- Email: [email protected]
Surgery or endoscopy for treatment of postcholecystectomy bile duct strictures?
The American Journal of Surgery 185 (2003) 532–535
Scientific paper
Surgery or endoscopy for treatment of postcholecystectomy bile duct strictures? Giovanni D. De Palma, M.D.,*, Giovanni Persico, M.D., Roberto Sottile, M.D., Alessandro Puzziello, M.D., Gianpaolo Iuliano, M.D., Vincenzo Salvati, M.D., Mario Donisi, M.D., Francesco Persico, M.D., Luigi Mastantuono, M.D., Marcello Persico, M.D., Stefania Masone, M.D. Department of Surgery and Advanced Technologies, University Federico II, School of Medicine, Naples, Italy Manuscript received April 19, 2002; revised manuscript November 2, 2002
Abstract Background: Surgery is considered the treatment of choice for postoperative biliary strictures. Recently, endoscopic stent placement has been proposed as an alternative to surgical management in selected patients. Methods: In this retrospective study, 157 patients with postoperative biliary strictures were included. Eighty patients (group A) were treated endoscopically and 77 by surgery (group B). Baseline characteristics of patients were comparable in both groups. Endoscopic therapy consisted of placement of endoprostheses, with trimonthly elective exchange for a 1-year period. Surgical therapy consisted of constructing a biliary-digestive anastomosis in normal ductal tissue. Data were evaluated according to intention-to treat analysis. Results: Successful treatment was achieved in 54% of group A and 73% of group B (P ⬍0.001). Overall 31% of patients developed complications in group A and 23% of patients in group B (P ⬍0.05). However, the rates of severe complications were comparable in both groups (11% versus 13%; P ⫽ not significant) In group A the mortality rate was 0% compared with 8% of group B (P ⬍0.05). Recurrent stenosis was evidenced in 6% of patients of group A and 5% of patients of group B. Conclusions: Surgery provides a better long-term outcome over the endoscopy, because of patients with total obstruction are not amenable to endoscopic approach. When successfully done, endoscopic results are similar to surgical results with less mortality. © 2003 Excerpta Medica, Inc. All rights reserved. Keywords: Bile duct stricture; Postoperative complications; Endoscopic stents; Surgery
The majority of benign bile duct strictures occur as a result of injuries sustained during the course of cholecystectomy with or without common bile duct exploration. After the introduction of laparoscopic cholecystectomy, there have been suggestions that the incidence of postcholecystectomy bile duct injuries has increased [1–3]. These injuries are often devastating, particularly since many patients are young, previously healthy women of childbearring age. Management of postoperative biliary strictures has traditionally been surgical, mainly Roux-en-Y hepaticojejunostomy. Recently, endoscopic stent placement has been proposed as an alternative to surgical management in se-
* Corresponding author. Tel.: ⫹0039-81-8262866; fax: ⫹0039-818262866. E-mail address: [email protected]
lected patients [4 –7]. Endoscopic stent placement is often the initial therapy in symptomatic patients with postoperative strictures because they are usually diagnosed at the time of endoscopic retrograde cholangiopancreatography (ERCP), at least in the era before magnetic resonance cholangiopancreatography (MRCP). The most important criticism of endoscopic management relates to the relatively short follow-up of these patients that left open the possibility of recurrent stricture formation years after the original therapy. Opponents argued that it only delayed the inevitable endpoint of surgery, at a later time under less favorable conditions, even though the results of surgical approach is not entirely satisfactory, with recurrent stricture formation in as many as 12% to 25% of patients [8 –11]. The aim of the present study was to compare the results
0002-9610/03/$ – see front matter © 2003 Excerpta Medica, Inc. All rights reserved. doi:10.1016/S0002-9610(03)00082-5
G.D. De Palma et al / The American Journal of Surgery 185 (2003) 532–535
of surgery and endoscopy for postoperative biliary strictures in one institution, over the same period of time and the same outcome definitions. The study was conducted according to an intention-to treat analysis; this provides a realistic estimation of the outcome of endoscopic versus surgical treatment, and supplies the basis for establishing a proper comparison between options in terms of long-term results.
