgery Roma 98 6th International Congress of European Association for Endoscopic Surgery May 31June 6, 1998 Rome, Italy
This World Congress is being organized with the cooperation of IFSES. 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 email: ega worldendo98@uni.net http://www.ega.it/worldendo98

385

Second International Congress of Thorax Surgery June 2427, 1998 Palazzo della Cultura e del Congressi, Bologna, Italy
For further information, please contact the Scientific Secretariat: Dott. A. Bini Department of Thorax Surgery Saint Orsola University Hospital Via Massarenti, 9 40138 Bologna, Italy Tel: (+3951) 344068-6363287 Fax: (+3951) 305092

be held September 2426, 1998, at the Sorento Palace Conference Center in Naples, Italy. For further information, please contact: Office of Continuing Medical Education George Washington University Medical Center 2300 K Street, NW Washington, DC 20037, USA Tel: (202) 994-4285 or Dottore Vincenzo Landolfi Second University Universita BDI Napoli Cattedra di Chirurgia dell Apparoato DiGenrente Primary PoliclinicoPiazza Miraglia 3 80138 Napoli, Italy Tel: 011-39-81-566-5279 Fax: 011-39-81-459-137 e-mail: comvplan@syren

International Course in Laparoscopic Colorectal Surgery September 24, 1998 Trondheim, Norway Course Directors: R. Mrvik, MD, R. Bergamaschi, MD Host Chairman: H.E. Myrvold, MD Guest Faculty: S.D. Wexner, MD
For further information, please contact: National Center for Advanced Laparoscopic Surgery Trondheim University and Regional Hospital 7006 Trondheim, Norway Tel: +47-73-999888 Fax: +47-73-999889

Third Congress of the International Federation for the Surgery of Obesity (IFSO) September 35, 1998 Holiday Inn Crowne Plaza, Bruges, Belgium (12th International Symposium on Obesity Surgery)
International papers and symposia covering the field of laparoscopic and open bariatric surgery. For further information, please contact: Third Congress of the IFSO Secretariat Post Office Box 80 B-8310 Bruges Sint Kruix, Belgium or J.W.M. Greve, M.D. Fax +31-43-387-5473 email: jgreve@shbe.azm.nl or George S.M. Cowan, Jr., M.D. Professor of Surgery University of Tennessee 956 Court Avenue Suite A 212 Memphis, TN 38163, USA Tel: (901) 448-6781 Fax: (901) 448-4688 email: bpitts@utmem1.utmem.edu

Tenth International Conference of the Society for Minimally Invasive Therapy September 35, 1998 London, England
Host Chairman: Mr. J. Wickham For further information, please contact: The Society for Minimally Invasive Therapy 2nd Floor, New Guys House Guys Hospital St. Thomas Street London, SE1 9RT, England Tel: +44 (0)171 955 4478 Fax: +44 (0)171 955 4477 email: j.wickham@umds.ac.uk

Current Trends in Colon and Rectal Surgery September 2426, 1998 Naples, Italy
Current Trends in Colon and Rectal Surgery, sponsored by the University of Naples, George Washington University, and the Cleveland Clinic, will

Surg Endosc (1998) 12: 331334

Springer-Verlag New York Inc. 1998

Laparoscopic liver surgery

A report on 28 patients
J. Marks,1 J. Mouiel,2 N. Katkhouda,2 J. Gugenheim,2 P. Fabiani2
1

Department of Surgery, Allegheny University of the Health Sciences, Hahnemann Division, Broad and Vine, M. S. 413, Philadelphia, PA 19102-1192, USA 2 Department of Surgery, Hopital Saint-Roch, University of Nice Sophia-Antipolis, B.P. 319-06006-Nice Cedex 1, France Received: 10 May 1996/Accepted: 26 July 1996

Abstract Background: An effort was made to evaluate the indications, safety, and therapeutic efficacy of laparoscopic liver surgery. Methods: Between 1989 and 1996, 28 patients, 23 to 88 years old were operated upon laparoscopically. Pathology consisted of simple cyst (ten), polycystic liver disease (seven), hydatid cyst (three, two of them calcified), abscess (one), focal nodular hyperplasia (six), and metastatic breast cancer (one). Results: Operations included 17 fenestrations, three pericystectomies, and eight resections (two lateral lobes). Operative time was 45 to 525 min with only four cases longer than 4 h. There was a 21% morbidity rate. There were no mortalities. Follow-up was 167 months with one asymptomatic recurrence. Conclusions: Laparoscopic hepatic surgery can be performed safely with good results by surgeons with hepatic and laparoscopic experience when careful selection criteria are followed. We advocate the four-hands technique for simultaneous dissection and control of bleeding and bile ducts during resections. Key words: Laparoscopic surgery Hepatic surgery Liver Four-hands technique

geons, however, are as comfortable with open hepatic surgery as they are with the gallbladder, hernia, appendix, or stomach. Moreover, the equipment required is expensive and probably best purchased jointly with another department, perhaps urology or gynecology, to defray the cost. Furthermore, only a limited number of lesions, depending upon their location and etiology, can be approached in this manner. However, having satisfied these criteria, the laparoscope can certainly be applied to the liver. We report here our experience with laparoscopic liver surgery in a select group of 28 patients from an active practice of hepatic surgery. The four-hands technique for hepatic resection is described. Methods
Between 1989 and 1995, 28 patients, 20 of them women, underwent laparoscopic hepatic surgery at Saint Roch Hospital, University of Nice Sophia-Antipolis. The ages of the patients ranged from 23 to 88 years (mean of 54 years). There were lesions in the right lobe in four cases, the left lobe in 17 cases, and bilateral in the other seven cases. The preoperative diagnosis included symptomatic solitary hepatic cyst (ten), symptomatic polycystic liver disease (PCLD) (seven), hydatid cyst (two) (one calcified), focal nodular hyperplasia (FNH) (three), adenoma (three), abscess (one), metastatic breast cancer (one), and calcified gallbaldder (one). The size of the lesions ranged from 2.5 to 22 cm with a mean of 8.4 cm. Presenting complaints included: pain (14), compressive symptoms (seven), abnormal x-rays (five), and sepsis (two). Preoperative radiographic evaluation varied considerably depending upon the diagnosis and proposed procedure. Evaluation included CAT scan (18), ultrasound (15), MRI (four), arteriography (three), and/or EGD (5). Nine patients had three or more studies.

Following the introduction of laparoscopic cholecystectomy in 1987, the laparoscopic approach has been applied to the full spectrum of abdominal procedures [68, 11, 15, 16]. Despite this proliferation, the liver has been virtually ignored and, perhaps, with good reason. To safely perform liver surgery laparoscopically, the surgeon must be both an accomplished laparoscopist and hepatic surgeon. Few surCorrespondence to: J. Marks

Technique
The procedures were performed with the patient under general anesthesia with oro- or nasogastric decompression and a pneumoperitoneum of 1214 mmHg. The patients were in the French position, a modified lithotomy with minimal flexion of the hips, and the primary surgeon positioned between the legs. The first assistant or second surgeon was on the patients left side. The scrub nurse was between them. For fenestrations, we used a four-trocar configuration. A 10-mm port at the umbilicus housed the 0 laparoscope. A 5-mm trocar was placed just below the xiphoid process to

332 Table 1. Proceduresa Fenestrations Pericystectomy Cyst sterilization and partial cystectomy Wedge resection Left lateral lobectomy
a b

16/17b 2/2 1/1 5/6 1/2

Numerator represents successful laparoscopically completed cases. Two patients required minilaparotomy for control of bleeding.

Table 2. Pathology Solitary hepatic cyst Polycystic liver disease Hydatid cyst Active Calcified FNH Abscess Metastatic breast cancer 10 7 3 1 2 6 1 1

Fig. 1. Trochar position: four-hands technique.

the right or the left of the falciform ligament, depending on the location of the cyst. This port was used to expose the liver, often using an irrigationaspiration probe. Two other 5-mm or 10-mm ports, in the right and left flank, allowed the surgeon to puncture the cyst dome, aspirate its contents, and excise the cyst wall in a careful sequential fashion to facilitate hemostasis. For more extensive procedures, a strong light source (300-W xenon) and high-quality 0 and 30 scopes are required. To perform resections safely with a minimum of wasted motion, we advocate the four-hands technique. This uses four to six trocars (Fig. 1) and allows for the primary surgeon to expose and dissect the liver while surgeon 2 obtains control and transects the blood vessels and bile ducts. The procedure entails the same components as in open hepatic surgery. First, the patient is explored, both visually and ultrasonographically. Mobilization of the liver and hilar dissection are performed as necessary to obtain vascular control. Division and ligation of the round ligament followed by freeing of the falciform and the right or left triangular ligaments allow access to perform thorough exploration, resection, and hemostasis. Should bleeding become a problem, this can be controlled by directly clamping the liver or performing a Pringle maneuver. Dissection is begun by scoring Glissons capsule with the electrocautery or Nd:Yag laser. Parenchymal dissection is performed using the ultrasonic dissector (Tetrad Corporation) or the laparoscopic equivalent of the finger fracture technique, kellyclasia: using a dissecting forcep to grasp the tissue and gently compress it, leaving only the bile ducts and vessels to be ligated with clips or ties and transected. Following resection, the mass is placed in an impermeable specimen bag for removal. Cholangiography is very useful to detect possible bile leaks. The raw surface of the liver is then inspected, coagulated by Nd:Yag laser, and covered with fibrin glue (Tissucol, Immuno, Vienna, Austria). The specimen is extracted either by partial morceillation, dilatation at the umbilicus, enlarging another port site, or by a small McBurney or subcostal incision.

and two procedures with patients undergoing extensive fenestrations for polycystic liver disease. Three of the cases were converted to open procedures (11%). Two of these were done immediately following exploration: one because the mass was too close to the inferior vena cava on ultrasound and the other because the giant cyst stretched too far posteriorly to allow for safe vascular control. In the third patient with FNH, the dissection was well underway when it was determined that the mass extended too close to the vascular pedicle to use the Ultrasonic CUSA (Pfizer Howmedica, Boulder, CO, USA) safely. Intraoperative complications consisted solely of two cases of hemorrhage in patients with PCLD requiring a 5-cm minilaparotomy to oversew the bleeding point. These two patients and one woman with FNH, who autodonated her blood before a wedge resection, were each transfused two units and represent the only patients receiving transfusions (11%). Morbidity consisted of a urinary tract infection, phlebitis, and in the PCLD group, two cases of ascites (one treated immediately and one requiring drainage), one intraabdominal infection treated with p.o. antibiotics, and a delayed abscess requiring a reoperation 5 months later in a patient whose cyst had been sclerosed preoperatively with alcohol. There was no mortality. Hospital stay averaged 7.7 days with a range of 144 days. Follow-up ranged from 1 to 67 months. There was one recurrence (4%) found on routine ultrasound 5 months postoperatively in an asymptomatic patient with PCLD. Discussion The indications and preparation for laparoscopic liver surgery remain the same as in open hepatic surgery. Unchanged, too, are the component parts of the procedure: exploration, mobilization and vascular control, parenchymnal dissection, hemostasis, and specimen extraction. Visualization is excellent with the laparoscope, but the addition of laparoscopic ultrasound has been shown to alter intraoperative plans in up to 66% of cases when compared to laparoscopic exploration alone [2]. While not necessary for all fenestrations, we strongly advocate its use when con-

Results The procedures performed with the associated pathology are shown in Tables 1, 2, and 3. Ten patients also underwent cholecystectomy with intraoperative cholangiography. Mean operative time was 179 min (range 45525 min). Only four operations lasted longer than 4 h: the lateral lobectomy, the sterilization and evacuation of a hydatid cyst,

333 Table 3. Procedures performed for associated pathology Symptomatic hepatic cyst 10 Polycyctic liver disease 6/7a 1 2 1 10 7 1 1 2 3/4a 1/2a 6 1 1 Hydatid disease Metastatic breast CA

Max Fenestration Sterilization, evacuation, and partial cystectomy Pericystectomy Wedge resection Left lateral lobectomy Total
a

Abscess

Calcified

FNH

Total 16/17a 1 2 5/6a 1/2a 28

Numerator completed laparoscopically.

templating more extensive procedures. Adequate mobilization and hilar dissection are essential to minimize the risks of bleeding and air embolus. We introduced here the concept of the four-hands technique. This is a major improvement, and we strongly advocate this for laparoscopic resections. This surgery is technically challenging and time consuming. By having surgeon 1 dissect while surgeon 2 keeps the visual field clear with the irrigation and aspiration, the procedure can proceed in a step-wise and orderly fashion. This allows for surgeon 2 to clamp or clip vessels prior to surgeon 1 cutting them, a very important point in laparoscopic hepatic surgery. Pneumoperitoneum coupled with the partial transection of a large uncontrolled hepatic intraparenchymal vessel could not only result in hemorrhage but also in a potentially large CO2 embolus. While we have not experienced this complication, we believe it is due to a continual recognition of this as a potential disaster. In fact, two of our three conversions were due to the proximity of major vascular structures. No problems occurred with bilomas postoperatively. The high-quality optics allow for easier visualization of uncontrolled bile ducts. In addition, one must have a willingness to proceed with a cholangiogram to clarify the anatomy or identify leaks if questions exist. By combining careful dissection and vascular control with the use of laser cautery and Tissucol to the cut surface of the liver, problems with postoperative bleeding or hematomas were entirely avoided. We attribute our low incidence of intraoperative complications and our absence of mortality not only to great technical care but to a low threshold to convert to an open procedure. As in all laparoscopic procedures, this does not represent a failure but rather good surgical judgment. Our experience, as well as that of other authors [5, 9, 10, 14], indicates that laparoscopic hepatic surgery, while technically difficult, can be performed safely with good results with careful patient selection. Attention to the etiology of the lesion and its location is essential. Ideal candidates have a large solitary cyst or a symptomatic benign mass located superficially, laterally, or far enough from the pedicle to allow direct clamping of the liver or access to the hilum to perform a Pringle maneuver should bleeding occur. This explains the preponderance of left lobe lesions in our series. Contraindications to this technique include patients with cirrhosis, hepatocellular carcinoma, or posterior or centrally located lesions. While we have utilized this approach for solitary small metastatic disease, hydatid disease, hepatic abscess, and PCLD, these should be viewed with a great deal of circumspection. Problems exist to varying degrees

should any of these lesions be spilled. Port site recurrences have not yet been reported for hepatic cancers, but this remains a concern when using laparoscopy in any patient with cancer [17]. This is of special concern when considering pairing this approach with cryoablation [4]. With echinococcal cysts, the risk of spillage is also obvious, though less problematic with calcified cysts. Techniques and instruments for minimizing this problem are described [1, 3, 12, 13, 15], but one must be highly selective with this surgery. If one does use a laparoscopic approach for hydatid disease, we recommend a cholangiogram to rule out a connection with the biliary system. While fenestration of polycystic liver disease has been described by others both by open and laparoscopic approaches, we have found this disease particularly bothersome. Transcystic fenestration of deeper cysts makes the control of bleeding very difficult. Suddenly one is working very centrally. Indeed, our only two operative complications were venous bleeding from central cysts in patients with polycystic liver disease requiring a 5-cm minilaparotomy for control. Furthermore, twothirds of our morbidities occurred in this group. Many obstacles exist to laparoscopic hepatic surgery. The required equipment is expensive. The surgery is often difficult. However, laparoscopic liver surgery is a viable option when careful patient selection is performed. In the hands of a skilled hepatic and laparoscopic surgeon who is equipped with the proper tools, these procedures are safe, effective, and reproducible. References
1. Alper A, Emre A, Hazar H, Ozden I, Bilge O, Acarli K, Ariogul O (1995) Laparoscopic surgery of hepatic hydatid disease: initial results and early follow-up of 16 patients. World J Surg 19(5): 725728 2. Barbot D, Marks J, Feld R, Liu J, Rosato F (1995) Improved staging of liver tumors using laparoscopic intraoperative ultrasound. J Surg Oncol (accepted for publication) 3. Cappuccino H, Campanile F, Knecht J (1994) Laparoscopy-guided drainage of hepatic abscess. Surg Laparosc Endosc 4(3): 234237 4. Cuschieri A, Crosthwaite G, Shimi S, Pietrabissa A, Joypaul V, Tair I, Naziri W (1995) Hepatic cryotherapy for liver tumors: development and clinical evaluation of a high-efficiency insulated multineedle probe system for open and laparoscopic use. Surg Endosc 9(5): 483 489 5. Ferzli G, David A, Kiel T (1995) Laparoscopic resection of a large hepatic tumor. Surg Endosc 9(6): 733735 6. Fitzgibbons RJ Jr, Camps J, Nguyen N, Litke BS, Annibali R, Salerno GM (1995) Laparoscopic inguinal herniorrhaphy: results of a multicenter trial. Ann Surg 21(1): 313 7. Frazee RC, Roberts JW, Symmonds RE, et al (1994) A prospective randomized trial comparing open versus laparoscopic appendectomy. Ann Surg 219:725731

334 8. Katkhouda N, Mouiel J (1991) A new technique of surgical treatment of chronic duodenal ulcer without laparoscopy by videocoelioscopy. Am J Surg 161: 361364 9. Libutti SK, Starker PM (1994) Laparoscopic resection of a nonparasitic liver cyst. Surg Endosc 8(9): 11051107 10. Morino M, DeGiuli M, Festa V, Garrone C (1994) Laparoscopic management of symptomatic nonparasitic cysts of the liver: Indications and results. Ann Surg 219(2): 157164 11. Phillips EH, Franklin M, Carroll BJ, Fallas MJ, Ramos R, Rosenthal D (1992) Laparoscopic colectomy. Ann Surg 216: 703707 12. Robles PJ, Lara JG, Lancaster B (1994) Drainage of hepatic amebic abscess successfully treated by laparoscopy. J Laparoendosc Surg 4(6): 451454 13. Rogiers X, Bloechle C, Broelsch CE (1995) Safe decompression of hepatic hydatid cyst with a laparoscopic surgiport. Br J Surg 82(8): 1111 Schwartz DS, Gwertzman G, Kaleya RN, Gliedman ML (1994) Laparoscopic unroofing of multiple benign liver cysts with intraperitoneal drainage: a case report. J Laparoendosc Surg 4(2): 157160 Soper NHJ, Stockman PT, Dunnegan DL, Ashely SW (1992) Laparoscopic cholecystectomythe new gold standard. Arch Surg 127: 917923 Weerts JM, Dallemagne B, Hamoir E, Demarche M, Markiewicz S, Jehaes C, Lombard R, Demoulin JL, Etienne M, Ferron PE, Fontaine F, Gillard V, Delforge M (1993) Laparoscopic Nissen fundoplication: detailed analysis of 132 patients. Surg Laparosc Endosc 3: 359364 Wexner SD, Cohen SM (1995) Port site metastases after laparoscopic colorectal surgery for cure of malignancy. Br J Surg 82: 295298

14. 15. 16.

17.

Surg Endosc (1998) 12: 315321

Springer-Verlag New York Inc. 1998

Bile duct injury after laparoscopic cholecystectomy

The United States experience
B. V. MacFadyen, Jr.,1 R. Vecchio,2 A. E. Ricardo,1 C. R. Mathis1
1 2

Department of Surgery, The University of Texas Medical School, 6431 Fannin, #4292, Houston, Texas 77030, USA Department of Surgery, University of Catania, Via Carnazza n. 2, 95030 Tremestieri Etneo, Italy

Received: 24 September 1996/Accepted: 28 July 1997

Abstract Background: Forty series reporting experience with laparoscopic cholecystectomy in the United States from 1989 to 1995 were reviewed. A total of 114,005 cases were analyzed and 561 major bile duct injuries (0.50%) and 401 bile leaks from the cystic duct or liver bed (0.38%) were recorded. Intraoperative cholangiography (IOC) was attempted in 41.5% of the laparoscopic cholecystectomies and was successful in 82.7%. In major bile duct injuries, the common bile duct/common hepatic duct were the most frequently injured (61.1%) and only 1.4% of the patients had complete transection. Methods: When reported, most of the bile duct injuries were managed surgically with a biliary-enteric anastomosis (41.8%) or via laparotomy and t-tube or stent placement (27.5%). The long-term success rate could not be determined because of the small number of series reporting this information. The management for bile leaks usually consisted of a drainage procedure (55.3%) performed endoscopically, percutaneously, or operatively. Results: The morbidity for laparoscopic cholecystectomy, excluding bile duct injuries or leaks, was 5.4% and the overall mortality was 0.06%. It was also noted that the conversion rate to an open procedure was 2.16%. Conclusions: It is concluded based on this review of laparoscopic cholecystectomies that the morbidity and mortality rates are similar to open surgery. In addition, the rate of bile duct injuries and leaks is higher than in open cholecystectomy. Furthermore, bile duct injuries can be minimized by lateral retraction of the gallbladder neck and careful dissection of Calots triangle, the cystic ductgallbladder junction, and the cystic ductcommon bile duct junction. Key words: Bile duct injury Bile leak Laparoscopic cholecystectomy

Approximately 500,000600,000 cholecystectomies are performed each year in the United States [48], accounting for almost 25% of a general surgery practice. Since the clinical introduction of laparoscopic cholecystectomy (LC) in 1989, it has become the technique of choice for gallbladder removal. Initially, the utilization of this technique was pushed by the media, by patient demand, by industry, and by physician competition. Rapid adoption into clinical surgical practice, however, preceded studies on its effectiveness and safety. Early series of LC reported an incidence of major complications ranging from 1.0 to 8.0% [6, 38, 41, 52, 66, 68]. These problems included hemorrhage, wound infection, trocar and Veress needle injuries, bile duct injury and leakage, and organ system failure. Among the major complications of LC, bile duct injuries are the most difficult challenge because they can result in considerable patient disability. In early published series, the rate of laparoscopic-related bile duct injuries and leaks has been reported to range from 0.0 to 2.0% [41, 49, 52, 64, 68, 74], which is higher than the 0.1% to 0.25% for open cholecystectomy [2, 27, 44]. The reported number of iatrogenic bile duct injuries in various series has often excluded the denominator, which is the total number of laparoscopic cholecystectomies performed. This study was undertaken to determine the actual incidence of bile duct injuries and bile leaks.

Methodology
A Medline review of large series of laparoscopic cholecystectomies from the U.S. literature from January 1991 to December 1995 was undertaken. Within this evaluation, five series came from preoperative studies and 11 papers were reported from multi-institutional reviews. There were a total of 114,005 cases and the following data were assessed: the incidence of bile duct injuries and bile leaks, the type and location of these lesions, the time of recognition, the results of early and delayed bile duct repair procedures, their therapeutic management and results, and possible etiologic factors responsible for these problems. In addition, the data were evaluated on the number of conversions to an open procedure, the utilization of intraoperative cholangiography (IOC), morbidity, and mortality.

Correspondence to: B. V. MacFadyen, Jr.

316 Table 1. Laparoscopic cholecystectomy series Reference no. 68 66 38a 6 8 23 53 36 45 5 33 72 13 30a 61 60 18 25a 65 67 29a 62 16 22 21 19a 12b 31 42a 11b 20a 49a 71 32 47b 1a 43a,b 4 59 70 Total
a b a,b

Table 2. Incidence of bile duct injury and bile leak after laparoscopic cholecystectomy Male/female 0.34 NS NS NS NS NS 0.26 0.19 0.28 NS 0.56 NS 0.23 0.22 0.29 0.29 0.34 0.33 0.28 NS 0.64 0.31 0.37 0.29 NS NS 0.24 NS 0.21 NS NS 0.34 0.49 NS NS NS 0.20 0.21 0.21 NS Bile duct injury 561/112,532 cases (0.50%) Bile leak 401/105,438 cases (0.38%)

No. patients 1,518 500 1,983 800 418 350 283 300 82 375 50 381 90 304 152 60 622 261 618 280 300 271 250 111 100 9,597 514 100 762 650 77,604 4,640 1,107 1,000 2,427 2,671 950 201 500 823 114,005

Mean ageyears (range) 47 (898) 46 (1786) (1398) NSc NS NS 46.1 (1285) 49 (2185) NS 51 (2072) 45 15 48.6 (2077) 50 (1483) 43.2 (1783) 31 (2280) 48 (1797) 49 (1794) 47 (1582) NS 47 (1880) 50.8 (1591) 49 (1597) 43.3 (2472) 43.6 (1680) NS 47.2 (1593) 46 (1584) 49.9 (1496) NS NS NS 56 (1094) NS NS NS 48.8 (1397) 44.2 45.5 (1787) NS 47.0

Multicenter. Prospective study. c NS: not specified.

The biliary complications were categorized by differentiating bile duct injuries from bile leaks. Bile duct injuries (BDI) were further subdivided into direct laceration (partial or complete transection of the bile duct), and their location as to whether they involved the common bile duct (CBD), the common hepatic duct, the right hepatic duct, or one of the intrahepatic branches of the liver. Cystic duct avulsion and necrosis were included because they also produced an associated common bile duct injury. Bile leaks were classified as originating from the cystic duct, the accessory duct between the gallbladder and liver (Duct of Luschka), the gallbladder fossa, or from liver biopsy sites.

Results Table 1 records the series of 114,005 laparoscopic cholecystectomies (LC) that were analyzed. The indications for LC included acute and chronic cholecystitis and cholelithiasis, acalculous cholecystitis, motility disorders of the gallbladder, and gallstone pancreatitis. The most common preoperative diagnostic procedure prior to LC was ultrasonography; however, endoscopic retrograde cholangiopan-

creatography with endoscopic sphincterotomy (ERCP-ES) and/or percutaneous transhepatic cholangiography (PTC) was obtained when choledocholithiasis was suspected. In only 20,262 patients was the underlying pathology stated, and it was only possible in this group to determine the incidence of acute cholecystitis (2,346 patients11.6%) [1, 5, 6, 1113, 18, 2123, 25, 2931, 33, 36, 38, 43, 45, 49, 53, 5962, 65, 66, 68, 71]. The rate of conversion to an open cholecystectomy occurred in 2,201 of 102,085 patients (2.16%) [1, 46, 8, 1113, 16, 18, 2023, 25, 2931, 36, 38, 42, 43, 45, 47, 49, 53, 5962, 6568, 71, 72] and acute inflammation accounted for 44.7% of these conversions followed by fibrosis, adhesions, unclear anatomy, common bile duct exploration, and bile duct injuries. A total of 561 bile duct injuries were reported in 112,532 cases (0.5%) [1, 46, 8, 12, 13, 16, 1823, 25, 2933, 36, 38, 42, 43, 45, 47, 49, 53, 5962, 6568, 71, 72] (Table 2). It should be noted that if a study did not mention bile duct injury, it was not assumed to be zero and these cases were excluded from the calculation. Table 3 records the site of bile duct injury in the 561 patients [1, 46, 8, 12, 13, 16, 1823, 25, 2933, 36, 38, 42, 43, 45, 47, 49, 53, 5962, 6568, 71, 72]. The common bile duct and the common hepatic duct were injured in 343 of the 561 cases (61.1%), followed by cystic duct injury, which occurred in 101 of 561 cases (18%) and an aberrant duct in 48 of the 561 cases (8.6%). Complete transection of the common bile occurred in eight of 561 cases (1.4%) and injury to the left or right hepatic ducts was described in 10 cases (1.8%). This is contradictory to previous reports of bile duct injury. However, in most series reviewed in this paper, although the location of the injury was documented, the type of CBD injury was not specified in the majority of cases, so there may be more complete transections than are noted in this report. However, when the type of injury was recorded, it was most commonly a small hole or laceration, cystic duct avulsion from the CBD, or improperly applied metal clips. The corrective therapy reported in 561 bile duct injuries is listed in Table 4 [1, 46, 8, 12, 13, 16, 1823, 25, 2933, 36, 38, 42, 43, 45, 47, 49, 53, 5962, 6568, 71, 72]. The specific management of these injuries was reported in only 182 of those cases (32.4%), and the remaining cases were not sufficiently specified to analyze (379). Of the 182 cases specifying treatment of a bile duct injury, the vast majority were treated surgically (97.3%) and most of those were either a biliary-enteric anastomosis (76), choledochoduodenostomy (eight), or end-to-end anastomosis (eight). The remaining surgical repairs consisted of T-tube or stent placement (50), primary suture repair without drainage (28), or simply removal of a misapplied clip (seven). In only five cases was nonsurgical therapy reported which consisted of either endoscopic or percutaneous stent placement. In only 66 of the 561 (11.8%) cases was the time of

317 Table 3. Location of bile duct injury (N 561) Location CBD/CHD Hole, laceration, misapplied clips, cannulation of CBD, other not specified Cystic duct Hole, laceration, avulsion Abberant duct RHD/LHD Hole, laceration, misapplied clips, other not specified Transection of CBD Ampullary disruption Subtotal Site not specified Number (%) 343 (61.1) 101 (18.0) 48 (8.6) 10 (1.8) 8 (1.4) 1 (0.2) 511 50 (8.9) Table 5. Immediate vs. delayed repair of bile duct injuries Repair done Primary repair with or without drainage Biliary-enteric anastamosis Choledochoduodenostomy Clip removal Not specified Total Immediate 26 8 2 1 7 44 Delayed 2 16 1 1 2 22

Abberent duct, anomalous duct; CBD, common bile duct; CHD, common hepatic duct; LHD, left hepatic duct; RHD, right hepatic duct.

Table 4. Repair of 561 bile duct injuries Procedure done Biliary-enteric anastamosis (Hepatico- or choledochojejunostomy) Laparotomy with T-tube or stent Laparotomy with primary suture repair Choledochoduodenostomy End-to-end anastamosis Clip removal Endoscopic or percutaneous stent placement Subtotal Not specified Total Number 76 50 28 8 8 7 5 182 (32.4%) 379 561

recognition of injury reported [1, 5, 8, 12, 18, 29, 30, 32, 33, 38, 43, 47, 49, 53, 5961, 65, 68]. Of these cases, 44 (66.6%) were recognized at the time of LC and the remainder were diagnosed postoperatively. The difference in the approach to management of bile duct injuries discovered immediately or delayed is evident in Table 5 [1, 5, 8, 12, 18, 29, 30, 32, 33, 38, 43, 47, 49, 53, 5961, 65, 68]. It is interesting to note that the majority of cases diagnosed immediately underwent primary repair of the injury with or without drainage (59%), whereas those whose injury was diagnosed postoperatively usually had a biliary-enteric anastomosis (72.7%). It is difficult to analyze a success rate for any of these treatments for bile duct injury since the majority of reports did not specify immediate, short-term, or long-term success from the therapy. However, some trends can be noted. Of the 182 cases that specified treatment of a bile duct injury, only nine early failures were reported (4.9%). All nine reported failures had undergone surgical management of their injury, although it should be noted that 97.3% of the total cases had undergone a surgical procedure. However, the method by which the failures were treated varied. Twothirds of the failed cases did not undergo repeat operation and were successfully managed endoscopically or percutaneously. One-third of the failures did require another laparotomy with a successful immediate outcome. The cause of the failure in seven of the nine cases was most often diagnosed as postoperative stricture formation (71.4%) and secondarily as bile leakage (28.6%) [1, 6, 29, 32, 47, 49, 65].

