Figure 1. Receiver operating characteristic curve of the duration of biliary obstruction for discriminating advanced hepatic fibrosis indicates that the duration threshold for best-compromise sensitivity-specificity is approximately 6.25 months, with corresponding sensitivity and specificity of 86%. Area under the curve, 0.924; P<.001.
Figure 2. Receiver operating characteristic curve of basal alanine aminotransferase levels for predicting advanced hepatic fibrosis reveals that the alanine aminotransferase threshold for best-compromise sensitivity-specificity is approximately 60.5 U/L, with corresponding sensitivity of 59% and specificity of 60%. Area under the curve, 0.648; P= .04.
Figure 3. Receiver operating characteristic curve of the time to normalization of alanine aminotransferase levels for predicting advanced hepatic fibrosis indicates that the cutoff value for best-compromise sensitivity-specificity is approximately 10.5 days, with corresponding sensitivity of 93% and specificity of 63%. Area under the curve, 0.843; P<.001.
Negi SS, Sakhuja P, Malhotra V, Chaudhary A. Factors Predicting Advanced Hepatic Fibrosis in Patients With Postcholecystectomy Bile Duct Strictures. Arch Surg. 2004;139(3):299-303. doi:10.1001/archsurg.139.3.299
Simplified evaluation based on clinical and biochemical variables might predict the degree of hepatic fibrosis in patients with postcholecystectomy bile duct strictures.
Prospective cohort study.
Tertiary care referral and teaching hospital.
Sixty-four patients with postcholecystectomy bile duct strictures undergoing definitive repair. Prospectively collected information included demographics, disease-related characteristics, and serial liver function tests. Hepatic histologic features (fibrosis, cholestasis, portal inflammation, and ductular proliferation) were independently graded by 2 pathologists masked to clinical data using a previously validated scale. Patients were dichotomized into groups based on degree of hepatic fibrosis. Univariate and multivariate analyses were performed.
Main Outcome Measure
Identification of variables that predict the presence of advanced hepatic fibrosis (grade 2-3).
Thirty-five patients (55%) had early hepatic fibrosis (grade 0-1), and the remaining 29 (45%) had advanced fibrosis (grade 2-3). Univariate analysis demonstrated that duration of biliary obstruction, presence of portal hypertension, basal alanine aminotransferase (ALT) levels, and time to normalization of serum total bilirubin, ALT, and alkaline phosphatase levels after surgical drainage were statistically significantly associated with the presence of advanced hepatic fibrosis. However, multivariate analysis revealed that only duration of biliary obstruction (odds ratio [OR], 1.6048; P = .009), basal ALT levels (OR, 0.9634; P = .02), and time to normalization of ALT levels after surgical drainage (OR, 1.6680; P = .006) were significant predictors of advanced hepatic fibrosis.
Duration of biliary obstruction, basal ALT level, and time to normalization of ALT level after surgical repair are independent predictors of advanced hepatic fibrosis (grade 2-3) in patients with postcholecystectomy bile duct strictures.
Iatrogenic bile duct injury is the most common cause of benign bile duct stricture. The growing popularity of laparoscopic cholecystectomy and its establishment as a gold standard of therapy for symptomatic cholelithiasis has renewed interest in the subject because of the overriding concern that the procedure might be associated with a higher incidence of bile duct injury. The biliary obstruction produced by these strictures initiates complex hepatic histopathologic changes, which can result in progressive hepatic fibrosis and secondary biliary cirrhosis.1 These changes affect the immediate and long-term outcomes of definitive surgical repair2 and, therefore, have prognostic and medicolegal implications. Hence, there is a need to identify factors, preferably based on simplified evaluation of clinical and biochemical variables, that might predict the presence of such changes.
