- Email: [email protected]
Intrahepatic Biliary Enteric Bypass Provides Effective Palliation in Selected Patients with Malignant Obstruction at the Hepatic Duct Confluence
Intrahepatic Biliary Enteric Bypass Provides Effective Palliation in Selected Patients with Malignant Obstruction at the Hepatic Duct Confluence William R. Jarnagin, MD, Edmund Burke, MD, Christina Powers, BSN, Yuman Fong, MD, Leslie H. Blumgart, MD, New York, New York
BACKGROUND: Palliating the effects of biliary obstruction is a major goal of therapy in patients with cancer at the hepatic duct confluence. This study was undertaken to evaluate the effectiveness of intrahepatic biliary-enteric bypass to either the segment III duct or the right sectoral hepatic ducts in patients with unresectable hilar cholangiocarcinoma or gallbladder carcinoma. METHODS: From December 1991 to October 1996, 55 consecutive bypass procedures were prospectively evaluated in patients with unresectable hilar cholangiocarcinoma or gallbladder cancer. Patients were divided into three groups based on the primary tumor and the type of bypass performed: group 1A, cholangiocarcinoma/segment III bypass (n 5 20); group 1B, cholangiocarcinoma/right sectoral hepatic duct bypass (n 5 14); group 2, gallbladder cancer/ segment III bypass (n 5 21). RESULTS: Mean hospital stay (14 6 2 days) and mean blood loss (629 6 84 mL) were similar among the three groups. Perioperative death occurred in 6 patients (11%): 0 in group 1A, 3 each in groups 1B and 2. All survivors had relief of jaundice and pruritis after bypass. Complications occurred in 25 patients (45%). Preoperative transhepatic biliary drainage, performed in 14 patients prior to referral, was associated with a higher incidence of contaminated bile, greater operative blood loss, and postoperative biliary leak that was less likely to resolve spontaneously. Median survival in patients with cholangiocarcinoma (groups 1A and 1B) was 52 weeks and was unaffected by the type of bypass performed. By contrast, median survival in patients with gallbladder cancer (group 3) was 20 weeks; all but 3 died within 32 weeks of surgery. In patients with cholangiocarcinoma, the 1-year bypass patency was 80% in group 1A (segment III
From the Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York. Requests for reprints should be addressed to William R. Jarnagin, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, New York 10021. Manuscript submitted October 6, 1997 and accepted in revised form January 12, 1998.
© 1998 by Excerpta Medica, Inc. All rights reserved.
bypass) and 60% in group 1B (right sectoral hepatic duct bypass). Overall, there were 9 late bypass failures (18%) requiring reintervention. CONCLUSIONS: Intrahepatic biliary-enteric bypass effectively relieves symptoms due to malignant hilar obstruction. In patients with cholangiocarcinoma, segment III bypass provides excellent palliation with relatively few late complications and can be performed with minimal morbidity and mortality. Bypass to the right sectoral hepatic ducts, on the other hand, is associated with significant procedure-related morbidity and mortality and more late complications. Patients with gallbladder cancer, because of their poor survival, are probably better palliated by percutaneous biliary stenting. Am J Surg. 1998;175:453– 460. © 1998 by Excerpta Medica, Inc.
M
alignant obstruction of the hepatic ducts represents a formidable management problem. Most proximal biliary obstructions are the result of primary tumors of the biliary tract, namely cholangiocarcinoma or gallbladder carcinoma. Complete excision is the only potentially curative therapy for these tumors, but is often not possible for lesions involving the biliary confluence.1,2 Palliating the effects of biliary obstruction is therefore the primary objective in the majority. The palliative management options include intrahepatic biliary-enteric bypass or transhepatic percutaneous stents. Endoscopic stenting for hilar malignancies is associated with a high failure rate.3 Percutaneous stenting, although less invasive than biliary bypass, is associated with significant morbidity and mortality.4 In addition, stents are prone to occlusion by encrustation or tumor ingrowth. This problem appears to be greater for stents in the hepatic ducts compared with those in the distal bile duct.5–7 Intrahepatic biliary-enteric bypass has an advantage in this regard since the anastomosis can be placed some distance from the primary tumor, but requires a major operative procedure with associated recovery and morbidity. In this study, we report our experience with intrahepatic biliary-enteric bypass in patients with malignant obstruction of the hepatic ductal confluence resulting from primary tumors of the biliary tree. Perioperative morbidity and mortality, the long-term effectiveness of surgery in relieving jaundice and symptoms, and survival are analyzed. 0002-9610/98/$19.00 PII S0002-9610(98)00084-1
453
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
PATIENTS AND METHODS From December 1990 through October 1996, a total of 58 intrahepatic biliary-enteric bypass procedures were performed. Thirty-four patients had cholangiocarcinoma, 21 had gallbladder carcinoma, 1 had metastatic colorectal cancer, and 2 others had benign biliary strictures. The latter 3 patients were not included in this review, leaving a total of 55 cases for analysis. The patients were divided into three groups based on the primary malignancy and the type of bypass performed: group 1A, cholangiocarcinoma/ segment III bypass (20 patients); group 1B, cholangiocarcinoma/right sectoral hepatic duct bypass (14 patients); and group 2; gallbladder carcinoma/segment III bypass (21 patients). All preoperative, intraoperative and follow-up data were obtained from a prospective hepatobiliary database and review of the complete medical record of each patient. The tumor stage, according to the American Joint Committee on Cancer TNM classification, was based on the intraoperative findings. In all cases, hilar obstruction of the biliary tree was confirmed by preoperative radiographic studies. The standard evaluation included an abdominal computed tomography (CT) scan to assess tumor size and the presence or absence of liver atrophy, Duplex ultrasound to assess vascular patency, and direct imaging of the biliary tree with either endoscopic retrograde cholangiography, percutaneous transhepatic cholangiography, or magnetic resonance cholangiopancreatography. Forty-eight (87%) patients were explored with the intention to perform a curative resection, whereas in 7 cases (13%) a palliative bypass was planned preoperatively. Tissue confirmation of malignancy was obtained at laparotomy. In patients previously biopsied, the pathology slides were reexamined. The technical aspects of segment III and right sectoral hepatic duct bypass are well described.8 Briefly, the segment III duct is exposed by first dividing the bridge of liver tissue (if present) connecting segment III to the quadrate lobe. The base of the ligamentum teres is then freed by dividing the connections at its left base. The segment III duct can then be identified above and behind the segmental portal vein branch. Access to the duct is often enhanced by making an hepatotomy just to the left of the falciform ligament. The duct is then incised over a distance of 1 cm, just beyond the point of division into the segment II and segment III ducts. The portal venous supply to segment III, although adjacent to the duct, can usually be preserved.8 To expose the right anterior or posterior sectoral ducts, the main right portal pedicle is first controlled through hepatotomies made at the base of the gallbladder fossa and the caudate process. The overlying hepatic parenchyma is then divided to reveal the division of the right sectoral pedicles. The relevant duct (usually the anterior sectoral duct) is then identified and opened longitudinally.8 Sideto-side biliary-enteric anastomosis is then performed to a Roux-en-Y loop of jejunum using interrupted absorbable sutures, according to the technique of Blumgart and Kelley.9 Alternatively, the biliary-enteric anastomosis can be created further distally on the Roux limb, leaving a longer defunctionalized segment of jejunum. The end of 454
this jejunal segment can be secured to the abdominal wall, above the fascia, allowing easy postoperative access to the biliary tree.8 Twenty-four patients were reconstructed using this technique and 14 of these received postoperative, intraluminal radiotherapy with iridium-192 catheters passed through the jejunal loop.10 Closed drainage was used in all patients. Drains were removed when the daily output was approximately 30 cc or less of nonbilious fluid. Biliary anastomotic leaks were usually obvious clinically and were defined as persistent bilious drainage for more than 10 days. Serum bilirubin and alkaline phosphatase levels were recorded preoperatively and within the first 4 weeks following discharge. Follow-up data were obtained by review of the medical record or by telephone interview. Readmissions for biliary sepsis, recurrent biliary obstruction or surgical complications, and any related procedures were recorded. Readmissions for palliative care of preterminal patients were excluded. Late bypass failure was defined as recurrent biliary obstruction that required an intervention (biliary stent or reoperation) to correct. Patients with jaundice at the time of death were assumed to have a nonfunctioning bypass unless patency was demonstrated radiographically. Statistical analyses were performed using SPSS for Windows, version 7.0 (SPSS, Chicago, Illinois). Continuous variables were compared using Student’s t test and categorical variables were compared by chi square. Cumulative survival and bypass patency were calculated using the method of Kaplan-Meier and compared using the log rank test. Numeric data are expressed as the mean 6 standard error of the mean (SEM) except where indicated. P values less than or equal to 0.05 were considered significant.