The compilation of series in this review reported 401 bile leaks in 105,438 laparoscopic cholecystectomies (0.38%) [1, 46, 1113, 16, 18, 2023, 25, 2931, 33, 36, 38, 42, 43, 53, 5962, 6568, 7072] (Table 2). Again, it should be noted that if no mention was made of a bile leak, it was not assumed to be zero and these cases were excluded from the calculation. Of the 401 bile leaks recorded, a specific location was identified in 185 of those cases (Table 6) and it was most commonly identified as being from the cystic duct stump followed by a leak from the liver bed and then from an accessory duct of Luschka [1, 46, 1113, 16, 18, 2023, 25, 2931, 33, 36, 38, 42, 43, 53, 5962, 6568, 7072]. It should be noted that 20 leaks were reported from the common bile duct or common hepatic duct. The actual pathology in these cases is unknown to these authors and could actually represent bile duct injury. However, we are reporting what was described by the original authors of those series. The treatment of the 401 bile leaks was similarly evaluated (Table 7) [1, 46, 1113, 16, 18, 2023, 25, 2931, 33, 36, 38, 42, 43, 53, 5962, 6568, 7072]. In only 103 cases (25.6%) was there sufficient data given regarding management. Of these 103 patients, the most common procedure was drainage of the leak either by percutaneous, or surgical, or endoscopic methods (74.8%). The remaining 26 patients in this group underwent ligation of the cystic duct stump (11), ligation of an accessory duct (three), or expectant management with spontaneous resolution (12). Once again, since the majority of cases were not sufficiently detailed, it is difficult to analyze this data in terms of success rates. Of the 103 cases reporting therapy for bile leaks, there were only three failures indicated (2.9%) [11, 31, 70]. Potential predisposing factors for iatrogenic bile duct injuries and bile leaks are noted in Table 8. Anatomical variations, such as a short or absent cystic duct, a cystic duct arising from the right hepatic duct, and an aberrant or accessory right hepatic duct along with acute or chronic inflammation and tenting of the common bile duct are probably the most common causes of injury. Additionally, attempts to control hemorrhage or inaccurate placement of metal clips on the cystic duct are other potential predisposing risk factors. Within this large series of 114,005 cases, IOC was attempted in 41.5% (4,637 of 11,175) with a success rate of 82.7% (6097%) [46, 12, 16, 18, 3032, 36, 38, 45, 61, 65, 66, 68, 71]. Additionally, several studies reported performing IOC; however, it is uncertain whether these were successes or failures [1, 2123, 42, 49, 59, 62]. The use of laser or electrocautery dissection was evaluated in 12,686 cases [5, 1113, 16, 22, 25, 29, 30, 36, 38, 49,

318 Table 6. Location of bile leak (N 401) Location Cystic duct stump Liver bed/gallbladder fossa Accessory duct of Luschka CBD/CHD Liver biopsy site SUBTOTAL Site not specified No. (%) 102 (25.4) 38 (9.5) 24 (6.0) 20 (5.0) 1 (0.2) 185 216 (53.9)

Table 7. Repair of 401 bile leaks Procedure Drainage Percutaneous41 Laparotomy10 Laparoscopy3 Drain placed at LC-4 Cystic duct ligation Accessory duct ligation ERCP with stent placement ERCP with nasobiliary tube placement ERCP with sphincterotomy Expectant management Subtotal Laparotomymanagement not specified Laparoscopymanagement not specified ERCPtherapy not specified Therapy not specified Total No. 57

tention (0.23%), cardiac disease (0.17%), retained CBD stones (0.15%), persistent abdominal pain (0.14%), and small- or large-bowel injury (0.1%). Therefore, it is concluded that the complication rate is similar to open cholecystectomy [44]. In this series, 108,837 patients were reviewed in which mortality was specifically reported [1, 46, 1113, 16, 18 23, 25, 2931, 33, 36, 38, 42, 43, 45, 49, 53, 6062, 65, 67, 68, 71, 72]. A total of 70 deaths occurred in this group (0.06%) of which 32 (0.03%) were recorded as being directly related to LC. The most commonly reported causes of death overall were heart disease, cerebrovascular accidents, and pulmonary disease. In those patients whose death was related to LC, the most common causes were bile duct injury (20) or sepsis due to a missed injury to the small bowel or colon (seven). Discussion

11 3 10 5 5 12 103 (25.6%) 150 18 17 113 401

Table 8. Potential predisposing factors in bile duct injuries and bile leaks Anatomical variations Absent or short cystic duct Cystic duct arising from RHD Aberrant or accessory RHD Acute or chronic inflammation Attempt to control hemorrhage Improper technique Tenting of CBD Improper placement of metal clips on the cystic duct Hole made in cystic or common ducts during dissection

5961, 66, 68, 72]. Electrocautery was used in 10,112 dissections (79.7%) and a potassium titanyl phosphate (KTP) or neodymium:yttrium aluminum-garnet (Nd:YAG) laser was used in 2,574 patients (20.3%). It was not possible to differentiate whether the laser or electrocautery was responsible for bile duct injury or leak in these reported series. In reviewing this series of laparoscopic cholecystectomies, 26,450 cases were analyzed whose reports specifically recorded morbidity data [1, 5, 6, 12, 13, 16, 18, 19, 2123, 25, 2931, 33, 36, 38, 42, 43, 45, 53, 5962, 6568, 71, 72]. Morbidity, excluding bile duct injuries and leaks, occurred in 1,430 cases for an overall complication rate of 5.4%. The most commonly reported problems were shoulder pain in 226 patients (0.8%), persistent nausea and vomiting in 177 patients (0.67%), bleeding in 123 cases (0.47%), wound infection in 118 patients (0.45%), prolonged ileus in 100 cases (0.38%), fever of unknown origin in 90 patients (0.34%), and pulmonary complications in 73 patients (0.28%). Less frequent complications included urinary re-

Although bile duct injuries can occur during various surgical procedures such as gastrectomy, hepatectomy, or portocaval shunt, 80% of the injuries develop during biliary tract surgery, especially cholecystectomy [9]. Early papers on laparoscopic cholecystectomy [41, 49, 52, 64, 68, 74] reported a high incidence of bile duct lesions ranging from 0.0 to 2.0%. In this review the incidence of bile duct injury is higher than in open cholecystectomy (0.10.25%) [2, 27, 44] but it should be noted that this is a 6-year review including reports during the early learning phase of this procedure. From the data we have reviewed, it cannot be determined whether bile duct injury and bile leaks in large series overlap with earlier reports by the same author. In addition, later series often include data from 1989 to 1995, thus potentially including higher and lower rates of injury. This problem can more likely be better evaluated by a prospective study in which the years studied are completely separated. In this report, the occurrences of bile duct injuries (0.50%) and bile leaks (0.38%) were separated thus making a distinction between these two problems. It appears that the early reported high incidence of bile duct injury could be due to the inclusion of bile leaks together with bile duct injuries. The high incidence of bile duct lesions in the early reported series could also be related to the learning curve in laparoscopic cholecystectomy. Data collected by Woods et al. [73] suggest this hypothesis since they noticed that with increasing experience there was a decreased number of injuries. In the series reported by Andren-Sandberg et al. [2] in open cholecystectomy, 85% of the injuries were caused by a surgeon who had performed less than 100 cholecystectomies. In laparoscopic cholecystectomy, the Southern Surgeons Club [68] has shown a 2.2% incidence of bile duct injuries during the first 13 cases operated on by each surgical group and a decrease to 0.1% in subsequent patients. In the series reported by Deziel et al. [20], a significant difference (p < 0.001) in laparoscopic-related bile duct injuries was reported when comparing the institutions that had performed more or less than 100 operations. Therefore, it may be concluded that the learning curve may be longer than the previously considered 1020 laparoscopic cholecystectomies. Besides the experience of the surgeon, anatomical varia-

319

tions play an important role in the development of bile duct injuries. According to Smadja and Blumgart [63], the cystic duct drains diagonally and laterally into the common bile duct in 75% of the cases, whereas in 20% of the patients, the cystic duct runs parallel to the bile duct. In the remaining 5%, the cystic duct passes posterior to the common bile duct before entering the distal left side of the common bile duct. In a large series including 600 patients, Berci [7] found similar anatomical variations but with different incidences. Other congenital anomalies have been reported and all of this data indicates the high degree of anatomical variability in this region. These factors are also complicated by acute cholecystitis and chronic fibrosis, which emphasizes the necessity of careful surgical dissection. Moossa et al. [46] support the idea that tissue dissection with the laser is more dangerous than electrocautery. In five of six patients in their series, a segment of bile duct was vaporized by the laser. The prospective analysis of the Southern Surgeons Club [68], however, did not show any difference in the incidence of bile duct injury using either electrocautery or laser. In this regard, the use of either laser or electrocautery or even the application of metal clips around the bile duct can compromise its blood supply by damaging the nutrient arteries running in the 3:00 and 9:00 oclock positions of the bile duct thus leading to bile duct injury or late stenosis [40]. In addition, various technical maneuvers are important to prevent bile duct injury. These considerations include adequate dissection of Calots triangle and accurate identification of the gallbladdercystic duct junction as well as the cystic ductcommon bile duct junction. It is also important to laterally retract the gallbladder neck so as to minimize tenting of the common bile duct, which can lead to its injury [10, 58]. In addition, acute cholecystitis and a short cystic duct are other important potential predisposing factors for bile duct injury. Another possible cause of bile duct injury is incorrect placement of metal clips impinging on the wall of the common bile duct. This may lead to partial or complete bile duct transection and/or late stricture formation. Complete bile duct transection has been reported in some series to occur in 45.5 to 72.7% of all bile duct injuries [10, 28, 58]. These reports specifically address CBD injury in major tertiary referral centers, and lesser bile duct injuries may not have been referred to that institution. However, in this series, only eight of 561 bile duct injuries (1.4%) were noted to be complete transection. Therefore, it should be noted that based on our review, the majority of injuries may be managed with procedures less complicated than a biliary enteric anastamosis. The role of intraoperative cholangiography (IOC) in preventing bile duct injuries has been previously discussed [12, 24, 57, 73]. Some investigators [12] have argued that IOC may increase the incidence of injury to the common bile duct, whereas other authors [24, 57, 73] believe that IOC is useful in identifying anatomical variations and bile duct injury. In this series, 41.5% of the patients had IOC performed with an 82.7% success rate. However, it cannot be concluded from this data whether or not the performance of IOC offers early detection of bile duct injuries, although it is very useful in identifying anatomical variations and choledocholithiasis. Since IOC is performed early in the

Table 9. Bismuth classification of biliary stricture 1. 2. 3. 4. 5. Low common hepatic duct stricturehepatic duct stump > 2 cm Mid common hepatic duct stricturehepatic duct stump < 2 cm Hilar stricture with no residual common hepatic ducthilar confluence intact Destruction of hilar confluenceright and left ducts separated Involvement of aberrant right sectoral duct alone or including common duct

gallbladder dissection, it may be postulated that in those patients who have complete transection of the common bile duct that it may not be detected because the injury occurs after the IOC is performed. Bile duct injuries that result in bile duct stricture have been classified by Bismuth [9] into five types according to their location in relationship to the hepatic duct bifurcation (Table 9). As he has reported, type II strictures were the most common (2738%). Type III injuries accounted for 2033%, type I 1826%, type IV 1416%, and type V 07% [39]. In a recent multicenter study of 81 patients with laparoscopic bile duct injury [73], similar frequencies of strictures were encountered. Moossa [46] reported in his series of six patients that 50% were type II strictures. In our review, it was found that the common bile duct/common hepatic duct were the most frequently injured structures (61.1%), which is consistent with other reported series [9, 46, 73]. Immediate recognition and repair of bile duct injury is believed to be associated with the best long-term result. Recently, a group of investigators [28] compared laparoscopic and open cholecystectomy cases in regard to the incidence of early vs delayed detection and short-term results after treatment. They did not find any significant difference in the incidence of bile duct lesions, nor was there any difference in short-term results of therapy. Bile duct lesions recognized postoperatively have been most frequently managed with biliary-enteric anastomosis. Nonsurgical treatment by ERCP, or papillotomy and placement of an endoprosthesis, or by percutaneous transhepatic cholangiography with stent placement are other options which can be used in less complicated cases or in patients who are poor operative candidates. In a series of 29 patients by Huibregtse et al. [34], 27 (93%) were successfully treated endoscopically. Greenen et al. [26], also reported excellent results in 88% of his patients with a mean follow up of 4.5 years. On the other hand, Pitt et al. [56] compared surgical repair with percutaneous transhepatic techniques and found a much higher success rate with surgery (88% vs 55%, respectively). This data indicates that the wide variety of treatment options and which therapy is chosen may depend on the resources at ones institution. In the present series, bile leaks were most frequently managed by percutaneous, endoscopic, or laparoscopy/ laparotomy drainage techniques. Recently, Peters et al. [54] reported their results with the endoscopic treatment of laparoscopic-related bile leaks. Among their 15 patients, endoscopic treatment was performed in nine cases, and all of these cases were successfully treated. In another series by Kozarek [37] evaluating the endoscopic management of bile leaks, common bile duct strictures, and fistulas, 86.5% of the cases had successful resolution of their problem. It appears from this data that the best initial treatment is endo-

320

scopic and/or percutaneous drainage, but surgical intervention may be necessary in more complicated cases. In conclusion, the incidences of bile duct injuries and leaks in this series are slightly higher than for open cholecystectomy. With increased surgical experience in laparoscopy, it is anticipated that these incidences will continue to decrease. It is very important that these complications be identified and treated early as immediate repair is most likely to achieve the best long-term results. If there is any doubt regarding the operative findings, the surgeon should not hesitate to convert a laparoscopic cholecystectomy to an open procedure. More importantly, however, every effort should be made to prevent the occurrence of bile duct injuries by thorough knowledge of the anatomy and by careful surgical technique.

References
1. Airan M, Appel M, Berci G, et al (1992) Retrospective and prospective multi-institutional laparoscopic cholecystectomy study organized by the Society of American Gastrointestinal Endoscopic Surgeons. Surg Endosc 6: 169176 2. Andren-Sandberg A, Alinder G, Bengmark S (1985) Accidental lesions of the common bile duct at cholecystectomy. Pre- and perioperative factors of importance. Ann Surg 201(3): 328332 3. Andren-Sandberg A, Johansson S, Bengmark S (1985) Accidental lesions of the common bile duct at cholecystectomy: results of treatment. Ann Surg 201: 452455 4. Arnaud JP, Bergamaschi R, Casa C, et al (1993) Coelioscopic cholecystectomy: experience with 201 initial patients. Surg Laparosc Endosc 3: 4446 5. Bailey RW, Zucker KA, Flowers JL, et al (1991) Laparoscopic cholecystectomy. Ann Surg 214(4): 531541 6. Baird DR, Wilson JP, Mason EM, et al (1992) An early review of 800 laparoscopic cholecystectomies at a university-affiliated community teaching hospital. Am Surg 58(3): 206210 7. Berci G (1992) Biliary ductal anatomy and anomalies. The role of intraoperative cholangiography during laparoscopic cholecystectomy. Surg Clin North Am 72(5): 10691075 8. Berci G, Sackier JM (1991) The Los Angeles experience with laparoscopic cholecystectomy. Am J Surg 161: 382384 9. Bismuth H, Lazorthes F (1981) Le traumatismes operatoires de la voie biliare principale. J Chir (Paris) 118: 601693 10. Branum G, Schmitt C, Baillie J, et al (1993) Management of major biliary complications after laparoscopic cholecystectomy. Ann Surg 217(5): 532541 11. Brooks DC, Becker JM, Connors PJ, et al (1993) Management of bile leaks following laparoscopic cholecystectomy. Surg Endosc 7: 292 295 12. Clair DG, Carr-Locke DL, Becker JM, et al (1993) Routine cholangiography is not warranted during laparoscopic cholecystectomy. Arch Surg 128: 551555 13. Cooperman AM (1991) Laparoscopic cholecystectomy: results of an early experience. Am J Gastroenterol 86(6): 694696 14. Cox MR, Wilson TG, Jeans PL, et al (1994) Minimizing the risk of bile duct injury at laparoscopic cholecystectomy. World J Surg 18: 422427 15. Csendes A, Diaz JC, Burdiles P, et al (1989) Late results of immediate primary end to end repair in accidental section of the common bile duct. Surg Gynecol Obstet 168: 125130 16. Dashow L, Friedman I, Kempner R, et al (1992) Initial experience with laparoscopic cholecystectomy at the Beth Israel Medical Center. Surg Gynecol Obstet 175: 2530 17. Davidoff AM, Pappas TN, Murray EA, et al (1992) Mechanisms of major biliary injury during laparoscopic cholecystectomy. Ann Surg 215(3): 196202 18. Davis CJ, Arregui ME, Nagan RF, et al (1992) Laparoscopic cholecystectomy: the St. Vincent experience. Surg Laparosc Endosc 2(1): 6468

19. Deveney KE (1993) The early experience with laparoscopic cholecystectomy in Oregon. Arch Surg 128: 627632 20. Deziel DJ, Millikan KW, Economou SG, Doolas, et al (1993) Complications of laparoscopic cholecystomy: a national survey of 4,292 hospitals and an analysis of 77,604 cases. Am J Surg 165: 914 21. Ferguson CM (1992) Electrosurgical laparoscopic cholecystectomy. Am Surg 58(2): 9699 22. Ferzli G, Kloss DA (1991) Laparoscopic cholecystectomy: 111 consecutive cases. Am J Gastroenterol 86(9): 11761178 23. Fitzgibbons RJ Jr, Schmid S, Santoscoy R, et al (1991) Open laparoscopy for laparoscopic cholecystectomy. Surg Laparosc Endosc 1(4): 216222 24. Flowers JL, Zucker KA, Graham SM, et al (1992) Laparoscopic cholangiography, results and indications. Ann Surg 215(3): 209216 25. Frazee RC, Thames T, Appel M, et al (1991) Laparoscopic cholecystectomy: a multicenter study. J Laparoendosc Surg 1(3): 157159 26. Geenen DJ, Geenen JE, Hogan WJ, et al (1989) Endoscopic therapy for benign bile duct strictures. Gastrointest Endosc 55: 567571 27. Gilliland Tm, Traverso LW (1990) Modern standards for comparison of cholecystectomy with alternative treatments for symptomatic cholelithiasis with emphasis on long-term relief of symptoms. Surg Gynecol Obstet 170: 3944 28. Gouma DJ, Go PMNYH (1994) Bile duct injury during laparoscopic and conventional cholecystectomy. J Am Coll Surg 178: 229233 29. Graber JN, Schultz LS, Pietrafitta JJ, et al (1992) Complications of laparoscopic cholecystectomy: a prospective review of an initial 100 consecutive cases. Lasers Surg Med 12: 9297 30. Graves HA, Ballinger JF, Anderson WJ (1991) Appraisal of laparoscopic cholecystectomy. Ann Surg 213(6): 655664 31. Harris BC (1993) Retrospective comparison of outcome of 100 consecutive open cholecystectomies and 100 consecutive laparoscopic cholecystectomies. South Med J 86(9): 993996 32. Hawasli A (1993) Does routine cystic duct cholangiogram during laparoscopic cholecystectomy prevent common bile duct injury? Surg Laparosc Encosc 3(4): 290295 33. Hawasli A, Lloyd LR (1991) Laparoscopic cholecystectomy, the learning curve: report of 50 patients. Am Surg 57(8): 542545 34. Huibregtse RM, Katon RM, Tytgat GNJ (1986) Endoscopic treatment of postoperative biliary stricture. Endoscopy 18: 133137 35. Hunter JG (1991) Avoidance of bile duct injury during laparoscopic cholecystectomy. Am J Surg 162: 7176 36. Ko ST, Airan MC (1991) Review of 300 consecutive laparoscopic cholecystectomies: Development, evolution and results. Surg Endosc 5: 103108 37. Kozarek RA, Ball TJ, Patterson DJ, et al (1994) Endoscopic treatment of biliary injury in the era of laparoscopic cholecystectomy. Gastrointest Endosc 40: 1016 38. Larson GM, Vitale GC, Casey J, et al (1992) Multipractice analysis of laparoscopic cholecystectomy in 1,983 patients. Am J Surg 163: 221 226 39. Lillemoe KD, Pitt HA, Cameron JL (1990) Postoperative bile duct strictures. Surg Clin North Am 70: 13551380 40. Lillemoe KD, Pitt HA, Cameron JL (1992) Current management of benign bile duct strictures. In: Cameron JL (ed) Advances in surgery. Mosby Year Book, St. Louis, pp 119174 41. Litwin DEM, Girotti MJ, Poulin EC, et al (1992) Laparoscopic cholecystectomy: TransCanada experience with 2201 cases. Can J Surg 35: 291296 42. Martin M, Abrams M, Arkin R, et al (1993) Safe laparoscopic cholecystectomy in a community setting, N 762. Surg Endosc 7: 300303 43. McGee JMC, Randel MA, Morgan RM, et al (1992) Laparoscopic cholecystectomy: an initial community experience. J Laparoendosc Surg 2: 293302 44. McSherry CK (1989) Cholecystectomy: the gold standard. Am J Surg 158: 174178 45. Meador JH, Nowzaradan Y, Matzelle W (1991) Laparoscopic cholecystectomy: report of 82 cases. South Med J 84(2): 186189 46. Moossa AR, Easter DW, van Sonneberg E, et al (1992) Laparoscopic injuries to the bile duct, a cause for concern. Ann Surg 215(3): 203 208 47. Morgenstern L, McGrath M, Carroll BJ, et al (1995) Continuing hazards of the learning curve in laparoscopic cholecystectomy. Am Surg 61: 914918 48. National Inpatient Profile (1989) Healthcare knowledge system, Ann Arbor, Michigan, pp 360363

321 49. Orlando R III, Russell JC, Lynch J, et al (1993) Laparoscopic cholecystectomy, a statewide experience. Arch Surg 128: 494499 50. Pain JA, Knight M, Smith RS (1988) Long-term results of the mucosal graft operation for benign bile duct strictures, In Netherland J SurgAbstract of Section World Congress of Hepato-Biliary-Pancreatic Surgery, p 170 51. Pellegrini CA, Tromas MJ, Way LW (1994) Recurrent biliary stricture: pattern of recurrence and outcome of surgical therapy. Am J Surg 147: 175180 52. Peters JH, Ellison EC, Innex JT, et al (1991) Safety and efficacy of laparoscopic cholecystectomy: a prospective analysis of 100 initial patients. Ann Surg 213: 312 53. Peters JH, Gibbons GD, Innes JT, et al (1991) Complications of laparoscopic cholecystectomy. Surgery 110: 769778 54. Peters JH, Ollila D, Nichols KE, et al (1994) Diagnosis and management of bile leaks following laparoscopic cholecystectomy. Surg Laparosc Endosc 4(3): 163170 55. Pitt HA, Miyamoto T, Parapatis SK, et al (1982) Factors influencing outcome in patients with postoperative biliary stricture. Am J Surg 144: 1421 56. Pitt HA, Kaufman SL, Coleman J, et al (1989) Benign postoperative biliary strictures: operate or dilatate? Ann Surg 210: 417427 57. Rosenthal RJ, Steigerwald SD, Imig R, et al (1994) Role of intraoperative cholangiography during endoscopic cholecystectomy. Surg Laparosc Endosc 4(3): 171174 58. Rossi RL, Schirmer WJ, Braasch JW, et al (1992) Laparoscopic bile duct injuries, risk factors, recognition and repair. Arch Surg 127: 596 602 59. Rubio PA (1993) Laparoscopic cholecystectomy: experience in 500 consecutive cases. Int Surg 78: 277279 60. Salky BA, Bauer JJ, Kreel I, et al (1991) Laparoscopic cholecystectomy: an initial report. Gastrointest Endosc 37(1): 14 61. Schirmer BD, Edge SB, Dix J, et al (1991) Laparoscopic cholecystectomy. Ann Surg 213(6): 665677 62. Sim RR, Nowicky DJ, McAlhany JC Jr, et al (1992) Laparoscopic cholecystectomy in a community hospital setting. Surg Gynecol Obstet 175: 161166 63. Smadja C, Blumgart LH (1988) The biliary tract and the anatomy of biliary exposure. In: Blumgart LH (ed) Surgery of the liver and biliary tract, vol 1. Churchill Livingstone, Edinburgh, pp 1122, 1988. 64. Soper NJ (1991) Laparoscopic cholecystectomy. Curr Probl Surg 28: 587655 65. Soper NJ, Stockmann PT, Dunnegan DL, et al (1992) Laparoscopic cholecystectomy, the new gold standard. Arch Surg 127: 917923 66. Spaw AT, Reddick EJ, Olsen DO (1991) Laparoscopic laser cholecystectomy: analysis of 500 procedures. Surg Laparosc Endosc 1(1): 27 67. Stoker ME, Vose J, OMara P, et al (1992) Laparoscopic cholecystectomy, a clinical and financial analysis of 280 operations. Arch Surg 127: 589595 68. The Southern Surgeons Club (1991) a prospective analysis of 1518 laparoscopic cholecystectomies. New Eng J Med 324(16): 10731078 69. Vecchio R, Ferrara M, Pucci L, et al. IL trattamento delle lesioni iatrogene della via biliare principale. Minerva Chir (in press) 70. Walker AT, Brooks DC, Tumeh SS, et al (1993) Bile duct disruption after laparoscopic cholecystectomy. Semin Ultrasound CT MRI 14: 346355 71. Williams LF Jr, Chapman WC, Bonau RA, et al (1993) Comparison of laparoscopic cholecystectomy with open cholecystectomy in a single center. Am J Surg 165: 459465 72. Wolfe BM, Gardiner BN, Leary BF, et al (1991) Endoscopic cholecystectomy, an analysis of complications. Arch Surg 126: 11911198 73. Woods MS, Traverso LW, Kozarek RA, et al (1994) Characteristics of biliary tract complications during laparoscopic cholecystectomy: a multi-institutional study. Am J Surg 167: 2734 74. Zucker KA, Bailey RW, Gadacz TR, et al (1991) Laparoscopic guided cholecystectomy. Am J Surg 161: 3644

Surg Endosc (1998) 12: 335337

Springer-Verlag New York Inc. 1998

Total radiated power, infrared output, and heat generation by cold light sources at the distal end of endoscopes and fiber optic bundle of light cables
C. Hensman,1 G. B. Hanna,1 T. Drew,2 H. Moseley,2 A. Cuschieri1
1 2

Department of Surgery, Ninewells Hospital and Medical School, University of Dundee, Dundee, Tayside DD1 9SY, Scotland Directorate of Medical Physics, Ninewells Hospital and Medical School, Dundee, Tayside DD1 9SY, Scotland

Received: 10 March 1997/Accepted: 1 August 1997

Abstract Background: Skin burns and ignition of drapes have been reported with the use of cold light sources. The aim of the study was to document the temperature generated by cold light sources and to correlate this with the total radiated power and infrared output. Methods: The temperature, total radiated power, and infrared output were measured as a function of time at the end of the endoscope (which is inserted into the operative field) and the end of the fiber optic bundle of the light cable (which connects the cable to the light port of the endoscope) using halogen and xenon light sources. Results: The highest temperature recorded at the end of the endoscope was 95C. The temperature measured at the optical fiber location of the endoscope was higher than at its lens surface (p < 0.0001). At the end of the fiber optic bundle of light cables, the temperature reached 225C within 15 s. The temperature recorded at the optical fiber location of all endoscopes and light cables studied rose significantly over a period of 10 min to reach its maximum (p < 0.0001) and then leveled off for the duration of the study (30 min). The infrared output accounted only for 10% of the total radiated power. Conclusions: High temperatures are reached by 10 min at the end of fiber optic bundle of light cables and endoscopes with both halogen and xenon light sources. This heat generation is largely due to the radiated power in the visible light spectrum. Key words: Infrared output Heat generation Cold light source

In 1877, Nitze constructed a cystoscope with an internal light source consisting of a heated platinum wire and a cooling system. Later on, the platinum wire was replaced by Edisons incandescent lamp to provide better illumination without the need for a cooling system [3]. The introduction of optical light fibers in 1955 permitted the use of highintensity external light sources without exposing the internal organs to high temperature [4, 5]. Halogen or xenon light sources with in-built heat filtering systems, referred to as cold light sources, are currently used in minimal access surgery (MAS). The total radiated power of cold light sources consists mainly of electromagnetic waves in the visible light and infrared spectrum as halogen and xenon lamps are not efficient emitters of UV radiation [6]. Thermal hazards such as skin burns and ignition of drapes have been reported with the use of such cold light sources [1]. To date, there has been very little reported data on the temperature generated by cold light sources used in MAS. The aim of the study was to document the temperature generated by cold light sources at the end of endoscopes and fiber optic bundle of light cables and to correlate the heat generation at these sites with the total radiated power and infrared output. Materials and methods
The illumination system consisted of halogen and xenon light sources connected to endoscopes via standard fiber optic cables. The temperature, total radiated power, and infrared output were recorded at the end of the endoscopes and the fiber optic bundle of light cables.

Equipment

Correspondence to: A. Cuschieri

Video-endoscopic system. The light sources used in the study were (1) halogen video cold fountain (model 450 BV) containing a 250-W halogen lamp and (2) xenon video cold light fountain (model 615 B) containing a 175-W xenon lamp.

336

Results Endoscopes The highest temperature recorded was 95C. The temperature measured at the optical fiber location was higher than at the lens surface (p < 0.0001, t-test for independent samples). The temperature recorded at the optical fiber location rose significantly in all endoscopes up to 10 min (p < 0.0001, t-test for paired samples) and was then followed by a steady state for the period studied, next 20 min (p 0.7, t-test for paired samples). The difference in the temperature and total radiated power between different endoscopes was not accounted for by a significant difference in infrared output (Table 1). The temperature recorded at 10 min correlated better with total radiation power (correlation coefficient 0.6, p < 0.0001) than with infrared output (correlation coefficient 0.4, p < 0.001). Light cables The highest recorded temperature at the end of the fiber optic bundle of light cables was 239C. The temperature reached 225C within 15 s. Again, there was a significant rise in the temperature recorded during the first 10 min (p < 0001, t-test for paired samples). Table 2 shows the mean and standard deviation of the temperature, total radiated power, and infrared output at the end of the fiber optic bundle of light cables. Light sources The temperature, total radiation power, and infrared output at the fiber optic bundle of light cables are shown in Table 3. Although the halogen light source employed a more powerful lamp (250 W) than that of xenon source (175 W), there was no significant difference in the temperature or total radiated power at the light cables or endoscopes between the two light sources. However, the infrared output at the fiber optic bundle of the light cables and endoscopes was higher with the xenon light source than with the halogen source (p < 0.0001, t-test for independent samples). Discussion The present study has documented high temperatures produced by cold light sources. These reach 95C at the end of the endoscope and 239C at the fiber optic bundle of light cables. Most of the temperature rise occurs during the first 15 s of opening the iris of the light source. The heat generation is largely caused by the radiated power in the visible spectrum as infrared component accounts for only 10% of the light energy. Furthermore, the temperature rise correlates better with total radiation power than with infrared output. This means that cold light sources have efficient heat filters to reduce the infrared output. As the temperature is caused by the radiated power in the visible spectrum, the heat production by current light sources and fiber optic light cables can be reduced only at the expense of illumination. Further research is required into the light absorption characteristics of fiber optic light cables especially at the distal end in order to produce adequate illumination with less heat production. The incorporation of the charged coupled devices (CCDs) at the distal end of the endoscope (optoelectronic endoscope) by permitting use of less powerful light sources will reduce heat generation with adequate il-

Fig. 1. Experimental setup. In experiment I, measurements were obtained at the end of the endoscope, while in experiment II measurements were taken at the end of the fiber optic bundle of the light cable without being connected to the endoscope.