Sixty-four patients with postcholecystectomy bile duct strictures undergoing hepaticojejunostomy in the Department of Gastrointestinal Surgery, Gobind Ballabh Pant Hospital, University of Delhi, between January 1, 1998, and March 31, 2003, were included in this study. Individuals with positive hepatitis B or C viral markers, with associated portal vein injury, receiving concomitant cholestatic drugs, or known to consume alcohol were excluded. Prospectively collected information included age, sex, type of cholecystectomy (laparoscopic or open), duration of biliary obstruction, level of obstruction (Bismuth classification), degree of obstruction (partial or complete), and the presence of cholangitis, pruritus, previous biliary-enteric anastomosis, infected bile, intraductal calculi, cholangitic liver abscess, portal hypertension, atrophy-hypertrophy complex, or postoperative complications. The duration of biliary obstruction was estimated from the period of awareness of clinical jaundice. The level and degree of obstruction were decided on the basis of findings from magnetic resonance cholangiopancreatography in 53 patients (83%) and percutaneous transhepatic cholangiography in the remaining 11 patients (17%). The level of biliary obstruction was classified as "high" in patients with Bismuth types 3 and 4 biliary stricture, and those with types 1 and 2 stricture were classified as having "low" obstruction. The diagnosis of portal hypertension was based on clinical findings in conjunction with those from endoscopy or Doppler ultrasound and was confirmed with intraoperative findings. Liver atrophy was diagnosed on the basis of findings from cholangiography, computed tomography, or magnetic resonance imaging and was defined as at least a 50% reduction in the size of the hepatic lobe or segment. The presence of infected bile was confirmed by culture findings from intraoperatively obtained bile specimens. Serum total bilirubin, serum alanine aminotransferase (ALT), and serum alkaline phosphatase (SAP) levels were measured immediately before surgery (basal values); serially on the 1st, 4th, 14th, and 28th days after hepaticojejunostomy; and thereafter at monthly intervals until normalization of corresponding values. The time to normalization of liver function test results after surgical decompression was also estimated. The reference value for serum bilirubin was less than 1.2 mg/dL (20.5 µmol/L), for ALT was less than 35 U/L, and for SAP was less than 260 U/L.
Liver biopsy was performed during laparotomy in both lobes of the liver. The biopsy samples were fixed in 10% neutral buffered formalin solution and embedded in paraffin, and 4-µm-thin sections were obtained. The sections were stained with hematoxylin-eosin, Masson trichrome, reticulin, and orcein. Hall staining for bile pigment was performed if required. The histologic features were independently graded by 2 pathologists (P.S. and V.M.) masked to clinical data using a scale previously validated in a study of primary biliary cirrhosis3 and later used to study regression of hepatic histologic changes after biliary drainage.4 Fibrosis was graded on a scale from 0 to 3, with 0 denoting the absence of fibrosis; 1, portal and periportal fibrosis; 2, the presence of numerous septa; and 3, cirrhosis. Portal and periportal inflammation was graded on a scale from 0 to 3, with 0 denoting the absence of inflammation; 1, mild inflammation (cells in less than one third of portal tracts); 2, moderate inflammation (cells in one third to two thirds of portal tracts); and 3, severe inflammation (dense packing of cells in more than two thirds of portal tracts). Ductular proliferation was graded on a scale from 0 to 2, with 0 denoting absent or mild; 1, moderate; and 2, severe. Cholestasis was graded on a scale from 0 to 3, with 0 denoting the absence of cholestasis; 1, bile accumulation in centrolobular hepatocytes; 2, bile accumulation in centrolobular and periportal hepatocytes or in portal tracts; and 3, the presence of bile infarcts (bile accumulation with hepatocyte necrosis surrounded by foamy histiocytes).
The data were analyzed using a commercially available statistical software package (SPSS for Windows version 9.0; SPSS Inc, Chicago, Ill). Data are given as mean ± SEM and ranges. A bivariate correlation matrix was constructed initially to determine the relationships among hepatic histologic variables. Since the primary outcome measure was identification of variables that predict the presence of advanced hepatic fibrosis (grades 2 and 3),5 patients were accordingly dichotomized into groups for further analysis. Univariate analysis was performed using the χ2 test for categorical variables and the 2-tailed, unpaired t test for continuous variables. Backward stepwise logistic regression analysis was performed using variables found to be statistically significant on univariate analysis, and odds ratios were estimated using the final regression model, with calculation of 95% confidence intervals (CIs). The diagnostic values of variables identified by multivariate analysis were compared using receiver operating characteristic curves, and the area under the curve was calculated using 2-sided P values. An area under the curve of 1.0 indicated a predictor variable with perfect discrimination between patients with and without advanced hepatic fibrosis, whereas an area under the curve of 0.5 or less indicated a predictor variable with no discriminatory power. A P≤.05 (2-tailed) was considered statistically significant.
