S106 Management of Post-Cholecystectomy Bile Leaks: Predictors for Persistent Leak After Initial ERCP

Introduction: Endoscopic retrograde cholangiopancreatography (ERCP) is a 1st-line treatment for post-cholecystectomy bile leaks (PCBL). Despite appropriate initial ERCP interventions that optimize transpapillary bile flow, PCBL can persist. We aim to evaluate baseline clinical factors associated with persistent PCBL after initial ERCP. Methods: We created a retrospective database of patients with PCBL referred for ERCP at Indiana University Health (IUH) University Hospital. Data collected included endoscopic reports, fluoroscopic imaging, patient demographics, type of PCBL, ERCP timing and interventions, technical success, and treatment outcomes. A high-grade bile leak (HG-BL) was defined as visualization of contrast extravasation from the bile duct before filling of intrahepatic biliary branches with contrast. The PCBL was deemed to be persistent if > 1 interventional ERCP was required for the leak to resolve on cholangiogram. (Figure) Results: From 2011 to 2021, 369 cases of PCBL were referred for ERCP. After excluding cases with transected bile ducts (n=21) and patients lost to follow up (n=15), 333 cases were included in data analysis (Image 1). 21 patients received their initial ERCP at an outside hospital. All patients received biliary sphincterotomy with stenting unless there was coagulopathy or Roux-en-Y gastric bypass (n=12). Univariate logistic regression analysis identified male gender, steroids use at time of ERCP, presence of duodenal stricture, a HG-BL, Strasberg class D PCBL, presence of biloma, presence of abdominal drain, presence of biliary stricture, and initial ERCP performed at outside hospital to be significant variables for persistent PCBL (Table). On multivariate analysis, presence of a HG-BL (OR 7.08, CI 1.96 – 25.58, p = 0.003) and initial ERCP performed at an outside facility (OR 29.14, CI 3.35 – 253.63, p=0.002) remained significantly associated with persistent PCBL. Being a female gender (OR 0.30, CI 0.11 – 0.85, P = 0.022) remained significant for higher odds of PCBL resolution after initial ERCP. Conclusion: Our study suggests that a HG-BL in a male patient may justify more aggressive ERCP interventions or more prolonged stent indwell time. As a tertiary referral hospital, we had a significant referral bias in our cohort, which could have acted as a confounder causing statistical significance for lower initial ERCP success rates at outside medical facilities. The relatively high volume of ERCP cases at our center may also have contributed to better ERCP outcomes.Figure 1.: Flowchart of the Study Table 1. - Univariate analysis of clinical variables for persistent PCBL after initial ERCP treatment Variables Initial ERCP failure OR (95% CI) P value Patient Characteristics Female 25/204 (12.25%) 0.49 (0.27 – 0.86) 0.013 Age 55 years old or more 34/175 (19.43%) 1.60 (0.90-2.86) 0.112 Median BMI (IQR) 32.0 (26.0, 36.0) 1.00 (0.98 – 1.03) 0.874 On diabetic treatment 13/58 (22.41%) 1.81 (0.93 – 3.53) 0.083 On steroids 4/6 (66.67%) 10.18 (1.82 – 57.00) 0.008 Major papilla within duodenal diverticulum 11/57 (19.30%) 1.46 (0.73 – 2.93) 0.282 Duodenal stricture/stenosis 5/10 (50.00%) 5.17 (1.45 – 18.48) 0.011 Location and timing of ERCP Procedure ERCP performed > 3 days 48/273 (17.58%) 1.00 (0.47 – 2.12) 0.994 Initial ERCP performed at outside facility 17/21 (80.95%) 30.94 (9.96 -96.06) < 0.001 Bile leak characteristics Strasberg Class A bile leaks 44/302 (14.57%) 0.26 (0.12 – 0.54) < 0.001 Strasberg Class D bile leaks 10/31 (32.26%) 2.70 (1.19 – 6.12) 0.018 Bile leak from cystic duct* 34/192 (17.71%) 1.25 (0.70 – 2.24) 0.446 Bile leak from duct of Luschka 4/104 (3.85%) 0.21 (0.09 – 0.50) < 0.001 Bile leak from common bile duct/common hepatic duct 7/25 (28.00%) 2.66 (1.13 – 6.27) 0.025 High-grade bile leak 25/63 (39.68%) 4.80 (2.58 – 8.95) < 0.001 Concomitant biloma present 27/115 (23.48%) 2.06 (1.13 – 3.76) 0.018 Percutaneous abdominal drain present 42/204 (20.59%) 2.02 (1.07 – 3.82) 0.030 Presence of biliary stones 5/67 (7.46%) 0.34 (0.13 – 0.88) 0.026 Presence of biliary stricture 7/21 (33.33%) 2.63 (1.01 – 6.84) 0.048 ERCP interventions Biliary sphincterotomy performed 56/325 (17.23%) 1.04 (0.22 – 4.88) 0.960 Biliary sphincterotomy alone 2/6 (33.33%) 2.46 (0.44 – 13.73) 0.306 Biliary stenting alone 2/12 (16.67%) 0.96 (0.21 – 4.51) 0.960 Bridging biliary stents 20/134 (14.93%) 0.78 (0.43 – 1.43) 0.426 Multiple biliary plastic stents 6/55 (10.91%) 0.55 (0.22 – 1.36) 0.198 Self-expandable metal stent 2/11 (18.18%) 1.09 (0.23 – 5.21) 0.910 *Reference group. IQR, interquartile range; IUH, Indiana University Health.