The fiber optic cables, all 180 cm in length, were (1) 3.5 mm (model 495 NL), (2) 4.8 mm (model 495 NB), and (3) 4.8 mm (model 495 NB DP). The following endoscopes were studied: (1) a Hopkins II 0 direction of view (model 26003 AA), (2) a Hopkins I 0 direction of view (model 26033 AP), (3) a Hopkins I 0 direction of view (model 26033 APA), (4) a Hopkins I 30 (model 26033 BPA), and (5) a Hopkins I 45 direction of view (model 26033 FPA). All endoscopes were 10 mm in diameter. The endoscopes, light sources, and fiber optic cables were all new equipment (Karl Storz, Tuttlingen, Germany).

Total radiated power and infrared measuring system. The total radiation power was measured by the Ophir Energy system and Thermopile (model DGX, Optronics Ltd., Tel Aviv, Israel). An infrared pass filter which transmits only infrared radiation (model 03MCS005, Melles Griot, Cambridge, England) was utilized in conjunction with the thermopile to measure infrared output.

Temperature recording. The system consisted of a thermocouple (RS, Brighton, UK) connected to a computer card and software (Amplicon PC 73 A, Boston Technology, Boston, USA) configured on a Dell316SX personal computer (Dell Computer Corporation, Berkshire, England) running on a digital operating system.

Experimental procedure
Two experiments were performed. In the first experiment, temperature, total radiated power, and infrared output were recorded at the distal end of the endoscopes using both halogen and xenon light sources with 3.5-mm and 4.8-mm light cables. The temperature at the end of the endoscopes was measured both at the lens surface and at the optical fiber location (Fig. 1). In the second experiment, the temperature, total radiated power, and infrared output were measured at the end of the fiber optic bundle of three light cables connected to halogen and xenon light sources (Fig. 1). These measurements at the end of the fiber optic bundle were obtained with the light cable not connected to the light port of the endoscope. In both experiments, endoscopes and cables were secured to a clamp stand, and the thermocouple for the temperature recording was placed on the site being studied. On each occasion the temperature measurement was obtained immediately following disconnection of the light cable from the light source in order to prevent light absorption directly heating the thermocouple. The temperature was recorded at 15-s and 10-, 20-, and 30-min intervals from the opening of the iris of the light source. Measurement of total radiated power and infrared output was obtained at 10 min. During these experiments, the room temperature was kept constant at 26C. Both experiments were conducted in triplicate.

337 Table 1. Temperature, total radiated power, and infrared output at the end of endoscopes, mean and (standard deviation)a Endoscopes Temperature at 15 s (C) fiber location Temperature at 10 min (C) at fiber location Temperature at 10 min (C) at lens surface Total radiated power (Watts) at 10 min Infrared output (Watts) at 10 min 0 H I (AP) 0 H II (AA) 0 H I (APA) 30 H I (BPA) 45 H I (FPA) p* value <0.05 <0.05 0.2 <0.05 0.5

64.96 (15.43) 69.74 (17.95) 35.61 (1.02) 1.0408 (0.4079) 0.0808 (0.0676)

60.54 (16.56) 66.14 (19.57) 35.93 (0.85) 1.3450 (0.5012) 0.1042 (0.0944)

67.08 (13.89) 77.03 (19.57) 36.38 (1.14) 1.5058 (0.7689) 0.1442 (0.0989)

88.63 (42.30) 95.04 (41.92) 36.68 (1.45) 1.1242 (0.4081) 0.09442 (0.0776)

59.04 (14.77) 65.20 (14.46) 36.15 (1.08) 0.8858 (0.5010) 0.0925 (0.1067)

* p value on applying one-way ANOVA. (H) refers to Hopkins followed by letters between parentheses to indicate the model.

Table 2. Temperature, total radiated power, and infrared output at the fiber optic bundle of light cables, mean and (standard deviation) Light cables Temperature at 15 s (C) Temperature at 10 min (C) Total radiated power (Watts) at 10 min Infrared output (Watts) at 10 min * p value on applying one-way ANOVA. 3.5 mm 163.75 (7.50) 174.00 (10.98) 1.5025 (0.2743) 0.1300 (0.0983) 4.8 mm 180.50 (22.53) 191.75 (32.62) 3.5800 (0.2743) 0.3175 (0.1935) 4.8 mm (DP) 225.00 (0.58) 239.25 (13.74) 3.7300 (0.4516) 0.3675 (0.2165) p* value <0.01 <0.01 <0.0001 0.2

Table 3. Temperature, total radiated power, and infrared output at the fiber optic bundle of light cables, mean and (standard deviation)a Light sources Temperature at 15 s (C) Temperature at 10 min (C) Total radiated power (Watts) at 10 min Infrared output (Watts) at 10 min Halogen 186.00 (22.1359) 204.00 (30.82) 2.6450 (1.0693) 0.1250 (0.0635) Xenon 177.17 (18.5302) 199.33 (41.02) 3.2300 (1.0693) 0.4183 (0.1607) p* value 0.4 0.7 0.4 <0.001

* p value of applying one-way ANOVA. Halogen light source employs a 250-W lamp while xenon source uses 175-W lamp.

lumination of the operative field. The ultimate solution to the heat problem is the design of a self-illuminating endoscope based on an array of light-emitting diodes. This is a much more efficient system and will dispense with the need for powerful extrinsic light sources and connecting light cables. The data from this study have also demonstrated more heat generation at the optical fiber location near the distal end of the endoscope than at the lens surface. This temperature differential accounts for the fogging of the lens by condensation from intraperitoneal water vapor on the relatively cool surface of the endoscope lens. Fogging obscures vision and disrupts the progress of the procedure. The most commonly used methods to overcome this problem include the use of antifogging agents and pre-warming with dry laparoscope heaters or simple immersion in warm sterile water. The hydrolaparoscope (Circon AC-MI, Stanford,

CA, USA) and stereoendoscope used in transanal endoscopic microsurgery (Wolf, Knittlingen, Germany) have integrated irrigating channels designed to clear the optic when condensation occurs. The Tubingen multifunction laparoscope utilizes a separate electrical heating element to warm the optic together with warm humidified CO2 jet which streams across the lens to dry it [2] but this is a complex and unwieldy setup and has not proven popular. A simpler approach based on the observations from the present study is to deploy some of the thermal energy produced at the distal end of the endoscope to warm the surface of the lens and prevent fogging. This may be achieved by directing some of the optical fibre bundles onto the periphery of the lens surface, acting as an automatic lens warming system. There are other clinical implications of the study since such high-temperature levels may result in skin burns or ignition of drapes. Both surgeons and nurses should be made aware of thermal hazards associated with the use of video-endoscopic systems. References
1. Bellina JH, Haas M (1984) Cold light sourcesare they really cold? J Reprod Med 29: 275277 2. Bessell JR, Fleming E, Kunert W, Buess G (1996) Maintenance of clear vision during laparoscopic surgery. Min Invas Ther 5:450455 3. Casper DL (1906) Physical methods of examination. In: Casper DL, Bonney CW (eds) A text-book of genito-urinary diseases. Rebman, London, pp 1154 4. Heel ACS (1954) A new method of transporting optical images without aberrations. Nature 173: 39 5. Hopkins HH, Kapany NS (1954) A flexible fibrescope. Nature 173: 3941 6. Mosely H (1983) Sources of ultraviolet radiation. In: Mosely H (ed) Non ionising radiation. Adam Higler, Bristol, UK, pp 110113

Original articles
Surg Endosc (1998) 12: 294300 Springer-Verlag New York Inc. 1998

Biliary tract complications in laparoscopic cholecystectomy

A multicenter study of 148 biliary tract injuries in 26,440 operations
J. Rego ly-Me rei, M. Iha sz, Z. Szeberin, J. Sa ndor, M. Ma te
3rd Surgical Department, Semmelweis Medical University, H-1096, Budapest, Nagyva rad te r 1, Hungary Received: 1 March 1996/Accepted: 26 November 1996

Abstract Background: The higher risk of biliary tract injury is considered the most significant disadvantage of laparoscopic cholecystectomy. Methods: A national multicenter retrospective study was performed to determine the frequency, etiology, and treatment of biliary tract injuries between January 1, 1991, and December 31, 1994. Follow-up was by questionnaire. Results: Some 148 biliary tract complications were observed during 26,440 laparoscopic cholecystectomies. There was no significant correlation found between the number of LCs performed in one institute and the incidence of biliary tract injuries and postoperative bile leakage, but in the 2nd year of practice, the incidence of both complications decreased. In institutes with more conversions, more cases of bile leakage were also observed. A significant positive relationship was found between biliary tract injuries and postoperative bile leaks. There was no significant relationship between usage of intravenous and intraoperative cholangiography and ERCP. In univariant analysis of the type of injury, the primary treatment modality did not affect the outcome of injury or entail the necessity of reoperation. Obscure anatomy leads to significantly more main bile duct injuries. According to multivariant analysis, the outcome is significantly influenced unfavorably by the necessity of repeated interventions and advanced age. Conclusions: The definitely higher risk of bile duct injury mentioned in early studies was not confirmed. Key words: Laparoscopic cholecystectomy Biliary tract injury Postoperative bile leak Biliary tract reconstruction Conversion

Laparoscopic cholecystectomy (LC) has been become widespread since 1989, and today it is the standard method for treating cholecystolithiasis. The first procedures in Hungary were performed in December 1990 and it is now used in almost 90 institutes there. Compared to traditional open cholecystectomy (OC), the higher risk of biliary tract injury is considered the most significant disadvantage of this method. The incidence of these injuries was higher in the early reports (19901992) than in later ones (19931995), as experience in LC, patient selection criteria, and operative technique fundamentally determine the frequency of lesions [3, 4, 6, 810, 14, 16, 1922, 26]. The role of nonoperative, invasive radiological, and endoscopic methods in the treatment of biliary tract injuries and the comparison of their success with the operative results are debated issues. Is primary suture over a T-tube or bilio-enteric anastomosis the treatment of choice at reconstruction? Which factors determine the outcome of injury and the risk of repeated operations? These questions should be investigated in multicenter studies with a large number of patients. Data of institutes with different experience in performing LC give a more realistic picture of the method than if we only examine the results of departments with great experience.

Patients and method

In February 1995, questionnaries were sent to 119 Hungarian surgical departments with the assistance of the Hungarian Surgical Society. The objective of this retrospective study was to determine the frequency, etiology, and treatment of biliary tract complications of LC between January 1, 1991, and December 31, 1994. According to the 105 responses were received, LC was not performed in 16 institutes; therefore we analyzed the data of 89 surgical departments. It comprises 95% of all laparoscopic cholecystectomies performed in the same time period in Hungary. Our main emphasis was laid on studying the possible connection among preand intraoperative cholangiography, preoperative ERCP, frequency of conversion, and the biliary tract injuries and spontaneously ceased postoperative biliary leaks. The symptoms, the diagnostic methods, the localization of injuries, the treatment modalities, the type of reoperations, the outcome, and the current status of the patients were analyzed. We registered 148

Correspondence to: J. Rego ly-Me rei

295 Table 1. Summary of Hungarian data Institutional Nationwide data Number of LCs Conversions Preoperative cholangiograms Preoperative ERCPs Intraoperative cholangiograms Biliary tract injuries Biliary leaks 26,440 5.9% 6.9% 4.1% 6.9% 0.6% 1.8% Average SD 297 418 6.5 5.6% 8.5 24.8% 3.8 7.3% 3.9 13.6% 0.7 0.95% 2.2 3% Minimum 3 0% 0% 0% 0% 0% 0% Maximum 2405 32.4% 100% 40.2% 94.5% 4.5% 21.6%

biliary tract lesions in 26,440 LCs in the above-mentioned period, but the real number of injuries could be estimated only in the long-term follow-up. The data of the questionnaires were analyzed using an IBM-compatible personal computer. Lotus software was used to log and monitor the data and SPCC/PC+ was used for analysis. Statistical correlations were examined using regression analysis. The Student t-test was used for the analysis of significant differences in the case of continuously changing factors, and the Fischer exact test 2-test were used in the case of discrete variables. The multivariant examination was performed using Wilks discriminant analysis. The most important topics examined are: 1. The role of experience in LCs, the incidence of conversions, the use of intraoperative cholangiography and preoperative diagnostic tests (preoperative cholangiography and ERCP) in the prevention of biliary complications 2. The effects of location of the lesion, their primary repair, and the time of the diagnosis on the outcome (frequency of reoperations, mortality, postoperative complaints, etc.) 3. The suspected mechanism of injuries, the method of diagnosis, and the symptoms

Results

Analysis of institutional data In the period examined, 26,440 cholecystectomies were performed laparoscopically (the total number of cholecystectomies was 55,605) in 89 institutes; the LCs were 48% of all cholecystectomies. Although the national average is 297 418 LCs institute, there were less than 50 procedures performed in 21 departments and less than 300 in another 29, meaning that these institutes are in the learning phase. Twenty-nine departments have had more than 300; three departments performed more than 1,000 and two more than 2,000 operations. The great variety of data made the analysis difficult, so in several instances we performed the analysis using different statistical methods to reduce the possibility of bias. In Table 1 we summarize LCs, the nationwide and institutional rate (mean SD) of conversion, pre- and intraoperative cholangiography, ERCP, and biliary tract injuries and postoperative bile leaks. It is to be noted that preoperative cholangiography was used in only 32, preoperative ERCP in only 50, intraoperative cholangiography in only 33 institutes. Conversions were necessary in 81 departments; biliary tract injury was reported from 56 and postoperative bile leak from 68 departments. The data in an annual breakdown showed that the use of preoperative intravenous cholangiography decreased continuously (15% in 1991, 5.1% in

1994), while the frequency of ERCP increased somewhat (1.4% and 4.5%, respectively, in the same years). There was no significant connection between percentage of LCs and the rate of conversion among the institutes. There was a wide range of figures among the different institutes (conversion: 0 to 32.4%, average SD: 6.5 5.5%); therefore, the regression curve between the absolute number of LCs performed in 1 year and conversions was examined also. It was realized that the regression coefficient of the curve decreased significantly in the 2nd year of practice (p < 0.05, t 2.082), meaning that as more experience was gained, less conversions were associated with the same number of operations (Fig. 1). No significant correlation was found between the percentage of biliary tract injuries and the absolute number of LCs performed in the same institution, the incidence of conversions, the frequency of intra- and preoperative cholangiography, and ERCP. However, when the correlation between the absolute number of biliary tract injuries and the absolute number of LCs performed was evaluated, we found that although the higher number of procedures is connected with a higher risk of lesions, the coefficient of the regression curve decreased significantly in the 2nd year compared with the 1st year, and in the 3rd year there is no more significant correlation between the two parameters (Fig. 2). A significant correlation was found between biliary tract injuries and the postoperative bile leaks that ceased spontaneously (p < 0.05, r 0.228). The latter also showed positive correlation with the frequency of conversions (p < 0.001, r 0.356), but it was not in significant correlation with the number of LCs performed, with the incidence of the use of intra- and preoperative cholangiographies, or with ERCP. Evaluating the absolute numbers of LCs performed and bile leaks in connection with the year of practice, we found similar correlations, as in the case of biliary tract injuries, i.e., the coefficient of the regression curve decreased as experience was gained (1st year/2nd year: p < 0.001, t 10.222, 1st year/3rd year: p < 0.001, t 9.9009 (Fig. 3).

Detailed analysis of 148 biliary tract complications In the course of the 26,440 LCs, 29 male and 119 female patients suffered biliary tract complications, and the average age was 48.7 14.7 years (1684 years). The injuries were detected intraoperatively in 34.5% (n 51) and conversion was performed in 50 cases. In one case, a small, point-like injury was drained and the patient was closely observed.

296 Table 2. Types of biliary tract complications (n 148) N Cystic duct lesion Partial injury of main bile ducts Complete transection of main bile ducts Excision of bile duct Stricture Accessory bile duct injury Combined injurya Unidentified Total 39 52 30 10 8 3 4 2 148 % 26.4 35.1 20.3 6.8 5.4 2.1 2.8 1.4 100

a Injury + stricture (two cases); excision + lesion of the hepatic artery (one case); stricture + accessory bile duct lesion (one case)

Table 3. Primary treatment of biliary tract injuries I N Patients with drains in place, observation EST + stent Percutaneous drainage EST + Stent + percutaneous drainage Conversion Reoperation Diagnosed only at autopsy Total 3 1 2 1 50 90 1 148 % 2.0 0.7 1.4 0.7 33.8 60.8 0.7 100

Fig. 1. The correlation between the number of conversions and the number of LCs performed in 1 year. Fig. 2. The correlation between the number of biliary tract injuries and the number of LCs performed in 1 year. Fig. 3. The correlation between the number of biliary leaks terminated spontaneously and the number of LCs performed in 1 year.

The injury was postoperatively detected in 65.5% of the cases (n 97) on average postoperative day 6 7.6 (164). The different types of biliary complications are shown in Table 2. The partial (type 2) and complete (type 3) injuries of the main biliary ducts were detected intraoperatively significantly more frequently than injuries of the cystic duct or confluence (type 1). The cases in the latter group (n 39) were usually diagnosed in the postoperative period (conversion: six, postoperative detection:33)(type 1/type2: p < 0.001, type1/type3: p < 0.05). There was no significant correlation between the type of injury and the delay of diagnosis, except in combined injuries, where the diagnosis is made later (but in the latter group the number of cases is inadequate for statistical analysis).

Clinical signs in 61.9% (spontaneous pain 45.4%, tenderness 51.5%, fever 26.8%, nausea 24.7%, vomiting 15.5%, acute abdomen, guarding 14.4%, other 4%) and bile leak via a drain in 63.9% indicated lesion. In spite of the diagnostic methods (sonography, ERCP, PTC, CT, HIDA cholescintigraphy), in 14.4% of the cases the real etiology of the symptoms was revealed only at the time of reoperation and in one patient only by autopsy. The primary treatment of the patients is shown in Table 3. There were significantly more reoperations and less conversions in the cases of cystic duct lesions than in the cases of main bile duct injuries. No significant difference, however, were found between the different types of main bile duct injuries (partial lesion, complete transection, excision) with respect to the incidence of conversion and reoperation. In the 148 patients, 50 conversions and 90 reoperations were performed primarily; the distribution of these operations is summarized in Table 4. There was not enough data in 5% of the patients for accurate classification. In the 39 cystic lesions six patients (15.4%) were treated by conversion and 31 (79.5%) by reoperation. One patientwith an intraoperatively inserted drainwas only observed, while another one was drained percutaneously. Cystic stump revision and closure alone were performed with open surgery in 22 (56.4%) and laparoscopically in one (2.6%) case. Reconstruction over biliary drainage was necessary in eight patients (20.5%), while biliodigestive anastomosis was done in two (5.1%) patients. Evacuation of fluid accumulation and CBD stone extraction combined with biliary drainage were performed each in one case, respectively. The type of reintervention was not clear in two (5.1%) cases. In cases of partial CBD lesions, reconstruction over a T-tube or other type of drain was performed primarily in 52 (69.2%) cases, and bilioenteric anastomosis was performed

297 Table 4. Primary treatment of biliary tract injuries II (Conversion + reoperation) N Exploration alone Evacuation or drainage alone Treatment of accessory bile ducts Removal or change of clips Treatment of cystic duct, drainage, or suture of cystic duct angle Suture of bile duct T-tube or other drainage Some form of biliodigestive anastomosis Removal of bile duct stone + T-tube Hepaticojejunostomy + reconstruction of hepatic artery Laparoscopic cystic duct management Unidentified Total 1 3 3 3 22 59 37 2 1 2 7 140 % 0.7 2.1 2.1 2.1 15.7 42.0 26.4 1.4 0.7 1.4 5.0 100

Table 5. Secondary treatment of biliary tract injuries: n 22 (15%) N EST EST + stent Biliodigestive anastomosis Redo of previous anastomosis T-tube ligation or suture of cystic stump Other interventions for stricture Abscess drainage Duodenum suture + biliary tract reconstruction + T-tube Lysis of adhesions Vagotomy + excision of ulcer Total 1 1 11 1 3 1 1 1 1 1 22 % 4.5 4.5 50.0 4.5 13.6 4.5 4.5 4.5 4.5 4.5 100

in three (5.8%) cases. Evacuation, CBD stone extraction combined with biliary drainage, laparoscopic reconstruction, EST combined with stent implantation, and percutaneous drainage were done in one case each. In this group we did not find any data about the type of reoperation in four cases (7.7%). Of the 30 complete CBD transections bilioenteric anastomosis was performed in 19 (63.3%) patients and reconstructions over a T-tube were done in ten (33.3%). In one patient we did not have any information. Bilioenteric anastomosis dominated in the treatment of excisions (8/10) and strictures (4/8). A second intervention became necessary in 22 patients (15%) (Table 5), and in half of these (n 11) biliodigestive anastomosis was the choice treatment. Gastrointestinal bleeding and adhesions were the indication for surgery in two patients. EST with or without stent was performed in two patients. Only one patient required reoperation among the 22 cases with cystic stump suture, drainage, or reclipping. Repeated intervention became necessary in eight cases after biliary reconstruction+drainage and in six patients after previous biliodigestive anastomosis. This does not represent a significant difference between the type of primary surgery and the necessity for reintervention. A third intervention was needed in six patients (4.1%). In one case a stent was placed; the type of procedure is not known in one case, while of the remaining four patients, one biliodigestive anastomosis was performed and the anastomosis was repeated in three cases.

Patients recovered completely in 75.7% of the cases following a primary procedure and in 7.4% after a reintervention (83.1%). Seven patients expired (4.7%); two of them had reintervention. Nine-and-one-half percent (3.4% had one procedure, 6.1% had reintervention) still have complaints (meteorism, cholangitis, anastomotic stricture, positive liver function tests). Four patients (2.7%) were transferred to another institute and no data was available on their follow-up. Age influenced the outcome significantly; the average age of the cured (p < 0.05, t 2.213) and the survivals with complaints (p < 0.01, t 3.380) was lower than that of the expired. It took 18.7 20.3 (6180) days on average to recover following one intervention and the same figure was 31.7 33.1 (590) following reintervention. This difference does not attain statistical significance because of the wide deviation. There was no significant correlation found between the outcome (cured/expired and patients with or without complaints) and the type of injury (cystic duct lesion, partial or complete lesion of main bile duct, excision, etc.) using univariant analysis. Also based on univariant analysis, the type of the primary procedure (cystic duct closure alone, bile duct suture over a drain, biliodigestive anastomosis) did not significantly influence the ratio of cure and mortality, but significantly more patients had complaints following biliodigestive anastomosis than following cystic duct closure (p < 0.05). Evaluating the correlation between the time of cure and primary procedure, there is no significant difference between cystic duct closure and bile duct suture with or without drain, but the recovery is significantly longer in cases of biliodigestive anastomosis compared to cystic duct closure (p < 0.01, t 2.908). There is no significant difference between recovery periods in the cases of bile duct suture and biliodigestive anastomosis. Multivariant analysis was employed to examine the combined influence of age, the type of injury (cystic duct lesion, partial or complete transection of a biliary tract, excision, suture), the type of the primary procedure (cystic duct closure, biliary tract reconstruction drainage, biliodigestive anastomosis), the time of diagnosis (intraoperative-conversion, postoperative-reoperation), and the repeated intervention (yes or no) on the outcome. If the patients were placed in two separate groups of cured and symptomatic + expired, repeated intervention and delayed diagnosis (reoperation and no conversion) proved to be significant factors. If they were grouped as surviving patients (with or without complaints) and expired, age in the first place and repeated procedure in the second exhibited a significant negative effect on the outcome. If we analyze the data in three separate groups of surviving patients without complaints, surviving patients with complaints, and expired, the same two factors were found to be significant, and repeated procedure takes first place and age second. The suspected etiology of biliary tract injuries is listed in Table 6. However, the reason could not be identified in 4.1% of the cases. Identification problems were found in almost half of the cases (43.2%), and combined mechanisms were supposed in 10.1%. Analyzing the correlation between the etiology and the type of injury, it was self-explanatory that difficulties in applying clips caused significantly more

298 Table 6. Suspected etiology of biliary tract injuries N Malposition or spontaneously slipped clips Injury caused by clips Thermal injurya Identification difficultiesb Technical failure Bile duct stone Carcinoma of the gallbladder Combined mechanism Other Unidentified Total 21 12 19 64 5 4 1 15 1 6 148 % 14.2 8.1 12.8 43.2 3.4 2.7 0.7 10.1 0.7 4.1 100

a Thermal origin was significantly more often seen in partial lesion than in complete transsection b Identification difficulties were significantly more often met in partial or complete lesions than in the injury of the region of the cystic duct

cystic duct complications than main bile duct injuries (p < 0.001, 2 14.56). Coagulation caused significantly more partial biliary duct injury than complete transsection (p < 0.05, 2 4.178). Obscure anatomy caused significantly more partial (p < 0.001, 2 12.33) and complete (p < 0.001, 2 34.23) biliary tract lesions than cystic duct injuries or insufficiency and also played a major role in complete dissections compared to partial lesions (p < 0.001, 2 11.99). In the 148 patients we analyzed, the effect of intravenous cholangiography and ERCP (preoperative diagnosis) and intraoperative cholangiography separately as well. There was no significant correlation found (similarly to the institutional results) between the usage of the preoperative tests and the incidence of the biliary tract complications. Intraoperative cholangiography was performed significantly more often in the case of injuries (p < 0.05, 2 9.15), but most likely the cholangiograms were already performed with the suspicion of bile duct injuries. In the institutional data it was shown that intraoperative cholangiography does not effect the incidence of biliary tract injuries. Seven patients were lost who suffered biliary tract injury (4.7% mortality). In two cases gallbladder carcinoma was partially responsible for the fatal outcome. The cause of death was cardiorespiratory failure in one case, hepatic coma in another; the three others died of a septic condition, peritonitis (in one case accompanied with bleeding). We would like to emphasize that one patient was returned to a surgical department a few days following the surgery in a moribund condition, and he expired within hours. Autopsy revealed that the cause of death was peritonitis caused by biliary tract injury. Discussion The incidence of biliary tract injuries during LCs ranges from 0.2% to 0.9% [3, 6, 7, 9, 10, 13, 16, 21, 22, 26], but in the early period following its introduction figures of 1.4 to 2.9% were also reported [4, 12]. The Hungarian average of 0.6% is comparable with international data. Only 2438% of the biliary tract injuries are detected at the time of the surgery, and the delay in the diagnosis of lesions varies from 1 to 246 days, most frequently 2 to 30 days [1, 3, 4, 9, 20,

23, 25, 28]. The early diagnosis improves if intraoperative cholangiography is used routinely [28]. The incidence of conversion changes from 0.9 to 13% in the literature, mostly 4 to 4.5%, and one-third of them are performed acutely [511, 21, 22, 26]. Biliary tract injury is the reason for conversion in 0.2 to 0.3% of the cases. The symptoms of biliary tract injury appear often only after discharge (the average hospital stay is 24 to 72 h), so family doctors should be educated to recognize this complication. There are different classifications of the biliary tract injuries in the literature. The Bismuth classification is based on localization [16, 28]; another emphasizes the type of complication (injury, stricture, cholangitis, biliary cirrhosis [20]; and the thirdwhich is probably more useful in the case of LCs-distinguishes partial/tangential lesions, injuries caused by clips, complete transections excisions and lesions of the right hepatic duct [20, 23]. In our own practice, we applied the last classification with moderate modification (see Table 2). The role of sonography in preoperative diagnostic tests is well known as it decreases operative risks and avoids biliary tract injuries [17]. Some teams routinely use preoperative intravenous cholangiography [8, 11, 18, 26] but the majority avoid this test [10]. The incidence of preoperative ERCP depends on whether laparoscopic revision of the common bile duct can be performed in the institute [11, 13, 15, 18, 26]. Both methods decrease the incidence of unexpected biliary tract stones found during LCs, but we did not find any evidence that the incidence of biliary tract complications and postoperative bile leaks decreases if these two methods are used more often. Opinions differ about the routine and selective application of intraoperative cholangiography [2, 5, 68, 9, 11, 18, 21, 22, 26]. Dubois et al. [6] state that intraoperative cholangiography does not facilitate the prevention of biliary tract complications, and they only perform it if there is a higher risk of a common bile duct stone. Others say that the intraoperative cholangiogram helps to avoid the completing partial injuries, that lesions can be identified earlier, and that conversion can be performed. The prognosis is better than in the case of injuries recognized only postoperatively [28]. Obviously injuries may occur following cholangiography due to a malpositioned clip or during the dissection of the gallbladder from the liver bed. Berci [2] emphasized that the indications for intraoperative cholangiography are different in traditional and laparoscopic surgery. In the latter case the main purpose is to avoid iatrogenic lesions. In the Hungarian data there was no significant correlation between the use of intraoperative cholangiography and the incidence of biliary tract injuries and bile leaks. We should think of biliary tract injury if atypical pain, abdominal distension, vomiting, ileus, or cholangitis is present in the postoperative period or if peritoneal signs, anorexia, or pathologic laboratory results are detected [1, 4, 19, 25]. In our data, 14.4% of the patients with injuries had acute abdominal symptomsperitonitis. Biliary flow via the intraoperatively placed drain is a very demonstrative sign (63.9%). While sonography and CT prove pathologic fluid accumulation, HIDA, ERCP, and PTC show the lesion itself and the last two methods can also localize the injuries [1, 3, 5, 25, 27, 28]. Several things are responsible for biliary tract injuries

299

[1, 6, 8, 14, 19]. Obscure anatomy is the dominating factor (43.2%) in our practice. It leads to significantly more partial or complete lesions of the main bile ducts than cystic duct injury and is also responsible for more complete transection than partial lesion. Similar factors are responsible for postoperative bile leaks, as well. We also should mention perforations caused by cholangiographic catheters and the possible presence of the ducts of Luschka [10, 27]. While this complication occurs in 0 to 0.6% during traditional cholecystectomies, the incidence is 0.3 to 0.9% in connection with LCs [8, 10]. The similar etiology of postoperative bile leak and biliary tract lesion is supported by the Hungarian data. Intraabdominal drainage helps in the early diagnosis of complications and it can be therapeutic, too [10]; however, there are different opinions in the literature about its selective or routine application [58, 22]. The best results can be achieved when reintervention is performed before the appearance of peritonitis and septic complication. It is not clear, however, what type of procedure should be performed. Some teams emphasize the success of invasive radiological, endoscopic interventions and relaparoscopy [3, 4, 24]. If the biliary tract has a good outflow, the lesion can heal following the diversion [25]. Others report less favorable results, but they think that due to invasive procedures the patient will reach a better condition by the time of the definitive surgery [16, 23, 25]. The lesions of the cystic duct and liver bed can be treated well by nonoperative methods [5, 28] and percutaneous drainage is almost always successful in the cases of encapsulated fluid accumulations [25]. In the Hungarian data, nonsurgical interventions were applied as primary treatment in only 2.8% of the patients, so conclusions cannot be drawn. There is always an indication for bilioenteric anastomosis in the case of late strictures [16], but it is questionable whether primary bile duct suture ( biliary tract drainage) or biliodigestive anastomosis is more advantegous in the early treatment. Schol et al. [20] state that the time of diagnosis is also important besides the severity of injury. End-to-end anastomosis can be performed if a diagnosis is made promptly, and bilioenteric anastomosis is recommended in the case of delayed treatment. In our data the primary procedure was determined by the type of the injury. In our data, bile duct suture over a drain was performed in 59 patients as primary treatment and biliodigestive anastomosis was applied in 37 cases. There was no significant difference between the two groups in the outcome and incidence of reintervention, but the neccessity of repeated interventions made a significantly negative influence on the outcome. Higher age made the prognosis of injury significantly worse by uni- and multivariant analysis as well; however, the type of injury and the primary treatment did not influence significantly the ratio of the cured and the expired patients. The follow-up studies of biliary tract injuries report a 10 to 30% incidence of delayed strictures [16, 27]. These results are based on the experience gained in traditional cholecystectomies as the time since the introduction of LC is insufficient to evaluate the real incidence. Raute et al. [16] state that 65 to 85% of the strictures present in the first 2 years and only 5 to 10% of them appear later than 10 years. The supposed 1 to 2% incidence of biliary tract injuries

in the early phase of LC was not apparent in our experience. On the other hand, 15% of the lesions required reintervention, and the mortality was 4.7%, so further prospective studies are necessary to determine the ideal surgical technique to decrease the incidence of injuries and the best primary reconstructive protocols.
gnes Zo Acknowledgments. We thank A ra ndi for her hard work in the statistical analysis. We thank our colleagues in all participating surgical departments for supplying data on patients and for taking part in this national survey.