This study consisted of 55 females (86%) and 9 males (14%) with a mean age of 36.86 ± 1.29 years (range, 14-69 years). The bile duct injury was sustained during open cholecystectomy in 48 patients (75%) and during laparoscopic cholecystectomy in the remaining 16 patients (25%). The mean duration of biliary obstruction was 16.62 ± 3.43 months (range, 1-120 months); 1 patient (2%) had Bismuth type 1 stricture, 23 (36%) had type 2, 35 (55%) had type 3, and 5 (8%) had type 4. The level of biliary obstruction was classified as high in 40 patients (63%), and the degree of obstruction was partial in 12 patients (19%). Consequent to previous failed attempts at surgical repair, 20 patients (31%) had a biliary-enteric anastomosis performed for a biliary stricture that had stenosed, thereby necessitating reoperation. Episodic cholangitis was the initial feature seen in 58 patients (91%), and 37 patients (58%) had associated pruritus. However, none of the patients had active cholangitis at the time of hospital admission. At the time of surgical repair, 13 patients (20%) had associated portal hypertension, 8 (12%) had atrophy-hypertrophy complex, 4 (6%) had cholangitic liver abscesses, 7 (11%) had hepatolithiasis, and 25 (39%) had infected bile. All the patients underwent hepaticojejunostomy, and none developed cholangitis during follow-up. One patient died consequent to hepatic failure (this patient had grade 3 hepatic fibrosis with associated portal hypertension). Twelve patients (19%) developed at least 1 postoperative complication, including wound infection in 6 (9%), bile leak in 4 (6%), ascitic leak in 3 (5%), intra-abdominal collection in 1 (2%), and stress bleed in 1 (2%).
The mean basal level of serum total bilirubin was 6.56 ± 0.64 mg/dL (112.2 ± 10.9 µmol/L) (range, 1.50-26.80 mg/dL [25.6-458.3 µmol/L]), of ALT was 87.55 ± 8.88 U/L (range, 17-290 U/L), and of SAP was 626.78 ± 70.31 U/L (range, 115-2878 U/L). After definitive surgical repair, serum total bilirubin levels normalized after a mean of 18.01 ± 3.21 days (range, 4-168 days), ALT levels normalized after 20.19 ± 3.48 days (range, 4-168 days), and SAP levels normalized after 24.53 ± 5.56 days (range, 4-252 days).
