Long-term prognosis and risk factors associated with post-ERCP pneumobilia in patients with common bile duct stones

Post-ERCP pneumobilia is not uncommon; however, studies focusing on the long-term prognosis of patients with post-ERCP pneumobilia are limited. This study aimed to explore long-term prognosis and risk factors associated with post-ERCP pneumobilia in patients with common bile duct stones (CBDSs). We conducted a retrospective analysis of 1380 patients who underwent ERCP for CBDSs at our hospital from January 2010 to December 2017. Patients were selected based on inclusion and exclusion criteria and divided into pneumobilia and nonpneumobilia groups, followed by propensity score matching. The matched groups were then compared in terms of incidence rates of both single and multiple recurrences of CBDSs, acute cholangitis, and acute cholecystitis. Multivariate logistic regression analysis was used to explore risk factors associated with pneumobilia. After propensity matching, there was no significant difference in the rate of single recurrence of CBDSs (22.5% vs 30%; P = .446) between the pneumobilia and nonpneumobilia groups. However, the incidences of multiple recurrences of CBDSs (32.5% vs 12.5%; P = .032) and acute cholangitis without stone recurrence (32.5% vs 2.5%; P = 1 cm) (odds ratio [OR], 2.48; 95% confidence interval [CI], 1.03-3.76; P = .043), endoscopic sphincterotomy with moderate incision (OR, 5.38; 95% CI, 1.14-25.47; P = .034) and with large incision (OR, 8.7; 95% CI, 1.83-41.46; P = .007) were identified as independent risk factors for pneumobilia after initial ERCP. Patients with post-ERCP pneumobilia have increased risk of multiple recurrences of CBDSs and acute cholangitis without stone recurrence. Independent risk factors for pneumobilia include peripapillary diverticulum, a dilated CBD (>1 cm), and endoscopic sphincterotomy with moderate and large incisions. A normal-sized CBD appears to serve as a secondary barrier against enterobiliary reflux, necessitating further research for confirmation.