RESULTS Demographics Demographic data for patients in each group are shown in Table I. Men constituted a majority of the patients with cholangiocarcinoma, while the converse was true in patients with gallbladder carcinoma. The mean age was similar in all groups (mean 52 6 2 years, range 35 to 83). In 7 patients the tumor stage could not be adequately assessed because of severe inflammation from prior procedures. Of the remainder, 58% had stage IV disease. There was no significant difference in disease stage between patients with gallbladder cancer (71% stage IV) and cholangiocarcinoma (50% stage IV). Morbidity and Mortality The overall procedure-related morbidity was 45% and the 30-day mortality was 11% (Table II). Of the 6 perioperative deaths, none occurred in group 1A and 3 each occurred in groups 1B and 2 (P 5 0.03 and P 5 0.08, respectively). The causes of death were as follows: sepsis (2), upper gastrointestinal hemorrhage (1), myocardial infarction (1), renal failure (1), hepatic failure (1). Overall, the procedure-related mortality for segment III bypass (groups 1A and 2) was lower than for right sectoral hepatic duct bypass (group 1B; 7% versus 21%, P 5 0.14). Perioperative complications were higher in group 1A (55%) and group 1B (57%) compared with group 2 (29%), but these differences were not significantly different. Sim-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
TABLE I Patient Demographics
1A. Segment III Bypass
Group 1 Hilar Cholangiocarcinoma 1B. R. Sectoral Duct Bypass
Total Cholangiocarcinoma
Segment III Bypass
Total
20 11 9 62 6 3 0 0 5 9 6
14 10 4 62 6 2 0 0 5 8 1
34 21 13 62 6 2 0 0 10 17 7
21 9 12 61 6 3 0 0 6 15 0
55 30 25 62 6 2 0 0 16 32 7
Number Men Women Age (years) Stage I II III IV Indeterminate
Group 2 Gallbladder Carcinoma
Stage, according to the American Joint Committee on Cancer TNM classification, was determined at laparotomy.
TABLE II Perioperative Morbidity and Mortality Group 1 Group 2 Hilar Cholangiocarcinoma Gallbladder Carcinoma 1A. 1B. Total Segment III Bypass R. Sectoral Duct Bypass Cholangiocarcinoma Segment III Bypass Estimated blood loss (mL) Positive intraoperative bile cultures Length of stay (days) Postoperative complications (total) Bile leak Wound infection Biloma/abscess Renal failure Pancreatitis Entero-cutaneous fistula 30-Day mortality
Total
772 6 159
698 6 143
738 6 106
477 6 133
629 6 84
7 (35%) 13 6 1
6 (33%) 18 6 6
13 (38%) 15 6 2
4 (19%) 12 6 2
17 (31%) 14 6 2
11 (55%) 6 3 0 2 0
8 (57%) 4 1 2 0 1
19 (56%) 10 4 2 2 1
6 (29%) 3 1 1 0 0
25 (45%) 13 5 3 2 1
0 0*
0 3 (21.4%)*
0 3 (9%)
1 3 (14.3%)*
1 6 (11%)
Estimated blood loss and length of stay are expressed as the mean 6 standard deviation. Complications are expressed as the total number of complications divided by the number of patients in each group; 7 patients experienced more than one complication. The complication rate was higher in groups 1A and 1B compared with group 2 but the differences were not significant. * P 5 0.03, group 1A versus group 1B; P 5 0.08, group 1A versus 2.
ilarly, biliary reconstruction using an extended jejunal loop did not result in greater complications. Postoperative biliary-enteric anastomotic leak occurred in 13 patients (23%), and accounted for 52% of the total complications. Four of these patients required percutaneous stenting to control postoperative biliary leaks, all of whom had been subjected to transhepatic drainage before surgery (P 5 0.0001). The remainder resolved spontaneously. One patient required reoperation for a complication (upper gastrointestinal hemorrhage). The mean estimated blood loss for all procedures was 629 6 84 mL and did not differ significantly among the three groups (Table II). The mean length of stay from the time of surgery was 14 6 2 days, and although higher in patients who underwent a right sectoral hepatic duct bypass (group 2, 18 6 6 days), this difference was not signif-
icantly different. A postoperative biliary leak significantly prolonged the hospital stay beyond 10 days (P 5 0.04). Consequences of Preoperative Biliary Drainage Thirty patients (55%) were subjected to preoperative biliary tract procedures prior to referral (Table III). Percutaneous transhepatic cholangiogram (PTC) with external biliary drainage was the most common procedure (n 5 14) but a significant number underwent either endoscopic retrograde cholangiopancreatography (ERCP) with attempted stent placement (n 5 9) or surgery (cholecystectomy or placement of a T-tube or both, n 5 11). Sixteen patients (53%) experienced one or more complications related to these procedures. Ten patients either presented with or were previously treated for cholangitis; 4 others developed an infected bile collection. Preoperative trans-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
455
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
TABLE III Preoperative Biliary Tract Procedures Performed Prior to Referral and Resulting Complications
1A. Segment III Bypass Preoperative biliary interventions (patients) PTC ERCP Surgery Complications from preoperative interventions (patients) Cholangitis Biloma/abscess Tumor in tube tract
Group 1 Hilar Cholangiocarcinoma 1B. R. Sectoral Duct Total Bypass Cholangiocarcinoma
Group 2 Gallbladder Carcinoma
Segment III Bypass
Total
14 (70%) 10 2 4
9 (64%) 2 3 4
23 (68%) 12 5 8
7 (33%) 2 4 3
30 (55%) 14 9 11
8 (57%) 6 1 1
6 (67%) 4 1 1
14 (61%) 10 2 2
2 (29%) 0 2 0
16 (53%) 10 4 2
Four patients underwent multiple procedures. PTC 5 percutaneous transhepatic cholangiogram with either external biliary drainage or internal/external stent; ERCP 5 endoscopic retrograde cholangiopancreatography with stent placement; Surgery 5 8 patients underwent laparotomy and cholecystectomy, 3 patients had placement of T-tubes.
TABLE IV Effect of Preoperative Percutaneous Transhepatic Biliary Drainage on Intraoperative and Postoperative Results
Positive intraoperative bile cultures Operative blood loss .500 mL Postoperative intervention for bile leak Postoperative complications (total)
Preoperative Percutaneous Transhepatic Biliary Drainage (n 5 14)
No Preoperative Biliary Drainage (n 5 25)
P
64%
8%
0.001
79%
14%
0.007
29%
0%
0.0001
50%
39%
NS
NS 5 not significant.
hepatic drainage caused tumor seeding along the tube tract in 2 patients. Positive intraoperative bile cultures were more frequent after preoperative percutaneous transhepatic drainage (PTC) than in patients not previously drained (64% versus 8%, P 5 0.001). Operative blood loss was also greater (EBL .500 mL, 79% preoperative PTC versus 14% no preoperative drainage, P 5 0.007). Furthermore, in preoperatively drained patients, postoperative biliary leak was less likely to resolve spontaneously and more often required percutaneous stenting for control (29% versus 0%, P 5 0.0001). Postoperative morbidity was higher in preoperatively drained patients but this difference was not significant (50% versus 39%, P 5 0.4; Table IV). Relief of Symptoms All surviving patients experienced relief of jaundice and related symptoms following bypass. Pruritis resolved in 10 of 10 patients reporting this symptom preoperatively. The overall decrease in the serum bilirubin level was 54 6 8% at the time of initial follow-up (within 4 weeks from discharge). The preoperative serum bilirubin level in pa456
tients with gallbladder cancer (group 3) was higher than that in the other groups, but the rate of decline was similar (Figure 1). The decline in serum alkaline phosphatase levels was much less dramatic than that for bilirubin, and in many cases remained static or increased. Neither the preoperative nor the postoperative serum bilirubin and alkaline phosphatase, nor the percentage decrease in these parameters, was a useful predictor of adverse outcome. Survival Overall survival was significantly greater in patients with cholangiocarcinoma (groups 1A and 1B) compared with those with gallbladder cancer (P 5 0.0001). Median survival in patients with cholangiocarcinoma was 52 weeks (mean 5 62 6 7), and 6 patients are alive at 6, 10, 25, 30, 68, and 104 (2 patients) weeks. There was no significant difference in survival between groups 1A and 1B (74 6 14 weeks versus 57 6 9 weeks, P 5 0.2). On the other hand, median survival in patients with gallbladder cancer was 20 weeks (mean 21 6 4). Two patients are alive at 4 weeks from the time of surgery. One patient survived beyond 1 year (57 weeks), while all of the remaining patients died
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
Figure 2. Cumulative survival stratified by primary tumor. Median survival was significantly greater for patients with cholangiocarcinoma than for patients with gallbladder carcinoma (52 weeks versus 20 weeks, P 5 0.0001). There was no difference in survival between groups 1A (hilar cholangiocarcinoma/segment III bypass) and 1B (hilar cholangiocarcinoma/right sectoral hepatic duct bypass).