Methods We conducted a chart review of patients referred for surgical or endoscopic treatment of biliary postoperative strictures from January 1, 1990, to June 30, 2001. Postoperative bile duct strictures included patients with strictures after open or laparoscopic cholecystectomy, or common bile duct exploration. Patients with bile duct strictures discovered and repaired during surgery were excluded from the study. Our study groups consisted of two separate populations. The endoscopic group (group A) consisted of 80 patients; the surgical group (group B) consisted of 77 patients who underwent surgical repair as primary procedure or after unsuccessful endoscopic treatment. Because so many patients were referred from various institutions, the choice of treatment (endoscopic or surgical) frequently reflected a different approach to biliary postoperative strictures used by the referring physicians. Moreover, in many cases, the endoscopic approach was selected by the patients or surgeons because of they were reluctant to operate given that a surgical complication created the problem in the first place. Variables recorded included demographic data, presenting symptoms, interval between the time of surgery and referral, and level of the stenosis. The level of the bile duct stricture was classified according to the Bismuth grading system [12]. Patients characteristics were comparable in both groups. The endoscopic procedure consisted of the insertion of polyethylene stent bridging the stenosis. The treatment protocol specified placement of two 10F stents. For multiple stent insertion, an endoscopic sphincterotomy was performed to facilitate stent placement. If the stenosis was too tight to allow passage of the catheter, dilating catheters of gradually increasing diameter were passed over the guidewire and through the stenosis, sometimes with supplemental balloon dilatation. There was no standardized protocol concerning balloon dilation or catheter dilation before stent placement. With tight stenoses sometimes only a 7F or a single 10F could be placed. In such cases, the single stent was exchanged for two 10F stents after 6 weeks. Stents were electively exchanged every 3 months, for a 1-year period. The stents were not replaced if the stenosis was considered to be adequately dilated based on the following criteria as subjectively assessed by the endoscopist: (1) adequate dilation of the stenosis on cholangiography; (2) satisfactory drainage of contrast from the biliary tree; and (3) passage of
533
a diagnostic catheter and extraction balloon through the stenosis without encountering significant resistance. Surgical procedure consisted of various procedures of biliary enteric repair, mainly Roux-en-Y hepaticojejunostomy. Surgical reconstruction was generally performed as an elective procedure. If there was evidence of ongoing bile leak or sepsis, percutaneous or endoscopic biliary stent placement was attempted as the first step, with drainage of bile collection, as indicate. Follow-up was obtained through hospital charts (patients were initially seen for follow-up at 3- to 6-month intervals and subsequently at yearly intervals if asymptomatic). Additional information was obtained by readmitting the patients to the hospital. The history was taken and physical examinations and liver function tests were performed, as were follow-up ultrasound, hepatoiminoadiacetic acid scanning, cholangiographic studies or MRCP, when indicated. The patient’s current status was rated as excellent if there were no symptoms attributable to the biliary tract injury, or good if mild symptoms, not requiring invasive investigation or treatment, were present. Patients classified as either excellent or good were considered to be treatment success. Patients were considered treatment failures if an invasive procedure, either radiologic, endoscopic or surgical, was necessary to treat ongoing symptoms or strictures recurrence. Results are reported as mean ⫾ SD. The Pearson chisquare test with Yates’ correction and the Fisher exact test for categorical comparison of data were used as appropriate. The t test was used for continuous data comparison. Significance was accepted at the 5% level. The calculations were performed with SPSS statistical software (SPSS, Chicago, Illinois)
Results Baseline characteristics of patients, presenting symptoms, surgical history and level of the stenosis, for both groups, are shown in Table 1. In group A, ERCP with stent insertion was attempted in 80 patients. A diagnostic cholangiogram was obtained in all patients. Total obstruction in the bile duct was evidenced in 17 (21%) patients. Successful stent insertion was achieved in 58 of 80 patients (72%). Reasons for unsuccessful stent insertion were as follows: total obstruction in the bile duct (n ⫽ 17; 19%), successful proximal cholangiography but inability to pass a guidewire though stenosis (n ⫽ 5; 6%). Further management for the 22 patients in whom stent insertion was unsuccessful, was as follow: surgery (n ⫽ 19), percutaneous stent plus surgery (n ⫽ 3). Procedure-related complications occurred in 11 of 80 (14%) cases: fever/ cholangitis (n ⫽ 7), pancreatitis (n ⫽ 3), and postsphincterotomy bleeding (n ⫽ 1). Fourteen (17%) patients developed complications during the period with stent in situ: stent clogging (n ⫽ 9), and fever/cholangitis (n ⫽ 5), related to
534
G.D. De Palma et al / The American Journal of Surgery 185 (2003) 532–535
Table 1 Baseline characteristics of patients
Number Male/Female Mean age (years ⫾ SD) Presenting symptoms* Jaundice Cholangitis Cholestatic liver function tests Bile leakage Surgical history Open cholecystectomy Laparoscopic Cholecystecomy Open cholecystectomy and CBDE Location of the stenosis Bismuth I Bismuth II Bismuth III Bismuth IV Bismuth V
Group A
Group B
80 18/62 46 ⫾ 18.5
77 22/55 48 ⫾ 19.2
50 (63%) 18 (23%) 21 (26%) 9 (11%)
46 (60%) 20 (26%) 16 (21.%) 13 (17%)
31 49 —
34 38 5
9 (11%) 49 (61%) 14 (17%) 7 (9%) 1 (1%)
9 (12%) 52 (67%) 11 (14%) 5 (6%) —
* Many patients experienced more than one symptom. CBDE ⫽ common bile duct exploration.
