Incidence and Predictors of Bleeding During and Following ERCP

While very effective, endoscopic retrograde cholangiopancreatography (ERCP) is widely known to have the highest adverse event (AE) profile among all commonly performed endoscopic procedures, with a collective AE rate of >10%. Common AEs include post-ERCP pancreatitis, bleeding, cholangitis, cholecystitis, perforation, and cardiopulmonary events. The incidence of bleeding during ERCP and following ERCP has been estimated using retrospective sources, but granular predictors of bleeding remain unknown, including the use of direct-acting anticoagulants and discontinuation and resumption patterns surrounding their use.

It is of critical priority to patients, practitioners, and health administrators to investigate factors associated with all AEs and unplanned healthcare encounters (UHEs) following ERCP, especially given that most ERCPs are performed on an outpatient basis. The per-admission costs of post-ERCP UHEs are substantial. Thus, researchers must prioritize the study of ERCP outcomes, striving to both identify and modify factors leading to AEs and UHEs.

(2) Research Question and Objectives In this study, we will aim to assess the incidence and predictors of intra-procedural bleeding during ERCP, and clinically significant post-procedural bleeding following ERCP.

(3) Methods Design: This is a multicenter prospective cohort study. The primary exposure of interest will be patient use of antithrombotic medications including antiplatelet agents and/or anticoagulant agents. In addition to these variables, other parameters we will assess include: the presence and timing of pharmacologic pancreatitis prophylaxis, extent and timing of trainee involvement, the number and timing of common bile duct (CBD) cannulation attempts, the depth, timing, trajectory and number of pancreatic duct (PD) cannulation(s), the presence and extent of PD opacification, the size(s) of sphincterotomy and/or sphincteroplasty, intra-procedural pathology, and the composition, caliber and length of any PD or CBD stent(s).

Outcomes: The primary outcomes will be clinically significant post-ERCP bleeding (CSPEB), using established definitions, and intra-procedural bleeding, defined as bleeding requiring endoscopic management during ERCP. Secondary outcomes (defined a priori) will include bleeding severity, overall and specific AEs (pancreatitis, cholangitis, cardio-pulmonary events), cannulation time and success rate, as well as overall procedure time and success rate.

Sample Size and Power: Using anticipated CSPEB rates of 2.0% for non-anticoagulant users and 5.0% for anticoagulant users, a minimum of 588 patients in each arm will be required to demonstrate this difference with 80% power and alpha of 0.05.

Statistical Analysis: Variables will be compared using Student's t-test for measured variables and chi-squared test for categorical variables. P values < 0.05 will be considered significant. We will use multivariable logistic regression to assess associations between risk factors and having bleeding versus not having bleeding. Clinically relevant subgroup analyses will also be performed by relevant patient-, endoscopist-, and procedure-related characteristics. Odds ratios per outcome will be reported with 95% CIs.