References
1. Asbun HJ, Rossi RL, Lowell JA, Munson L (1993) Bile duct injury during laparoscopic cholecystectomy: mechanism of injury, prevention and management. World J Surg 17: 547552 2. Berci G, Sackier JM, Paz-Partlow M (1991) Routine or selected intraoperative cholangiography during laparoscopic cholecystectomy. Am J Surg 161: 355360 3. Bezzi M, Silecchia G, Orsi F, Materia A, Salvatori FM, Fiocca F, Fantini A, Basso N, Rossi P (1995) Complications after laparoscopic cholecystectomy. Coordinated radiologic, endoscopic and surgical treatments. Surg Endosc 9: 2936 4. Brooks DC, Becker JM, Connors PJ, Carr-Locke DL (1993) Management of bile leaks following cholecystectomy. Surg Endosc 7: 292 295 5. Cox MR, Wilson TG, Jeans PL, Padbury RTA, Toouli J (1994) Minimizing the risk of bile duct injury at laparoscopic cholecystectomy. World J Surg 18: 422427 6. Dubois F, Berthelot G, Levard H (1995) Coelioscopic cholecystectomy: experience with 2006 cases. World J Surg 19: 748752 7. Ezer P, Antal A, Schmidt P, Gula csi I, Pavlovics G, Korompai F (1995) A laparoszko pos cholecystectomia (LC) sora n beko vetkezo epeu tse ru le sek megelo ze se nek leheto se gei (The possible prevention of biliary tract injuries during laparoscopic cholecystectomy). Magy Seb 48: 369376 8. Faust H, Ladwig, D, Reichel K (1994) Die laparoskopische Cholecystektomie als Standardeingriff bei symptomatischen Cholecystolithiasis. Erfahrungen bei 1277 Patienten. Chirurg 65: 194199 9. Go PMNYH, Schol F, Gouma DJ (1993) Laparoscopic cholecystectomy in the Netherlands. Br J Surg 80: 11801183 10. Iha sz M, Rego ly-Me rei J, Fazekas T, Ba torfi J, Ba lint A, Za borszky A, Po sfai G (1995) Hazai tapasztalatok a laparoszko pos cholecystectomia szo vo dme nyeiro l: 71 inte zetben ve gzett 13.833 mu te t elemze se (Our national experiences with complications of laparoscopic cholecystectomy. Evaluation of 13,883 procedures in 71 institutes). Magy Seb 48: 122 11. Jatzko GR, Lisborg PH, Pertl AM, Stettner HM (1995) Multivariate comparison of complications after laparoscopic cholecystectomy and open cholecystectomy. Ann Surg 221: 381386 12. Kozarek R, Gannan R, Baerg R, Wagonfeld J, Ball T (1992) Bile leak after laparoscopic cholecystectomy. Diagnostic and therapeutic application of endoscopic retrograde cholangiopancreatography. Arch Intern Med 152: 10401043 13. Lezoche E, Paganini AM (1995) Single-stage laparoscopic treatment of gallstones and common bile duct stones in 120 unselected, consecutive patients. Surg Endosc 9: 10701075 14. Peiper M, Emmermann A, Rogiers X, Bro lsch CE (1994) Stenosierung des Ductus Choledochus durch Metall-Clips nach laparoskopischer Cholecystektome. Chirurg 65: 217220 15. Phillips EH, Caroll BJ, Pearlstein AR, Daykhovsky L, Fallas MJ (1993) Laparoscopic choledochoscopy and extraction of common bile duct stones. World J Surg 17: 2228 16. Raute M, Podlech P, Aschke WJ, Manegold BC, Trede M, Chiri B (1993) Management of bile duct injuries and strictures following cholecystectomy. World J Surg 17: 553562 17. Rego ly-Me rei J, Iha sz M, Fazekas T, Za borszky A, Ba torfi J, Barta T, Bereczky M, Szeberin Z (1995) A sonographia szerepe a laparoscopos cholecystectomia ban (The role of sonography in the laparoscopic cholecystectomy). Orv Hetil 136: 13711379 18. Rieger R, Salzbacher H, Woisetschla ger R, Schrenk P, Wayand W

300 (1994) Selective use of ERCP in patients undergoing laparoscopic cholecystectomy. World J Surg 18: 900905 Rossi RL, Schirmer WJ, Braasch JW, Sanders LB, Munson JL (1992) Laparoscopic bile duct injuries. Risk factors, recognition and repair. Arch Surg 127: 596602 Schol FBG, Go PMNYH, Gouma DJ (1995) Outcome of 49 repairs of bile duct injuries after laparoscopic cholecystectomy. World J Surg 19: 753757 Schrenk P, Woisetschla ger R, Wayand WU (1995) Laparoscopic cholecystectomy: cause of conversions in 1300 patients and analysis of risk factors. Surg Endosc 9: 2528 Stahlschmidt M, Lotz GW, Moergel K, Maurer T (1992) Ergebnisse der konventionellen und laparoskopischen Cholecystektomie. Z Gastroenterol 30: 713716 Stewart L, Way LW (1995) Bile duct injuries during laparoscopic cholecystectomy. Factors that influence the results of treatment. Arch Surg 130: 11231129 Traverso LW, Kozarek RA, Ball TJ, Brandabur JJ, Hunter JA, Jolly PC, Patterson DJ, Ryan JA, Thirlby RC, Wechter DG (1993) Endoscopic retrograde cholangiopancreatography after laparoscopic cholecystectomy. Am J Surg 165: 581586 van Sonnenberg E, DAgostino HB, Easter DW, Sanchez RB, Christensen RA, Kerlan RK Jr, Moossa AR (1993) Complication of laparoscopic cholecystectomy: coordinated radiologic and surgical management in 21 patients. Radiology 188: 399404 Wayand WU, Woisetschla ger R, Gitter T (1993) Laparoskopische sterreich. O sterreichisches Register in 1991. Cholecystektomie in O Chirurg 64: 303306 Woods MS, Shellito JL, Santoscoy GS, Hagan RC, Kilgore WR, Traverso LW, Kozarek RA, Brandabur JJ (1994) Cystic duct leaks in laparoscopic cholecystectomy. Am J Surg 168: 560563 Woods MS, Traverso LW, Kozarek RA, Donohue JH, Fletcher DR, Hunter JG, Oddsottir M, Rossi RL, Tsao J, Windsor J (1995) Biliary tract complications of laparoscopic cholecystectomy are detected more frequently with routine intraoperative cholangiography. Surg Endosc 9: 10761080

19. 20. 21. 22. 23. 24.

25.

26. 27. 28.

Editorial
Surg Endosc (1998) 12: 291293 Springer-Verlag New York Inc. 1998

Complications of laparoscopic cholecystectomy

There are several interesting articles in this issue, which call for some comment. In the multicenter study from Hungary by Rego ly-Me rei et al. [3], in 4 years 26,440 laparoscopic cholecystectomies (LCs) were reported and the information was obtained via questionnaires. In the same time period, more than 27,000 additional gallbladders were removed in a standard open way. I cannot explain why LC did not gain wider acceptance (the operators inertia?, the economy?, special local circumstances?). There are gigabites of data available with an enormous variety of statistical approaches, but there are some interesting statements: There was no significant correlation found between the percentage of biliary tract injuries and the absolute number of LCs performed in the same institutions . . . The authors stated at a later stage that the incidence of complications decreased with experience. The sampling range was enormous. If my interpretation is correct, 21 institutions in the 4-year period performed 50 LCs (112 per year) and in 29 other hospitals, fewer than 300 LC (75 cases per year) were performed. In another 29 hospitals, more than 300 LCs were performed. Three hospitals had more than 1,000 LC and two hospitals performed more than 2,000 LC procedures in the same 4-year time period. I do not know how these large fluctuations of cases per hospitals could be brought to a common denominator if it comes to analysis. Only 34.5% of the injuries were discovered during the first operation, whereas more than two-thirds were seen in an average of 67 postoperative days. The total number of ductal injuries was 148. The bottom line of the article is the 0.6% incidence of biliary tract injuries, which is higher, in my opinion, than in the pre-LC era, even in Hungary. The type of injuries were: complete transections, 20.3%; partial injuries, 35.1%; cystic duct lesion, 26.4%; excision of the duct: 6.8%. The rest of the injuries were minor. The cystic duct lesion is not clearly described. Are we talking about a slippage of a clip? The rate of reoperation was 60.8% and the dominant repair was suturing of the bile duct over a T-tube (42% of cases). Whether this technique of repair is optimal can be addressed only if longer and accurate followup data are available. The various contributors were blaming identification difficulties in 43.2% and problems in verifying the cystic duct common bile duct junction. The conversion rate, because of visible bile leak (drain) was 63.9%. The average performance of a cholangiogram was only 6.9%, which is low. The authors recommend further prospective studies to determine the ideal surgical technique to decrease the incidence of injuries and the best primary reconstructive protocols. Before commenting on this huge number of data and evaluation of 26,000 + LCs, I would like to mention the other article. Carroll et al. [1] in this same issue reported on common bile duct injuries during laparoscopic cholecystectomy that resulted in litigation. The reason I would like to discuss these two articles together is because they contain many similar problems. In Carrolls report, the authors analyzed 46 ductal injuries that resulted in litigation. Fifteen were transections, but only 20% of the injuries were discovered during the primary operation. The average delay in diagnosis was ten days. These complications occurred whether the operation was performed by inexperienced or experienced laparoscopic surgeons, but these particular data are incomplete. The statement of decreasing injuries with the learning curve experience can be questioned. Four injuries occurred between 50100 cases and five after experiencing over 100 cases. It is not uncommon that the learning period or learning curve explanation is debated. In the references 78,000 cases with a bile duct injury of 0.36 to 0.47% were quoted. In 16 of these 46 cases, cholangiograms were performed, but in 11 cholangiograms, the hold up of contrast material was obvious, but the surgeon misinterpreted (overlooked) the findings. The settlement cost varied from $30,000 to $1.3 million with an average of $221,000. The following questions arose while reading these two articles:

Ethical issues It is true that ductal injuries will always occur if we operate on the biliary system and we will probably never achieve a 0% incidence. However, all attempts should be made to decrease ductal injuries to a minimum. What should we learn from the last nine years to change our way of thinking or to make changes in practicing biliary surgery if it comes to LC? The first message is an ethical one. Are we informing the patient appropriately? There is no doubt that the incidence of ductal injuries was increased (by two to three times) since LC was introduced. Are we telling the patient that there will be less postoperative discomfort, shorter hospitalization stay, faster return to activities, but there is a

292

(slim) possibility that a significant complication with severe sequelae could occur more frequently as compared with an open operation? After an honest explanation of the pros and cons, should we not ask the patient about his or her choice? Patients are entitled to know that LC could have a higher complication rate, particularly with regard to ductal injuries.

Intraoperative cholangiography Intraoperative cholangiography has been debated since Mirizzi described it in 1933. I do not intend to go into endless pro and cons, but LC has changed the indications as well as the usefulness and importance of cholangiography. I understood the reluctance of the inpatient surgeons to wait 1520 minutes to receive the two or three films when a large percentage is noninformative or technically unacceptable. 1. We embrace modern high technology quickly in working from a TV screen with a new remote surgical technique instead of the open exploration and naked eye vision. Approximately 30% of abdominal surgery is performed today with laparoscopy. We have difficulties in accepting that there are digitized fluoroscopes available in which one does not extend the operating time more than 510 minutes because the image is immediately visible. The digitized system improved the quality of the display (information) significantly. I cannot underscore the value (in 10% of cases) of discovering anomalies of surgical importance in a timely fashion. A good example is the short cystic duct. If this is not recognized, the normal caliber CBD is pulled laterally and is therefore tented. This can mislead the operator to misinterpret the so-called cystic duct, which is really a CBD, which will be double-clipped and transected. This appearance can be discovered immediately with fluorocholangiography (cholangio grasper proximity to the main duct). In 2% of cases, the short cystic duct can drain into the right hepatic duct or a spiral cystic duct crosses very near the common hepatic duct. Injuries can be avoided if attention is drawn in time to dangerous anomalies at the primary operation. 2. The problems are enhanced if the case is difficult or is an acute one where efforts made to clarify the anatomy in these cases are time-consuming. A cystic duct or cholecysto-cholangiogram could be of great help in identifying the structures. The cholangiogram will give you some hints. If you extend the operating room time only 10 minutes, the effort is worthwhile. Do not hesitate to convertand still perform a cholangiogram. 3. If the injury is discovered by reading the films properly, this should indicate the need for immediate exploration. Surgeons who cannot read gross changes on a cholangiogram (for instance stop of contrast material or extravasation) should not operate on the biliary system. The patient is much better off if immediately operated upon than to be re-explored a few days later, attempting then to repair the injury in an abdomen with biliary peritonitis. 4. A large number or the majority of preoperative of ERCPs can be avoided if routine fluorocholangiography is em-

ployed. This endoscopic procedure has a morbidity, and entails a significant additional cost. There is only one indication for a preoperative ERC: the high-risk patient with a severe comorbid condition (septicemia, cholangitis, or jaundice), where ERC should be the first step. 5. The majority of surgeons abrogated the removal of CBD stones and refer it to the endoscopist for a second procedure with additional morbidity and mortality, not speaking about the cost. There is no excuse for a surgeon not acquiring the additional skill or the equipment. Choledocholithiasis is a surgical disease. If performed laparoscopically, fluorocholangraphy is obligatory, but this is a separate story. 6. There are also some lessons to be learned from the articles about the postoperative period. Only 1020% of injuries are discovered during the primary operation. The majority are operated late because of severe symptoms (peritonitis, jaundice, etc.). Stricture formation can occur later (62448 months) after the primary operation. The lesson is that if the patient on the first postoperative day has only vague symptoms, for instance, did not want to get out of bed in the morning, does not eat breakfast, does not feel 100%, this may indicate injury. Some routine laboratory examinations can be noninformative. The surgeon should immediately think of the possibility or try to exclude the presence of bile leak by ordering a Hida scan, and if positive, follow this with an emergent ERCP and immediate operation. If a normal caliber duct is transected or an injury found in an infected abdomen, the surgeon should secure a tube in the proximal lumen of the transected duct and drain the bile in a container outside. Drain also the abdominal cavity. The surgeon wins time and can perform the repair under better conditions at a later stageif the surgeon feels competent to do this special procedureor he or she can refer the case to a tertiary center. Another article of interest is one by Scha fer et al. [4], who made a retrospective analysis of 10,000 LCs with spilled stones. In 5.7%, this problem was discovered. It seems that in other studies it occurred more frequently. The complications of these events are abscess formation or, at a later stage, fistula formation. Therefore, all attempts should be made to recover spilled stones. Place the gallbladder in a bag during the pull-through maneuver, which could be another source. The authors discuss the indication for conversion, which should be considered if (large) or innumerable calculi are lost and are difficult to retrieve. The formation of abscesses in the authors experience was low (0.08%), but can occur. Again the patient should be informed of this event and the problem should be explained to draw attention to it. If early symptoms should occur, the patient should immediately report this to the surgeon. There is no question that the recommendation to retrieve all stonesif possibleis advisable. Dr. Marks et al. [2] reported an experimental study of the advantages of biliary stenting in cystic duct leakage without sphincterotomy. This study was performed in dogs and showed that those groups without sphincterotomy but only stented fared better than the group with sphincterotomy. Cystic stump leakage is unfortunately common after

293

LC and stenting alone (with less mobidity), which would decrease the resistance of the sphincter, facilitating free bile flow, and which would be sufficient prevention. Spontaneous closure of the cystic stump will occur. We should be aware that minimal access surgery can sometimes carry a maximal risk. We should know it; the patient should know it. Every attempt should be made to decrease this dreadful complication, avoiding the creation of biliary cripples by educating ourselves, developing better judgment, and having a more open mind for the use of routine fluorocholangiography and early converison. Recommended reading Berci G, Cuschieri A (1997) Bile ducts and ductal stones. Saunders, Philadelphia

References
1. Carroll BJ, Birth M, Phillips EH (1998) Common bile duct injuries during laparoscopic cholecystectomy that result in litigation. Surg Endosc 12: 310314 2. Marks JM, Ponsky JL, Shillingstad RB, Singh J (1998) Biliary stenting is more effective than sphincterotomy in the resolution of biliary leaks. Surg Endosc 12: 327330 3. Rego ly-Me rei J, Iha sz M, Szeberin Z, Sa ndor J, Ma te M (1998) Biliary tract complications in laparoscopic cholecystectomy: a multicenter study of 148 biliary tract injuries in 26,440 operations. Surg Endosc 12: 294300 4. Scha fer M, Suter C, Klaiber Ch, Wehrli H, Frei E, Kra henbu hl L (1998) Spilled gallstones after laparoscopic cholecystectomy: a relevant problem? A retrospective analysis of 10,174 laparoscopic cholecystectomies. Surg Endosc 12: 305309

George Berci
Cedars-Sinai Medical Center Los Angeles, CA 90048, USA

Surg Endosc (1998) 12: 338341

Springer-Verlag New York Inc. 1998

Bilateral thoracoscopic stapled volume reduction for bullous vs diffuse emphysema

J. K. Champion, J. B. McKernan
Dunwoody Medical Center, 4575 North Shallowford Road, Atlanta, GA 30338, USA Received: 14 May 1997/Accepted: 6 August 1997

Abstract Background: We compared our results with bullous vs diffuse emphysema by performing a bilateral thoracoscopic stapled volume reduction technique in 15 patients (age 45 80, 10 males, five females). Methods: Eight patients demonstrated bullous emphysema and seven patients diffuse emphysema. Lung reduction was performed with a bilateral thoracoscopic stapled technique utilizing bovine pericardium in the supine position. Results: Comparison of the bullous versus diffuse groups revealed no significant differences in means for the following variables: length of air leak (7.5 vs 3.3 days); length of stay (8.1 vs 6.5 days); pre-op FEV1 (23% vs 22%); pre-op dyspnea index (3.4 vs 3.6). At 3 months the bullous subset had a highly significant improvement (p < 0.007) in FEV1 (88%) compared with the diffuse subset FEV1 (59%). Conclusions: These early results suggest that patients with bullous emphysema are at no greater risk and demonstrate a significantly greater improvement in FEV1 than patients with diffuse emphysema. Key words: Thoracoscopic Emphysema Lung volume reduction

ploying a simultaneous stapled technique via a median sternotomy [4]. By surgically reducing the hyperinflated lung volume associated with advanced chronic obstructive pulmonary disease (COPD), significant improvement in patients pulmonary function and quality of life has been demonstrated [4, 5]. Advances in videoscopic instrumentation and technology allowed a thoracoscopic approach to be employed by a number of surgeons with outcomes equivalent to the open technique [1, 6, 7, 9, 10, 12]. Patient selection criteria vary, however, and several centers exclude patients with bullae greater than 5 cm [1, 7, 10, 11]. It has been stated that patients with bullous disease develop severe air leaks which prolong hospitalization and lead to an unsuccessful result. In this study, we compared our results in performing a bilateral thoracoscopic stapled volume reduction in patients with diffuse vs bullous emphysema to determine whether multiple bullae greater than 5 cm. should exclude patients as candidates for surgery. Materials and methods Patient population
From November 1995 through April 1996, 15 patients underwent bilateral thoracoscopic stapled volume reduction. There were 10 males and five females, ages 4580. All patients had end-stage emphysema with significant functional impairment due to dyspnea and had been on maximal medical management. Supplemental oxygen, at rest, was required by 53%. Steroids were used by 40% of candidates. Preoperative evaluation consisted of pulmonary function tests, arterial blood gas, chest X-ray with computed tomography (CT), selected ventilation-perfusion (V/Q) scans, and evaluation of dyspnea index by Modified Medical Research Council Dyspnea Scale [14]. Eight patients demonstrated multiple bullae greater than 5 cm without compression of surrounding emphysematous lung on CT and were classified as having bullous emphysema. Seven patients exhibited diffuse emphysema alone. Diffuse emphysema was defined as generalized parenchymal destruction with no distinct bullae. Eligible patients demonstrated hyperinflation with a heterogeneous pattern of parenchymal involvement on radiologic evaluation. Eligibility criteria are listed in Table 1.

Lung volume reduction (LVR) surgery for emphysema was first reported by Brantigan in 1960 utilizing sequential staged open thoracotomies and suturing [2]. The associated mortality (16%) reported with this technique discouraged its widespread adoption despite the symptomatic improvement reported by the majority of survivors. The modern era of lung volume reduction began with the report of Cooper and associates in 1992 when they described their results emPresented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 2021 March 1997 Correspondence to: J. K. Champion, 130 Vann Street Suite 220, Marietta, GA 30060, USA

Operative technique
Patients were admitted the morning of surgery. A double-lumen endotracheal tube and general anesthesia were employed to allow isolated venti-

339 Table 1. Eligibility criteria for thoracoscopic stapled volume reduction Chest X-ray with hyperinflation Heterogeneous pattern on chest CT or V/Q scan Forced expiratory volume in 1 s (FEV1) <35% Total lung capacity (TLC) >125%1 Respiratory volume (RV) >250% Nonsmoker pCO2 < 65 mm Marked impairment in physical activity despite maximal medical management No pulmonary hypertension >35 mm Mean No major coexisting medical problems lation to either lung. A flexible bronchoscopy is routinely performed to rule out an occult intrabronchial malignancy and to confirm proper positioning of the double-lumen endotracheal tube. Patients were placed in the supine position, which in our experience is better tolerated hemodynamically than a lateral decubitus position and reduces operating time since both thoraces are accessible without repositioning and redraping. An initial thorax is chosen to begin, and the lung was deflated with one-lung ventilation to the contralateral thorax. A three-port videoscopic technique was employed to access each pleural space. The pleural space was initially entered in the anterior axillary line at the level of the nipple with a 12-mm Optiview trocar (Ethicon Endosurgery, Cincinnati, Ohio) and 10-mm 0 videoscope. This allows direct visualization as the pleura is entered to avoid adhesions and prevent inadvertent lung injury which can increase air leaks. After the pleural space was entered, inspection identified the most diseased areas of lung as they remained hyperinflated, while more normal areas collapsed from absorptive atecletasis. Insufflation to 5 mm was begun to aid in more rapid collapse of the lung and was then discontinued. No adverse hemodynamic effects have been noted with insufflation. Two 5-mm ports were then inserted in the anterior midclavicular line in the second and sixth interspaces. The 10-mm 0 videoscope was then exchanged for a 5-mm 30 videoscope which was inserted in the lower anterior port. The optical trocar was removed and a 20-mm flexible Thoracoport (Ethicon Endosurgery, Cincinnati, Ohio) was reinserted at the same lateral site. Utilizing the CT scan and visual inspection, we identified the area for planned resection, attempting to remove 2030% of the lung volume of each lung. A 45-mm thoracoscopic stapler (Ethicon Endosurgery, Cincinnati, Ohio) with bovine pericardial strips was then utilized to perform the excision-plication along a linear continuous line. An area approximately 5 cm 15 cm was removed, which usually requires five stapler firings per side. The excised tissue was removed via the Thoracoport and examined pathologically. The videoscope was transferred to the upper 5-mm anterior port and a single 24-Fr straight chest tube was inserted via the lower 5-mm port and positioned at the apex of the chest. The lung was inflated under direct visualization and inspected for air leaks, and the chest tube was connected to an underwater seal. Suction of 10 cmH2O is utilized only for massive SQ emphysema or for a pneumothorax greater than 30%. The same procedure is then repeated on the contralateral lung. Operative time averaged 90 min. All patients were extubated in the operating room and none required mechanical ventilation. Patients were monitored in ICU for 2472 until hemodynamically stable. Physical therapy and ambulation, twice daily, were begun on postop day 1. The chest tube was connected to a Heimlich valve for prolonged air leaks (>5 days). Intensive chest physiotherapy was utilized and perioperative antibiotics were given for 35 days. Pain management was by either thoracic epidural analgesia or by Patient-Controlled Anesthesia (PCA)intravenous narcotics by patient choice for the first 72 h, then converted to oral medications. Table 2. Complications of thoracoscopic stapled volume reduction Complication Air leak over 7 days Arrhythmia Panic attack Urinary retention No. of patients (%) 4 (27) 1 (6) 1 (6) 8 (54)

leak over 7 days in duration. Upon discharge, patients were instructed on a home exercise program without formal pulmonary rehabilitation. At 3 months postoperatively all patients underwent reevaluation with pulmonary function tests and assessment of dyspnea index. Preoperative pulmonary function tests and dyspnea index are listed in Table 3 along with in-hospital variables of length of air leak, number of stapling firings, and length of stay. Comparison of the means for the bullous vs diffuse groups revealed no significant differences. Postoperative evaluation at 3 months revealed a highly significant improvement in pulmonary function and dyspnea index for both groups (Table 4). However, the bullous subset had a highly significant (p < 0.007) improvement in FEV1 compared with the diffuse subset. Discussion As early results of lung volume reduction surgery were reported, questions arose as how to best define the limits for the procedure [13]. The primary areas of controversy include: (1) What is the extent of preoperative evaluation? (2) What parameters should be used to identify who is a candidate? (3) What surgical approach should be used? (4) What is the role of pulmonary rehabilitation? Our approach has been to simplify the evaluation and attempt to refine the technique so that it may be offered to the greatest number of patients who may benefit, and to provide this in a community setting. The extent of preoperative evaluation is important to properly identify those patients who may benefit from lung reduction surgery, but excessive testing could result in an unnecessary socioeconomic burden due to the large number of potential candidates eligible to be screened. Our approach has been to simplify the preoperative evaluation by routinely performing only pulmonary function tests, arterial blood gases, and chest X-ray with computed tomography. Ventilation-perfusion lung scans were utilized only in selected patients with diffuse emphysema to identify target areas for resection. Two diffuse emphysema patients had a pattern on CT scan that was ambiguous; therefore, a V/Q scan confirmed a heterogenous pattern suitable for resection. In patients with bullous emphysema, the CT scan was sufficient to direct the resection. A cardiac evaluation was obtained only if a history of unstable heart disease was present. There appears to be general agreement among reported series on the selection criteria listed in Table 1 [1, 57, 9, 12]. We disagreed, however, that patients with multiple bullae >5 cm without compression of underlying lung should be excluded from consideration [1, 7, 10, 11]. It has been established that patients with giant bullae that occupy over

Statistical analysis
Students paired t-tests were used to evaluate changes between groups in pulmonary function, length of stay, and length of air leak. Wilcoxon matched-pairs signed rank test was utilized to examine differences in dyspnea index between groups. A p value of less than 0.05 was considered significant.