stent clogging. According to Cotton et al [13] 16 patients had mild complications and 9 patients severe complications. Ten patients were lost to follow-up. The follow-up in the remaining 70 cases was 5.9 years on average (range 0.9 to 10.6). Seven (9%) patients developed symptoms of stricture recurrence. The ERCP showed a recurrence of stricture in 5 (6%) cases and stones in the remaining. Stricture recurred 11.8 months on average (range 1 to 48) after removal of the stent. In group B, surgical management consisted of a Roux en-Y hepaticojejunostomy in 52 of 77 cases. In 11 of 77 cases repair was to the hepatic duct at hilus using the approach of Hepp and Couinaud to the left hepatic duct, while in 5 of 77 the segment III duct was approached, using the ligamentum teres approach described by Blumgart and Kelley. In 9 of 77 patients a choledochoduodenostomy was performed. There were 18 (23%) complications with 6 (8%) cases of related mortality: myocardial infarction (n ⫽ 2), diffuse peritonitis (n ⫽ 2), pulmonary embolism (n ⫽ 2), stress ulcer (n ⫽ 2), variceal bleeding (n ⫽ 1), atelectasis of the lower pulmonary lobe (n ⫽ 1), and wound infection (n ⫽ 8). Four patients were lost to follow-up. The follow-up in
Table 3 Further management of patients, in both groups A and B, on whom treatment was unsuccessful Number Unsuccessful stent insertion surgery percutaneous stent plus surgery Recurrence after stent removal surgery expectant management Recurrence after surgery percutaneous dilation surgery
22 19 3 5 4 1 4 3 1
the remaining 67 cases was 7.7 years on average (range 1 to 11.6 years). In 4 (5%) patients symptoms developed of stricture recurrence. Stricture recurred 6.8 years on average (range 2.1 to 11.8) after surgical intervention. All strictures recurrenced at the bilioenteric anastomosis. In summary, successful treatment was achieved in 43 of 80 (54%) patients in group A and in 56 of 77 (73%) in group B (Table 2). Further management for the patients, in both groups A and B, in whom treatment was unsuccessful, is shown in Table 3.
Comments There are no, at present, randomized controlled clinical trials comparing surgical and endoscopic treatment of postoperative biliary stenoses. Thus, the debate regarding which should be the preferred treatment option is generally based on the comparison between the outcomes of independent cohorts of patients. A retrospective study of surgical versus endoscopic therapy showed that both approaches had similar long term success rate, but this assumption may be misleading [14]. There are no an intention-to-treat studies, and this is of paramount importance because a noteworthy proportion of patient will be excluded from the endoscopic treatment because of complete ductal obstruction. The present intention-to treat analysis assesses the results obtained with surgical and endoscopic treatment in a cohort of 157 patients with postoperative biliary stricture treated in a single institution during the last 2 decades. This provides
Table 2 Outcome of endoscopic versus surgical treatment
Successful drainage Procedure-related complications Mortality Recurrence of stenosis Successful treatment NS ⫽ not significant.