Figure 1. Mean serum bilirubin (A) and alkaline phosphatase (B) levels for each group before and after bypass. Panel C shows the percentage decrease in serum bilirubin and alkaline phosphatase for each group after surgery.
within 32 weeks of bypass. The Kaplan-Meier actuarial survival curves, stratified by primary tumor, are shown in Figure 2. Age greater than 60 years was the only factor that significantly reduced mean survival in patients with gallbladder cancer (16 6 3 weeks versus 26 6 7 weeks, P 5 0.05). In patients with cholangiocarcinoma, stage IV disease (41 6 6 weeks versus 76 6 11 weeks, P 5 0.008) and postoperative biliary-enteric anastomotic leak (31 6 4 weeks versus 69 6 8 weeks, P 5 0.02) significantly reduced survival, while age greater than 60 years was a much less significant factor (53 6 8 weeks versus 78 6 13 weeks, P 5 0.1). Of the 14 patients who received postoperative intraluminal radiotherapy, 13 had cholangiocarcinoma. While mean survival was greater in this group compared with patients with cholangiocarcinoma not given radiotherapy, the difference was not significant (73 6 11 weeks versus 54 6 9 weeks, P 5 0.2). Late Complications Twenty-three patients (47%) required readmission to the hospital for biliary tract sepsis, biliary tract interventions or surgical complications; readmissions for comfort care of preterminal patients were excluded (Table V). Nine patients were readmitted more than one time. When compared with segment III bypass (group 1A), right sectoral hepatic duct bypass (group 1B) was associated with more readmissions per patient (1.6 6 0.4 versus 0.8 6 0.2, P 5 0.1) and more readmission hospital days per patient (14.6 6 4 versus 5.1 6 2, P 5 0.04). Readmissions for
biliary tract sepsis were also more common in these patients compared with group 1A (P 5 0.05). Fifteen patients (31%) required postoperative biliary tract intervention at some point in their course—14 transhepatic stents and 1 reoperation and revision of the anastomosis. Nine procedures (including 1 reoperation) were performed for late bypass failure secondary to either stricture or progression of disease, 4 stents were placed to help control anastomotic leak, and 2 stents were placed for cholangitis in an undrained hepatic lobe. The average time to occlusion in all patients whose bypass did fail was 23 6 6 weeks. Patients who underwent a right sectoral hepatic duct bypass (group 1B) constituted the majority of those requiring reintervention for late bypass failure (6 patients, 55% versus 3 patients, 15% in group 1A, P 5 0.06; Table V). Bypass Patency The cumulative patency of all bypasses performed for hilar cholangiocarcinoma, stratified by group, is shown in Figure 3. The actuarial patency was greater for segment III bypass (Group 1A) than for right sectoral hepatic duct bypass (Group 1B) but the difference was not significant (P 5 0.1). Eighty percent of segment III bypasses performed for cholangiocarcinoma were patent at 1 year compared with 60% for right sectoral hepatic duct bypass. Patients with gallbladder cancer were not included in analyses of long-term bypass performance because nearly all died of their disease by 32 weeks. Mean follow-up for all patients was 42 6 5 weeks and was similar to the overall survival of 49 6 6 weeks.
COMMENTS Malignant obstruction of the hepatic ductal confluence remains a difficult clinical problem. Cholangiocarcinoma
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
457
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
TABLE V Summary of Readmissions and Late Bypass Failures after Surgery Group 1 Group 2 Hilar Cholangiocarcinoma Gallbladder Carcinoma 1A. Segment III 1B. Total Bypass R. Sectoral Duct Bypass Cholangiocarcinoma Segment III Bypass All readmissions Number of patients Readmissions per patient (mean) Readmission hospital days per patient (mean) Readmissions for biliary tract sepsis Number of patients Readmissions per patient (mean) Biliary interventions for late bypass failure Number of patients
Total
10 (50%)
8 (73%)
18 (58%)
5 (28%)
23 (47%)
0.8 6 0.2
1.6 6 0.4
1 6 0.2
0.3 6 0.1
0.8 6 0.2
5.1 6 2*
14.6 6 4*
8.1 6 2
3 6 1.6
6.3 6 1.4
7 (35%)
6 (55%)
13 (42%)
4 (22%)
17 (35%)
0.4 6 0.1†
1.0 6 0.3†
0.6 6 0.2
0.2 6 0.1
0.5 6 0.1
3 (15%)‡
6 (55%)‡
9 (29%)
0
9 (18%)
Readmissions for biliary tract sepsis, biliary tract interventions, and surgical complications were included. Readmissions for comfort care of terminally ill patients were excluded. Two patients readmitted with biliary tract sepsis required percutaneous drainage, while the remainder were successfully treated with antibiotics. Patients with gallbladder cancer, because of their significantly shorter survival, are not included in the statistical analysis of readmissions and late bypass failures. * P 5 0.04 group 1A versus group 1B. † P 5 0.05 group 1A versus group 1B. ‡ P 5 0.06 group 1A versus group 1B.
Figure 3. Cumulative patency of biliary-enteric bypass in all patients with cholangiocarcinoma. Both median and 1-year patency were greater for segment III bypass (group 1A) than for right sectoral hepatic duct bypass (group 1B) but the differences did not reach statistical significance.
and gallbladder cancer, the two most common causes, are amenable to curative resection in about one third of cases compared with 80% to 90% for periampullary tumors.11 Therefore, in patients with malignant obstruction of the hepatic ductal confluence, the goal of therapy is often palliation. Determining irresectability can be difficult, however, and often requires bile duct transection to establish with certainty. This is especially true in patients who have had previous biliary tract surgery or biliary intubation. In these cases, intrahepatic bypass is necessary to restore biliary continuity. 458
Biliary bypass and percutaneous stenting remain the principal options for palliating malignant hilar obstruction. In considering either of these approaches, the patient’s general state of health and life expectancy must be the foremost concerns. However, the procedure’s efficacy in relieving symptoms, the procedure-related morbidity and mortality, and its durability must also be considered. Additionally, and perhaps most importantly, patient inconvenience must be minimized. Any palliative intervention that introduces a new complication, such as frequent return trips to the hospital or an external drain that requires close attention, represents a great burden to the patient. For palliative biliary drainage, cholangitis and recurrent obstruction are the principal complications that require further intervention and must be minimized to preserve quality of life. Percutaneous biliary endoprostheses are the most effective alternative to surgical bypass for tumors of the hilum; endoscopic stents are generally ineffective for lesions in this location. The relatively recent introduction of selfexpandable metallic stents, with a greater patency than their plastic counterparts, has made the percutaneous approach somewhat more attractive.12 Whether percutaneous stents should supplant surgical bypass for unresectable hilar cancers, as suggested by some authors, remains an open question. Indeed, there have been four prospective studies comparing biliary stents with surgical bypass for distal bile duct obstruction.13 These studies demonstrate that surgery is associated with greater early morbidity and mortality but greater long-term patency and a lower incidence of recurrent jaundice. In a separate study, van den Bosch et al14 analyzed outcomes in 148 patients with distal bile duct obstruction from pancreatic carcinoma, concluding that surgical bypass provides optimal palliation for
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
patients surviving more than 6 months while biliary stenting is preferred for those surviving less than 6 months. Prospective comparisons between surgery and stents do not exist for proximal bile duct obstruction, but several reports suggest a similar conclusion. Firstly, the reported morbidity and mortality of stenting hilar cancers ranges from 7% to 31% and 12% to 14%, respectively, considerably higher than distal bile duct stents and similar to many surgical series.4 –7 Moreover, several studies have shown that hilar stents have a significant incidence of late complications. Becker et al6 reported a 46% 12-month patency for metallic stents placed for hilar malignancies versus 89% for those placed for distal bile duct tumors. The overall incidence of hilar stent occlusion ranges from 24% to 38%.4 –7 The long-term patency of intrahepatic biliaryenteric bypass, on the other hand, is not well reported in the literature. In this study, long-term bypass patency in patients with cholangiocarcinoma is clearly documented, and the results suggests that segment III bypass is superior to biliary stents. Several surgical techniques have been described for intrahepatic biliary bypass. The two most common approaches are bypass to the segment III duct and the right sectoral hepatic ducts.8 The former, because of the more constant anatomy and long extrahepatic course of the left hepatic duct, is technically easier and preferred. However, the type of bypass is usually dictated by the location of the tumor. In general, segment III bypass is performed unless the left lobe is atrophic or heavily involved with tumor or if the primary lesion extends to the umbilical fissure of the liver. In patients with unresectable gallbladder cancer, segment III bypass is usually the only option since the right hepatic ducts are frequently involved with tumor. Normalization of bilirubin levels and relief of jaundice can be achieved if at least one third of the functioning liver parenchyma is adequately drained. Lack of communication between the right and left hepatic ducts, therefore, does not affect the results, provided that the undrained liver has not been percutaneously drained or otherwise contaminated. In this circumstance, there is a high risk of persistent biliary fistula and cholangitis.15 This study shows that intrahepatic biliary-enteric bypass effectively relieves jaundice and symptoms from malignant hilar obstruction. All patients were effectively palliated following surgery. Most patients experienced a prompt and steady decline in serum bilirubin, which approached normal within 4 weeks of bypass. The change in serum alkaline phosphatase was less dramatic and often remained static or increased. This observation, not unexpected, is most likely the result of persistent ductal obstruction in the undrained liver. Neither the preoperative bilirubin and alkaline phosphatase nor the postoperative change had any prognostic value. Several points emerge from analysis of the three subgroups. First, the median survival for patients with gallbladder cancer was 5 months. The 28% morbidity and 16.6% mortality associated with segment III bypass in these patients might be considered excessive in light of their poor survival. These patients are, therefore, unlikely to reap the long-term benefits of surgical bypass and are probably best palliated by percutaneous biliary endoprostheses.