Results There were no operative deaths. Complications are listed in Table 2. The major postop complication was a prolonged air

340 Table 3. Comparison of preoperative and in-hospital variables between groups Bullous Stapler firings Length of air leak Length of stay Pre-op FEV1 volume Pre-op FEV1 percent Pre-op dyspnea index 11 7.5 days (range: 135 days) 8.1 days (range: 313 days) 673 cc 23% (range: 1430%) 3.4 Diffuse 9.7 3.3 days (range: 19 days) 6.5 days (range: 310 days) 585 cc 22% (range: 1332%) 3.6

Table 4. Comparison of postoperative pulmonary function and dyspnea index Bullous Post-op FEV1 volume Post-op FEV1 Post-op dyspnea index 1,265 cc 44% (88% improvement) 1.3 Diffuse 930 cc 35% (59% improvement) 1.4

30% of a thorax and compress underlying lung tissue can benefit from resection [3, 8]. Our previous experience in thoracoscopically resecting bullae in emphysema patients after a spontaneous pneumothorax was encouraging, as patients reported subjective symptomatic improvement in preop dyspnea. We therefore included a bullous subgroup in our study, anticipating they might experience a higher incidence of air leak and length of stay. Our data indicated both groups postop had a highly significant improvement in FEV1 and dyspnea index. It was surprising, however, to see that the bullous group improvement (88%) was highly significant (p < 0.007) compared to the diffuse group (59%). Resecting 2030% of volume occupying bullae may, theoretically, remove proportionally less pulmonary parenchyma than the same resection in diffuse emphysema. Indeed, our diffuse groups [2] 59% improvement in FEV1 compares with Coopers 51% after median sternotomy [4, 5]. Our data further demonstrated that there was no significant difference in length of air leak, number of stapler firings, or length of hospital stay between the two groups. Our incidence of prolonged air leaks (27%) was approximately half of the incidence (46%) reported by Cooper [5]. We could not attribute any increase in air leaks in our study to the type emphysema or to steroid usage. There were two prolonged air leaks in both the diffuse and bullous subgroups. While 40% (two bullous, four diffuse) of our patients utilized steroids, only one prolonged air leak (diffuse patient) occurred in the steroid users, for an overall incidence of 17% (1/6). Indeed, the 30% prolonged air leak rate was higher in our nonsteroid group (3/9). A variety of surgical approaches have been advocated. Controversy exists as to whether an open or thoracoscopic incision is better and as to whether a unilateral, bilateral simultaneous, or bilateral staged procedure offers advantages. Our approach was to adopt a bilateral simultaneous thoracoscopic approach in the supine position, which we believe is less traumatic and therefore better tolerated. The

goal of removing 2030% of each lung by a stapled bovine pericardial technique can be accomplished thoracoscopically today. We believe the videoscopic technique we utilized involves less handling and manipulation of the lung, which may allow for reduced air leaks. We adopted a bilateral simultaneous approach with the philosophy of attempting to improve the patients as much as possible with one procedure. McKenna and associates found their results with a bilateral thoracoscopic approach were comparable to median sternotomy and had no increased mortality or morbidity [9]. Economically, the savings favor a single bilateral procedure over two separate admissions. The quality of improvement in patients pulmonary function with a bilateral volume reduction (57%) vs a unilateral (31%) is significant. The recovery from just one procedure allows patients to maximize recovery in a shorter time than staged procedures performed on separate dates, with the overall goal of lung reduction surgery to improve the patients quality of life. We omit a formal pulmonary rehabilitation program, which is one area of controversy in our approach to lung volume reduction surgery. Some authors exclude any patient from consideration unless they can complete pulmonary rehabilitation and perform 30 min of sustained exercise on a treadmill or bike [11]. We felt this requirement would restrict access to a potentially therapeutic modality and select out only the most physically fit patient, who it can be argued is least in need of the procedure. We utilized no pulmonary rehabilitation preoperatively but insisted on early ambulation and exercise postoperatively. The patients were instructed in the importance of exercise and given goals for twice-daily exercise at home. This was monitored as an outpatient and gradually increased to a goal of 30 min sustained exercise within 3 months of discharge. Our patients experienced no mortality or ventilatory support, and our functional results with improvement in FEV1 and dyspnea index appear comparable with other studies which utilized rehab [1, 7, 10, 11]. We believe postop exercise is important for good functional results but may be accomplished economically at home with proper instruction. Finally, all of our procedures were performed in a community hospital rather than academic teaching institutions. Our hospital charges averaged $30,000 compared with the $50,000 reported by some centers [13]. Ultimately, questions regarding surgical approach, role of pulmonary rehabilitation, and further refining of selection criteria may be clarified in the future with pending randomized prospective studies proposed by Medicare. This report demonstrates that for a bilateral thoracoscopic stapled volume reduction, patients with bullous emphysema are at no greater risk and demonstrate a significantly greater improvement in pulmonary function than patients with diffuse emphysema.

References
1. Bingisser R, Zollinger A, Hauser M, Bloch K, Erich R, Weder W Bilateral volume reduction for diffuse emphysema by video assisted thoracoscopy. J Thorac Cardiovasc Surg 112: 875882 2. Brantigan OC, Kress MB, Mueller EA (1961) The surgical approach to pulmonary emphysema. Dis Chest 39: 485501 3. Connolly JE, Wilson AF (1989) The current status of surgery for bullous emphysema. J Thorac Cardiovasc Surg 97: 351361

341 4. Cooper JD, Trulock EP, Triantafillou AN, Patterson GA, Pohl MS, Deloney PA, Sundaresan RS, Roper CL (1995) Bilateral pneumectomy (volume reduction) for chronic obstructive pulmonary disease. J Thorac Cardiovasc Surg 109: 106119 5. Cooper JD, Patterson GA, Sundaresan RS, Trulock EP, Yusen RD, Pohl MS, Lefrak SS (1996) Results of 150 consecutive bilateral lung volume reduction procedures in patients with severe emphysema. J Thorac Cardiovasc Surg 112: 13191330 6. Eugene J, Ott RA, Gogia HS, et al. (1995) Video-thoracic surgery for treatment of endstage bullous emphysema and chronic obstructive pulmonary disease. Am Surg 61: 933936 7. Keenan RJ, Landreneau RJ, Sciurba FC, Ferson PF, Holbert JM, Brown ML, Fetterman LS, Bowers CM (1996) Unilateral thoracoscopic surgical approach for diffuse emphysema. J Thorac Cardiovasc Surg 111: 308315 8. Laros CD, Gelissen HJ, Bergstein PG (1986) Bullectomy for giant bullae in emphysema. J Thorac Cardiovasc Surg 35: 480487 9. McKenna RJ, Brenner M, Fischel RJ, Gelb AF (1996) Should lung volume reduction for emphysema be unilateral or bilateral. J Thorac Cardiovasc Surg 112: 11311339 10. McKenna RJ, Brenner M, Gelb AF, Mullin M, Singh N, Peters H, Panzera J, Calmese J, Schein MJ (1996) A randomized prospective trial of stapled lung reduction versus laser bullectomy for diffuse emphysema. J Thorac Cardiovasc Surg 111: 317322 11. Miller JJ, Lee RB, Mansoor KA (1996) Lung volume reduction surgery: lessons learned. Ann Thorac Surg 61: 14641469 12. Naunheim KS, Keller CA, Krucylak PE, Singh A, Ruppel G, Osterloh JF (1996) Unilateral video assisted thoracic surgical lung reduction. Ann Thorac Surg 61: 10921098 13. Rusch VW (1996) Lung reduction surgery: a true advance? J Thorac Cardiovasc Surg 111: 293295 14. Task Group on Screening for Respiratory Disease in Occupational Settings (1982) Official statement of the American thoracic society. Am Rev Respir Dis 126: 952956

Discussion Dr. Greene: In my discussion with individuals doing this procedure, the real commitment is ICU care and the cost involved. These patients can really take over your entire unit if youre not careful. Could you comment on the length of ICU management in this population of patients? Dr. Champion: Our routine is to place the patients in the ICU at least overnight. The average stay is usually one to three days. We looked at our cost for a procedure. For these first 15 patients we averaged $29,000 for the hospitalization, compared to $55,000 involved with an open sternotomy. Certainly postoperatively, because of the changes in V/Q shunt, theres a real roller-coaster for the first 72 hours as the body is adjusting. The pO2s run anywhere from 40% up to 90%, and part of this is just training the nursing staff to allow the patients to adjust and not to overreact. We basically just ignore the pO2 postop.

Surg Endosc (1998) 12: 322326

Springer-Verlag New York Inc. 1998

How, when, and why bile duct injury occurs

A comparison between open and laparoscopic cholecystectomy
E. M. Targarona,1,3 C. Marco,2 C. Balague ,1 J. Rodriguez,2 E. Cugat,2 C. Hoyuela,2 E. Veloso,1 M. Trias1,3
1 2

Service of Surgery, Hospital Clinic, Villarroel 170, 08036 Barcelona, Spain Service of Surgery, Hospital de Mutua de Terrassa, Garca Humet 2, Terrassa, University of Barcelona, Barcelona, Spain 3 Service of Surgery, Hospital de S. Pau. Avda. P. Claret 167, 08025 Barcelona, Spain Received: 6 June 1997/Accepted: 8 September 1997

Abstract Background: Bile duct injury (BDI) is a severe complication of laparoscopic cholecystectomy (LC). There is general agreement about the increase of this complication after LC vs open cholecystectomy (OC), but comparative studies are scarce. The aim of this paper has been to compare the incidence and clinical features of BDI after LC vs open procedures. Materials and methods: 3,051 OC, performed from June 1977 to December 1988 were retrospectively analyzed and compared with 1,630 LCs performed from June 91 to August 96, for which data were prospectively recorded. Age, sex, type of BDI, performance of intraoperative cholangiography (IOC), underlying biliary pathology, morbidity, mortality, and late morbidity were all analyzed. Results: BDI incidence was higher in group II (LC) (N: 16, 0.95%) than in group I, (OC, N: 19, 0.6%). BDI incidence was also higher in the group of patients in which it was necessary to convert to an open procedure (3/109, 2.7%, p < 0.05). BDIs were more frequently diagnosed intraoperatively in group I (OC, 18/19) than in group II (LC, 12/16). In both groups, BDI was more prevalent in cases operated by staff surgeons than residents, mainly in complicated gallbladder patients, with a bile duct of less than 7-mm diameter. Morbidity, postoperative stay, mortality, and late morbidity were similar after a BDI in both types of approach. Conclusion: (1) BDI increases with LC. (2) BDI after LC carries a similar postoperative morbidity and mortality to those after OC. (3) Incidence of BDI in converted cases increases significantly and this constitutes a high-risk group. Key words: Gallstones Open cholecystectomy Laparoscopic cholecystectomy Bile duct injury

Because of the benign nature of biliary surgery for gallstones, it should be followed by minimal morbidity and mortality. With the developmen of laparoscopic cholecystectomy (LC), a smooth and quick recovery is added. However, severe complications can occur and bile duct injury (BDI) is devastating after cholecystectomy. Soon after reports of early series of laparoscopic cholecystectomy, several authors commented on the quick rise in the incidence of severe BDI [6, 8, 11, 15] (Table 1). The special characteristics of laparoscopic surgery and the modifications of the standard surgical rules used for years in biliary surgery (routine colangiography or drainage, transcystic bile duct exploration) have modified the incidence, mechanisms, and clinical features of BDI [2, 5, 10, 12, 17, 19, 20]. However, reports of BDI after LC are subject to many biases. Complications tend to be underreported, especially when a new technique is being developed. Early reports come from personal experience or from specialized centers, which may downplay the real incidence in the general surgical community [2, 5, 9, 19]. Surveys frequently fail to receive a sufficient number of answers, and they mainly cover severe cases [10, 20]. Few reports have tried to compare the incidence and features of BDI in open vs laparoscopic surgery. In 1992 the present authors performed a retrospective analysis of the incidence of BDI in a series of 3,051 cholecystectomies. The results of this series have been used as a reference for comparison with a prospective series of 1,600 LCs performed in a similar setting [21, 22]. Materials and methods Open operation
Some 3,051 open cholecystectomies performed between June 1977 and December 1989 in the Department of Surgery of the Hospital de Mutua de Terrassa were retrospectively analyzed. Intraoperative cholangiography (IOC) and subhepatic were routinely used. All the reports of the operations were scanned for a description of an extrahepatic bile duct injury, defined as the partial or total involuntary section of the duct with or without loss of substance. Rupture of the

Presented as a free paper at the Fifth Meeting of the E.A.E.S., Istanbul, Turkey, 1721 June 1997 Correspondence to: E. M. Targarona

323 Table 1. Incidence of BDI in cumulative series of laparoscopic cholecystectomies McMahon [13] n Open cholecystectomy Laparoscopic cholecystectomy (single centre experiences (>300 cases)) Laparoscopic cholecystectomy (multicentric series) Open cholecystectomy (multicentric series and surveys) Laparoscopic cholecystectomy (multicentric series and surveys) 66,163 11,978 136,816 nBDI 160 36 634 % 0.2 0.3 0.5 Range (00.5) (01.8) (0.10.9) 25,544 124,433 81 650 0.32 0.52 (00.9) (02.3) n Strasberg [21] n BDI % Range

Table 2. Comparative analysis of BDI between open and laparoscopic cholecystectomy Laparoscopic Open (n:3051) No. bile duct injuries Transection bile duct Age Sex (m/f) Diagnosis of BDI by intraoperative cholangiography Complicated clinical presentation of gallstones Diameter bile duct <7 mm Morbidity Mortality Postoperative stay (days) Late morbidity BDI, bile duct injury; LC, laparoscopic cholecystectomy. 19 (0.6%) 6 (0.2%) 65 17 8/11 67% 70% 63% 50% 5% 25 15 35% Completed LC (n:1,561) 13 (0.8%) 4 (0.24%) 59 14 5/8 40% 53% 77% 46% 7.7% 23 26 17% Converted LC (n:109 [6.5%]) 3 (2.7%) 1 (10.9%) 72 9 2/1 66% 33% 33% 33% 28 24 50% Overall LC (n:1,670) 16 (0.95%) 5 (0.29%) 61 14 7/9 33% 56% 68% 43% 12% 24 25 25%

gallbladder was not considered a BDI. Dehiscence after biliodigestive anastomosis, postoperative pancreatitis, and biliary leak after T-tube removal were excluded. Results of this series have been published elsewhere [21, 22].

operations were classified into three BDI risk categories. Risk I: Cholecystectomy without inflammatory signs, Risk II: Cholecystectomy associated to bile duct exploration. Risk III: Acute cholecystitis, scleroatrophic gallbladder, or when the surgeon stated the technical difficulty of identifying Calots triangle structures. Results were analyzed using the chi2 test or Students t-test.

Laparoscopic cholecystectomy
This group included 1,670 laparoscopic cholecystectomies; 1,142 were operated on in the Hospital de Mutua de Terrassa between June 1991 and June 1996. The other 528 patients underwent operations between January 1992 and July 1996 in the Hospital Clinic of Barcelona and data was stored in a prospective database. Intraoperative cholangiography was performed selectively. Patients suspected of having bile duct stones received a preoperative ERCP. No liver bed drainage was used routinely. For analysis of the laparoscopy group, it was divided between patients with completed LC and patients whose LC was converted to open surgery. Gender, age, preoperative diagnosis, operative diagnosis, experience of the surgeon, technical difficulty encountered in the operative field, intraoperative cholangiography, diameter of the bile duct, anatomy of the bile tree, early and late morbidity, and mortality were all recorded.

Results Open operation During the study period, 3,051 cholecystectomies for gallstones were performed. In 685 cases, a surgical exploration of the bile duct was added (23%). Intraoperative cholangiography was performed in 93.4% of patients. Nineteen BDIs were found, an incidence of 0.6% (19/3,051). Mean age was 65 17 years (2984). Gender distribution was similar (eight males and 11 females) (Table 2). Three BDIs occurred at the common bile duct and seven at the hepatic duct. Three others occurred in the right hepatic duct, and one each in the left hepatic duct, bifurcation, and cystic and hepatic confluence. In 18 patients, the BDI was diagnosed intraoperatively; and in one case, in the immediate postoperative period (high-output biliary fistula). Anomalous anatomy was not found in any case. In 12 out of the 19 cases, the diameter of the bile duct was under 10 mm. It is noteworthy that in six of seven patients with a severe BDI, the bile duct size was less than 7 mm. In 12 out of the 19 patients (66%), IOC helped to find the BDI, but in the

Definitions
BDI has been classified in two categories, depending of the tme of diagnosis [10]: Group 1. BDI diagnosed during the surgical procedure: 1.a. Severe: This included transection of the bile duct or loss of substance of more than 25% of the circumference. 1.b. Mild: Longitudinal tears or puntiform defects with a minor loss of substance. Group 2. BDI injury diagnosed in the early postoperative period, usually as biliary leakage, bile peritonitis, or jaundice and cholangitis. In order to analyze the impact of technical difficulty on BDI, the

324 Table 3. Bile duct injury incidence according to the existence of intraoperative difficulties Open cholecystectomy N Cholecystectomy without incidence Cholecystectomy plus bile duct examination Difficult cholecystectomy 4/1839 6/683 8/524 % 0.2 0.9 1.5 Laparoscopic cholecystectomy N 8/1069 0/30 8/601 % 0.74 0 1.3 Complete transection Partial section Longitudinal tear Puntiform injury Posterior wall injury Cystichepatic junction Unknown Table 5. Types of bile duct injury Open cholecystectomy (n:19) 6 1 4 4 3 1 Laparoscopic cholecystectomy (n:16) 5 2 4 5

Table 4. Etiology of bile duct injury Open cholecystectomy (n:19) Cystic duct confusion Hartmanns pouch dissection Difficult hemostasia Cystic duct cannulation Choledochotomy Instrumental perforation Others 4 9 1 1 1 2 Laparoscopic cholecystectomy (n:16) 5 4 3 4

Table 6. Technique for repairing the bile duct injury Open cholecystectomy (n:19) End-to-end anastomosis Hepaticojejunostomy Suture of a T tube T tube Conservative Suture simple Choledocoduodenostomy Cystic ligature 3 3 7 1 4 1 Laparoscopic cholecystectomy (n:16) 3 4 3 4 2

others the clue was the appearance of bile in the operative field. Operative findings in this group were acute cholecystitis in ten, bile duct stones in two, atrophic gallbladder in one, and chronic cholecystitis in the other six. BDI injury occurred more frequently in patients with a complex intraoperative field than during a noninflammatory cholecystectomy (p < 0.05) (Table 3); and also when trained surgeons rather than surgical residents did the operations (16/2,051 [0.8%] vs 3/1,000 [0.2%], p: NS). The mechanism of BDI and treatment are summarized in Tables 46. Nine of the patients (50%) developed postoperative complications and one patient died after an hepaticojejunostomy. Fourteen of these patients were followed up for a mean of 33 months (5 months to 9 years). Five of them developed late complications (cholangitis [two], splenic abscess [one], bile duct stenosis [one], and choledocholithiasis [two]). Two of them underwent a second operation for bile duct stones and a splenic abscess was drained percutaneously.

Laparoscopic surgery Completed LC. Of the 1,670 LCs attempted, 1,561 were successfully performed, with a conversion rate of 6.5% (109 patients). IOC was performed on 426 (25%) of these patients. Preoperative ERCP was performed on 334 (20%) patients suspected to harbor bile duct stones, and intraoperative exploration of the bile duct (transcystic or choledochotomy) was done in 30 (1.7%) patients (Table 2). Thirteen BDI injuries were identified (0.8%). Mean age was 59 14 years. Nine cases were diagnosed perioperatively (69%). Five of them were severe BDI and four were

minor injuries. IOC as performed in 37% of these cases and helped to the diagnosis of the BDI in two of five cases. Transection of the bile duct occurred in four cases and partial transection in one, in all of them when it was confused with the cystic duct. In one case, the hepatic duct was injured when dissecting Hartmanns pouch; in two cases, the confluence of cystic and hepatic duct was injured on attempting IOC. In a fourth case, the confluence of a short and parallel cystic duct that was inserted in the right hepatic duct was injured with the clip applier when attempting IOC. In all these cases, BDI was diagnosed intraoperatively: by IOC in two, by the presence of bile in the operative field, or by the observation of the uncorrectly clipped structure in the hepatic hilum. Five of these patients were converted, and the bile duct was repaired (three hepaticojejunostomies and two end-to-end reconstructions over a T tube). The other four cases were not converted, and were resolved with the laparoscopic insertion of a T tube (Tables 46). Four cases were diagnosed in the immediate postoperative period due to bile peritonitis and all of them were reoperated. In one case, an orifice in the hepatic bile duct was observed, and in the other three, a dislodgement of a clip and an orifice near the clip in the cystichepatic junction was identified. Two of them were treated with sutures over a T tube and two with ligature of the cystic stump. In the group as a whole, preoperative and intraoperative diagnosis in eight of the patients with BDI were noncomplicated gallbladders; meanwhile, complicated manifestations of gallstones (acute biliary pancreatitis [PAB], acute cholecystitis, bile duct stones) appeared in the other eight (Table 3). In 77% of cases, the bile duct was below 7 mm in diameter. Anomalies of the biliary tree related to the development of a BDI existed in two cases: a low confluence of the hepatic ducts with the cystic duct draining into the right hepatic duct; and a Mirizzi syndrome that facili-

325

tated the confusion of the cystic duct with the bile duct due to the presence of a stone located in the bilobiliary fistula. All BDIs occurred in cases operated by surgeons well trained in laparoscopic surgery with experience in over 50 LCs. Forty-six percent of the patients developed early postoperative morbidity: bile leakage (two), impossibility of retrieving the T tube due to intraabdominal adhesions that required a minilap for retrieval, and dehiscence of hepaticojejunostomy. One patient died after multiorgan failure. Mean stay was 23 23 days (range 6103). After a mean follow-up of 36 months, two patients developed late morbidity (incisional hernia and bile duct stenosis that required hepaticojejunostomy). Converted patients. In three out of 109 (2.7%) patients that were converted a severe BDI occurred (Table 2). The preoperative diagnoses were acute cholecystistis, biliary pancreatitis, and bile duct stones. In all three patients, the reason for conversion was a severe distortion of anatomy. In two patients, a complete transection of the bile duct occurred, and in one, a partial section. Two of them were treated with reconstruction over a T tube, and one with Roux Y hepaticojejunostomy. One patient died after reoperation due to hemorrhage, and one developed a biliary fistula that closed spontaneously. Postoperative stay was 28 24 days. One patient developed an incisional hernia that need surgical repair several months later. Discussion BDIs are severe complications of biliary surgery mainly because they cause serious morbidity and are potentially lethal. This effect is increased when BDI occurs after LC, a surgical procedure designed to increase postoperative comfort and to shorten the postoperative period. Soon after widespread use of LC, several authors detected a higher rate of BDI referral to specialized centers [6, 8, 11]. However, objective data to support this finding, just as some years ago with open BDI, is not easy to find. Analysis from cumulative series seems to indicate that LC carries a higher risk of BD injury (Table 1). A more accurate and simple way of identifying rise in BDI incidence is to compare operations performed in similar settings with similar indications, as this paper does. To date, five papers have reported data [4, 7, 13, 14, 16, 18]. Brune et al. [4] reported a similar incidence of BDI (0.2%) after LC or OC in an analysis of an institution, but there was a bias because LC included simpler cases than OC. Morgenstern et al. [13, 14] showed, for a series of 3,000 LCs, twice the incidence of BDI (0.5 vs 0.2) after LC compared to after OC. Three surveys (Connecticut, Holland, and Norway) [7, 16, 18] showed higher incidence of BDI after LC, but analysis of one of them over a 3-year period (1991-93) [16] revealed higher initial incidence of BDI, which declined in the 2nd year and was comparable to OC in the 3rd year. These findings suggest a learning curve adaptation. An evident fact is that the spectrum of BDI has changed. The use of diathermic hooks with more dissection in the

Calot triangle, the use of clips for cystic duct ligature and, mainly, the special features of laparoscopic surgery have modified the characteristics of the BD injuries. The most serious lesion is transection of the bile duct. The cause is usually misidentification of the cystic duct. The lesion can be linked to resection of a segment of bile duct and to vascular injury. This lesion occurred in our series two times more often after LC than in the open group, just as other studies found [7, 10, 13]. The lesion is usually a consequence of the inadequate exposure of the confluence of the cystic duct and Hartmanns pouch. Although all six cases were diagnosed during operation, IOC enables the lesion to be detected before the duct is severed. The role of IOC in preventing BDI is controversial. Some authors suggest that in spite of the existence of a BDI, IOC often shows as normal or is not properly interpreted by the surgeon [3, 9]. In Morgensterns series [13, 14], IOC diagnosed 11 of 12 BDIs. Woods et al. [25], in a cumulative report on 177 BDIs after LC, showed that more BDIs were diagnosed perioperatively in patients on whom an IOC was performed and that these patients needed fewer surgical repairs than patients diagnosed later. In our series, IOC diagnosed 67% of cases in the open group and two out of five in the laparoscopic group. However, BDI during LC can occur after a normal IOC. Dissection of the adherent Hartmanns pouch, excessive electrocautery dissection near the bile duct, or difficulties in the management of the cystic duct stump can induce BDI. During open surgery, IOC was a rare source of BDI [24], but during LC, lesions of the cystic duct or at the confluence can occur because the catheters for IOC are straighter and more rigid and the insertion maneuvers are more difficult. In some series of BDI after LC, up to 20% are caused by an attempted IOC [23]. In this series, a tear in the bile duct occurred while dissecting in order to have enough length of cystic duct. It was repaired with the laparoscopic insertion of a T tube. Immediate IOC showed the anomaly of a short cystic duct that connected to a right hepatic duct. BDI can be diagnosed in the immediate postoperative period. The avoidance of routine drainage and the more liberal use of cautery or clips may cause delayed leaks not seen during LC. In our series, one patient was diagnosed in the early postoperative period in the open group (0.03%), but there were four cases in the LC group (0.23%). All these cases needed reoperation due to bile peritonitis. A selective policy of drainage is advisable, especially in cases with difficult dissection or wide, raw, or denuded subhepatic areas. Two factors that can cause BDI during open surgery are commonly accepted: local anatomy and the experience of the surgeon. In the open series, BDI increased sharply both in difficult cases and in the hands of more experienced surgeons, probably because difficult operations were performed by these surgeons. The analysis of our series, also shows that technical difficulty arising from anatomy distortion and scarring affected the incidence of BDI. Russell et al. [16], in a state-wide analysis, also showed that previous acute cholecystitis (odds ratio:3.3) and biliary pancreatitis (odds ratio:3.6) increased significantly the risk of BDI (p < 0.01). Surgeons experience in LC is more difficult to assess accurately because it involves the learning curve of the en-

326

tire surgical team. Some initial data revealed that the most important risk of BDI was during the first dozen LCs; the analysis of the Connecticut LC records revealed a trend toward lower incidence of severe BDI after the wide implementation of LC [16]. An interesting analysis of 8,800 LCs performed by The Southern Surgeons Club [23] showed that the risk of BDI caused by an experienced surgeon fell from 1.7% during the first case to 0.17% after the first 50. But analysis of the present and other data indicate that risk of a BDI exists despite adequate training due to the special features of laparoscopic surgery. Morgenstern et al. [14] observed in a series of 3,000 LCs performed in the Cedars of Sinai Hospital a similar BDI rate (0.5%) in the first 1,500 LCs compared to the second series of 1,500. In the present studys series of LC, most BDI took place under surgeons who had done more than 50 LCs and sometimes many more. This strongly suggests that the risk of BDI is always present due to the features of laparoscopic surgery and in spite of the skill of the surgeon. Modifications of biliary anatomy can cause BDI, but are not frequent. In the open series, no case was considered secondary to biliary anomaly, but in two cases in the LC group a biliary anomaly favored BDI (cystic duct drainage in right hepatic duct and Mirizzi syndrome). In the study by Andren-Sandberg et al. [1] an anatomic anomaly was considered responsible for BDI only in 20% of BDI patients. As in the open group, severe BDI during LC occurred more frequently in nondilated ducts (<7 mm). No structure should be severed without confirmation that it is the cystic duct: In this context, selective IOC is likely to be extremely helpful. Most lesions in both groups were diagnosed perioperatively and repaired. No cases in this series were diagnosed more than 1 week after LC, either as stricture or stenosis. It is noteworthy that three cases were resolved by laparoscopy with the simple insertion of a T tube and IOC to monitor, and that a BDI does not neccessarily imply conversion. However, in both series (OC and LC) it is clear that BDI, in spite of primary repair, was associated with high immediate (50 vs 43%) and late morbidity (35 vs 25%) and prolonged hospital stay (25 vs 24 days). An important finding is that the incidence of BDI in converted patients was significantly higher than after OC or LC (2.5 vs 0.5 vs 0.8, p < 0.05). It has been suggested that conversion is a risk factor for complications after LC [23]. The incidence can probably be explained by the fact that the group of converted patients included acute situations (bleeding) in which the surgeon had to act quickly, or more probably because these were those more difficult cases with a distorted or inflamed anatomy. References
1. Andren-Sandberg A, Alinder G, Bengmark S (1985) Accidental lesions of the common bile duct at cholecystectomy: pre and perioperative factors of importance. Ann Surg 201: 328332

2. Asbun HJ, Rossi RL, Lowell JA, Munson JL (1993) Bile duct injury during laparoscopic cholecystectomy: mechanism of injury, prevention and management. World J Surg 17: 547552 3. Barkun JS, Fried GM, Barkun AN, Sigman HH, Hinchey EJ, Garzon J (1993) Cholecystectomy without operative cholangiography. Implications for common bile duct injury and retained common bile duct stones. Ann Surg 218: 371379 4. Brune IB, Scho nleben K, Omran S (1994) Complications after laparoscopic and conventional cholecystectomy: A comparative study. HPB Surg 8: 1925 5. Davidoff AM, Pappas TN, Murray EA, Hilleren DJ, Johnson RD, Baker ME (1992) Mechanism of major biliary injury during laparoscopic cholecystectomy. Ann Surg 215: 196202 6. Garden OJ (1991) Iatrogenic injury to the bile duct. Br J Surg 78: 14121413 7. Gouma DJ, Go PMNYH (1994) Bile duct injury during laparoscopic and conventional cholecystectomy. J Am Coll Surg 178: 229233 8. Keith RG (1993) Is the increasing frequency of laparoscopic bile duct injury justifiable? Can J Surg 36: 501502 9. Lorimer JW, Fairfull RJ (1995) Intraoperative cholangiography is not essential to avoid duct injuries during laparoscopic cholecystectomy. Am J Surg 169: 344347 10. McMahon AJ, Fullarton G, Baxter JN, ODwyer PJ (1995) Bile duct injury and bile leakage in laparoscopic cholecystectomy. Br J Surg 82: 307313 11. Moossa AR, Easter DW, Van Sonnenberg E, Casola G, DAgostino H (1992) Laparoscopic injuries to the bile duct. Ann Surg 215: 203208 12. Morgenstern L, Berci G, Pasternak H (1993) Bile leakage after biliary tract surgery. Surg Endosc 7: 432438 13. Morgenstern L, Wong L, Berci G (1992) Twelve hundred consecutive cholecystectomies before the laparoscopic era: Morbidity and mortality and general observations. Arch Surg 127: 400403 14. Morgenstern L, McGrath MF, Carroll BJ, Paz-Partlow M, Berci G (1995) Continuing hazards of the learning curve in laparoscopic cholecystectomy. Am Surg 61: 914918 15. Richardson MC, Bell G, Fullarton GM, West Scotland Laparoscopic Cholecystectomy Audit Group (WSLCAG) (1996) Incidence and nature of bile duct injuries following laparoscopic cholecystectomy: an audit of 5913 cases. Br J Surg 83: 13561360 16. Russell JC, Walsh SJ, Mattie AS, Lynch JT (1996) Bile duct injuries, 19891993. A statewide experience. Arch Surg 131: 382388 17. Schol FPG, Go PMNYH, Gouma DJ (1994) Risk factors for bile duct injury in laparoscopic cholecystectomy. Br J Surg 81: 17861788 18. Solheim K, Buanes T (1995) Bile duct injury in laparoscopic cholecystectomy. Int Surg 80: 361364 19. Stewart L, Way LW (1995) Bile duct injuries during laparoscopic cholecystectomy. Arch Surg 130: 11231129 20. Strasberg SM, Hertl M, Soper NJ (1995) An analysis of the problem of biliary injury during laparoscopic cholecystectomy. J Am Coll Surg 180: 101125 21. Targarona EM, Garcia-Olivares E, Mun oz E, Centeno J, Otal C, Marco ML (1991) Lesio n quiru rgica de la v a biliar. Cir Esp 49: 2027 22. Targarona EM, Garcia Olivares E, Marco C (1992) When and why bile duct injury occurs. HPB Surg 5(Suppl): 8 23. The Southern Surgeons Club, Moore MJ, Bennett CL (1995) The learning curve for laparoscopic cholecystectomy. Am J Surg 170: 55 59 24. White TT, Hart MJ (1985) Cholangiography and small duct injury. Am J Surg 149: 640643 25. Woods MS, Traverso LW, Kozarek RA, Donohue JH, Fletcher DR, Hunter JG (1995) Biliary complications of laparoscopic cholecystectomy are detected more frequently with routine intraoperative cholangiography. Surg Endosc 9: 10761080

Symposium
Co-chairs: Bruce Schirmer, L. William Traverso
Surg Endosc (1998) 12: 361373 Springer-Verlag New York Inc. 1998

SSAT/SAGES minimally invasive surgery

Advanced laparoscopic hepatobiliary surgery
Intraoperative screening for common duct stones: ultrasound or cholangiography? Nathanial J. Soper, St. Louis, MO, USA ERCP: when preop, when postop, ever intraop? Richard A. Kozarek, Seattle, WA, USA Transcystic CBD exploration Edward H. Phillips, Los Angeles, CA, USA Choledochotomy and CBD exploration Joseph B. Petelin, Shawnee Mission, KS, USA Laparoscopic treatment of hepatic cystic disease Lawrence Way, San Francisco, CA, USA Laparoscopic screening for RUQ tumors and hepatic metastases Fredrick L. Greene, Columbia, SC, USA Laparoscopic biliary bypass for unresectable malignancy Lee L. Swanstrom, Portland, OR, USA Laparoscopic hepatic resection Ricardo Rossi, Burlington, MA, USA

Introduction The Society for Surgery of the Alimentary Tract (SSAT) and the Society of American Gastrointestinal Endoscopic Surgeons (SAGES) jointly sponsored the following series of talks given during the 1997 Digestive Disease Week held in Washington, D.C. Experts with experience in minimally invasive surgery of the hepatobiliary tract, all members of either one or both of the sponsoring societies, presented information on topics considered to be at the cutting edge of this approach to diseases of the liver and biliary tract. These experts were chosen on the basis of their experience; often they are the recognized foremost authority on their particular topic in the country. As a result, the discussions were enhanced by the presenters individual clinical

experiences and understanding of the nuances of these treatments. These summaries attempt to convey the core knowledge presented during the symposium. Along with the selected readings at the end of each summary, they are an excellent overview of the topic and allow the reader to review the latest information with respect to the dynamic applications of minimally invasive surgery to this area of general surgery. Bruce Schirmer
Department of Surgery University of Virginia Health Science Center Box 181 Charlottesville, VA 22908, USA

362

Intraoperative screening for common bile duct stones

Ultrasound or cholangiography?
Intraoperative screening of the common bile duct for stones during laparoscopic cholecystectomy can be done using cholangiography or ultrasonography. Application of one of these modalities for screening of the duct may be performed either routinely or selectively, a debate that has been reawakened by the emergence of laparoscopic cholecystectomy. Recently, a number of studies have compared intraoperative cholangiography (IOC) to laparoscopic intracorporeal ultrasonography (LICU) for screening of the common bile duct during laparoscopic cholecystectomy. When performing cholangiography, radio-iodinated contrast material may be injected either into the gallbladder directly or into the cystic duct, and radiographic images may be obtained by static or dynamic (fluoroscopic) techniques. A consensus is emerging that fluoroscopic cholangiography performed with intubation of the cystic duct should be the primary method for IOC. The overall success rate for obtaining adequate images by laparoscopic intraoperative cholangiography varies from 60 to 98% with false-positive and false-negative rates of less than 5%. Potential negatives of cholangiography are the need to intubate the cystic duct, with the possibility of ductal injury occurring as a direct result of the intubation, and the time and cost expended on the examination itself. During open cholecystectomy, several studies showed intracorporeal ultrasonography to be more accurate than operative cholangiography in assessing the common bile duct (CBD) for stones (9799% vs 8994%). However, few surgeons adopted ultrasound for this purpose. Laparoscopic intracorporeal ultrasound has been used in several centers to scan the biliary tree during laparoscopic cholecystectomy. With intracorporeal ultrasonography, the transducer may be of higher frequency and improved resolution compared to that used with transabdominal ultrasonography. Ultrasonography may be used to map the extrahepatic ductal system, identify anomalies of ductal and vascular structures, quan-

tify ductal diameter, and visualize CBD stones and sludge. In experienced hands, laparoscopic intracorporeal ultrasonography appears to be more sensitive than cholangiography for demonstrating choledocholithiasis, whereas cholangiography may be more accurate for demonstrating biliary anatomy. Thus, the two studies appear to be complementary for the delineation of biliary disease. Several studies have shown laparoscopic ultrasonography to screen the bile duct much more rapidly than cholangiography and LICU costs approximately $145 less than IOC per examination (unpublished results). Despite these promising data, more clinical experience will be necessary to establish the appropriate role for laparoscopic intracorporeal ultrasonography for screening the common bile duct during laparoscopic cholecystectomy.