Group A (endoscopy)
Group B (surgery)
P value
58/80 (72%) 25/80 (31%) 0/88 (0%) 5/80 (6%) 43/80 (54%)
77/77 (100%) 18/77 (23%) 6/77 (8%) 4/77 (5%) 56/77 (73%)
⬍ 0.001 ⬍ 0.05 ⬍ 0.05 NS ⬍ .001
G.D. De Palma et al / The American Journal of Surgery 185 (2003) 532–535
a realistic estimation of the outcome of endoscopic treatment, and supplies the basis for establishing a proper comparison between options in terms of long-term results. According to intention-to-treat analysis, results of our study show that surgical treatment provides a better longterm outcome over the endoscopic approach. (73% versus 54% P ⬍0.001). Endoscopic treatment of total ductal obstruction is not possible. Because ERCP was required for diagnosis (MRCP was not available during most of the study period) treatment of these patients was regarded as unsuccessful. Similarly, the following patients were considered failures: patients in whom initial stent insertion was unsuccessful, patients lost to follow-up during the stent in-situ phase and treated elsewhere without adherence to protocol, and patients with restenosis after stent removal. This is the “worst case scenario” according to the intention-to-treat principle (all results and follow-up are based on the original total cohort of patients submitted to ERCP). If these patients are arbitrarily excluded and results evaluated according to per-protocol analysis (only patients who received successful stent insertion), the overall success rate increases to 74% (43 of 58), which is the same as surgery. Overall, treatment-related complications rate was significantly higher in the endoscopic group (31% versus 23% P ⬍0.05), However, many of these complications were mild, and mainly related to the sphincterotomy (stent-insertion phase) or the stent clogging, during stent in-situ phase. If these patients were excluded, the rate of severe complication in the endoscopic group was 11% (9 of 80 cases), which is the same as surgery (10 of 77 cases, 13%). In the endoscopic group mortality rate was 0% compared with 8% of the surgical group (P ⬍0.05). With respect to 3 of 6 deaths, however, a causal relationship with surgical management may be disputed: 2 patients died of myocardial infarction, and 1 died of variceal bleeding (biliary cirrhosis was recognized at initial presentation). If these patients are arbitrarily excluded, the rate of mortality in the surgical group was 3.9%. Recurrent stenosis was evidenced in 5 of 80 (6%) patients of the endoscopic group and 4 of 77 (5%) patients of the surgical group. Restenosis after endoscopic treatment developed earlier compared with surgical approach (11.8 months versus 6.8 years; P ⬍0.00001). It is our impression that stenoses in the more proximal segments of the biliary system were more difficult to treat by endoscopic stenting, and, in our opinion, primary surgical treatment should be advised for patients with hilar stenoses. However, in the present study the number of patients with restenosis was too small to correlate bile duct segment as a risk factor with restenosis. Multivariate analysis of large cohorts of patients will be necessary to resolve this issue.
535
In conclusion, results of our study show that, surgery provides a better long-term outcome over the endoscopy, because patients with total obstruction are not amenable to the endoscopic approach. When successfully done, endoscopic results are similar to surgical results with less mortality. Although the endoscopic approach requires multiple sessions, most of our patients and referring physicians prefer endoscopic treatment to surgery. Prior endoscopic treatment does not preclude surgery, whereas endoscopic treatment is impossible once a Roux-en-Y loop has been constructed. Moreover, recurrent stenoses can also be successfully treated by repeated stent insertion. These considerations and the absence of treatment-related mortality lead us to propose endoscopic approach as a first-line treatment for postoperative biliary strictures. Indications for surgery are complete transection, failed previous repairs, and failures of endoscopic therapy. References [1] Davidoff AM, Branum GD, Meyers WC. Clinical features and mechanisms of major laparoscopic biliary injury. Semin Ultrasound CT MR 1993;14:338 – 45. [2] The Southern Surgeons Club. A prospective analysis of 1518 laparoscopic cholecystectomies. N Engl J Med 1991;324:1073– 8. [3] Deziel DJ, Millikan KW, Economou S, et al. Complications of laparoscopic cholecystectomy: a national survey of 4292 hospitals and analysis of 77,604 cases. Am J Surg 1993;165:9 –14. [4] Berkelhammer C, Kortan P, Haber GB. Endoscopic biliary prostheses as treatment for benign postoperative bile duct strictures. Gastrointest Endosc 1989;35:95–101. [5] Davids PH, Rauws EA, Coene PP, et al. Endoscopic stenting for postoperative biliary strictures. Gastrointest Endosc 1992;38:12–18. [6] Bergman JJGHM, Burgemeister L, Bruno MJ, et al. Long-term follow up after biliary stent placement for postoperative bile duct stenosis. Gastrointest Endosc 2001;54:154 – 61. [7] Costamagna G, Pandolfi M, Mutignani M, et al. Long-term results of endoscopic management of postoperative bile duct strictures with increasing numbers of stents. Gastrointest Endosc 2001;54:162– 8. [8] Lee JG, Leung JW. Long-term follow-up after biliary stent placement for postoperative bile duct stenosis. Gastrointest Endosc 2001;54: 272– 4. [9] Bottger T, Junginger T. Long term results after surgical treatment of iatrogenic injury of the bile ducts. Eur J Surg 1991;157:477– 80. [10] Ross CB, H’Doubler WZ, Sharp KW, Potts III. Recent experience with benign biliary strictures. Am Surg 1989;55:64 –70. [11] Frattaroli FM, Reggio D, Illomei G, Pappalardo G. Benign biliary strictures: a review of 21 years of experience. J Am Coll Surg 1996;183:506 –13. [12] Bismuth H. Postoperative strictures of the bile duct. In: Blumgart LH, editor. The biliary tract. Edinburgh: Churchill Livingstone, 1982, p 209 –18. [13] Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37:383–93. [14] Davids PH, Tanka AK, Rauws EA, et al. Benign biliary strictures. Surgery or endoscopy? Ann Surg 1993;217:237– 43.