Survival in patients with cholangiocarcinoma, on the other hand, was 12 months. The perioperative morbidity in this group can, therefore, be justified if good long-term results are achieved. Late complications, particularly cholangitis and recurrent biliary obstruction, are the bane of all palliative biliary procedures. In this study, we used readmissions for biliary tract-related problems and bypass patency over the life of the patient to assess long-term functional results. Overall, 47% of patients required readmission for treatment of biliary sepsis or recurrent biliary obstruction. Patients with gallbladder cancer had deceptively few late complications when compared with the other groups, which is almost certainly attributable to their poor survival. In patients with cholangiocarcinoma, segment III bypass provided superior long-term palliation than right sectoral hepatic duct bypass. Not only were readmissions for biliary tract-related complications fewer and shorter for these patients, but bypass patency was better. Late bypass failure and the need for further biliary tract interventions were significantly less common after segment III bypass than after right sectoral hepatic duct bypass. This observation may reflect the relatively greater technical difficulty of approaching the right sectoral hepatic ducts. The procedure-related mortality of 11% is within the range of other reports.2,16,17 On the other hand, the mortality associated with segment III bypass, 0% in patients with cholangiocarcinoma and 7% overall, is lower than that reported in many other surgical series. In fact, the reported mortality for biliary stents is often higher. Biliary anastomotic leak was the most common postoperative complication, accounting for 50% of all complications and significantly prolonged the hospital stay. The incidence of biliary leak was 23%. Although two thirds of all leaks resolved without any intervention, the presence of a biliary leak had a significant negative impact on survival in patients with cholangiocarcinoma. Anastomotic stricture is a potential sequela of biliary anastomotic leak. This mechanism, while possibly playing a role, is not the sole reason for the survival difference since patients who developed a biliary leak did not have a significantly increased incidence of biliary sepsis or late bypass failure. More than half of the patients in this study were referred after having had one or more biliary tract interventions. Percutaneous transhepatic drainage was the most common preoperative procedure and significantly increased the perioperative morbidity. Not surprisingly, contaminated bile at the time of surgery was a very common finding in these patients, although postoperative infectious complications were not significantly increased. Preoperative percutaneous drainage altered the natural history of biliary-enteric anastomotic leak by preventing spontaneous closure in some patients. Furthermore, the intense inflammatory response caused by biliary stents made dissection within the portal area very difficult, as evidenced by the significantly greater operative blood loss in stented patients. Many of these drainage catheters were placed for unclear indications, and two thirds failed to achieve their objective of relieving jaundice. In 2 cases, percutaneous drainage resulted in spread of the tumor along the tube tract. It is well-recognized that routine preoperative biliary
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
459
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
drainage does not improve outcome after surgery for obstructive jaundice; in fact, there is strong evidence that morbidity and mortality are increased.18 –20 MacPherson et al21 noted an increased operative mortality in previously drained patients undergoing resection of proximal bile duct cancers. Heslin et al22 reported that, in patients with distal bile duct obstruction, preoperative biliary stenting increases the incidence of postoperative infectious complications. In addition, the pathogens responsible for these infections are usually the same organisms colonizing the biliary tree.23 Moreover, the resectability of many hilar cancers can only be determined at laparotomy, a task made much more difficult in the presence of a biliary stent. Routine preoperative biliary drainage, in the absence of cholangitis or severe pruritis, should therefore be avoided. Clearly, both intrahepatic biliary-enteric bypass and percutaneous stenting have a role in palliating patients with malignant hilar obstruction. The best results will be obtained by selective use of each technique. This study suggest that the decision must take into account the natural history of the underlying disease. Those patients with advanced disease from an aggressive cancer, such as gallbladder carcinoma, have a limited survival and are probably best palliated with biliary stents. Cholangiocarcinoma, on the other hand, is a more indolent disease. In these patients, segment III bypass can be performed with an acceptable morbidity and a mortality lower than that reported for biliary stents. Moreover, this procedure provides excellent palliation with good long-term function and relatively few late complications. Bypass to the right sectoral hepatic ducts is associated with more early and late complications, and is therefore a less attractive option. Whether surgical bypass is superior to percutaneous biliary stenting can only be answered by a randomized, prospective study.
REFERENCES 1. Blumgart LH, Benjamin IS, Hadjis NS, Beazly R. Surgical approaches to cholangiocarcinoma at confluence of bile ducts. Lancet. 1984;i:66 –70. 2. Bismuth H, Castaing D, Traynor O. Resection or palliation: priority of surgery in the treatment of hilar cancer. World J Surg. 1988;12:39 – 47. 3. Cheung K-L, Lai ECS. Endoscopic stenting for malignant biliary obstruction. Arch Surg. 1995;130:204 –207. 4. Glattli A, Stain SC, Baer HU, et al. Unresectable malignant biliary obstruction: treatment by self-expandable biliary endoprostheses. HPB Surg. 1993;6:175–184. 5. Stoker J, Lameris JS. Complications of percutaneously inserted biliary Wallstents. JVIR. 1993;4:767–772. 6. Becker CD, Glattli A, Maibach R, Baer HU. Percutaneous palliation of malignant obstructive jaundice with Wallstent endoprosthesis: follow-up and reintervention in patients with hilar and non-hilar obstruction. JVIR. 1993;4:597– 604.
460
7. Rossi P, Bexxi M, Rossi M, et al. Metallic stents in malignant biliary obstruction: results of a multicenter European study of 240 patients. JVIR. 1994;5:279 –285. 8. Blumgart LH, Baer HU. Hilar and intrahepatic biliary-enteric anastomosis. In: Blumgart LH, ed. Surgery of the Liver and Biliary Tract. 2nd ed. Edinburgh: Churchill Livingstone; 1051–1067. 9. Blumgart LH, Kelley CJ. Hepaticojejunostomy in benign and malignant high bile duct stricture: approaches to the left hepatic ducts. Br J Surg. 1984;71:257–261. 10. Kuvshinoff BW, Armstrong JG, Fong Y, et al. Palliation of irresectable hilar cholangiocarcinoma with biliary drainage and radiotherapy. Br J Surg. 1995;82:1522–1525. 11. Nakeeb A, Pitt HA, Sohn TA, et al. Cholangiocarcinoma: a spectrum of intrahepatic, perihilar and distal tumors. Ann Surg. 1996;224:463– 475. 12. Schmassmann A, van Gunten E, Knuchel J, et al. Wallstents versus plastic stents in malignant biliary obstruction: effects of stent patency of the first and second stent on patient compliance and survival. Am J Gastroenterol. 1996;91:654 – 659. 13. Lillemoe KD, Pitt HA. Palliation: surgical and otherwise. Cancer. 1996;78:605– 614. 14. van den Bosch R, van der Schelling GP, Klinkenbijl JHG, et al. Guidelines for the application of surgery and endoprostheses in the palliation of obstructive jaundice in advanced cancer of the pancreas. Ann Surg. 1994;219:18 –24. 15. Baer HU, Rhymer M, Stain SC, et al. The effect of communication between the right and left liver on the outcome of surgical drainage for jaundice due to malignant obstruction at the hilus of the liver. HPB Surg. 1994;8:27–31. 16. Traynor O, Castaing D, Bismuth H. Left intrahepatic cholangio-enteric anastomosis (round ligament approach): an effective palliative treatment for hilar cancers. Br J Surg. 1987;74:952–954. 17. Myburgh JA. Resection and bypass for malignant obstruction of the bile duct. World J Surg. 1995;19:108 –112. 18. Sirinek KR, Levine BA. Percutaneous transhepatic cholangiography and biliary decompression. Arch Surg. 1989;124:885– 888. 19. Pitt HA, Gomes AS, Lois JF, et al. Does preoperative percutaneous biliary drainage reduce operative risk or increase hospital cost? Ann Surg. 1985;201:545–553. 20. Karsten TM, Davids PHP, van Gulik TM, et al. Effects of biliary endoprostheses on the extrahepatic bile ducts in relation to subsequent operation of the biliary tract. J Am Coll Surg. 1994;178: 343–352. 21. MacPherson GAD, Benjamin IS, Hodgson HJF, et al. Preoperative percutaneous transhepatic biliary drainage; the results of a controlled trial. Br J Surg. 1984;74:371–375. 22. Heslin MJ, Brooks AD, Hochwald SN, et al. Preoperative biliary stenting is associated with increased complications after pancreaticoduodenectomy. Arch Surg. In press. 23. Karsten TM, Allema JH, Reinders M, van Gulik TM, et al. Preoperative biliary drainage, colonisation of bile and postoperative complications in patients with tumours of the pancreatic head: a retrospective analysis of 241 consecutive patients. Eur J Surg. 1996;162:881– 888.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
BACKGROUND: Palliating the effects of biliary obstruction is a major goal of therapy in patients with cancer at the hepatic duct confluence. This study was undertaken to evaluate the effectiveness of intrahepatic biliary-enteric bypass to either the segment III duct or the right sectoral hepatic ducts in patients with unresectable hilar cholangiocarcinoma or gallbladder carcinoma. METHODS: From December 1991 to October 1996, 55 consecutive bypass procedures were prospectively evaluated in patients with unresectable hilar cholangiocarcinoma or gallbladder cancer. Patients were divided into three groups based on the primary tumor and the type of bypass performed: group 1A, cholangiocarcinoma/segment III bypass (n 5 20); group 1B, cholangiocarcinoma/right sectoral hepatic duct bypass (n 5 14); group 2, gallbladder cancer/ segment III bypass (n 5 21). RESULTS: Mean hospital stay (14 6 2 days) and mean blood loss (629 6 84 mL) were similar among the three groups. Perioperative death occurred in 6 patients (11%): 0 in group 1A, 3 each in groups 1B and 2. All survivors had relief of jaundice and pruritis after bypass. Complications occurred in 25 patients (45%). Preoperative transhepatic biliary drainage, performed in 14 patients prior to referral, was associated with a higher incidence of contaminated bile, greater operative blood loss, and postoperative biliary leak that was less likely to resolve spontaneously. Median survival in patients with cholangiocarcinoma (groups 1A and 1B) was 52 weeks and was unaffected by the type of bypass performed. By contrast, median survival in patients with gallbladder cancer (group 3) was 20 weeks; all but 3 died within 32 weeks of surgery. In patients with cholangiocarcinoma, the 1-year bypass patency was 80% in group 1A (segment III
From the Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York. Requests for reprints should be addressed to William R. Jarnagin, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, New York 10021. Manuscript submitted October 6, 1997 and accepted in revised form January 12, 1998.
