Endoscopic management of postcholecystectomy biliary leak: When and how? A nationwide study

Abstract

ERCP is considered the first-line therapy for biliary duct leaks (BDLs). However, the optimal ERCP timing and endotherapy methods remain controversial. Our aim was to evaluate these factors as predictors of poor clinical outcomes after BDLs. Adults who underwent ERCP for BDLs after cholecystectomy were identified from the Nationwide Inpatient Sample from 2000 to 2014. ERCP was classified as emergent, urgent, and expectant if it was done within 1 day, after 2 to 3 days, or >3 days after BDLs, respectively. Endotherapy was classified into sphincterotomy, stent, or combination. Post-ERCP adverse events (AEs) were defined as requiring pressor infusion, endotracheal intubation, invasive monitoring, or hemodialysis. Early endotherapy failure was defined as the need for salvage surgical or radiology-percutaneous biliary intervention after ERCP. A total of 1028 patients with a median age of 56 years were included. ERCP was done emergently (19%), urgently (30%), and expectantly (51%). Endotherapy procedures were sphincterotomy (24%), biliary stent (24%), and combination (52%). Post-ERCP AEs were 11%, 10%, and 9% for emergent, urgent, and expectant ERCP, respectively (P= .577). In-hospital mortality showed a U-shape trend of 5%, 0%, and 2% for emergent, urgent, and expectant ERCP, respectively (P< .001). Combination and stent monotherapy had lower failure rates of 3% and 4%, respectively as compared with sphincterotomy monotherapy with failure rate of 11% (P< .001). When multivariate analysis was used, both combination (odds ratio, .2; 95% confidence interval, .1-.5) and stent monotherapy (odds ratio, .4; 95% confidence interval, .2-.9) were less likely to fail as compared with sphincterotomy monotherapy. There were no statistically significant differences between combination therapy and stent monotherapy in the univariate and the multivariate analyses. Although limited by retrospective design and the possibility of selection bias, this analysis suggests that the timing of ERCP is not a significant predictor of post-ERCP AEs after BDLs. Furthermore, combination or stent monotherapy had lower failure rates as compared with sphincterotomy monotherapy.