Scientific paper
Surgery or endoscopy for treatment of postcholecystectomy bile duct strictures? Giovanni D. De Palma, M.D.,*, Giovanni Persico, M.D., Roberto Sottile, M.D., Alessandro Puzziello, M.D., Gianpaolo Iuliano, M.D., Vincenzo Salvati, M.D., Mario Donisi, M.D., Francesco Persico, M.D., Luigi Mastantuono, M.D., Marcello Persico, M.D., Stefania Masone, M.D. Department of Surgery and Advanced Technologies, University Federico II, School of Medicine, Naples, Italy Manuscript received April 19, 2002; revised manuscript November 2, 2002
Abstract Background: Surgery is considered the treatment of choice for postoperative biliary strictures. Recently, endoscopic stent placement has been proposed as an alternative to surgical management in selected patients. Methods: In this retrospective study, 157 patients with postoperative biliary strictures were included. Eighty patients (group A) were treated endoscopically and 77 by surgery (group B). Baseline characteristics of patients were comparable in both groups. Endoscopic therapy consisted of placement of endoprostheses, with trimonthly elective exchange for a 1-year period. Surgical therapy consisted of constructing a biliary-digestive anastomosis in normal ductal tissue. Data were evaluated according to intention-to treat analysis. Results: Successful treatment was achieved in 54% of group A and 73% of group B (P ⬍0.001). Overall 31% of patients developed complications in group A and 23% of patients in group B (P ⬍0.05). However, the rates of severe complications were comparable in both groups (11% versus 13%; P ⫽ not significant) In group A the mortality rate was 0% compared with 8% of group B (P ⬍0.05). Recurrent stenosis was evidenced in 6% of patients of group A and 5% of patients of group B. Conclusions: Surgery provides a better long-term outcome over the endoscopy, because of patients with total obstruction are not amenable to endoscopic approach. When successfully done, endoscopic results are similar to surgical results with less mortality. © 2003 Excerpta Medica, Inc. All rights reserved. Keywords: Bile duct stricture; Postoperative complications; Endoscopic stents; Surgery
The majority of benign bile duct strictures occur as a result of injuries sustained during the course of cholecystectomy with or without common bile duct exploration. After the introduction of laparoscopic cholecystectomy, there have been suggestions that the incidence of postcholecystectomy bile duct injuries has increased [1–3]. These injuries are often devastating, particularly since many patients are young, previously healthy women of childbearring age. Management of postoperative biliary strictures has traditionally been surgical, mainly Roux-en-Y hepaticojejunostomy. Recently, endoscopic stent placement has been proposed as an alternative to surgical management in se-
* Corresponding author. Tel.: ⫹0039-81-8262866; fax: ⫹0039-818262866. E-mail address: [email protected]
lected patients [4 –7]. Endoscopic stent placement is often the initial therapy in symptomatic patients with postoperative strictures because they are usually diagnosed at the time of endoscopic retrograde cholangiopancreatography (ERCP), at least in the era before magnetic resonance cholangiopancreatography (MRCP). The most important criticism of endoscopic management relates to the relatively short follow-up of these patients that left open the possibility of recurrent stricture formation years after the original therapy. Opponents argued that it only delayed the inevitable endpoint of surgery, at a later time under less favorable conditions, even though the results of surgical approach is not entirely satisfactory, with recurrent stricture formation in as many as 12% to 25% of patients [8 –11]. The aim of the present study was to compare the results
0002-9610/03/$ – see front matter © 2003 Excerpta Medica, Inc. All rights reserved. doi:10.1016/S0002-9610(03)00082-5
G.D. De Palma et al / The American Journal of Surgery 185 (2003) 532–535
of surgery and endoscopy for postoperative biliary strictures in one institution, over the same period of time and the same outcome definitions. The study was conducted according to an intention-to treat analysis; this provides a realistic estimation of the outcome of endoscopic versus surgical treatment, and supplies the basis for establishing a proper comparison between options in terms of long-term results.