© 1998 by Excerpta Medica, Inc. All rights reserved.
bypass) and 60% in group 1B (right sectoral hepatic duct bypass). Overall, there were 9 late bypass failures (18%) requiring reintervention. CONCLUSIONS: Intrahepatic biliary-enteric bypass effectively relieves symptoms due to malignant hilar obstruction. In patients with cholangiocarcinoma, segment III bypass provides excellent palliation with relatively few late complications and can be performed with minimal morbidity and mortality. Bypass to the right sectoral hepatic ducts, on the other hand, is associated with significant procedure-related morbidity and mortality and more late complications. Patients with gallbladder cancer, because of their poor survival, are probably better palliated by percutaneous biliary stenting. Am J Surg. 1998;175:453– 460. © 1998 by Excerpta Medica, Inc.
M
alignant obstruction of the hepatic ducts represents a formidable management problem. Most proximal biliary obstructions are the result of primary tumors of the biliary tract, namely cholangiocarcinoma or gallbladder carcinoma. Complete excision is the only potentially curative therapy for these tumors, but is often not possible for lesions involving the biliary confluence.1,2 Palliating the effects of biliary obstruction is therefore the primary objective in the majority. The palliative management options include intrahepatic biliary-enteric bypass or transhepatic percutaneous stents. Endoscopic stenting for hilar malignancies is associated with a high failure rate.3 Percutaneous stenting, although less invasive than biliary bypass, is associated with significant morbidity and mortality.4 In addition, stents are prone to occlusion by encrustation or tumor ingrowth. This problem appears to be greater for stents in the hepatic ducts compared with those in the distal bile duct.5–7 Intrahepatic biliary-enteric bypass has an advantage in this regard since the anastomosis can be placed some distance from the primary tumor, but requires a major operative procedure with associated recovery and morbidity. In this study, we report our experience with intrahepatic biliary-enteric bypass in patients with malignant obstruction of the hepatic ductal confluence resulting from primary tumors of the biliary tree. Perioperative morbidity and mortality, the long-term effectiveness of surgery in relieving jaundice and symptoms, and survival are analyzed. 0002-9610/98/$19.00 PII S0002-9610(98)00084-1
453
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
PATIENTS AND METHODS From December 1990 through October 1996, a total of 58 intrahepatic biliary-enteric bypass procedures were performed. Thirty-four patients had cholangiocarcinoma, 21 had gallbladder carcinoma, 1 had metastatic colorectal cancer, and 2 others had benign biliary strictures. The latter 3 patients were not included in this review, leaving a total of 55 cases for analysis. The patients were divided into three groups based on the primary malignancy and the type of bypass performed: group 1A, cholangiocarcinoma/ segment III bypass (20 patients); group 1B, cholangiocarcinoma/right sectoral hepatic duct bypass (14 patients); and group 2; gallbladder carcinoma/segment III bypass (21 patients). All preoperative, intraoperative and follow-up data were obtained from a prospective hepatobiliary database and review of the complete medical record of each patient. The tumor stage, according to the American Joint Committee on Cancer TNM classification, was based on the intraoperative findings. In all cases, hilar obstruction of the biliary tree was confirmed by preoperative radiographic studies. The standard evaluation included an abdominal computed tomography (CT) scan to assess tumor size and the presence or absence of liver atrophy, Duplex ultrasound to assess vascular patency, and direct imaging of the biliary tree with either endoscopic retrograde cholangiography, percutaneous transhepatic cholangiography, or magnetic resonance cholangiopancreatography. Forty-eight (87%) patients were explored with the intention to perform a curative resection, whereas in 7 cases (13%) a palliative bypass was planned preoperatively. Tissue confirmation of malignancy was obtained at laparotomy. In patients previously biopsied, the pathology slides were reexamined. The technical aspects of segment III and right sectoral hepatic duct bypass are well described.8 Briefly, the segment III duct is exposed by first dividing the bridge of liver tissue (if present) connecting segment III to the quadrate lobe. The base of the ligamentum teres is then freed by dividing the connections at its left base. The segment III duct can then be identified above and behind the segmental portal vein branch. Access to the duct is often enhanced by making an hepatotomy just to the left of the falciform ligament. The duct is then incised over a distance of 1 cm, just beyond the point of division into the segment II and segment III ducts. The portal venous supply to segment III, although adjacent to the duct, can usually be preserved.8 To expose the right anterior or posterior sectoral ducts, the main right portal pedicle is first controlled through hepatotomies made at the base of the gallbladder fossa and the caudate process. The overlying hepatic parenchyma is then divided to reveal the division of the right sectoral pedicles. The relevant duct (usually the anterior sectoral duct) is then identified and opened longitudinally.8 Sideto-side biliary-enteric anastomosis is then performed to a Roux-en-Y loop of jejunum using interrupted absorbable sutures, according to the technique of Blumgart and Kelley.9 Alternatively, the biliary-enteric anastomosis can be created further distally on the Roux limb, leaving a longer defunctionalized segment of jejunum. The end of 454
this jejunal segment can be secured to the abdominal wall, above the fascia, allowing easy postoperative access to the biliary tree.8 Twenty-four patients were reconstructed using this technique and 14 of these received postoperative, intraluminal radiotherapy with iridium-192 catheters passed through the jejunal loop.10 Closed drainage was used in all patients. Drains were removed when the daily output was approximately 30 cc or less of nonbilious fluid. Biliary anastomotic leaks were usually obvious clinically and were defined as persistent bilious drainage for more than 10 days. Serum bilirubin and alkaline phosphatase levels were recorded preoperatively and within the first 4 weeks following discharge. Follow-up data were obtained by review of the medical record or by telephone interview. Readmissions for biliary sepsis, recurrent biliary obstruction or surgical complications, and any related procedures were recorded. Readmissions for palliative care of preterminal patients were excluded. Late bypass failure was defined as recurrent biliary obstruction that required an intervention (biliary stent or reoperation) to correct. Patients with jaundice at the time of death were assumed to have a nonfunctioning bypass unless patency was demonstrated radiographically. Statistical analyses were performed using SPSS for Windows, version 7.0 (SPSS, Chicago, Illinois). Continuous variables were compared using Student’s t test and categorical variables were compared by chi square. Cumulative survival and bypass patency were calculated using the method of Kaplan-Meier and compared using the log rank test. Numeric data are expressed as the mean 6 standard error of the mean (SEM) except where indicated. P values less than or equal to 0.05 were considered significant.