The grading of hepatic histologic features by the 2 pathologists was concordant for all the patients except 4; the final histologic grading in these patients was agreed on after consensus. The grade of hepatic fibrosis demonstrated positive correlation with the grade of portal inflammation (r = 0.547; P<.001), ductular proliferation (r = 0.373; P = .002), and cholestasis (r = 0.268; P = .03). Hepatic fibrosis was graded as absent (grade 0) in 5 patients (8%), portal or periportal (grade 1) in 30 (47%), severe with numerous septa (grade 2) in 22 (34%), and cirrhosis (grade 3) in 7 (11%). Based on the degree of hepatic fibrosis, 35 patients (55%) had early fibrosis, and the remaining 29 (45%) had advanced hepatic fibrosis. Univariate analysis demonstrated that the duration of biliary obstruction (P<.001), portal hypertension (P<.001), basal ALT levels (P = .01), and times to normalization of serum total bilirubin (P = .02), ALT (P = .003), and SAP (P = .02) levels were significantly associated with the presence of advanced hepatic fibrosis (Table 1, Table 2, and Table 3). However, subsequent logistic regression analysis revealed that the duration of biliary obstruction, basal ALT levels, and time to normalization of ALT levels after surgical decompression were the only statistically significant predictors of advanced hepatic fibrosis (Table 4). For discrimination of advanced hepatic fibrosis, the area under the receiver operating characteristic curve was 0.924 ± 0.034 (95% CI, 0.856-0.991; P<.001) for duration of biliary obstruction, 0.648 ± 0.072 (95% CI, 0.507-0.788; P = .04) for basal ALT levels, and 0.843 ± 0.048 (95% CI, 0.749-0.937; P<.001) for time to normalization of ALT levels after surgical decompression (Figure 1, Figure 2, and Figure 3). The analysis of receiver operating characteristic curves further revealed that the threshold values of predictor variables for best-compromise sensitivity-specificity was 6.25 months for the duration of biliary obstruction (sensitivity and specificity, 86%), 60.5 U/L for basal ALT levels (sensitivity, 59%; specificity, 60%), and 10.5 days for time to normalization of ALT levels (sensitivity, 93%; specificity, 63%). The mean duration of biliary obstruction associated with grade 1 hepatic fibrosis was 3.87 ± 0.38 months (range, 1-9 months), with grade 2 fibrosis was 22.45 ± 6.17 months (range, 2-107 months), and with grade 3 fibrosis was 62 ± 14.53 months (range, 18-120 months).
Prolonged biliary obstruction may lead to progressive hepatic fibrosis and secondary biliary cirrhosis. The presence of cirrhosis at the time of surgical repair of biliary strictures is considered an ominous sign and is the most important risk factor for postoperative morbidity and mortality consequent to liver failure or hemorrhage.2,6 Also, the results of definitive surgical repair are suboptimal in these patients. In a study2 of patients with recurrent biliary strictures, only 25% with cirrhosis had a satisfactory result after the first repair. Moreover, secondary biliary cirrhosis, despite a patent hepaticojejunostomy, is an important cause of late mortality in patients with postcholecystectomy bile duct strictures.7
Most information regarding structural and functional changes in the liver after extrahepatic bile duct obstruction has come from experimental studies. Morphologically, these changes are characterized by "biliary interface activity" or "biliary piecemeal necrosis," which includes cholestasis, ductular proliferation, portal inflammation, and fibrosis in varying combinations.1 We observed statistically significant positive correlations between hepatic fibrosis and both portal inflammation and ductular proliferation. Ductular proliferation with accompanying inflammatory infiltrate might play an important role in hepatic fibrogenesis.8 Unlike previous investigations,9,10 which included patients with diverse causes of biliary obstruction, our study focuses on patients with a single common cause. We observed hepatic histopathologic changes in most patients with postcholecystectomy bile duct strictures, with grade 1 hepatic fibrosis being reported in 47%, grade 2 in 34%, and grade 3 in 11%. In a study11 of 16 patients with postcholecystectomy bile duct strictures, none of the 14 liver biopsy results were reported to be normal, and the findings were categorized as mild in 43%, moderate in 43%, and severe in 14%.