Methods We conducted a chart review of patients referred for surgical or endoscopic treatment of biliary postoperative strictures from January 1, 1990, to June 30, 2001. Postoperative bile duct strictures included patients with strictures after open or laparoscopic cholecystectomy, or common bile duct exploration. Patients with bile duct strictures discovered and repaired during surgery were excluded from the study. Our study groups consisted of two separate populations. The endoscopic group (group A) consisted of 80 patients; the surgical group (group B) consisted of 77 patients who underwent surgical repair as primary procedure or after unsuccessful endoscopic treatment. Because so many patients were referred from various institutions, the choice of treatment (endoscopic or surgical) frequently reflected a different approach to biliary postoperative strictures used by the referring physicians. Moreover, in many cases, the endoscopic approach was selected by the patients or surgeons because of they were reluctant to operate given that a surgical complication created the problem in the first place. Variables recorded included demographic data, presenting symptoms, interval between the time of surgery and referral, and level of the stenosis. The level of the bile duct stricture was classified according to the Bismuth grading system [12]. Patients characteristics were comparable in both groups. The endoscopic procedure consisted of the insertion of polyethylene stent bridging the stenosis. The treatment protocol specified placement of two 10F stents. For multiple stent insertion, an endoscopic sphincterotomy was performed to facilitate stent placement. If the stenosis was too tight to allow passage of the catheter, dilating catheters of gradually increasing diameter were passed over the guidewire and through the stenosis, sometimes with supplemental balloon dilatation. There was no standardized protocol concerning balloon dilation or catheter dilation before stent placement. With tight stenoses sometimes only a 7F or a single 10F could be placed. In such cases, the single stent was exchanged for two 10F stents after 6 weeks. Stents were electively exchanged every 3 months, for a 1-year period. The stents were not replaced if the stenosis was considered to be adequately dilated based on the following criteria as subjectively assessed by the endoscopist: (1) adequate dilation of the stenosis on cholangiography; (2) satisfactory drainage of contrast from the biliary tree; and (3) passage of
533
a diagnostic catheter and extraction balloon through the stenosis without encountering significant resistance. Surgical procedure consisted of various procedures of biliary enteric repair, mainly Roux-en-Y hepaticojejunostomy. Surgical reconstruction was generally performed as an elective procedure. If there was evidence of ongoing bile leak or sepsis, percutaneous or endoscopic biliary stent placement was attempted as the first step, with drainage of bile collection, as indicate. Follow-up was obtained through hospital charts (patients were initially seen for follow-up at 3- to 6-month intervals and subsequently at yearly intervals if asymptomatic). Additional information was obtained by readmitting the patients to the hospital. The history was taken and physical examinations and liver function tests were performed, as were follow-up ultrasound, hepatoiminoadiacetic acid scanning, cholangiographic studies or MRCP, when indicated. The patient’s current status was rated as excellent if there were no symptoms attributable to the biliary tract injury, or good if mild symptoms, not requiring invasive investigation or treatment, were present. Patients classified as either excellent or good were considered to be treatment success. Patients were considered treatment failures if an invasive procedure, either radiologic, endoscopic or surgical, was necessary to treat ongoing symptoms or strictures recurrence. Results are reported as mean ⫾ SD. The Pearson chisquare test with Yates’ correction and the Fisher exact test for categorical comparison of data were used as appropriate. The t test was used for continuous data comparison. Significance was accepted at the 5% level. The calculations were performed with SPSS statistical software (SPSS, Chicago, Illinois)
Results Baseline characteristics of patients, presenting symptoms, surgical history and level of the stenosis, for both groups, are shown in Table 1. In group A, ERCP with stent insertion was attempted in 80 patients. A diagnostic cholangiogram was obtained in all patients. Total obstruction in the bile duct was evidenced in 17 (21%) patients. Successful stent insertion was achieved in 58 of 80 patients (72%). Reasons for unsuccessful stent insertion were as follows: total obstruction in the bile duct (n ⫽ 17; 19%), successful proximal cholangiography but inability to pass a guidewire though stenosis (n ⫽ 5; 6%). Further management for the 22 patients in whom stent insertion was unsuccessful, was as follow: surgery (n ⫽ 19), percutaneous stent plus surgery (n ⫽ 3). Procedure-related complications occurred in 11 of 80 (14%) cases: fever/ cholangitis (n ⫽ 7), pancreatitis (n ⫽ 3), and postsphincterotomy bleeding (n ⫽ 1). Fourteen (17%) patients developed complications during the period with stent in situ: stent clogging (n ⫽ 9), and fever/cholangitis (n ⫽ 5), related to
534
G.D. De Palma et al / The American Journal of Surgery 185 (2003) 532–535
Table 1 Baseline characteristics of patients
Number Male/Female Mean age (years ⫾ SD) Presenting symptoms* Jaundice Cholangitis Cholestatic liver function tests Bile leakage Surgical history Open cholecystectomy Laparoscopic Cholecystecomy Open cholecystectomy and CBDE Location of the stenosis Bismuth I Bismuth II Bismuth III Bismuth IV Bismuth V
Group A
Group B
80 18/62 46 ⫾ 18.5
77 22/55 48 ⫾ 19.2
50 (63%) 18 (23%) 21 (26%) 9 (11%)
46 (60%) 20 (26%) 16 (21.%) 13 (17%)
31 49 —
34 38 5
9 (11%) 49 (61%) 14 (17%) 7 (9%) 1 (1%)
9 (12%) 52 (67%) 11 (14%) 5 (6%) —
* Many patients experienced more than one symptom. CBDE ⫽ common bile duct exploration.