Selected readings
Cuschieri A, Shimi S, Banting S, Nathanson LK, Pietrabissa A (1994) Intraoperative cholangiography during laparoscopic cholecystectomy. Routine vs. selective policy. Surg Endosc 8: 302305 Jones DB, Soper NJ (1995) Results of a change to routine fluorocholangiography during laparoscopic cholecystectomy. Surgery 118: 693702 Orda R, Sayfan J, Strauss S, et al. (1994) Intraoperative ultrasonography as a routine screening procedure in biliary surgery. Hepatogastroenterology 41: 6164 Machi J, Siegel B, Zaren HA, et al. (1993) Operative ultrasonography during hepatobiliary and pancreatic surgery. World J Surg 17: 640646 Machi J, Sigel B, Zaren A, et al. (1993) Technique of ultrasound examination during laparoscopic cholecystectomy. Surg Endosc 7: 545549 Stiegmann GV, McIntyre RC, Pearlman NW (1994) Laparoscopic intracorporeal ultrasound: an alternative to cholangiography? Surg Endosc 8: 167172 Teefey SA, Soper NJ, Middleton WD, Balfe DM, Brink JA, Strasberg SM, Callery MP (1995) Imaging of the common bile duct during laparoscopic cholecystectomy: sonography versus videofluoroscopic cholangiography. Am J Roentgenology 165: 847851

Nathaniel J. Soper
Department of Surgery Washington University School of Medicine St. Louis, MO, USA

363

ERCP interaction with laparoscopic cholecystectomy

Alternatives Endoscopic retrograde cholangiopancreatography (ERCP) interaction with laparoscopic cholecystectomy (LC) is a seesaw which is contingent both upon local endoscopic and surgical expertise. In institutions in which endoscopists are uncertain of their diagnostic or therapeutic ERCP capabilities, any patient with the potential for a common bile duct calculus is likely to undergo a preoperative endoscopic cholangiogram with attendant attempt at some extraction if one is found. This is especially true if the surgeon performing the laparoscopic cholecystectomy has limited experience in performing intraoperative cholangiography (IOC), let alone transcystic or laparoscopic common bile duct exploration (LCBDE). Alternatives to preoperative ERCP for choledocholithiasis include intra- or postoperative ERCP or laparoscopic or open common bile duct exploration. Intraoperative biliary sphincterotomy, in turn, can be performed either conventionally or by means of an endoscopically visualized papillotome placed laparoscopically across the sphincter through the cystic duct. More commonly, however, ERCP and laparoscopic cholecystectomy interact in the postoperative setting in patients who develop symptoms (biliary colic, cholestasis, pancreatitis) following an uncomplicated laparoscopic cholecystectomy. Alternatively, IOC may demonstrate a bile duct stone for which retrieval is unsuccessful or not attempted.

Results Most reported series in which ERCP-LC interaction has been recorded have involved preoperative ERCP in patients with suspected common bile duct stones. Moreover, in a recently published prospective series evaluating endoscopic interaction with laparoscopic cholecystectomy, community endoscopists were 1.67 times more likely to perform pre- as opposed to postoperative ERCP. Prospectively comparing 780 ERCPs performed in the setting of LC, data from 14 biliary endoscopists at four academic medical centers was compared to that from 33 community-based gastroenterologists, 23 of whom had performed less than 500 total ERCPs. Four hundred ninety-three of these patients (63%) underwent preoperative ERCP for a variety of indications including abnormal liver function tests (90%), jaundice (50%), fever (33%), or pancreatitis (35%). Forty-three percent of these patients were found to have stones, suggesting that over half of these individuals incurred the risk and expense of the procedure with the only potential benefit being definition of biliary tract anomalies or preclusion of need for intraoperative cholangiography. Although there have been techniques described to effect prograde sphincterotomy through the cystic and common duct, most laparoscopists who utilize intraoperative ERCP approach the sphincter in a retrograde fashion. I find either technique cumbersome and logistically problematic. Intra-

operative ERCP can best be described as how to make a straightforward procedure difficult: Do it on-call in the OR; use suboptimal radiographic equipment; perform it in a supine position (rather than the usual prone position); and inflate the bowel, making the laparoscopy itself more difficult. In contrast to pre- or intraoperative ERCP, individuals and institutions with extensive endoscopic experience tend to shift ERCP interaction with LC to the postoperative setting. This is particularly true in institutions in which surgeons have experience with IOC and LCBDE. In the latter setting, successful stone retrieval precludes the risks associated with sphincterotomy or balloon dilation of the papilla. It also precludes the additive expenses of ERCP (room charge and professional fee) as well as the need for 12 days of additional hospital time. ERCPs in this setting should be limited to patients in whom attempted LCBDE proves unsuccessful or those patients who develop signs or symptoms of biliary calculi, leak, or stenoses postoperatively. This approach implies confidence in ones endoscopic capabilities because an unsuccessful LCBDE would then require open surgical intervention. In the previously cited review by Davis et al., 59% of ERCP interactions with LC occurred postoperatively in the four academic medical centers, whereas 69% of the ERCPs performed by community-based gastroenterologists were undertaken preoperatively. Our group, in turn, reported that 92% of ERCP-LC interactions in our institution were in the postoperative setting. In contrast to a 50% normal ERCP rate in series reporting a preponderance of preoperative exam, only 7% of our patient group had normal studies. Pathology included choledocholithiasis (31%), biliary leak and/or stricture (26%), cystic duct leak (6%), and papillary stenosis/microlithiasis (19%). In addition, six patients (11%) were studied for postoperative pancreatitis; four of these patients were found to have biliary gravel/microlithiasis in conjunction with sphincter dysfunction and two were felt to have passed common duct calculi.

Recommendations Given the above, and given the additive expense and risks of combining laparoscopic cholecystectomy with ERCP, I recommend that the latter be undertaken primarily in the postoperative setting. The latter presupposes that my laparoscopic colleagues have or will develop the capabilities to do both IOC and LCBDE. This will decrease the number of ERCPs needed because only patients with retained stones or postoperative symptoms will undergo a procedure. It will also decrease the potential problems that occasionally occur following long-term sphincter ablation. It does, however, imply an onus on the endoscopist: Be successful or the patient may require a third procedure. Each endoscopist has a variable degree of comfort level with this onus. Mine stops with Roux-en-y and perhaps Billroth II anatomy. I shift patients such as these into a preoperative study knowing that my failure and the failure of LCBDE to remove common duct stones can be overcome by conversion to an open procedure at time of LC. Moreover, ill patients with acute cholangitis and a subset with severe pancreatitis may have dramatic amelioration of their symptoms following success-

364

ful ERCP and stone removal. I tend to shift these patients into a preoperative setting also, reasoning that lifethreatening situation and need for urgent surgery can be controlled acutely, allowing a subsequent laparoscopic procedure electively. Ours is a codependent situation given that both the performance and the timing of ERCP in conjunction with LC will be contingent upon the players and their aptitudes. By necessity, practice patterns will be institutionally defined. Selected readings
Berci G, Morgenstern L (1994) Laparoscopic management of common bile duct stones: a multi-institutional SAGES study. Surg Endosc 8: 1168 1175 Davis WZ, Cotton PB, Arias R, et al. (1997) ERCP and sphincterotomy in the context of laparoscopic cholecystectomy. Academic and community practice patterns and results. Am J Gastroenterol 92: 597601 Esber EJ, Sherman S (1996) The interface of endoscopic retrograde cholangiopancreatography and laparoscopic cholecystectomy. Gastrointest Endosc Clin North Am 6: 5780 Kozarek RA (1993) Laparoscopic cholecystectomy: what to do with the common duct. Gastrointest Endosc 39: 99100 Kozarek RA, Gannan R, Baerg R, et al. (1992) Endoscopic approach to bile leak following laparoscopic cholecystectomy. Arch Intern Med 152: 10401043 Kozarek RA, Ball TJ, Patterson DJ, et al. (1994) Endoscopic biliary injury

in the era of laparoscopic cholecystectomy. Gastrointest Endosc 40: 1016 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 Rhodes M, Nathanson L, ORourke N, et al. (1995) Laparoscopic exploration of the common bile duct: lessons learned from 129 consecutive cases. Br J Surg 82: 666668 Rieger R, Wayand W (1995) Yield of prospective, noninvasive evaluation of the common bile duct combined with selective ERCP/sphincterotomy in 1390 consecutive laparoscopic cholecystectomy patients. Gastrointest Endosc 42: 612 Strasberg SM, Soper NJ (1995) Management of choledocholithiasis in the laparoscopic era. Gastroenterology 109: 320322 Strasberg SM, Callery MP, Soper NJ (1996) Laparoscopic surgery of the bile ducts. Gastrointest Endosc Clin North Am 6: 81106 Toouli J (1997) Preoperative endoscopic sphincterotomy. In: Berci G, Cuschieri A (eds). Bile ducts and bile duct stones. WB Saunders, Philadelphia, pp 116119 Traverso LW, Kozarek RA, Ball TJ, et al. (1993) ERCP after laparoscopic cholecystectomy. Am J Surg 165: 581586 Welbourn CRB, Mehta D, Armstrong CP, et al. (1995) Selective preoperative ERCP with sphincterotomy avoids bile duct exploration during laparoscopic cholecystectomy. Gut 37: 576579

Richard A. Kozarek
Section of Gastroenterology Virginia Mason Medical Center Seattle, WA, USA

365

Laparoscopic transcystic duct common bile duct exploration

Background Laparoscopic techniques of transcystic common bile duct exploration (LTCBDE) treat common duct stones in one session and avoid the potential complications of endoscopic sphincterotomy. These techniques include simple lavage, fluoroscopic-guided stone extraction using balloon catheters or wire baskets, biliary endoscopy with wire basket stone retrieval, intraoperative endoscopic or antegrade sphincterotomy, and lithotripsy techniques. The indications for LTCBDE include filling or equivocal defects on cholangiography, stone size less than 10 mm, fewer than nine stones, and possible tumor. The contraindications include stones proximal to the cystic duct entrance into the common bile duct and small/friable cystic duct. Choledochoscopic transcystic duct common bile duct exploration Choledochoscopic transcystic duct common bile duct (CBD) exploration is applicable in 8090% of patients with CBD stones. The choledochoscope (2.73.2 mm diameter) should have bidirectional deflection and a 1.2-mm working channel. It requires cystic duct dilation with a balloondilating catheter or ureteral dilation bougies. Stones are visualized and a wire basket is passed through the working channel to entrap the stone. An effort is made to pass the scope up into the liver, although this can only be performed in approximately 10% of cases. The technique is successful in cleaning the common duct of stones in over 90% of cases. Fluoroscopic basket stone retrieval Fluoroscopic basket stone retrieval is less expensive than choledochoscopy and is successful approximately 60% of the time. Special spiral wire baskets with flexible leaders must be used to avoid injury to the common bile duct. The advantage of not having to dilate the cystic duct is offset by the problem of extracting the basket with an entrapped stone through a nondilated cystic duct. Biliary balloon catheter trolling Biliary balloon catheter trolling is occasionally helpful in cases with a dilated cystic duct and stones that float. The drawback to this technique is the potential to pull the stone into the common hepatic duct, out of reach of an endoscope. Ampullary balloon dilation Ampullary balloon dilation is a controversial technique. Its advantage is that it can be employed when the cystic duct is extremely small and an endoscope cannot be inserted. A radial expanding balloon is placed under fluoroscopic guid-

ance to span the sphincter. Hyperamylasemia often follows, and severe pancreatitis may occur.

Other techniques Fitzgibbons recommends placing a ureteral catheter through the cystic duct into the duodenum. This catheter is used to perform guide-wire-assisted ERCP postoperatively. DePaulo and Zucker have treated small numbers of patients, with good results, using antegrade sphincterotomy via the cystic duct during laparoscopic cholecystectomy (LC). A gastroscope must be inserted orally to observe the orientation of the papillotome.

Summary and recommendations A recent prospective multicenter study of endoscopic sphincterotomy (ES) in 1,494 patients showed procedurerelated morbidity at 7.4% when ES was performed in conjunction with LC, procedure-related mortality of 0.5%, and total mortality of 2.2%. Therapeutic recommendations for patients with suspected common bile duct stones should consider these more inclusive morbidity figures. They should also reflect the immediate risks of pancreatitis, bleeding, and perforation and the delayed risk of stricture. These risks may exceed those of open common bile duct exploration in younger patients, especially with smalldiameter common ducts, but are similar to laparoscopic common bile duct exploration (CBDE) in patients over 65 years of age. Some reports indicate that open CBDE has almost no mortality in patients under age 60, but 24.3% mortality in those over age 60. The experience with LTCBDE of our group, De Paula and Petelin, shows that the approach is applicable in more than 85% of cases and is successful in 8595% of cases. Complications do occur (6% major and 12% minor), but these also include complications associated with LC. There were only two LTCBDE procedure-related complications in our series of 193 LTCBDEs, no mortality in patients under 65, and one death (<1%) in a patient over 65. Comparing these outcomes with either ES or open common bile duct exploration, the data show that patients under age 65 have improved outcomes with LTCBDE, and those over 65 have comparable outcomes with ES or LTCBDE, but only require a single procedure with LTCBDE. Choledochoscopy via the cystic duct appears to be the most effective (90%) and safest approach to the common duct. Nevertheless, it is possible to employ fluoroscopic basket retrieval in many cases (60%). Even irrigation and trolling with a biliary balloon catheter will be effective in some cases. Ampullary balloon dilation should be performed when the only alternative is ES. Leaving a catheter in the cystic duct or a transampullary biliary stent will ensure successful postoperative sphincterotomy. Antegrade sphincterotomy should be reserved for the 510% of complex cases that require a drainage procedure. Lithotripsy, either with electrohydraulic or laser energy, is best employed via a choledochotomy because of the resultant debris. Further technological advances will facilitate the application and adoption of laparoscopic approaches to the

366

common duct, which should become the primary strategy in the great majority of patients. Selected readings
Carroll BJ, Phillips EH, Chandra M, Fallas MJ (1993) Laparoscopic transcystic duct balloon dilation of the sphincter of Oddi. Surg Edosc 7: 514517 Curet M, Pitcher D, Martin D, Zucker K (1995) Laparoscopic antegrade sphincterotomy: a new technique for the management of complex choledocholithiasis. Ann Surg 221(2): 149155 DePaulo A, Hashiba K, Bafutto M, Zago R, Machado M (1993) Laparoscopic antegrade sphincterotomy. Surg Laparosc Endosc 3(3): 157160 De Paula AL, Shashiba K, Bafutto M (1994) Laparoscopic management of choledocholithiasis. Surg Endosc 8: 13991403 Fitzgibbons RJ, et al. (1995) An alternative technique for treatment of choledocholithiasis. Arch Surg 130: 638642 Freeman M, Nelson D, Sherman S, et al. (1994) Complications of endoscopic sphincterotomy (ES); a prospective multicenter 30-day outcome study. The Hennepin Count Medical Center, Minneapolis, MN and the MESH Study Group Abstract World Congress of Gastroenterology, October 27, 1994, Los Angeles, CA, USA Hunter JG, Soper NJ (1992) Laparoscopic management of bile duct stones. Surg Clin North Am 72(5): 10771080

Jones DB, Soper NJ (1996) The current management of common bile duct stones. Adv Surg 29: 271289 Lezoche E, Paganini M (1995) Single stage laparoscopic treatment of gallstones and common bile duct stones in 120 unselected, consecutive patients. Surg Endosc 9: 10701075 Morgenstern L, Wong L, Berci G (1992) Twelve hundred open cholecystectomies before the laparoscopic era. A standard for comparison. Arch Surg 127: 400403 Petelin J (1991) Laparoscopic approach to common bile duct pathology. Surg Laparosc Endosc 1: 3341 Phillips EH, Carroll BJ, Pearlstein AR, Daykhovsky L, Fallas MJ (1993) Laparoscopic choledochoscopy and extraction of common bile duct stones. World J Surg 17(1): 2228 Rhodes M, Nathanson L, ORourke N, Fielding G (1995) Laparoscopic exploration of the common bile duct: lessons learned from 129 consecutive cases. Br J Surg 82: 666668

Edward H. Phillips
Department of Surgery Cedars-Sinai Medical Center Los Angeles, CA, USA

367

Laparoscopic choledochotomy for treatment of common bile duct stones

Introduction According to some reports, up to 10% of patients who undergo laparoscopic cholecystectomy harbor common bile duct stones. Laparoscopic common bile duct exploration (LCDE), which in experienced centers yields ductal clearance rates >90%, has become the gold standard for singlestage management of choledocholithiasis and is the most efficient method of handling this problem. A number of techniques are available to accomplish choledocholithotomy. These include pharmacological and flushing techniques, balloon catheter manipulations, basket techniques with and without fluoroscopy, and choledochoscopic techniques. In the authors experience, and in many series, choledochoscopic exploration is used most frequently, as much as 63% of the time. Balloon extraction methods are usually combined with flushing and pharmacological techniques. These account for approximately 26% of ductal explorations. Basket techniques, utilized in 11% of the authors cases, may be employed either with the fluoroscope or without it, although the latter method requires a greater degree of skill and risk management by the surgeon in order to avoid perforation of the duct or capture of the papilla. Ductal exploration may be accomplished through the cystic duct or directly through the common bile duct. In most cases, the common duct may be explored through the cystic duct. However, in some situations choledochotomy is necessary or even preferred. The problem is deciding which approach is best for a given case. There are specific indications for the employment of each access route, and each is equally effective in clearing the ductal system in the appropriate setting. While the transcystic route is the least invasive and generally does not require any subsequent ductal manipulation or drainage procedure, the choledochotomy route requires either closure of the duct over a T-tube or primary closure of the choledochotomy without a T-tube.

Table 1. Factors influencing duct exploration approacha Factor One stone Multiple stones Stones 6-mm diameter Stones >6-mm diameter Intrahepatic stones Diameter of cystic duct <4 mm Diameter of cystic duct >4 mm Diameter of common duct <6 mm Diameter of common duct >6 mm Cystic duct entrancelateral Cystic duct entranceposterior Cystic duct entrancedistal Inflammationmild Inflammationmarked Suturing abilitypoor Suturing abilitygood
a

Transcystic + + + + + + + + + + +

Choledochotomy + + + + + + + + + + + + +

+, positive or neutral effect; , negative effect.

The choledochotomy and ductal exploration Most surgeons prefer a longitudinal choledochotomy. When performed laparoscopically, it is usually shorter than in open choledochotomy (generally no larger than the largest stone) and no stay sutures are required. External manipulations of the duct, balloon, and basket extraction techniques, and choledochoscopic maneuvers are all used to clear the duct. Occasionally, electrohydraulic lithotripsy is used under direct choledochoscopic vision to fragment impacted stones. Some authors have added intraoperative antegrade sphincterotomy in selected patients, but this is logistically complex and not widely practiced. T-tubes and T-tube management T-tubes are used in patients in whom there is concern for possible retained stones or debris, distal ductal spasm, inflammation, obstruction, pancreatitis, or general poor tissue quality. T-tubes provide access for ductal imaging in the postoperative period and a route for removal of residual common duct stones, should they be left after common bile duct exploration. Most authors prefer a 14-French latex Ttube (or larger) and ductal closure with an absorbable fine suture such as 4-0 or 5-0 polyglycolic acid. Although silicone T-tubes have been used by some authors in the USA and Europe, they are generally not preferred because they do not excite the degree of tissue reaction necessary to produce a tract to the surface in the case of persistent bile leakage after removal. Silicone T-tubes, however, have been associated with less bacterial contamination than latex T-tubes. Complications of T-tube placement and removal Management of T-tubes in the postoperative period may be associated with bacteremia, dislodgment of the tube, obstruction by the tube, or fracture of the tube. Removal of T-tubes postoperatively has been suggested as early as 4 days and as late as 6 weeks. Between these two extremes lies the most appropriate management plan. Some authors

Transcystic vs transductal approach Factors influencing the choice of access route for LCDE are included in Table 1. Note that it is usually a negative influence, precluding the use of a given access route, rather than a positive influence, permitting the use of one, that determines which route is preferred. In the authors series, 85% of ductal explorations were performed with the transcystic route. Fifteen percent of cases required choledochotomy. Most American series parallel this same approach. The transcystic method is the least invasive and is associated with shorter operating times, shorter length of stay, and the greatest patient satisfaction because there are usually no tubes left in situ after the procedure.

368

recommend broad-spectrum antibiotic coverage while the T-tube is in situ. T-tube cholangiography should be performed before removal of the tube. Removal of T-tubes has been associated with bile leaks, peritonitis, and reoperation. The incidence of these complications has generally been less than 5% in most laparoscopic series. Primary closure of the choledochotomy Primary closure of the common bile duct is performed in patients in whom the exploration successfully cleared all stones, and in which there is no evidence of distal edema, spasm, obstruction, or pancreatitis, and in whom tissue integrity appears substantial. Absorbable suture of 4-0 polyglycolic acid is usually used for the closure. This is accomplished in either a continuous fashion or with interrupted suture. A closed-system suction drain is placed adjacent to the closure in all cases. In the authors series, laparoscopic primary closure of the choledochotomy, used in 34% of cases involving choledochotomy, did not result in any complications. There was no incidence of bile leak, peritonitis, or clinical evidence of retained bile duct stones. Length of stay is shorter than that of patients with T-tubes. Patients report a higher degree of comfort and satisfaction than those in whom T-tubes had been placed. Summary 1. Laparoscopic choledocholithotomy via choledochotomy or the cystic duct is feasible and effective when used in the appropriate situations. 2. Choledochotomy closure with or without T-tube placement is feasible and effective. 3. When used in appropriately selected patients, choledochotomy closure without T-tube placement does not result in increased complications. 4. Primary closure of the common bile duct without T-tube placement is associated with higher patient satisfaction and shorter length of stay than T-tube placement. 5. Transductal common bile duct exploration is more invasive than transcystic ductal exploration but less invasive and less costly than open common bile duct exploration.

Selected readings
Bernstein DE, Goldberg RI, Unger SW (1994) Common bile duct obstruction following T-tube placement at laparoscopic cholecystectomy. Gastrointest Endosc 40(3): 362365 Carroll BJ, Phillips EH, Daykhovsky L, Grundfest WS, Gershman A, Fallas M, Chandra M (1992) Laparoscopic choledochoscopy: an effective approach to the common duct. J Laparoendosc Surg 2: 1521 DePaula A, Hashiba K, Bafutto M, Zago R, Machado M (1993) Laparoscopic antegrade sphincterotomy. Surg Laparosc Endosc 3(3): 157160 DePaula AL, Hashiba K, Bafutto M (1994) Laparoscopic management of choledocholithiasis. Surg Endosc 8: 13991403 Ferzli GS, Massaad A, Kiel T, et al. (1994) The utility of laparoscopic common bile duct exploration in the treatment of choledocholithiasis. Surg Endosc 8: 296298 Fletcher DR (1993) Common bile duct calculi at laparoscopic cholecystectomy: a technique for management. Aust NZ J Surg 63: 710714 Gillatt DA, May RE, Kenedy R, Longstaff AJ (1985) Complications of T-tube drainage of the common bile duct. Ann R Coll Surg Engl 67: 369371 Horgan PG, Campbell AC, Gray GR, Gillespie G (1989) Biliary leakage and peritonitis following removal of T tubes after bile duct exploration. Br J Surg 76: 12961297 Koivusalo A, Makisalo H, Talja A, et al. (1996) Bacterial adherence and biofilm formation on latex and silicone T-tubes in relation to bacterial contamination of bile. Scand J Gastroenterol 398403 Lygidakis NJ (1984) Incidence of bile infection in biliary lithiasis. Effects on postoperative bacteremia of choledochoduodenostomy, T-tube drainage, and primary closure of the common bile duct after choledochotomya prospective trial. Am Surg 50: 236240 Norrby S, Heuman R, Anderberg B, Sjodahl R (1988) Duration of T-tube drainage after exploration of the common bile duct. Acta Chir Scand 154: 113115 Petelin J (1991) Laparoscopic approach to common duct pathology. Surg Laparosc Endosc 1(1): 3341 Petelin J (1993) Laparoscopic approach to common duct pathology. Am J Surg 165: 487491 Stoker ME, Leveillee RJ, McCann JC, Maini BS (1991) Laparoscopic common bile duct exploration. J Laparoendosc Surg 1(5): 287293 Traverso LW (1996) A cost-effective approach to the treatment of common bile duct stones with surgical versus endoscopic techniques. In: Berci G, Cuschieri A (eds). Bile ducts and bile duct stones. WB Saunders, Philadelphia, pp 154160

Joseph B. Petelin
University of Kansas School of Medicine Kansas City, KS, USA

369

Laparoscopic treatment for right upper quadrant tumors and hepatic metastases
The laparoscope should be regarded as a significant tool for both the diagnosis and staging of abdominal cancer, especially tumors effecting the hepatobiliary tract. The laparoscope must be utilized in conjunction with other modalities which include radiologic imaging techniques, nuclear medicine scans, intraluminal endoscopic studies, and percutaneous ultrasound. When tumors of the right upper quadrant are assessed, the indication for laparoscopy must be viewed in relation to the need for celiotomy for overall management. Specifically, the diagnosis of a primary colorectal cancer generally will mandate resection either by formal open techniques or by newer laparoscopic methods. Preoperative laparoscopy for the assessment of colorectal cancer metastases is generally not undertaken except in unusual situations when liver involvement is noted without a primary cancer being identified. It may be important to preoperatively assess the liver, however, in order to discuss the potential of metastasectomy with the patient preoperatively as a method to enhance survival in advanced colorectal cancer. The concept of screening would indicate that there is a high-risk group of patients that may benefit from a particular radiologic or endoscopic study. When viewing this in terms of hepatic metastases, it is recognized that liver involvement occurs most frequently in gastrointestinal tumors, including colorectal cancer and pancreatic and gastric carcinoma. Less often, breast and lung cancer as well as lymphoma may involve the hepatic parenchyma and may lead to laparoscopic investigation. The important criteria to observe during laparoscopic evaluation are: (1) the number of metastases, (2) size of metastases, (3) location of extrahepatic disease, and (4) extent of hepatic involvement. The utilization of intraoperative ultrasound may enhance staging and, in fact, may change the surgical approach in as many as 20% of patients who are being evaluated for hepatic resection. Laparoscopy may be especially important in patients undergoing lymphoma staging in that inspection of the liver, appropriate biopsy of hepatic lobes, and nodal dissec-

tion can be undertaken. In addition to making a diagnosis of cancer, benign hepatic tumors may be found. The most common benign tumor is focal nodular hyperplasia (FNH). In addition, hepatocellular adenoma, relating to oral contraceptives and hemangiomas may be visualized. Primary hepatobiliary tumors occur most commonly in South America and Southeast Asia. In the United States, gallbladder carcinoma is the most common type of hepatobiliary tumor. Approximately 1% of gallbladders removed will harbor occult carcinoma. Certain groups in the United States, such as Native-American women in New Mexico, have a significantly higher rate of gallbladder cancer. The laparoscope may be used to confirm imaging studies, provide needle-directed biopsy of the liver or peritoneum, or help in surgical decision-making regarding the resectability of gallbladder cancer. Similarly, laparoscopy should be utilized preoperatively in the assessment of patients with hepatocellular cancer and cholangiocarcinoma of the extrahepatic biliary ducts. Patients with hepatolithiasis, cystic disease of the extrahepatic biliary tree, and those with sclerosing cholangitis represent high-risk groups for extrahepatic bile duct cancer.

