Aspirin exposure and its association with metal stent patency in malignant distal biliary obstruction: a large international multicenter propensity score–matched study

Background and Aims Stent dysfunction is common following endoscopic retrograde cholangiopancreatography (ERCP) with self-expanding metal stent (SEMS) insertion in malignant distal biliary obstruction (MDBO). Chronic aspirin exposure (ASA-E) has been previously shown to potentially decrease this risk. We aim to further ascertain the protective effect of ASA and to identify other predictors of stent dysfunction. Method Multicenter retrospective cohort study (9 Canada and 1 US). MDBO who underwent ERCP-SEMS between 01/2014-12/2019 were included and divided into two cohorts: ASA-E and ASA unexposed (ASA-U). Propensity score-matching (PSM) was performed to limit selection bias. Matched variables included: age, sex, tumor stage, and type of metal stent. The primary outcome was the hazard rate of stent dysfunction. A multivariable Cox proportional hazards model was used to identify independent predictors of stent dysfunction. Results 1,396 patients were assessed. Following PSM a total of 496 patients were analyzed (248 ASA-E, 248 ASA-U). ERCP-SEMS was associated with high clinical success of 82.2% in ASA-E and 81.2% in ASA-U, p=0.80. A total of 184 patients had stent dysfunction wth a mean stent patency time of 229.9 ± 306.2 days and 245.4 ± 241.4 days in ASA-E and ASA-U, resepectively (p=0.52). On multivariable analysis, ASA-E did not protect against stent dysfunction with a hazard ratio of 1.25 (95% confidence interval, CI: 0.96; 1.63). Pancreatic cancer etiology (HR 1.36, 95% CI:1.15-1.61) predicted stent dysfunction while cancer therapy was protective (HR 0.73, 95% CI: 0.55; 0.96). The use of chronic ASA was not associated with an increased risk for adverse events including bleeding, post-ERCP pancreatitis, and perforation. Conclusion In this large, multicenter study utilizing propensity score-matching, chronic exposure to ASA did not protect against stent dysfunction in MDBO. Instead, the analysis revealed that the etiology of pancreatic cancer was an independent predictor of stent dysfunction while cancer therapy was protective.