stent clogging. According to Cotton et al [13] 16 patients had mild complications and 9 patients severe complications. Ten patients were lost to follow-up. The follow-up in the remaining 70 cases was 5.9 years on average (range 0.9 to 10.6). Seven (9%) patients developed symptoms of stricture recurrence. The ERCP showed a recurrence of stricture in 5 (6%) cases and stones in the remaining. Stricture recurred 11.8 months on average (range 1 to 48) after removal of the stent. In group B, surgical management consisted of a Roux en-Y hepaticojejunostomy in 52 of 77 cases. In 11 of 77 cases repair was to the hepatic duct at hilus using the approach of Hepp and Couinaud to the left hepatic duct, while in 5 of 77 the segment III duct was approached, using the ligamentum teres approach described by Blumgart and Kelley. In 9 of 77 patients a choledochoduodenostomy was performed. There were 18 (23%) complications with 6 (8%) cases of related mortality: myocardial infarction (n ⫽ 2), diffuse peritonitis (n ⫽ 2), pulmonary embolism (n ⫽ 2), stress ulcer (n ⫽ 2), variceal bleeding (n ⫽ 1), atelectasis of the lower pulmonary lobe (n ⫽ 1), and wound infection (n ⫽ 8). Four patients were lost to follow-up. The follow-up in
Table 3 Further management of patients, in both groups A and B, on whom treatment was unsuccessful Number Unsuccessful stent insertion surgery percutaneous stent plus surgery Recurrence after stent removal surgery expectant management Recurrence after surgery percutaneous dilation surgery
22 19 3 5 4 1 4 3 1
the remaining 67 cases was 7.7 years on average (range 1 to 11.6 years). In 4 (5%) patients symptoms developed of stricture recurrence. Stricture recurred 6.8 years on average (range 2.1 to 11.8) after surgical intervention. All strictures recurrenced at the bilioenteric anastomosis. In summary, successful treatment was achieved in 43 of 80 (54%) patients in group A and in 56 of 77 (73%) in group B (Table 2). Further management for the patients, in both groups A and B, in whom treatment was unsuccessful, is shown in Table 3.
Comments There are no, at present, randomized controlled clinical trials comparing surgical and endoscopic treatment of postoperative biliary stenoses. Thus, the debate regarding which should be the preferred treatment option is generally based on the comparison between the outcomes of independent cohorts of patients. A retrospective study of surgical versus endoscopic therapy showed that both approaches had similar long term success rate, but this assumption may be misleading [14]. There are no an intention-to-treat studies, and this is of paramount importance because a noteworthy proportion of patient will be excluded from the endoscopic treatment because of complete ductal obstruction. The present intention-to treat analysis assesses the results obtained with surgical and endoscopic treatment in a cohort of 157 patients with postoperative biliary stricture treated in a single institution during the last 2 decades. This provides
Table 2 Outcome of endoscopic versus surgical treatment
Successful drainage Procedure-related complications Mortality Recurrence of stenosis Successful treatment NS ⫽ not significant.