RESULTS Demographics Demographic data for patients in each group are shown in Table I. Men constituted a majority of the patients with cholangiocarcinoma, while the converse was true in patients with gallbladder carcinoma. The mean age was similar in all groups (mean 52 6 2 years, range 35 to 83). In 7 patients the tumor stage could not be adequately assessed because of severe inflammation from prior procedures. Of the remainder, 58% had stage IV disease. There was no significant difference in disease stage between patients with gallbladder cancer (71% stage IV) and cholangiocarcinoma (50% stage IV). Morbidity and Mortality The overall procedure-related morbidity was 45% and the 30-day mortality was 11% (Table II). Of the 6 perioperative deaths, none occurred in group 1A and 3 each occurred in groups 1B and 2 (P 5 0.03 and P 5 0.08, respectively). The causes of death were as follows: sepsis (2), upper gastrointestinal hemorrhage (1), myocardial infarction (1), renal failure (1), hepatic failure (1). Overall, the procedure-related mortality for segment III bypass (groups 1A and 2) was lower than for right sectoral hepatic duct bypass (group 1B; 7% versus 21%, P 5 0.14). Perioperative complications were higher in group 1A (55%) and group 1B (57%) compared with group 2 (29%), but these differences were not significantly different. Sim-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
TABLE I Patient Demographics
1A. Segment III Bypass
Group 1 Hilar Cholangiocarcinoma 1B. R. Sectoral Duct Bypass
Total Cholangiocarcinoma
Segment III Bypass
Total
20 11 9 62 6 3 0 0 5 9 6
14 10 4 62 6 2 0 0 5 8 1
34 21 13 62 6 2 0 0 10 17 7
21 9 12 61 6 3 0 0 6 15 0
55 30 25 62 6 2 0 0 16 32 7
Number Men Women Age (years) Stage I II III IV Indeterminate
Group 2 Gallbladder Carcinoma
Stage, according to the American Joint Committee on Cancer TNM classification, was determined at laparotomy.
TABLE II Perioperative Morbidity and Mortality Group 1 Group 2 Hilar Cholangiocarcinoma Gallbladder Carcinoma 1A. 1B. Total Segment III Bypass R. Sectoral Duct Bypass Cholangiocarcinoma Segment III Bypass Estimated blood loss (mL) Positive intraoperative bile cultures Length of stay (days) Postoperative complications (total) Bile leak Wound infection Biloma/abscess Renal failure Pancreatitis Entero-cutaneous fistula 30-Day mortality
Total
772 6 159
698 6 143
738 6 106
477 6 133
629 6 84
7 (35%) 13 6 1
6 (33%) 18 6 6
13 (38%) 15 6 2
4 (19%) 12 6 2
17 (31%) 14 6 2
11 (55%) 6 3 0 2 0
8 (57%) 4 1 2 0 1
19 (56%) 10 4 2 2 1
6 (29%) 3 1 1 0 0
25 (45%) 13 5 3 2 1
0 0*
0 3 (21.4%)*
0 3 (9%)
1 3 (14.3%)*
1 6 (11%)
Estimated blood loss and length of stay are expressed as the mean 6 standard deviation. Complications are expressed as the total number of complications divided by the number of patients in each group; 7 patients experienced more than one complication. The complication rate was higher in groups 1A and 1B compared with group 2 but the differences were not significant. * P 5 0.03, group 1A versus group 1B; P 5 0.08, group 1A versus 2.
ilarly, biliary reconstruction using an extended jejunal loop did not result in greater complications. Postoperative biliary-enteric anastomotic leak occurred in 13 patients (23%), and accounted for 52% of the total complications. Four of these patients required percutaneous stenting to control postoperative biliary leaks, all of whom had been subjected to transhepatic drainage before surgery (P 5 0.0001). The remainder resolved spontaneously. One patient required reoperation for a complication (upper gastrointestinal hemorrhage). The mean estimated blood loss for all procedures was 629 6 84 mL and did not differ significantly among the three groups (Table II). The mean length of stay from the time of surgery was 14 6 2 days, and although higher in patients who underwent a right sectoral hepatic duct bypass (group 2, 18 6 6 days), this difference was not signif-
icantly different. A postoperative biliary leak significantly prolonged the hospital stay beyond 10 days (P 5 0.04). Consequences of Preoperative Biliary Drainage Thirty patients (55%) were subjected to preoperative biliary tract procedures prior to referral (Table III). Percutaneous transhepatic cholangiogram (PTC) with external biliary drainage was the most common procedure (n 5 14) but a significant number underwent either endoscopic retrograde cholangiopancreatography (ERCP) with attempted stent placement (n 5 9) or surgery (cholecystectomy or placement of a T-tube or both, n 5 11). Sixteen patients (53%) experienced one or more complications related to these procedures. Ten patients either presented with or were previously treated for cholangitis; 4 others developed an infected bile collection. Preoperative trans-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
455
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
TABLE III Preoperative Biliary Tract Procedures Performed Prior to Referral and Resulting Complications
1A. Segment III Bypass Preoperative biliary interventions (patients) PTC ERCP Surgery Complications from preoperative interventions (patients) Cholangitis Biloma/abscess Tumor in tube tract
Group 1 Hilar Cholangiocarcinoma 1B. R. Sectoral Duct Total Bypass Cholangiocarcinoma
Group 2 Gallbladder Carcinoma
Segment III Bypass
Total
14 (70%) 10 2 4
9 (64%) 2 3 4
23 (68%) 12 5 8
7 (33%) 2 4 3
30 (55%) 14 9 11
8 (57%) 6 1 1
6 (67%) 4 1 1
14 (61%) 10 2 2
2 (29%) 0 2 0
16 (53%) 10 4 2
Four patients underwent multiple procedures. PTC 5 percutaneous transhepatic cholangiogram with either external biliary drainage or internal/external stent; ERCP 5 endoscopic retrograde cholangiopancreatography with stent placement; Surgery 5 8 patients underwent laparotomy and cholecystectomy, 3 patients had placement of T-tubes.
TABLE IV Effect of Preoperative Percutaneous Transhepatic Biliary Drainage on Intraoperative and Postoperative Results
Positive intraoperative bile cultures Operative blood loss .500 mL Postoperative intervention for bile leak Postoperative complications (total)
Preoperative Percutaneous Transhepatic Biliary Drainage (n 5 14)
No Preoperative Biliary Drainage (n 5 25)
P
64%
8%
0.001
79%
14%
0.007
29%
0%
0.0001
50%
39%
NS
NS 5 not significant.
hepatic drainage caused tumor seeding along the tube tract in 2 patients. Positive intraoperative bile cultures were more frequent after preoperative percutaneous transhepatic drainage (PTC) than in patients not previously drained (64% versus 8%, P 5 0.001). Operative blood loss was also greater (EBL .500 mL, 79% preoperative PTC versus 14% no preoperative drainage, P 5 0.007). Furthermore, in preoperatively drained patients, postoperative biliary leak was less likely to resolve spontaneously and more often required percutaneous stenting for control (29% versus 0%, P 5 0.0001). Postoperative morbidity was higher in preoperatively drained patients but this difference was not significant (50% versus 39%, P 5 0.4; Table IV). Relief of Symptoms All surviving patients experienced relief of jaundice and related symptoms following bypass. Pruritis resolved in 10 of 10 patients reporting this symptom preoperatively. The overall decrease in the serum bilirubin level was 54 6 8% at the time of initial follow-up (within 4 weeks from discharge). The preoperative serum bilirubin level in pa456
tients with gallbladder cancer (group 3) was higher than that in the other groups, but the rate of decline was similar (Figure 1). The decline in serum alkaline phosphatase levels was much less dramatic than that for bilirubin, and in many cases remained static or increased. Neither the preoperative nor the postoperative serum bilirubin and alkaline phosphatase, nor the percentage decrease in these parameters, was a useful predictor of adverse outcome. Survival Overall survival was significantly greater in patients with cholangiocarcinoma (groups 1A and 1B) compared with those with gallbladder cancer (P 5 0.0001). Median survival in patients with cholangiocarcinoma was 52 weeks (mean 5 62 6 7), and 6 patients are alive at 6, 10, 25, 30, 68, and 104 (2 patients) weeks. There was no significant difference in survival between groups 1A and 1B (74 6 14 weeks versus 57 6 9 weeks, P 5 0.2). On the other hand, median survival in patients with gallbladder cancer was 20 weeks (mean 21 6 4). Two patients are alive at 4 weeks from the time of surgery. One patient survived beyond 1 year (57 weeks), while all of the remaining patients died
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
Figure 2. Cumulative survival stratified by primary tumor. Median survival was significantly greater for patients with cholangiocarcinoma than for patients with gallbladder carcinoma (52 weeks versus 20 weeks, P 5 0.0001). There was no difference in survival between groups 1A (hilar cholangiocarcinoma/segment III bypass) and 1B (hilar cholangiocarcinoma/right sectoral hepatic duct bypass).