Selected readings
Dagnini G, Marin G, Patella M, Zotti S (1984) Laparoscopy in the diagnosis of primary carcinoma of the gallbladder. Gastrointest Endosc 30: 289291 Eubanks S (1994) The role of laparoscopy in diagnosis and treatment of primary or metastatic liver cancer. Semin Surg Oncol 10: 404410 Greene FL, Rosin RD (1995) Minimal access surgical oncology. Radcliffe Medical Press, Oxford John TG, Greig JD, Crosbie JL, et al. (1994) Superior staging of liver tumors with laparoscopy and laparoscopic ultrasound. Ann Surg 220: 711719 Ravikumar TS (1996) Laparoscopic staging and intraoperative ultrasonography for liver tumor management. North Am Surg Oncol Clin 5: 271282

Frederick L. Green
Department of General Surgery Carolinas Medical Center Charlotte, NC, USA

370

Endoscopic surgery for pancreaticobiliary cancer

Introduction Pancreatic and extrahepatic biliary cancers are extremely aggressive tumors with a poor prognosis. Biological characteristics of this cancer create the need to achieve a balance in surgical treatment between operative agressiveness for cure and a low morbidity for palliative care. Unfortunately, this balance is easily lost: Many patients are not offered a surgical option based on the morbidity of standard laparotomy and the presumed (not always true) hopelessness of the diagnosis, or they are subjected to the morbidity of a laparotomy only to, not infrequently, find contraindications to curative resection. In the latter case, palliative bypass is the only potential benefit that can be offered to the patient during the laparotomy, but this palliation is compromised by the pain and long recovery from the surgery. As in so many other fields in gastrointestinal surgery, endoscopic procedures, both flexible and laparoscopic, are playing an increasing role in achieving a balance between the cruel cure and the kind palliation. Staging Diagnostic laparoscopy has been identified as an important tool in the accurate staging and determination of resectability in pancreatic and biliary cancers. This has been demonstrated by numerous investigators since the mid-70s with the primary benefit being the identification of small peritoneal metastases undetectable by radiologic imaging modalities. The low sophistication of ancillary instruments initially confined laparoscopy to a simple peritoneal survey and therefore the accuracy rate for determining resectability was somewhat low (60%). In the 1990s the use of laparoscopy for increasingly sophisticated surgeries gave pancreatic surgeons the tools needed to do a full exploration of the upper abdomen including the lesser sac and retroperitoneum and therefore markedly improved the diagnostic and staging accuracy rates for this modality. The recent addition of laparoscopic ultrasonography has further increased the usefulness of this approach with staging accuracy rates now being reported in the 90% range. Investigators who routinely use laparoscopy in conjunction with other imaging techniques can now with a high degree of accuracy (greater than 75%) avoid unneeded laparotomies which would only determine the patient was in fact unresectable. Similar advances in the technology of flexible endoscopy have also provided low-morbidity diagnostic capabilities as well. In particular, the use of endoscopic ultrasound has been shown to be highly accurate in determining invasion of the portal vein, the second most common reason for unresectability. This has resulted in a much lower morbidity than traditional angiography and allows the subsequent performance of flexible endoscopic palliation should that need exist. Palliation A substantial number of patients with pancreatobiliary cancers will eventually suffer biliary obstruction (75%) or

gastric outlet obstruction (21%) before their death. This markedly contributes to their discomfort and morbidity and certainly justifies palliative intervention. The use of endoscopically placed retrograde stents, and more recently expanded metal stents, offers a very minimally invasive palliation for malignant bile duct strictures. Unfortunately, the same is not true for gastric outlet obstruction, which would obviate flexible endoscopy for either approach. Flexible endoscopy also is not always successful and is not without its own associated risks and morbidity, which may be poorly tolerated in this patient group. Another drawback with the use of stents is their short life span (612 weeks) necessitating returns to the hospital for exchanges, often when the patients are at their very sickest terminal state. Since surgery, either open or laparoscopic, remains the gold standard for the accurate staging of patients, a single procedure to both staging and palliation is certainly a desirable goal. While gastrojejunostomy and cholecystoenterostomy or choledochoenterostomy are well-tolerated and effective means of achieving bypass and decompression, there are some patients who are not candidates. These include patients with such extensive tumor that the extrahepatic biliary tract is not accessible and those who have had previous biliary or gastric surgeries. The literature indicates that only 2258% of patients with malignant bile duct obstruction and 75% of gastric outlet obstruction meet the criteria for effective palliative bypass. Certainly if these bypasses can be achieved with laparoscopic means, effective palliation is accomplished while the morbidity and prolonged hospital stay of open laparotomy are avoided. The introduction of endoscopic stapling devices and the overall improvement and advancement of surgeons skills in the realm of suturing have made procedures such as cholecystojejunostomy and gastrojejunostomy quite feasible. More recently, choledochojejunostomy has been defined in animal models and may further broaden the repertoire of laparoscopic surgical bypass. Additional reports defining combinations of flexible endoscopy, laparoscopy, and even interventional radiology are beginning to appear. All of these can be considered minimally invasive endoscopic approaches to pancreatobiliary cancer which allow accurate staging, better determination of resectability, and an effective low-morbidity palliation for these unfortunate patients.

Selected readings
Cuschieri A, Hall AW, Clark J (1978) Value of laparoscopy in the diagnosis and management of pancreatic cancer. Gut 19: 672 Huibregtse K, Carr LD, Cremer M, et al. (1992) Biliary stent occlusiona problem solved with self-expanding metal stents? European Wallstent Study Group. Endoscopy. 24(5): 391394 John TG, Garden OJ (1993) Assessment of pancreatic cancer. In: Cuesta MA, Nagy AG (eds). Minimally invasive surgery in gastrointestinal cancer. New York: Churchill Livingstone, pp 95111 Rosch T, Braig C, Gain T, et al. (1992) Staging of pancreatic and ampullary carcinoma by endoscopic ultrasonography. Gastroenterology 102: 188 Schob OM, Schlumpf R, Schmid R, et al. (1995) Laparoscopic treatment of biliary and gastric outlet obstruction. Surg Laparosc Endosc 5(4): 288 295 Schob OM, Schmid R, Morimoto A, et al. (1997) Laparoscopic Roux-en-Y choledochojejunostomy. Am J Surg 173(4): 312319

371 Shimi S, Banting S, Cuschieri A (1992) Laparoscopy in the management of pancreatic cancer: endoscopic cholecystojejunostomy for advanced disease. Br J Surg 79(4): 317319 Singh SM, Longmire WP, Reber HA (1990) Surgical palliation for pancreatic cancerthe UCLA experience. Ann Surg 212:132139 Soulez G, Gagner M, Therasse E, et al. (1994) Malignant biliary obstruction: preliminary results of palliative treatment with hepaticogastrostomy under fluoroscopic, endoscopic, and laparoscopic guidance. Radiology 192: 241246 Tarnasky PR, England RE, Lail LM, et al. (1995) Cystic duct patency in malignant obstructive jaundice. Ann Surg 221(3): 265271 Warshaw AL, Pepper JB, Shipley AW (1986) Laparoscopy in the staging and planning of therapy for pancreatic cancer. Am J Surg 151: 76 Warshaw AL, Gu Z-Y, Wittenberg J, et al. (1990) Preoperative staging and assessment of resectability of pancreatic cancer. Arch Surg 125: 230

Lee L. Swanstrom
Department of Surgery Oregon Health Sciences University Portland, OR, USA

372

Laparoscopic surgery for solid lesions

Diagnostic and therapeutic procedures for the liver can now be done via an open, laparoscopic, laparoscopic assisted, percutaneous, or transvascular approach. Standard surgical instrumentation has improved (sutures, clips, cautery, hemostatic agents) and has been adapted for laparoscopic use (cryoablation, argon coagulator, endostaplers, echodoppler, injection). New technologies have been developed (hydrojet, the harmonic scalpel, microwave coagulation). The increasing expertise in advanced laparoscopy and continued technological advances are likely to provide the basis for expanding the use of laparoscopic procedures on the liver. Requirements for laparoscopic liver surgery are expertise in open liver surgery, expertise in two-handed laparoscopy, knowledge of laparoscopic instrumentation and technology, thorough understanding of the underlying disease and its biology, and knowledge of the alternative therapies available. Laparoscopic liver surgery has been used for diagnosis, biopsy and staging resections (anatomic and nonanatomic), enucleation, cryoablation, injection, coagulation and arterial catheter implantation. Nevertheless, the experience is still limited. For example, my personal series of resections is limited to nine cases. For this reason, we surveyed other groups through a questionnaire and personal communication. Data was obtained from Hotel-Dieu, Montreal (Gagner M., Pomp A.); Creighton University, Nebraska (Fitzgibbons R.); Washington University, Seattle (Pellegrini C., Helton S.); Mayo Clinic (Nagorney D.); Rush University, Chicago (Deziel D., Prinz R.C.); Emory University, Atlanta (Hunter J., Galloway, Branum E.); and Lahey Clinic, Boston (Rossi R., Tsao J.). The experience with adenomas and nodular hyperplasia was 19 cases with a median size of 6 cm. Two were treated by enucleation, 11 by nonanatomic resections, and six were only biopsied. The lesions were in anterior hepatic locations. Resective procedures took about 2 h to perform. There was bleeding in one case but none required conversion. Fifteen cases of hepatic hemangiomas were treated but only the Hotel Dieu and Lahey Clinic group managed lesions larger than 6 cm. The four largest hemangiomas were symptomatic. One was enucleated and three were treated by nonanatomic resections. There was no bleeding or need for conversion. The operations for two patients with hemangiomas over 6 cm were completed within 2 h. The data on primary hepatic tumors included 27 cases. Of these, 18 were biopsied and staged, two resected (one anatomic, one nonanatomic), three treated with alcohol injection, and one treated with cryoablation. There were two conversions, one for bleeding and one (Lahey Clinic case of 8 cm) because the right-sided lesion, thought initially to be benign preoperatively, was suspected during dissection to be malignant, favoring an open resective oncologic procedure. Information on 143 cases of metastatic disease to the liver was received. Biopsy and staging only was done in 113 cases. Nonanatomic resections were done in 19 cases, anatomic resection in one, cryoablation in ten, and alcohol in-

jection in three. There was one conversion for bleeding and one for a bile leak after cryoablation. The Lahey Clinic experience with three presumed cystadenomas is of interest. Two were enucleated successfully (6 and 12 cm) and confirmed histologically. However, the third case (4 cm) was felt during dissection to not be a typical cystadenoma. The procedure was converted; a nonanatomic resection was done; the final diagnosis was neurosarcoma with central necrosis. The majority of lesions were removed in plastic bags through enlarged umbilical incisions. However, other routes included minilaparotomies and colpotomy. Reports on laparoscopic resections are few. Kaneko et al. reported on 11 patients done with gasless laparoscopy. Six had nonanatomic resections and three left lateral segmentectomies. The indications were metastatic disease in three, hepatocellular carcinoma in four, hemangiomas in two, and others in two. Instruments used included endostaplers, microwave coagulation, ultrasonic dissector and the argon beam coagulator. The locations of the lesions were segment 2, or 2 and 3 in four cases, segment 4 in one case, segment 5 in one case, and segment 6 in two cases. One case required conversion for bleeding. Gugenheim et al. have reported on three successful nonanatomic resections for lesions of 4 cm or more in size, allowing earlier hospital discharge. Yamanaka et al., comparing microwave coagulation necrosis with open wedge resection for hepatocellular carcinoma, show that the former procedure can be performed in a shorter time, with less complications, minimal blood loss, and with similar survival as the latter, raising again the possible advantages of laparoscopic tumor destructive techniques. There is preliminary data that suggest advantages of the multimodal water jet dissector over the ultrasonic dissector. Cushieri et al. have developed a highefficiency cryosurgical unit with multiple probes. Laparoscopic-assisted techniques are being used for lesions in less accessible locations (segments 7, 8). Feliciotti et al. have reported on three cases of catheter implantation for arterial infusion using the gastroduodenal artery. Complications of laparoscopic liver surgery include gas embolism (from CO2 or argon), the possibility of incomplete tumor removal or destruction, bleeding, bile leak, and the possibility of tumor spread. In conclusion, we are in the early stages of evolving laparoscopic liver techniques, whose use is likely to increase. The merits and limitations of this approach compared to other treatments need further and continued evaluation. Those who become involved need expertise in open surgery and thorough knowledge of the underlying disease and of the alternative therapies. Laparoscopic techniques are good for diagnosis, staging and biopsy, especially if associated with endoscopic ultrasound. Based on experiences to date, laparoscopic procedures may be increasingly used for benign superficial lesions and selectively, for malignant tumors (resection, cryoablation, coagulation, injection, etc.). Destructive tumor techniques, if proven as effective as resection for certain lesions, could provide additional stimulus for a laparoscopic approach. It is suggested that S.S.A.T. and SAGES sponsor a database for laparoscopic liver cases.

373

Selected readings
Cushieri A, Crostwaite G, Shimi S, et al. (1995) Hepatic cryotherapy for liver tumors. Surg Endosc 9: 483489 Feliciotti F, Paganini A, Guerrier M, et al. (1996) Laparoscopic intraarterial catheter implantation. Surg Endosc 10: 449452 Gugenheim J, Mazza D, Katkouda B, et al. (1996) Laparoscopic resection of solid liver tumors. Br J Surg 83: 334335 Kaneko H, Takagi S, Shiba T (1996) Laparoscopic partial hepatectomy and left lateral segmentectomy. Surgery 120: 468475 Rau HG, Meyer G, Jauch KW, et al. (1996) Liver resection with the water jet. Chirurg 67: 546551

Yamanaka N, Tanaka T, Oriyama T, et al. (1996) Microwave coagulonecrotic therapy for hepatocellular carcinoma. World J Surg 20: 1076 1081

Ricardo Rossi
Facultad de Medicina Pontificia Universidad Catolica De Chila Santiago, Chile

Surg Endosc (1998) 12: 377

Springer-Verlag New York Inc. 1998

SAGES position statement on advanced laparoscopic training

The Society of American Gastrointestinal Endoscopic Surgeons (SAGES) endorses the following concepts for training in laparoscopic surgery. 1. Laparoscopic operations comprise a core component of general surgery. 2. Training of general surgeons in laparoscopic surgery should occur within the five-year general surgery residency. 3. General surgical training should include a defined number of basic* laparoscopic operations. 4. Advanced laparoscopic operations are a scarce commodity in most general surgical residency programs. Advanced laparoscopic operations should not, at this time, be designated as a defined category. 5. Program directors should be granted the flexibility to focus the residency experience in advanced laparoscopic surgery on those individuals who are committed to a career in general surgery.

This document was reviewed and approved by the Board of Governors of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), March, 1997. It was prepared by an ad hoc task force on Residency Integration.

*Basic laparoscopic surgery is comprised of diagnostic laparoscopy, laparoscopic cholecystectomy, and laparoscopic appendectomy. Advanced laparoscopic surgery consists of all other laparoscopic operations.

Society of American Gastrointestinal Endoscopic Surgeons (SAGES)

2716 Ocean Park Boulevard, Suite 3000 Santa Monica, CA 90405, USA

SAGES position papers

Surg Endosc (1998) 12: 374376 Springer-Verlag New York Inc. 1998

Integrating advanced laparoscopy into surgical residency training

Introduction Laparoscopic operations are a primary component of general surgery. The Society of American Gastrointestinal Endoscopic Surgeons (SAGES) is dedicated to the advancement of training in minimal access surgery to assure the safe performance of such operations. The purpose of this document is to further the integration of advanced laparoscopic surgical training into the curriculum of the general surgery residency. Opportunities to perform laparoscopic operations currently vary widely between surgical training programs. The proposals presented in this document are intended as current measures to improve training opportunities in a dynamic environment. As utilization of minimal access procedures increases in the future, residents will more readily learn the skills necessary to accomplish these operations safely. In that setting, the proposals below may no longer be necessary. core group of technical skills common to all advanced laparoscopic operations. Such skills are best acquired in the operating room or, alternatively, through skills laboratories involving surgical trainers, animal models, or other simulated operating conditions. Examples of such skills include two-handed dissection, intracorporeal suturing, and intraand extracorporeal knot tying. Mastery of these advanced laparoscopic skills by the resident is encouraged prior to initiating an experience performing advanced laparoscopic operations. Experience in the performance of a specific operation via celiotomy will also facilitate mastery of the similar laparoscopic procedure using a minimal access approach. Since many advanced laparoscopic skills are common to all advanced laparoscopic operations, experience in a specific operation enhances the acquisition of skills necessary to perform others. Therefore, it is the combined experience in advanced procedures that should be emphasized during training, rather than the mastery of any one individual procedure.

Curriculum guidelines in advanced laparoscopic surgery Basic laparoscopic operations include laparoscopic cholecystectomy, laparoscopic appendectomy, and diagnostic laparoscopy. All other laparoscopic operations are defined as advanced. Prior to learning or performing advanced laparoscopic operations, the resident must be familiar with and experienced in basic laparoscopy. Laparoscopic operations, as with all operations, are appropriately learned in the broad context of surgical science and practice. Critical educational components include: pathophysiology of disease, diagnosis, operative indications and contraindications, familiarity with alternative treatments, comprehensive principles of pre- and postoperative care, and understanding of the prevention, diagnosis, and treatment of complications. In addition, the relative advantages and disadvantages of both open and minimal access approaches must be known. SAGES Curriculum Guideline for Resident Education in GI Surgical Endoscopy, Laparoscopy, and Thoracoscopy contains an outline of knowledge and skills to be mastered in basic and advanced laparoscopy.

Methods to integrate advanced laparoscopy training into general surgery residency SAGES long-term goal is to facilitate complete integration of advanced laparoscopic training within each surgical program. Residents should ultimately learn these procedures in the operating room under the direction of skilled faculty instructors. Until such time as complete integration is possible, SAGES belies the following measures can help accomplish this goal:

Train faculty Train residents Provide guidelines for postresidency training for prospective faculty

I. Basic laparoscopic surgery Skills acquisition for advanced laparoscopic operations Training to learn advanced laparoscopic operations begins with acquisition of skills in basic laparoscopy. There is a Most programs provide adequate experience in basic laparoscopic surgery. To assure ongoing availability of basic training resources, the following is recommended:

375

Basic laparoscopy courses for residents SAGES plans to continue offering existing SAGES courses in basic laparoscopic surgery. These courses, offered twice yearly, are two-day courses for residents focusing on the basic principles and skills of endoscopic and laparoscopic surgery. A total of approximately 120 residents are enrolled in these courses yearly, which represents approximately 12% of the annual number of graduating residents. 2. Basic laparoscopy courses for faculty When laparoscopy for general surgery was introduced, SAGES organized a series of training the trainers courses, which provide on-site faculty training at 17 institutions. This intensive training provided a mechanism to train more than 250 surgical faculty. The need for training in basic laparoscopy has diminished. However, if a need for faculty training in basic laparoscopy still exists, SAGES will endeavor to provide a similar type of training.

3. 4.

II. Advanced laparoscopic surgery A. Faculty training 1. Courses: Hands-on courses are useful for conveying the techniques of laparoscopic operations to those who are proficient in the similar open operations. Faculty members interested in obtaining advanced laparoscopic training may benefit from advanced laparoscopic courses. SAGEs will provide an opportunity for faculty members to participate in advanced laparoscopic surgical courses. See Section B1. 2. Faculty mentoring: SAGES believes that faculty who have already acquired the fundamental skills in advanced laparoscopic surgery and who desire to learn a new or modified laparoscopic operation, will benefit from observing and interacting with a peer who is skilled and accomplished in that specific operation. SAGES will facilitate this process by developing a list of mentoring opportunities available for specific laparoscopic operations. 3. Fellowships: Postgraduate training in advanced laparoscopic surgery is another means by which faculty or faculty candidates may obtain experience. Such programs should not detract from the experience of the surgical residency training where they coexist. SAGES believes the main goal of such fellowships should be to train future faculty. B. Resident training 1. Courses: Courses in advanced laparoscopic procedures are one method of introducing skills. SAGES plans to offer ongoing courses for residents and, when necessary, an accompanying faculty member, for a laboratory experience in advanced laparoscopic surgery.

5.

Appropriate candidates for such courses are: residents who plan a career in general surgery residents who have already achieved a mastery of basic laparoscopic surgery residents who are unlikely, based on their programs current practice patterns, to obtain a significant experience in advanced laparoscopic surgical techniques. faculty from programs who do not have faculty to teach the procedure in question may elect to send a faculty member to the course. Skills labs: The creation of inanimate and animal training facilities by individual programs is encouraged to provide supplemental teaching of advanced laparoscopic surgical skills. SAGES will facilitate the acquisition of or access to advanced laparoscopic equipment and skills lab facilities. Additionally, SAGES plans to organize regional skills labs [1]. Needs assessment: SAGES will continue to assess the needs of residency programs both in terms of faculty training and overall program needs. Re-examination of residency training: Optimal training in a general surgery residency includes an adequate experience in both advanced open and laparoscopic procedures. Data suggest that case load is insufficient to produce such an experience. SAGES suggests that the appropriate leadership organizations consider re-examining the flexibility of the general surgery residency training in order to optimize the availability of such advanced cases for residents planning a career in general surgery. Educational resources: SAGES will continue to offer other educational endeavors such as postgraduate courses, annual meetings, an extensive video library, a syllabus on topics in endoscopic and laparoscopic surgery for residents, a curriculum guideline for residency training in endoscopic and laparoscopic surgery, guidelines for credentialing, training, and standards of practice. A separate candidate membership category exists for residents and fellows.

Summary SAGES was founded to further the advancement of surgeons performing gastrointestinal endoscopy. In part due to the societys previous efforts, training in flexible endoscopy is now recognized as a mandatory component of training in general surgery. SAGES has more recently promoted the safe practice of laparoscopic surgery, which since 1991 has evolved to be a standard component of the practice of most general surgeons. Consistent with the goal of continuing to provide the public with surgeons competent in performing all varieties of appropriate laparoscopic procedures, SAGES joins other major surgical organizations in addressing the concern that, while basic laparoscopic procedures are mastered during residency training in general surgery, advanced procedures and the skills required to initiate safely an experience in them often are not. Incorporation of advanced laparoscopic surgical skills into surgical residency and the safe incorporation and per-

376

formance of advanced laparoscopic operations into surgical practice are dynamic processes. With the above proposals, SAGES hopes to facilitate both processes. We have, herein, set forth suggested curriculum guidelines for resident training in laparoscopic surgery. We have outlined measures by which resident and faculty training in advanced laparoscopic surgical skills may be realized. Focusing on adequately training the trainers is essential in order to facilitate residency training in the future. As practice patterns evolve and the frequency of laparoscopic operations increases, the opportunities for residents to learn required skills will increase. As a result, they will be able to accomplish safely such procedures. In that setting, the proposed special efforts outlined above may no longer be necessary.

References
1. Society of American Gastrointestinal Endoscopic Surgeons (SAGES) (1994) Framework for postresidency surgical education and traininga SAGES guideline. Surg Endosc 8: 11371142 This document was reviewed and approved by the Board of Governors of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), March, 1997. It was prepared by an ad hoc task force on residency integration.

Society of American Gastrointestinal Endoscopic Surgeons (SAGES)

2716 Ocean Park Boulevard, Suite 3000 Santa Monica, CA 90405, USA

Surg Endosc (1998) 12: 378

Springer-Verlag New York Inc. 1998

SAGES statement on concentration in general surgery residency

The SAGES document entitled Integrating Advanced Laparoscopy into Surgical Residency Training and the companion position statement outline the Societys position on training in advanced laparoscopic surgery. To achieve the goals set forth in the document, SAGES recommends that surgical training programs adopt a policy that assures acquisition of advanced laparoscopic skills during residency training. Those surgeons-in-training who plan to practice the speciality of general surgery should be given preferential opportunity to apply those skills through specific operative experience at an appropriate time in their residency.
Adopted by SAGES Executive Committee, August, 1997. Approved by SAGES Board of Governors, October, 1997.

Society of American Gastrointestinal Endoscopic Surgeons (SAGES)

2716 Ocean Park Boulevard, Suite 3000 Santa Monica, CA 90405, USA

Surg Endosc (1998) 12: 381382

Springer-Verlag New York Inc. 1998

Granting of privileges for gastrointestinal endoscopy by surgeons

I. Principles of Credentialing Preamble This document is to serve as a guide for granting privileges in gastrointestinal endoscopy as an integral part of surgical practice. Surgeons who are experienced in operating on the gastrointestinal tract are familiar with anatomy, tissue tolerance, organ compliance, and pathological processes, and should readily develop endoscopic proficiency which should be assessed independent of the number of procedures performed. A. Purpose The purpose of this statement is to outline principles and to provide practical suggestions to assist hospital credentialing committees in the task of granting privileges to perform gastrointestinal endoscopy. In conjunction with the standard JCAHO guidelines for granting hospital privileges, implementation of these methods should help hospital staffs insure that endoscopy is performed only by individuals with appropriate competency, thus assuring high quality patient care and proper procedure utilization. B. Uniformity of standards Uniform standards should be developed, which apply to all hospital staff requesting privileges to perform endoscopy, and to all areas where endoscopy is performed within a given institution. Criteria must be established that are medically sound, not unreasonably stringent, and that are applicable in common to all those wishing to obtain privileges in each specific endoscopic procedure. The goal must be the delivery of high quality patient care. C. Specificity of credentialing Privileges should be granted for each major category of endoscopy separately. The ability to perform one endoscopic procedure does not imply adequate competency to perform another. Associated skills that are generally considered to be an integral part of an endoscopic category may be required before privileges for that category can be granted. D. Responsibility for credentialing The credentialing structure and process remains always the individual responsibility of each hospital. It should be the responsibility of the Department of Surgery, through its chief, to recommend individual surgeons for privileges in gastrointestinal endoscopy as for other procedures performed by members of his or her department. II. Training and determination of competence A. Formal residency training in gastroenterology or surgery The Accreditation Council for Graduate Medical Education mandated that: The program must provide experience to each resident in the performance of a variety of rigid and flexible endoscopic procedures, including laryngoscopy, bronchoscopy, esophagoscopy, gastroscopy, colonoscopy, diagnostic and therapeutic laparoscopy, and intraoperative choledochoscopy, as well as the study and performance of new and evolving endoscopic techniques (Directory of Residency Training Programs, Graduate Medical Education Directory 19971998). B. Endoscopic training and experience outside a formal residency program Equivalent training and/or experience obtained outside a formal program is recognized, but must be at least equal to that described above. Certification of experience by a skilled endoscopic practitioner must include a detailed description of the nature of informal training, the number of procedures performed with and without supervision, and the actual observed competency of the applicant for each endoscopic procedure for which privileges are requested. It is generally no longer acceptable for physicians to acquire equivalent endoscopic experience by performing unsupervised procedures when skilled endoscopists are available in the medical community. C. Determination of Competence 1. Completion of a surgical residency program which incorporates structured experience in gastrointestinal endoscopy. Competence should be documented by the instructor(s). 2. Proficiency in endoscopic procedure(s) and clinical judgment equivalent to that obtained in a residency program. Documentation and demonstration of competence is necessary. 3. Participation in gastrointestinal endoscopic training until competence in the specific procedure(s) is equivalent to 1.

382

4. The applicants endoscopic director should confirm in writing the training, experience (including the number of cases for each procedure for which privileges are requested), and actually observed level of competency. It is recognized that by virtue of completing a residency program in surgery, the surgeon endoscopist will have acquired at least five years of cognitive experience in anatomy, physiology, disease process, combined with the progressive development of visual and psychomotor skills, and the experience necessary for the performance of diagnostic and therapeutic procedures in the gastrointestinal tract. Such experience includes indications, complications and their management, and alternative approaches. It is recommended that the performance of at least 25 esophagogastroduodenoscopies, 50 colonoscopies, or 25 flexible sigmoidoscopies serve as minimal requirements for consideration of competence (in the specific procedure) in addition to the above training background. The training directors opinion and recommendation should be considered prima facie evidence for the trainees acceptance as an individual qualified in gastrointestinal endoscopy. Likewise, attendance at short endoscopy courses, which do not provide supervised, hands-on training with patients, is not an acceptable substitute in the development of equivalent competency. D. New Procedures Self training in new techniques in gastrointestinal endoscopy must take place on a background of basic surgical and endoscopic skills. The endoscopist should recognize when additional training is necessary. E. Proctoring Recognizing the limitations of written reports, proctoring of applicants for privileges in gastrointestinal endoscopy by a qualified, unbiased staff endoscopist may be desirable, especially when competency for a given procedure cannot be adequately verified by submitted written material. The procedural details of proctoring should be developed by the credentialing body of the hospital and provided to the applicant. Proctors may be chosen from existing endoscopy staff or solicited from endoscopic societies. The proctor should be responsible to the credentials committee, and not to the patient or to the individual being proctored. Documentation of the proctors evaluation should be submitted in writing to the credentials committee. Criteria of competency for each procedure should be established in advance. It is essential that proctoring be provided in an unbiased, confidential, and objective manner. A satisfactory mechanism for appeal must be established for individuals for whom privileges are denied or granted in a temporary or provisional manner. F. Monitoring of endoscopic performance To assist the hospital credentialing body in the ongoing renewal of privileges, there should be a mechanism for monitoring each surgical endoscopists procedural perfor-

mance. This should be done through existing quality assurance mechanisms or, alternatively, through a multidisciplinary endoscopy committee. This should include monitoring endoscopic utilization, diagnostic and therapeutic benefits to patients, complications, and tissue review in accordance with previously developed criteria. G. Continuing education Continuing medical education related to endoscopy should be required as part of the periodic renewal of endoscopic privileges. Attendance at appropriate local, national, or international meetings and courses is encouraged. H. The renewal of privileges For the renewal of privileges, an appropriate level of continuing clinical activity should be required, in addition to satisfactory performance as assessed by monitoring of procedural activity through existing quality assurance mechanisms, as well as continuing medical education relating to gastrointestinal endoscopy. References
1. Anderson ML, Heigh RI, McCoy GA, Parent K, Muhm JR, McKee GS, Eversman WG, Collins JM (1992) Accuracy of assessment of the extent of examination by experienced colonoscopists. Gastrointest Endosc 38: 560563 2. Barthel J, Hinojosa T, Shah, N (1995) Colonoscope length and procedure efficiency. J Clin Gastroenterol 21: 3032 3. Chak A, Cooper GS, Blades EW, Canto M, Sivak MV Jr (1996) Prospective assessment of colonoscopic intubation skills in trainees. Gastrointest Endosc 44: 5457 4. Church JM (1994) Complete colonoscopy: how often? And if not, why not? Am J Gastroenterol 89: 556560 5. Cosgrove JM, Cohen JR, Wait RB, Margolis IB (1995) Endoscopy training during general surgery residency. Surg Laparosc Endosc 5: 393395 6. Galandiuk S (1995) A surgical subspecialist enhances general surgical operative experience. Arch Surg 130: 11361138 7. Gruber M (1996) Performance of flexible sigmoidoscopy by a clinical nurse specialist. Gastroenterol Nurs 19: 105108 8. Hasseman JH, Lemmel GT, Emad RY, Douglas RK (1997) Failure of colonoscopy to detect colorectal cancer: evaluation of 47 cases in 20 hospitals. Gastrointest Endosc 45: 451455 9. Jentschura D, Raute M, Winter J, Henkel Th, Kraus M, Manegold BC (1994) Complications in endoscopy of the lower gastrointestinal tract (therapy and prognosis). Surg Endosc 8: 672676 10. Marshall B (1995) Technical proficiency of trainees performing colonoscopy: a learning curve. Gastrointest Endosc 42: 287291 11. Parry BR, Williams SM (1991) Competency and the colonoscopist: a learning curve. Australia/New Zealand J Surg 61: 419422 12. Rai S, Moran MR, Rai AM (May 1996) Colon and rectal cancer: epidemiology and investigation, are colonoscopies performed by subspecialists more expensive? Dis Colon Rectum 39 13. Saad JA, Pirie P, Sprafka JM (1994) Relationships between flexible sigmoidoscopy training during residency and subsequent sigmoidoscopy performance in practice. Family Med 26: 250253 This statement was reviewed and approved by the Board of Governors of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), May, 1997. It was prepared by the SAGES Committee on Credentialing. This is a revision of SAGES Publication #004 originally printed 10/89 and of SAGES Publication #0011 printed 1/92.