To the best of our knowledge, the role of clinical and biochemical factors in predicting the degree of hepatic fibrosis in patients with postcholecystectomy bile duct strictures has not been evaluated. The present study revealed that a long duration of biliary obstruction is the most important predictor of advanced hepatic fibrosis. Unlike in experimental studies, it is difficult to establish precisely when biliary obstruction develops in clinical cases, and the duration of biliary obstruction in the present study was estimated from the duration of clinical jaundice. Fibrotic changes were observed as early as 1 month after biliary obstruction, and the mean duration of biliary obstruction associated with the development of portal or periportal fibrosis (grade 1) was 3.87 ± 0.38 months, of severe fibrosis with numerous septa (grade 2) was 22.45 ± 6.17 months, and of cirrhosis (grade 3) was 62.00 ± 14.53 months. In an experimental study12 of extrahepatic biliary obstruction, it was demonstrated that there is a relative fall in the volume of hepatocytes and an increase in bile duct cells and biliary stroma that is directly related to the period of jaundice. Besides, hepatic fibroblastic activity has been known to occur within 7 days, and nodular regeneration was observed as early as 212 days after the onset of jaundice in patients with extrahepatic biliary obstruction.9 The time required for the development of secondary biliary cirrhosis depends on the cause of biliary obstruction, and the mean interval between the onset of biliary obstruction and the confirmation of cirrhosis has been reported to be 7.1 years in patients with common bile duct strictures, 4.6 years in those with common bile duct stones, and 0.8 years in patients with malignant biliary obstruction.10
Raised serum ALT levels in patients with biliary obstruction suggest ongoing hepatocyte necrosis. Our study identified raised basal ALT levels and time to normalization of ALT levels after definitive surgical repair of biliary strictures as independent predictors of advanced hepatic fibrosis. Because there was not a statistically significant difference in the incidence of clinical cholangitis or infected bile between early (grade 0-1) and advanced (grade 2-3) hepatic fibrosis, the higher ALT levels in patients with advanced hepatic fibrosis cannot be attributed to the presence of episodic cholangitis. Studies13,14 on recovery of conventional liver function test results after surgical decompression of biliary obstruction have shown that basal ALT levels might be slightly raised in these patients, and after biliary decompression rapid recovery to reference levels occurs by postoperative day 7. However, these studies excluded patients with secondary biliary cirrhosis, and, hence, the trend of the serum ALT level after biliary decompression in patients with extrahepatic biliary obstruction and advanced hepatic fibrosis cannot be extrapolated from these studies.
Portal hypertension was statistically significantly associated with advanced hepatic fibrosis on univariate analysis, but it did not emerge as an independent predictor on multivariate analysis. In patients with extrahepatic biliary obstruction, portal hypertension develops consequent to a combination of hepatocyte hyperplasia and increased portal fibrosis,15,16 although the relative contribution of these in the genesis of portal hypertension is still speculative. It is believed that the degree of biliary obstruction and the presence of associated infection might play an important role in hepatic fibrogenesis.1,10 However, we observed no association between the degree of hepatic fibrosis and the presence of complete or high biliary obstruction, clinical cholangitis, infected bile, hepatolithiasis, and cholangitic liver abscess. Although the presence of biliary sepsis might have a synergistic role in the generation of hepatic morphologic changes, it did not emerge as an independent predictor of such changes. In a study17 of 70 patients with mechanical biliary obstruction, no correlation was observed between the presence of clinical cholangitis and hepatic morphologic changes. Besides, it has also been demonstrated that the clinical presentation of the patient with extrahepatic biliary obstruction does not predict the incidence of infected bile.17,18
Most of these histopathologic changes, especially those in the early stages, are potentially reversible. In an experimental model of extrahepatic biliary obstruction, the reversibility of hepatic fibrosis and ductular proliferation has been reported after biliodigestive anastomosis.19,20 Moreover, significant regression of hepatic fibrosis, ductular proliferation, and portal inflammation was observed after biliary drainage in patients with chronic pancreatitis and common bile duct stenosis.4 Besides, recent experimental evidence suggests that interferon α-2a might promote the resolution of fibrosis and decrease bile duct proliferation in extrahepatic biliary obstruction.21
The results of this study suggest that the duration of biliary obstruction and the trend of the ALT level are independent predictors of hepatic fibrosis in patients with postcholecystectomy bile duct strictures. Those having a long duration of biliary obstruction, a high basal ALT level, and a long time to normalization of ALT levels after a biliary drainage procedure are likely to have advanced hepatic fibrosis. Such knowledge might be helpful in prognosticating patients and in predicting response to biliary drainage procedures. Because these morphologic changes have a high incidence, occur in severity in time bound fashion, and respond favorably to therapy (mostly early lesions), biliary decompression is advisable at the earliest opportune moment after the detection of biliary obstruction.
Corresponding author: Adarsh Chaudhary, MS, University of Delhi, D II/97, East Kidwai Nagar, New Delhi-110023, India (e-mail: firstname.lastname@example.org).
Accepted for publication October 28, 2003.