Figure 1. Mean serum bilirubin (A) and alkaline phosphatase (B) levels for each group before and after bypass. Panel C shows the percentage decrease in serum bilirubin and alkaline phosphatase for each group after surgery.
within 32 weeks of bypass. The Kaplan-Meier actuarial survival curves, stratified by primary tumor, are shown in Figure 2. Age greater than 60 years was the only factor that significantly reduced mean survival in patients with gallbladder cancer (16 6 3 weeks versus 26 6 7 weeks, P 5 0.05). In patients with cholangiocarcinoma, stage IV disease (41 6 6 weeks versus 76 6 11 weeks, P 5 0.008) and postoperative biliary-enteric anastomotic leak (31 6 4 weeks versus 69 6 8 weeks, P 5 0.02) significantly reduced survival, while age greater than 60 years was a much less significant factor (53 6 8 weeks versus 78 6 13 weeks, P 5 0.1). Of the 14 patients who received postoperative intraluminal radiotherapy, 13 had cholangiocarcinoma. While mean survival was greater in this group compared with patients with cholangiocarcinoma not given radiotherapy, the difference was not significant (73 6 11 weeks versus 54 6 9 weeks, P 5 0.2). Late Complications Twenty-three patients (47%) required readmission to the hospital for biliary tract sepsis, biliary tract interventions or surgical complications; readmissions for comfort care of preterminal patients were excluded (Table V). Nine patients were readmitted more than one time. When compared with segment III bypass (group 1A), right sectoral hepatic duct bypass (group 1B) was associated with more readmissions per patient (1.6 6 0.4 versus 0.8 6 0.2, P 5 0.1) and more readmission hospital days per patient (14.6 6 4 versus 5.1 6 2, P 5 0.04). Readmissions for
biliary tract sepsis were also more common in these patients compared with group 1A (P 5 0.05). Fifteen patients (31%) required postoperative biliary tract intervention at some point in their course—14 transhepatic stents and 1 reoperation and revision of the anastomosis. Nine procedures (including 1 reoperation) were performed for late bypass failure secondary to either stricture or progression of disease, 4 stents were placed to help control anastomotic leak, and 2 stents were placed for cholangitis in an undrained hepatic lobe. The average time to occlusion in all patients whose bypass did fail was 23 6 6 weeks. Patients who underwent a right sectoral hepatic duct bypass (group 1B) constituted the majority of those requiring reintervention for late bypass failure (6 patients, 55% versus 3 patients, 15% in group 1A, P 5 0.06; Table V). Bypass Patency The cumulative patency of all bypasses performed for hilar cholangiocarcinoma, stratified by group, is shown in Figure 3. The actuarial patency was greater for segment III bypass (Group 1A) than for right sectoral hepatic duct bypass (Group 1B) but the difference was not significant (P 5 0.1). Eighty percent of segment III bypasses performed for cholangiocarcinoma were patent at 1 year compared with 60% for right sectoral hepatic duct bypass. Patients with gallbladder cancer were not included in analyses of long-term bypass performance because nearly all died of their disease by 32 weeks. Mean follow-up for all patients was 42 6 5 weeks and was similar to the overall survival of 49 6 6 weeks.
COMMENTS Malignant obstruction of the hepatic ductal confluence remains a difficult clinical problem. Cholangiocarcinoma
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
457
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
TABLE V Summary of Readmissions and Late Bypass Failures after Surgery Group 1 Group 2 Hilar Cholangiocarcinoma Gallbladder Carcinoma 1A. Segment III 1B. Total Bypass R. Sectoral Duct Bypass Cholangiocarcinoma Segment III Bypass All readmissions Number of patients Readmissions per patient (mean) Readmission hospital days per patient (mean) Readmissions for biliary tract sepsis Number of patients Readmissions per patient (mean) Biliary interventions for late bypass failure Number of patients
Total
10 (50%)
8 (73%)
18 (58%)
5 (28%)
23 (47%)
0.8 6 0.2
1.6 6 0.4
1 6 0.2
0.3 6 0.1
0.8 6 0.2
5.1 6 2*
14.6 6 4*
8.1 6 2
3 6 1.6
6.3 6 1.4
7 (35%)
6 (55%)
13 (42%)
4 (22%)
17 (35%)
0.4 6 0.1†
1.0 6 0.3†
0.6 6 0.2
0.2 6 0.1
0.5 6 0.1
3 (15%)‡
6 (55%)‡
9 (29%)
0
9 (18%)
Readmissions for biliary tract sepsis, biliary tract interventions, and surgical complications were included. Readmissions for comfort care of terminally ill patients were excluded. Two patients readmitted with biliary tract sepsis required percutaneous drainage, while the remainder were successfully treated with antibiotics. Patients with gallbladder cancer, because of their significantly shorter survival, are not included in the statistical analysis of readmissions and late bypass failures. * P 5 0.04 group 1A versus group 1B. † P 5 0.05 group 1A versus group 1B. ‡ P 5 0.06 group 1A versus group 1B.
Figure 3. Cumulative patency of biliary-enteric bypass in all patients with cholangiocarcinoma. Both median and 1-year patency were greater for segment III bypass (group 1A) than for right sectoral hepatic duct bypass (group 1B) but the differences did not reach statistical significance.
and gallbladder cancer, the two most common causes, are amenable to curative resection in about one third of cases compared with 80% to 90% for periampullary tumors.11 Therefore, in patients with malignant obstruction of the hepatic ductal confluence, the goal of therapy is often palliation. Determining irresectability can be difficult, however, and often requires bile duct transection to establish with certainty. This is especially true in patients who have had previous biliary tract surgery or biliary intubation. In these cases, intrahepatic bypass is necessary to restore biliary continuity. 458
Biliary bypass and percutaneous stenting remain the principal options for palliating malignant hilar obstruction. In considering either of these approaches, the patient’s general state of health and life expectancy must be the foremost concerns. However, the procedure’s efficacy in relieving symptoms, the procedure-related morbidity and mortality, and its durability must also be considered. Additionally, and perhaps most importantly, patient inconvenience must be minimized. Any palliative intervention that introduces a new complication, such as frequent return trips to the hospital or an external drain that requires close attention, represents a great burden to the patient. For palliative biliary drainage, cholangitis and recurrent obstruction are the principal complications that require further intervention and must be minimized to preserve quality of life. Percutaneous biliary endoprostheses are the most effective alternative to surgical bypass for tumors of the hilum; endoscopic stents are generally ineffective for lesions in this location. The relatively recent introduction of selfexpandable metallic stents, with a greater patency than their plastic counterparts, has made the percutaneous approach somewhat more attractive.12 Whether percutaneous stents should supplant surgical bypass for unresectable hilar cancers, as suggested by some authors, remains an open question. Indeed, there have been four prospective studies comparing biliary stents with surgical bypass for distal bile duct obstruction.13 These studies demonstrate that surgery is associated with greater early morbidity and mortality but greater long-term patency and a lower incidence of recurrent jaundice. In a separate study, van den Bosch et al14 analyzed outcomes in 148 patients with distal bile duct obstruction from pancreatic carcinoma, concluding that surgical bypass provides optimal palliation for
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
patients surviving more than 6 months while biliary stenting is preferred for those surviving less than 6 months. Prospective comparisons between surgery and stents do not exist for proximal bile duct obstruction, but several reports suggest a similar conclusion. Firstly, the reported morbidity and mortality of stenting hilar cancers ranges from 7% to 31% and 12% to 14%, respectively, considerably higher than distal bile duct stents and similar to many surgical series.4 –7 Moreover, several studies have shown that hilar stents have a significant incidence of late complications. Becker et al6 reported a 46% 12-month patency for metallic stents placed for hilar malignancies versus 89% for those placed for distal bile duct tumors. The overall incidence of hilar stent occlusion ranges from 24% to 38%.4 –7 The long-term patency of intrahepatic biliaryenteric bypass, on the other hand, is not well reported in the literature. In this study, long-term bypass patency in patients with cholangiocarcinoma is clearly documented, and the results suggests that segment III bypass is superior to biliary stents. Several surgical techniques have been described for intrahepatic biliary bypass. The two most common approaches are bypass to the segment III duct and the right sectoral hepatic ducts.8 The former, because of the more constant anatomy and long extrahepatic course of the left hepatic duct, is technically easier and preferred. However, the type of bypass is usually dictated by the location of the tumor. In general, segment III bypass is performed unless the left lobe is atrophic or heavily involved with tumor or if the primary lesion extends to the umbilical fissure of the liver. In patients with unresectable gallbladder cancer, segment III bypass is usually the only option since the right hepatic ducts are frequently involved with tumor. Normalization of bilirubin levels and relief of jaundice can be achieved if at least one third of the functioning liver parenchyma is adequately drained. Lack of communication between the right and left hepatic ducts, therefore, does not affect the results, provided that the undrained liver has not been percutaneously drained or otherwise contaminated. In this circumstance, there is a high risk of persistent biliary fistula and cholangitis.15 This study shows that intrahepatic biliary-enteric bypass effectively relieves jaundice and symptoms from malignant hilar obstruction. All patients were effectively palliated following surgery. Most patients experienced a prompt and steady decline in serum bilirubin, which approached normal within 4 weeks of bypass. The change in serum alkaline phosphatase was less dramatic and often remained static or increased. This observation, not unexpected, is most likely the result of persistent ductal obstruction in the undrained liver. Neither the preoperative bilirubin and alkaline phosphatase nor the postoperative change had any prognostic value. Several points emerge from analysis of the three subgroups. First, the median survival for patients with gallbladder cancer was 5 months. The 28% morbidity and 16.6% mortality associated with segment III bypass in these patients might be considered excessive in light of their poor survival. These patients are, therefore, unlikely to reap the long-term benefits of surgical bypass and are probably best palliated by percutaneous biliary endoprostheses.