Society of American Gastrointestinal Endoscopic Surgeons (SAGES)

2716 Ocean Park Boulevard, Suite 3000 Santa Monica, CA 90405, USA

SAGES guidelines
Surg Endosc (1998) 12: 379380 Springer-Verlag New York Inc. 1998

Guidelines for granting of privileges for laparoscopic and/or thoracoscopic general surgery
I. Principles of privileging Preamble The Society of American Gastrointestinal Endoscopic Surgeons (SAGES) recommends the following guidelines for privileging qualified surgeons in the performance of general surgical procedures utilizing laparoscopy and/or thoracoscopy. The basic premise is that the surgeon must have the judgment, training, and capability of immediately proceeding to a traditional open thoracic or abdominal procedure when circumstances so indicate. This document is to serve as a guide for granting privileges in laparoscopic and/or thoracoscopic surgery as an integral part of surgical practice. Surgeons who are experienced in operating in the abdomen and/or chest are familiar with anatomy, tissue tolerance, organ compliance, and pathological processes, and should readily develop laparoscopic and/or thoracoscopic proficiency, which should be assessed regardless of the number of procedures performed. C. Responsibility for privileging The privileging structure and process remain the individual responsibility of each hospital. It should be the responsibility of the Department of Surgery, through its Chief, to recommend individual surgeons for privileges in laparoscopic and/or thoracoscopic general surgery as for other procedures performed by members of the department. II. Training and determination of competence A. Formal residency training in general and/or thoracic surgery Prerequisite training must include satisfactory completion with Board eligibility or certification from residency programs in general and/or thoracic surgery accredited by the Accreditation Council for Graduate Medical Education or the equivalent body if the program is based outside the United States or Canada. B. Determination of competence in laparoscopic and/or thoracoscopic surgery 1. A surgical residency program, which incorporates structured experience in laparoscopic and/or thoracoscopic general surgery, should be completed. The applicants Program Director should confirm in writing the training, experience, and actual observed level of competency, which could include case lists, as is done for other procedures in surgery. 2. The surgeon should demonstrate proficiency in laparoscopic and/or thoracic surgical procedures and clinical judgment equivalent to that obtained in a residency program. The requirements for documentation and demonstration of competence is determined by the appropriate credentialing and qualifications committee. 3. For those surgeons without residency training, which included laparoscopic and/or thoracoscopic surgery, or without documented prior experience in these areas, the training should include didactics, hands-on experience, participation as a first assistant, and performance of the operation under proctorship. The basic minimum requirements for training should be: a. completion of approved residency training in general

A. Purpose The purpose of this statement is to outline principles and provide practical suggestions to assist hospital privileging committees when granting privileges to perform laparoscopic and/or thoracoscopic surgery. In conjunction with the standard JCAHO guidelines for granting hospital privileges, implementation of these methods should help hospital staffs ensure that laparoscopic and/or thoracoscopic surgery is performed only by individuals with appropriate competence, thus assuring high quality patient care and proper procedure utilization.

B. Uniformity of standards Uniform standards should be developed, which apply to all hospital staff requesting privileges to perform laparoscopic and/or thoracoscopic general surgery. Criteria must be established which are medically sound but not unreasonably stringent, and which are universally applicable to all those wishing to obtain privileges. The goal must be the delivery of high-quality patient care.

380

and/or thoracic surgery, with privileging in the comparable open procedure for which laparoscopic and/or thoracoscopic privileges are being sought, b. training in laparoscopic general and/or thoracoscopic surgery by a surgeon experienced in laparoscopic/ thoracoscopic surgery or completion of a didactic course sponsored by an institution or society accredited by ACCME. Such a course should include instruction in handling and use of laparoscopic and/or thoracoscopic instrumentation, establishment of safe peritoneal and/or thoracic access, tissue handling, knot tying, equipment utilization, as well as hands-on experience in specific categories of procedures for which the applicant desires privileges. The individual must demonstrate to the satisfaction of an experienced physician course director/preceptor that he/she can perform a given procedure from beginning to end. Such proficiency for each category of procedure in question must be documented in writing by the physician course director. The course content and procedures taught should clearly include material specific to the category of procedure for which privileges are sought. Attendance at short courses that do not provide supervised hands-on training or documentation of proficiency is not an acceptable substitute, c. proctoring by a laparoscopic and/or thoracoscopic surgeon experienced in the same or similar procedure(s) until proficiency has been observed and documented in writing.

toring competence. This should be done through existing quality assurance mechanisms. This should include monitoring utilization, diagnostic and therapeutic benefits to patients, complications, and tissue review in accordance with previously developed criteria. E. Continuing education Continuing medical education related to laparoscopic and/or thoracoscopic surgery should be required as part of the periodic renewal of privileges. Attendance at appropriate local, national, or international meetings and courses is encouraged. F. Renewal of privileges For the renewal of privileges, an appropriate level of continuing clinical activity should be required. In addition to satisfactory performance as assessed by monitoring of procedural activity through existing quality assurance mechanisms, continuing medical education relating to laparoscopic and/or thoracoscopic surgery should also be required. References
1. Dent TL (March 1991) Clinical privileges for laparoscopic general surgery. Am J Surg 161: 399403 2. E.A.E.S. Guidelines (1994) Training and assessment of Competence. Surg Endosc 8: 721722 3. Greene FL (1991) Training, credentialing and privileging for minimally invasive surgery. Prob General Surg 8: 502:506 4. Jakimowicz J (1994) The European Association for Endoscopic Surgery, recommendations for training in laparoscopic surgery. Ann Chirurg Gynaecol 83: 137141 5. Laparoscopic Surgery (June 12, 1992) New York State Department of Health Memorandum, Series 92-20, Albany, New York 6. Ooi, LLPJ Training in laparoscopic surgeryhave we got it right yet? Ann Acad Med 25: 732736 7. Wexner SD, Weiss EG (October 1994) A recommended training schema for laparoscopic surgerythe future of laparoscopy in oncology/surgical oncology clinics of North America, vol. 3, no. 4, 759765 8. Wexner SD, Weiss EG (December 1994) Training and preparing for laparoscopic colectomy. Semin Colon Rectal Surg 5: 224227 9. SAGES (1991) Granting of privileges for laparoscopic general surgery. Am J Surg 161: 324325 10. Schwaitzberg SD, Connolly RJ, Sant GR, Reindollar R, Clevland RJ (1996) Planning, development, and execution of and international training program in laparoscopic surgery. Surg Laparosc Endosc 6: 1015 11. See WA, Cooper CS, Fisher RJ (1993) Predictors of laparoscopic complications after formal training in laparoscopic surgery. JAMA 270: This statement was reviewed and approved by the Board of Governors of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES) May, 1997. It was prepared by the SAGES Committee on Credentialing. This is a revision of SAGES publication #0005 printed 5/90 and of SAGES publication #0014 printed 1/92 and 10/92.

C. Proctoring Recognizing the limitations of written reports, proctoring of applicants for privileges in laparoscopic and/or thoracoscopic surgery by a qualified, unbiased staff surgeon experienced in general and/or thoracic and laparoscopic and/or thoracoscopic surgery is recommended. The procedural details of proctoring should be developed by the privileging body of the hospital and provided to the applicant. Proctors may be chosen from existing staff or solicited from surgical endoscopic societies. The proctor should be responsible to the privileging committee, and not to the patient or to the individual being proctored. Documentation of the proctors evaluation should be submitted in writing to the privileging committee. Criteria of competency for each procedure should be established in advance and should include evaluation of: familiarity with instrumentation and equipment, competence in their use, appropriateness of patient selection, clarity of dissection, safety, time taken to complete the procedure and successful completion of same. It is essential that proctoring be provided in unbiased, confidential, and objective manner. A satisfactory mechanism for appeal must be established for individuals for whom privileges are denied or granted in a temporary or provisional manner.

D. Monitoring of laparoscopic and/or thoracoscopic performance To assist the hospital privileging body in the ongoing renewal of privileges, there should be a mechanism for moni-

Society of American Gastrointestinal Endoscopic Surgeons (SAGES)

2716 Ocean Park Boulevard, Suite 3000 Santa Monica, CA 90405, USA

Surg Endosc (1998) 12: 342347

Springer-Verlag New York Inc. 1998

Ultrasonic epithelial ablation of the lower esophagus without stricture formation

A new technique for Barretts ablation
R. M. Bremner, R. J. Mason, C. G. Bremner, T. R. DeMeester, P. Chandrasoma, J. H. Peters, J. A. Hagen, M. Gadensta tter
Department of Surgery, University of Southern California School of Medicine, 1510 San Pablo Street, Suite 514, Los Angeles, CA 90033-4612, USA Received: 3 April 1997/Accepted: 6 October 1997

Abstract Background: The premalignant potential of Barretts esophagus has stimulated efforts to find a way to ablate the columnar epithelium in order to reheal the area with squamous epithelium, thus obviating the cancer risk. This study describes and evaluates a new technique using ultrasonic energy to ablate the epithelium of the lower esophagus in a porcine model. Methods: Eight young farm pigs were used to develop the technique of applying a laparoscopic Cavitron Ultrasonic Surgical Aspirator (CUSA) to the lower esophageal mucosa through an operating gastrostomy. A further 11 Yakutan minipigs then underwent CUSA epithelial ablation, followed by a laparoscopic Nissen fundoplication or postoperative acid suppression therapy. We then assessed the healing response in these subjects. Results: Optimal CUSA energy settings enabled complete ablation of the squamous epithelium with preservation of the muscularis mucosa and submucosa. The integrity of the aspirated cells was sufficient for cytological analysis. Healing occurred by squamous regeneration without stricture formation. Conclusions: The CUSA technique holds promise for complete ablation of the Barretts epithelium in a single setting. The unique tissue-selective nature of the ablative process allows complete mucosal reepithelialization without stricture formation. Key words: Barretts esophagusEpithelial ablation

columnar-lined Barretts esophagus is estimated at 0.8% per year. This means that of 100,000 patients with Barretts, 800 will develop adenocarcinoma. This risk is sufficient to encourage annual endoscopic surveillance in patients without dysplasia. An alternative approach to surveillance is to ablate the Barretts epithelium and encourage reepithelialization of the ablated area with squamous epithelium. Ablation has been performed with photodynamic therapy (PDT), laser therapy, and bipolar electrocoagulation [25, 7, 9]. Although all of these methods have had variable success, they also have demonstrated significant limitations that have hindered their wider application. This study evaluates a new technique that utilizes the tissue-selective nature of ultrasonic energy to ablate the epithelium of the lower esophagus in a porcine model. Methods CUSA ultrasonic generator
The ultrasonic energy used in this experiment is generated by a Cavitron Ultrasonic Surgical Aspirator, or CUSA (CUSA Model 200, Valleylab Inc, Boulder, CO, USA). This generator is commonly used for liver resections and removal of brain tumors. The energy is delivered to the esophageal mucosa using an instrument that has been modified for laparoscopic access. The instrument is 30.1 cm long and consists of an energy tip measuring 2.54 mm in diameter, coupled with an irrigation and suction channel (Fig. 1). The tip vibrates at 23 kHz. The operators ability to adjust the amplitude varies the energy transferred to the tissue. A patent cavipulse setting on the instrument causes an oscillation of pulses at lower frequencies, which enhances the selective tissue disruption and fragmentation.

The risk of developing adenocarcinoma in a segment of Pilot and acute studies

Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 1922 March 1997 Correspondence to: T. R. DeMeester An initial pilot study was performed in eight young farm pigs (6080 lb) to ascertain the optimal CUSA settings for ablation of esophageal squamous epithelium and to determine the optimal port site placement to access the esophageal lumen by the CUSA probe. By trial and error, it was

343

Fig. 1. Photograph of the laparoscopic CUSA instrument. Fig. 2. Diagram illustrating placement of the operating gastrostomy port. The laparoscopic CUSA is inserted through the port. An endoscope is used to direct the CUSA into the lumen of the esophagus.

Fig. 3. Histologic view of the interface between the normal epithelium and the area of ablation (10 magnification). Note the absence of any epithelial cells in the ablated area. The muscularis mucosa is completely intact, and there is no evidence of injury to the submucosa or muscularis propria. The muscularis propia is not visible in the magnification; it is deep to the area shown.

determined that placement of the port near the greater curvature of the stomach in the midline ensured easy access to the esophageal lumen and enabled the ultrasound tip to be applied to the full circumference of the luminal mucosa (Fig. 2). Optimal settings for epithelial ablation were determined by removing the esophagus from four animals, opening it longitudinally, and pinning the opened esophagus to a board. The CUSA was applied to the strips of esophageal mucosa using variable settings of amplitude and cavipulse. The extent and depth of ablation was determined by histologic examination. A cavipulse of 3 with 90100% amplitude enabled ablation of the epithelium while leaving the underlying muscularis mucosa and muscularis propria intact (Fig. 3). Higher cavipulse settings resulted in submucosal hemorrhage and occasional disruption of the muscularis mucosa; lower amplitudes or cavipulse settings resulted in incomplete ablation of the epithelium. The chosen settings were then used to ablate the epithelium of the esophagus in the intact animals. After esophageal epithelial ablation was attempted in four animals, the esophagus of each animal was removed and examined histopathologically to establish the completeness of ablation.

Operative studies
The mucosal ablation technique was used on 11 Yakutan minipigs, who were allowed to recover. Five animals underwent CUSA ablation of the distal 5 cm of the esophagus and were given Omeprazole 20 mg per day for 7 days to prevent acid injury to the raw esophageal surface. Six animals underwent CUSA ablation followed by laparoscopic Nissen fundoplication under the same anesthetic. The Nissen fundoplication was performed by enveloping the lower esophagus, which was intubated with a 30-Fr bougie with a tongue of gastric fundus, and then suturing it in place with two Ethibond sutures. It was not necessary to take down the short gastric vessels because the pig has a large, floppy fundus. The diaphragmatic crura were sutured to close the dissection defect at the hiatus. The animals were endoscoped at 24 weeks to assess healing of the mucosa. They were killed at 6 weeks using 6 cc of Euthasol (390 mg/ml pentobarbital) for euthanasia. The esophagus in each animal was resected and intubated with a Penrose drain containing barium sulfate for radiographic examination to identify any luminal narrowing or stricture formation. All specimens underwent histological examination.

Anesthesia was maintained with 0.52% isoflurane. The abdominal wall was cleaned with betadine solution and draped in a standard sterile fashion. An Olympus endoscope was passed through the esophagus into the stomach and directed anteriorly. The room was darkened, and the light from the tip of the endoscope was visualized through the anterior abdominal wall. A 16-gauge needle was passed percutaneously into the stomach using the light source as a target. The entrance of the needle into the stomach was monitored visually through the endoscope. An operating gastrostomy tube (Cook Surgical, Cook Inc., Bloomington, IN, USA) was placed into the stomach using a standard percutaneous endoscopic gastrostomy technique (Fig. 2). The balloon of the port was insufflated with air to secure the anterior gastric wall to the anterior abdominal wall. Correct placement of the operating gastrostomy port was critical to enable the passage of a rigid laparoscopic CUSA instrument through the gastroesophageal junction and into the lumen of the lower esophagus. If the placement was incorrect, the CUSA instrument could not be applied to the complete circumference of the esophageal lumen. An endoscope was placed in the upper esophagus to visualize the insertion of the laparoscopic CUSA instrument into the esophagus. The tip of the CUSA probe was guided in a sweeping motion over the mucosal lining, beginning cephalad and moving progressively caudad. The ablated epithelial cells were collected from the central aspiration port and preserved for cytology. The epithelium was ablated over a 5-cm circumferential segment of lower esophagus just proximal to the gastroesophageal junction. Epithelial ablation could be easily confirmed endoscopically by loss of the pearly-white appearance of the epithelium, thus revealing the underlying dark red muscularis mucosa. There was no visible blood loss associated with the ablation. Bleeding during the ablation usually indicated an injury to the muscularis mucosa and disruption of submucosal capillaries. It could be avoided by using the correct CUSA power settings and avoiding repeated contact of the ultrasonic tip with previously ablated areas. Care was taken to avoid excessive luminal insufflation, since it tended to cause distention of the small bowel. At the conclusion of the ablation process, the operating gastrostomy port was removed and the gastrotomy closed with a running silk suture through a slightly enlarged skin incision at the gastrostomy site. The fascia was closed with interrupted Prolene sutures, and the skin was closed with staples. The animals were recovered and given buphrenex for postoperative pain management. Oral feeding was started on the 1st postoperative day.

Operative technique Ethics

Prior to the procedure, the animals were fasted for 48 h to ensure an empty stomach, since Yakutan minipigs are known to have a prolonged gastric emptying time. Animals were anesthetized with 75 mg Telazol (tiletamine with zelozepam 100 mg/ml), 35 mg xylozine, and 0.2 mg glycopyrolate. This study was approved by the animal Use Committee of the University of Southern California (protocol number 8976).

344

Fig. 4. The microscopic specimen of the cell block recovered from the aspiration port during the ablation (100 magnification). The cells have maintained their integrity for cytologic examination. Fig. 5. Endoscopic pinch biopsy specimen obtained at 2 weeks after ablation showing the early regeneration of squamous epithelium with three to six cell layers (40 magnification).

Results Acute studies Complete ablation of the epithelium of the lower esophagus was possible through the operating gastrostomy port in all animals. In all specimens, only the epithelium was ablated, leaving a thin coagulum overlying the muscularis mucosa. Macroscopically and histologically, no islands of epithelium were left in the area of ablation. The muscularis mucosa was intact in all animals, and no injury was observed either in the submucosa or muscularis propria. The integrity of the aspirated cells was excellent and allowed for cytological evaluation (Fig. 4). Operative studies Two of the 11 animals died during the 6-week observation period. One animal who underwent laparoscopic Nissen fundoplication died suddenly during the night on the 7th postoperative day from a cecal volvulus with bowel infarction. Pathology of the esophageal specimen revealed granulation tissue superficial to the muscularis mucosa; there was no evidence of epithelial islands, confirming the effectiveness of the ablation technique. A second animal was noted at endoscopy 4 weeks after surgery to have herniated the fundoplication into the chest. The injury was due to failure to close the diaphragmatic crura at the time of fundoplication. We found it prudent to kill the animal at this time even though she had been eating well and had not lost weight. On inspection, the esophageal mucosa was completely reepithelialized and free from stricture. The remaining animals were endoscoped at 2-week intervals. At 2 weeks, endoscopy showed areas of translucent white epithelium covering the granulation tissue in the ablated area. Histology of pinch biopsies showed epithelialization with three to six layers of squamous cells (Fig. 5). At 4 weeks, epithelium covered the whole of the ablated areas, but it was slightly pink in color. By 6 weeks, complete healing had occurred without stricture formation, and the ablated area could not be identified. Radiographic exami-

nation of the barium-filled Penrose drain passed through the intact specimen showed no evidence of stricturing (Figs. 6, 7). On histological examination, all specimens were free of fibrosis in the submucosa or the muscularis propria, and it was difficult to distinguish normal from regenerated epithelium. Discussion Currently, there is no effective means of reversing the metaplastic process associated with Barretts esophagus. High doses of proton pump inhibitors or antireflux surgery have occasionally resulted in total or partial regression of the columnar epithelium, but the occurrence is rare and unpredictable. Further, acid suppression therapy has not been shown to prevent malignant progression of the columnar lining, and there is insufficient evidence at present that antireflux surgery is protective [8, 12]. Consequently, the International Society of Diseases of the Esophagus (ISDE) recommends that patients with Barretts metaplasia be endoscoped and biopsied at yearly intervals in order to detect the emergence of dysplasia, a sign of movement toward invasive cancer [10]. The cost of surveillance over a lifetime is significant. Added to this is the cost of esophagectomy if high-grade dysplasia or cancer develops. Consequently, alternative approaches to the problem are being investigated. Photodynamic therapy (PDT), laser ablation, and multipolar electrocoagulation have all been used in humans to remove columnar epithelium with variable success [15, 7, 9]. The results of published studies have shown that squamous reepithelialization of ablated areas occurs with squamous cells provided gastric acid secretion is suppressed. One limitation of these techniques is that the ablation of Barretts epithelium is incomplete. There are reports of squamous overgrowth of partially ablated Barretts epithelium, as well as development of adenocarcinoma beneath squamous reepithelialized areas [6]. Another drawback of these techniques is the inability to control the depth of tissue injury, with resultant stricture formation in 50% of patients. The great advantage of ultrasonic energy is the tissue-

345

Fig. 6. Necropsy specimens of resected esophagi after complete mucosal healing. A Photograph of the esophagi intubated with barium-filled Penrose drains of group 1 animals. B A radiograph of these esophagi demonstrates the absence of any luminal narrowing or stricture formation. C Four of the five esophagi from the animals that had a Nissen fundoplication again demonstrating the absence of stricture formation. The fifth animal was killed at a different time but showed similar findings.

selective nature of the ablative process. Tissue that has a high elastin or collagen content or that consists of muscle cells is resistant to damage by the ultrasound energy. This explains why the CUSA instrument can selectively ablate hepatic parenchymal cells while preserving the vascular and ductal tissues during liver resections. The aim of the present study was to develop a technique whereby ultrasonic energy could be used to ablate the esophageal epithelium without injury to the underlying tissue layers. The technique that we developed utilizes recent advances in both endoscopy and laparoscopy. Selective ablation of the distal esophageal epithelium through the operating gastrostomy port was easily accomplished once optimal CUSA settings were defined and the proper positioning of the gastrostomy site was determined. The impressive histology of the acute experiments showed that complete ablation of all squamous epithelium was possible without injury to deeper layers of the esophageal wall. This explains why healing occurred in the absence of stricture formation. It is assumed that the columnar epithelium that occurs in Barretts metaplasia will act similarly, but this still needs to be demonstrated.

A valuable feature of this technique is the ability to collect the ablated epithelium by aspiration so that cytologic examination can be performed. All techniques utilizing chemical or thermal means to destroy the epithelial cells make it impossible to detect occult carcinoma in the ablated epithelium. In effect the aspiration specimen of the ultrasonic ablative technique represents the ultimate biopsy. Although there was no objective difference in the healing of the animals with a Nissen fundoplication, this is not to be interpreted that a Nissen fundoplication should not be performed following ablation. The animals in this study had normal lower esophageal sphincters and therefore minimal esophageal exposure to gastric contents. This is not the situation in patients with Barretts esophagus where the presence of a defective sphincter is the rule (12). The structurally defective lower esophageal sphincter in patients with Barretts esophagus makes the medical management of reflux tenuous at best. Even if the epithelium was successfully ablated in these patients, lifelong acid suppression therapy would still be required, and the potential for re-injury and metaplasia would always remain. Consequently, without the addition of a Nissen fundoplication, the need for strict sur-

346 impact of therapy on extent of Barretts esophagus in 67 patients. Dig Dis Sci 35: 9396 Sampliner RE, Hixson LJ, Fennerty MB, Garewal HS (1993) Regression of Barretts esophagus by laser ablation in an anacid environment. Dig Dis Sci 38: 365368 Skinner DB, Siewert JR (1996) Results of the consensus on esophageal cancer held at the VIth World Congress of the International Society for Diseases of the Esophagus 1995. Dis Esophagus 9: 156 Stein HJ, Hoeft S, DeMeester TR (1993) Functional foregut abnormalities in Barretts esophagus. J Thorac Cardiovasc Surg 105: 107 111 Williamson WA, Ellis FH Jr, Gibb SP, Shahian DM, Aretz HT (1990) Effect of antireflux operation on Barretts mucosa [see Comments]. Ann Thorac Surg 49: 537541

9. 10. 11. 12.

Discussion Dr. Hunter: Dr. Bremner, tell me a little bit about how you anticipate the utilization of this, and where you see your next series of experiments going: Is this ready for human application, and if so, what patients with Barretts are those in whom you would consider utilizing this technique? Dr. Bremner: To answer your first question first, Dr. Hunter, we anticipate using this now in Barretts tissue and cadaveric resected esophageal specimens to insure that our assumption that this will work on Barretts epithelium is the same as it works on squamous mucosa. It certainly, in the porcine stomach, seems to work very well; in fact, it seems to ablate the columnar cells of the stomach even easier than the squamous cells. We anticipate that assumption to hold true. Then we anticipate taking it to human studies. Obviously we would like to try patients with smaller segments of Barretts initially. We anticipate that this would be a procedure that would be performed at the time of laparoscopic Nissen fundoplication, so under the same anesthetic ablating the mucosa and then performing the Nissen. That was one of the reasons for doing the Nissen fundoplication at the time of ablation in the porcine model here. We had wondered whether or not the raw surfaces of the epithelium, after ablation, wont coapt after the addition of a Nissen fundoplication in the lower esophagus. It was our impression, although we have no objective evidence, that the animals that underwent Nissen fundoplication healed more quickly than those without the addition of an antireflux procedure, and possibly the reason for that is that pigs normally have some physiologic reflux. We had acid suppression in these animals only for a week; possibly some of the physiologic reflux continued thereafter. We think it would need to be done at the same time as an antireflux procedure. Dr. Greene: I just had a question about the method of approach. Realizing that the ultrasonic application has to be done through a rigid system, would you anticipate that this could be done through a rigid endoscopic system from an oral approach, and would that benefit us, because we recognize that all Barretts is not around the GE junction. In fact, we see Barretts changes well up into the esophagus. I wonder what your thoughts would berather than transgastric, what would be the opportunity of an oral or even a rigid approach? Dr. Bremner: Thank you very much, Dr. Greene. Its a good question. Weve been able to ablate the epithelium a whole 13 cm up from the gastroesophageal junction, but thats not

Fig. 7. An opened necropsy specimen of a healed esophagus 6 weeks after mucosal ablation shows complete healing with no evidence of stricture formation.

veillance would continue. It is felt that the technique presented in this study holds promise for the permanent eradication of Barretts columnar epithelium and with it the risk of malignancy.
Acknowledgments. This work was supported in part by grants from SAGES and Valleylab Inc. (Boulder, CO, USA). We would like to acknowledge Harry Valenta of Valleylab for his technical assistance and enthusiasm with the project, and, Linda and Paul Kirkman of Animal Unit at USC for their care and help with the animals both intraoperatively and postoperatively.

References
1. Barr H, Shepherd NA, Dix A, Roberts DJ, Tan WC, Krasner N (1996) Eradication of high-grade dysplasia in columnar-lined (Barretts) oesophagus by photodynamic therapy with endogenously generated protoporphyrin IX [see Comments]. Lancet 348: 584585 2. Berenson MM, Johnson TD, Markowitz NR, Buchi KN, Samowitz WS (1993) Restoration of squamous mucosa after ablation of Barretts esophageal epithelium [see Comments]. Gastroenterology 104: 1686 1691 3. Brandt LJ, Blansky RL, Kauvar DR (1995) Repeat laser therapy of recurrent Barretts epithelium: success with anacidity [Letter]. Gastrointest Endosc 41: 267 4. Overholt BF, Panjehpour M (1995) Barretts esophagus: photodynamic therapy for ablation of dysplasia, reduction of specialized mucosa, and treatment of superficial esophageal cancer [see Comments]. Gastrointest Endosc 42: 6470 5. Overholt BF, Panjehpour M (1996) Photodynamic therapy for Barretts esophagus: clinical update. Am J Gastroenterol 91: 17191723 6. Sampliner RE, Fass R (1993) Partial regression of Barretts esophagus: an inadequate endpoint. Am J Gastroenterol 88: 20922094 7. Sampliner RE, Fennerty B, Garewal HS (1996) Reversal of Barretts esophagus with acid suppression and multipolar electrocoagulation: preliminary results. Gastrointest Endosc 44: 532535 8. Sampliner RE, Garewal HS, Fennerty MB, Aickin M (1990) Lack of

347

going to reach the entire esophagus in human Barretts esophagus. The limitation is the length of the instrument. At the moment it is 30.1 cm long, and so from the orogastric approach youre only going to get the upper esophagus. But certainly from above and below youd be able to get the entire esophagus, if you needed to. Through a rigid esophagoscope, I dont anticipate that with the present technology well be able to reach the lower esophagus through the mouth, but thats just a technical limitation at the moment. Dr. Mulvihill: I enjoyed this very much. I wondered if you had any information on how rapidly there was restitution of this mucosa? Dr. Bremner: We do, and its in the manuscript. We endoscoped these animals during the healing process. One of the

animals had an early death from a cecal volvulus, and that was at postop day number 7. That enabled us to look at the esophageal specimen in its entirety at a week. This confirmed that there was no epithelium left, and there was just a thin layer of granulation tissue over the muscularis mucosa. That is at one week. At two weeks we had some endoscopy specimens that showed that there are approximately three to six layers of squamous epithelium; just a thin layer of squamous epithelium already at two weeks over the entire surface. At four weeks it looked almost healed, but endoscopically there was slack pink change between the normal epithelium and the regenerated. At six weeks we could hardly see any difference endoscopically, and our pathologist said it was very difficult for him to see any difference between the normal squamous and the regenerated areas.