Group A (endoscopy)
Group B (surgery)
P value
58/80 (72%) 25/80 (31%) 0/88 (0%) 5/80 (6%) 43/80 (54%)
77/77 (100%) 18/77 (23%) 6/77 (8%) 4/77 (5%) 56/77 (73%)
⬍ 0.001 ⬍ 0.05 ⬍ 0.05 NS ⬍ .001
G.D. De Palma et al / The American Journal of Surgery 185 (2003) 532–535
a realistic estimation of the outcome of endoscopic treatment, and supplies the basis for establishing a proper comparison between options in terms of long-term results. According to intention-to-treat analysis, results of our study show that surgical treatment provides a better longterm outcome over the endoscopic approach. (73% versus 54% P ⬍0.001). Endoscopic treatment of total ductal obstruction is not possible. Because ERCP was required for diagnosis (MRCP was not available during most of the study period) treatment of these patients was regarded as unsuccessful. Similarly, the following patients were considered failures: patients in whom initial stent insertion was unsuccessful, patients lost to follow-up during the stent in-situ phase and treated elsewhere without adherence to protocol, and patients with restenosis after stent removal. This is the “worst case scenario” according to the intention-to-treat principle (all results and follow-up are based on the original total cohort of patients submitted to ERCP). If these patients are arbitrarily excluded and results evaluated according to per-protocol analysis (only patients who received successful stent insertion), the overall success rate increases to 74% (43 of 58), which is the same as surgery. Overall, treatment-related complications rate was significantly higher in the endoscopic group (31% versus 23% P ⬍0.05), However, many of these complications were mild, and mainly related to the sphincterotomy (stent-insertion phase) or the stent clogging, during stent in-situ phase. If these patients were excluded, the rate of severe complication in the endoscopic group was 11% (9 of 80 cases), which is the same as surgery (10 of 77 cases, 13%). In the endoscopic group mortality rate was 0% compared with 8% of the surgical group (P ⬍0.05). With respect to 3 of 6 deaths, however, a causal relationship with surgical management may be disputed: 2 patients died of myocardial infarction, and 1 died of variceal bleeding (biliary cirrhosis was recognized at initial presentation). If these patients are arbitrarily excluded, the rate of mortality in the surgical group was 3.9%. Recurrent stenosis was evidenced in 5 of 80 (6%) patients of the endoscopic group and 4 of 77 (5%) patients of the surgical group. Restenosis after endoscopic treatment developed earlier compared with surgical approach (11.8 months versus 6.8 years; P ⬍0.00001). It is our impression that stenoses in the more proximal segments of the biliary system were more difficult to treat by endoscopic stenting, and, in our opinion, primary surgical treatment should be advised for patients with hilar stenoses. However, in the present study the number of patients with restenosis was too small to correlate bile duct segment as a risk factor with restenosis. Multivariate analysis of large cohorts of patients will be necessary to resolve this issue.
535
In conclusion, results of our study show that, surgery provides a better long-term outcome over the endoscopy, because patients with total obstruction are not amenable to the endoscopic approach. When successfully done, endoscopic results are similar to surgical results with less mortality. Although the endoscopic approach requires multiple sessions, most of our patients and referring physicians prefer endoscopic treatment to surgery. Prior endoscopic treatment does not preclude surgery, whereas endoscopic treatment is impossible once a Roux-en-Y loop has been constructed. Moreover, recurrent stenoses can also be successfully treated by repeated stent insertion. These considerations and the absence of treatment-related mortality lead us to propose endoscopic approach as a first-line treatment for postoperative biliary strictures. Indications for surgery are complete transection, failed previous repairs, and failures of endoscopic therapy. References [1] Davidoff AM, Branum GD, Meyers WC. Clinical features and mechanisms of major laparoscopic biliary injury. Semin Ultrasound CT MR 1993;14:338 – 45. [2] The Southern Surgeons Club. A prospective analysis of 1518 laparoscopic cholecystectomies. N Engl J Med 1991;324:1073– 8. [3] Deziel DJ, Millikan KW, Economou S, et al. Complications of laparoscopic cholecystectomy: a national survey of 4292 hospitals and analysis of 77,604 cases. Am J Surg 1993;165:9 –14. [4] Berkelhammer C, Kortan P, Haber GB. Endoscopic biliary prostheses as treatment for benign postoperative bile duct strictures. Gastrointest Endosc 1989;35:95–101. [5] Davids PH, Rauws EA, Coene PP, et al. Endoscopic stenting for postoperative biliary strictures. Gastrointest Endosc 1992;38:12–18. [6] Bergman JJGHM, Burgemeister L, Bruno MJ, et al. Long-term follow up after biliary stent placement for postoperative bile duct stenosis. Gastrointest Endosc 2001;54:154 – 61. [7] Costamagna G, Pandolfi M, Mutignani M, et al. Long-term results of endoscopic management of postoperative bile duct strictures with increasing numbers of stents. Gastrointest Endosc 2001;54:162– 8. [8] Lee JG, Leung JW. Long-term follow-up after biliary stent placement for postoperative bile duct stenosis. Gastrointest Endosc 2001;54: 272– 4. [9] Bottger T, Junginger T. Long term results after surgical treatment of iatrogenic injury of the bile ducts. Eur J Surg 1991;157:477– 80. [10] Ross CB, H’Doubler WZ, Sharp KW, Potts III. Recent experience with benign biliary strictures. Am Surg 1989;55:64 –70. [11] Frattaroli FM, Reggio D, Illomei G, Pappalardo G. Benign biliary strictures: a review of 21 years of experience. J Am Coll Surg 1996;183:506 –13. [12] Bismuth H. Postoperative strictures of the bile duct. In: Blumgart LH, editor. The biliary tract. Edinburgh: Churchill Livingstone, 1982, p 209 –18. [13] Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37:383–93. [14] Davids PH, Tanka AK, Rauws EA, et al. Benign biliary strictures. Surgery or endoscopy? Ann Surg 1993;217:237– 43.