Survival in patients with cholangiocarcinoma, on the other hand, was 12 months. The perioperative morbidity in this group can, therefore, be justified if good long-term results are achieved. Late complications, particularly cholangitis and recurrent biliary obstruction, are the bane of all palliative biliary procedures. In this study, we used readmissions for biliary tract-related problems and bypass patency over the life of the patient to assess long-term functional results. Overall, 47% of patients required readmission for treatment of biliary sepsis or recurrent biliary obstruction. Patients with gallbladder cancer had deceptively few late complications when compared with the other groups, which is almost certainly attributable to their poor survival. In patients with cholangiocarcinoma, segment III bypass provided superior long-term palliation than right sectoral hepatic duct bypass. Not only were readmissions for biliary tract-related complications fewer and shorter for these patients, but bypass patency was better. Late bypass failure and the need for further biliary tract interventions were significantly less common after segment III bypass than after right sectoral hepatic duct bypass. This observation may reflect the relatively greater technical difficulty of approaching the right sectoral hepatic ducts. The procedure-related mortality of 11% is within the range of other reports.2,16,17 On the other hand, the mortality associated with segment III bypass, 0% in patients with cholangiocarcinoma and 7% overall, is lower than that reported in many other surgical series. In fact, the reported mortality for biliary stents is often higher. Biliary anastomotic leak was the most common postoperative complication, accounting for 50% of all complications and significantly prolonged the hospital stay. The incidence of biliary leak was 23%. Although two thirds of all leaks resolved without any intervention, the presence of a biliary leak had a significant negative impact on survival in patients with cholangiocarcinoma. Anastomotic stricture is a potential sequela of biliary anastomotic leak. This mechanism, while possibly playing a role, is not the sole reason for the survival difference since patients who developed a biliary leak did not have a significantly increased incidence of biliary sepsis or late bypass failure. More than half of the patients in this study were referred after having had one or more biliary tract interventions. Percutaneous transhepatic drainage was the most common preoperative procedure and significantly increased the perioperative morbidity. Not surprisingly, contaminated bile at the time of surgery was a very common finding in these patients, although postoperative infectious complications were not significantly increased. Preoperative percutaneous drainage altered the natural history of biliary-enteric anastomotic leak by preventing spontaneous closure in some patients. Furthermore, the intense inflammatory response caused by biliary stents made dissection within the portal area very difficult, as evidenced by the significantly greater operative blood loss in stented patients. Many of these drainage catheters were placed for unclear indications, and two thirds failed to achieve their objective of relieving jaundice. In 2 cases, percutaneous drainage resulted in spread of the tumor along the tube tract. It is well-recognized that routine preoperative biliary
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998
459
PALLIATIVE INTRAHEPATIC BILIARY ENTERIC BYPASS/JARNAGIN ET AL
drainage does not improve outcome after surgery for obstructive jaundice; in fact, there is strong evidence that morbidity and mortality are increased.18 –20 MacPherson et al21 noted an increased operative mortality in previously drained patients undergoing resection of proximal bile duct cancers. Heslin et al22 reported that, in patients with distal bile duct obstruction, preoperative biliary stenting increases the incidence of postoperative infectious complications. In addition, the pathogens responsible for these infections are usually the same organisms colonizing the biliary tree.23 Moreover, the resectability of many hilar cancers can only be determined at laparotomy, a task made much more difficult in the presence of a biliary stent. Routine preoperative biliary drainage, in the absence of cholangitis or severe pruritis, should therefore be avoided. Clearly, both intrahepatic biliary-enteric bypass and percutaneous stenting have a role in palliating patients with malignant hilar obstruction. The best results will be obtained by selective use of each technique. This study suggest that the decision must take into account the natural history of the underlying disease. Those patients with advanced disease from an aggressive cancer, such as gallbladder carcinoma, have a limited survival and are probably best palliated with biliary stents. Cholangiocarcinoma, on the other hand, is a more indolent disease. In these patients, segment III bypass can be performed with an acceptable morbidity and a mortality lower than that reported for biliary stents. Moreover, this procedure provides excellent palliation with good long-term function and relatively few late complications. Bypass to the right sectoral hepatic ducts is associated with more early and late complications, and is therefore a less attractive option. Whether surgical bypass is superior to percutaneous biliary stenting can only be answered by a randomized, prospective study.
REFERENCES 1. Blumgart LH, Benjamin IS, Hadjis NS, Beazly R. Surgical approaches to cholangiocarcinoma at confluence of bile ducts. Lancet. 1984;i:66 –70. 2. Bismuth H, Castaing D, Traynor O. Resection or palliation: priority of surgery in the treatment of hilar cancer. World J Surg. 1988;12:39 – 47. 3. Cheung K-L, Lai ECS. Endoscopic stenting for malignant biliary obstruction. Arch Surg. 1995;130:204 –207. 4. Glattli A, Stain SC, Baer HU, et al. Unresectable malignant biliary obstruction: treatment by self-expandable biliary endoprostheses. HPB Surg. 1993;6:175–184. 5. Stoker J, Lameris JS. Complications of percutaneously inserted biliary Wallstents. JVIR. 1993;4:767–772. 6. Becker CD, Glattli A, Maibach R, Baer HU. Percutaneous palliation of malignant obstructive jaundice with Wallstent endoprosthesis: follow-up and reintervention in patients with hilar and non-hilar obstruction. JVIR. 1993;4:597– 604.
460
7. Rossi P, Bexxi M, Rossi M, et al. Metallic stents in malignant biliary obstruction: results of a multicenter European study of 240 patients. JVIR. 1994;5:279 –285. 8. Blumgart LH, Baer HU. Hilar and intrahepatic biliary-enteric anastomosis. In: Blumgart LH, ed. Surgery of the Liver and Biliary Tract. 2nd ed. Edinburgh: Churchill Livingstone; 1051–1067. 9. Blumgart LH, Kelley CJ. Hepaticojejunostomy in benign and malignant high bile duct stricture: approaches to the left hepatic ducts. Br J Surg. 1984;71:257–261. 10. Kuvshinoff BW, Armstrong JG, Fong Y, et al. Palliation of irresectable hilar cholangiocarcinoma with biliary drainage and radiotherapy. Br J Surg. 1995;82:1522–1525. 11. Nakeeb A, Pitt HA, Sohn TA, et al. Cholangiocarcinoma: a spectrum of intrahepatic, perihilar and distal tumors. Ann Surg. 1996;224:463– 475. 12. Schmassmann A, van Gunten E, Knuchel J, et al. Wallstents versus plastic stents in malignant biliary obstruction: effects of stent patency of the first and second stent on patient compliance and survival. Am J Gastroenterol. 1996;91:654 – 659. 13. Lillemoe KD, Pitt HA. Palliation: surgical and otherwise. Cancer. 1996;78:605– 614. 14. van den Bosch R, van der Schelling GP, Klinkenbijl JHG, et al. Guidelines for the application of surgery and endoprostheses in the palliation of obstructive jaundice in advanced cancer of the pancreas. Ann Surg. 1994;219:18 –24. 15. Baer HU, Rhymer M, Stain SC, et al. The effect of communication between the right and left liver on the outcome of surgical drainage for jaundice due to malignant obstruction at the hilus of the liver. HPB Surg. 1994;8:27–31. 16. Traynor O, Castaing D, Bismuth H. Left intrahepatic cholangio-enteric anastomosis (round ligament approach): an effective palliative treatment for hilar cancers. Br J Surg. 1987;74:952–954. 17. Myburgh JA. Resection and bypass for malignant obstruction of the bile duct. World J Surg. 1995;19:108 –112. 18. Sirinek KR, Levine BA. Percutaneous transhepatic cholangiography and biliary decompression. Arch Surg. 1989;124:885– 888. 19. Pitt HA, Gomes AS, Lois JF, et al. Does preoperative percutaneous biliary drainage reduce operative risk or increase hospital cost? Ann Surg. 1985;201:545–553. 20. Karsten TM, Davids PHP, van Gulik TM, et al. Effects of biliary endoprostheses on the extrahepatic bile ducts in relation to subsequent operation of the biliary tract. J Am Coll Surg. 1994;178: 343–352. 21. MacPherson GAD, Benjamin IS, Hodgson HJF, et al. Preoperative percutaneous transhepatic biliary drainage; the results of a controlled trial. Br J Surg. 1984;74:371–375. 22. Heslin MJ, Brooks AD, Hochwald SN, et al. Preoperative biliary stenting is associated with increased complications after pancreaticoduodenectomy. Arch Surg. In press. 23. Karsten TM, Allema JH, Reinders M, van Gulik TM, et al. Preoperative biliary drainage, colonisation of bile and postoperative complications in patients with tumours of the pancreatic head: a retrospective analysis of 241 consecutive patients. Eur J Surg. 1996;162:881– 888.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 175 JUNE 1998