Annika Parmar
Introduction
Endoscopic Retrograde Cholangiopancreatography (ERCP) is a key therapeutic procedure in gallstone disease. Outpatient ERCP reduces inpatient pressures, but patient selection remains challenging. No validated criteria currently guide suitability for outpatient ERCP. This study aimed to identify factors associated with inpatient versus outpatient ERCP and to determine characteristics predictive of complications.
Methods
A retrospective cohort study included demographics, comorbidities, Clinical Frailty Scale (CFS), ASA classification, presenting diagnosis, and initial and peak bilirubin. Logistic regression assessed factors associated with inpatient ERCP; variables with p<0.10 on univariate analysis were included in multivariate models. ERCP-related complications were recorded, and univariate and multivariate regression analyses identified predictors of complications. Receiver Operating Characteristic (ROC) curves evaluated predictive performance, reported as area under the curve (AUC).
Results
Over one year, 112 patients met inclusion criteria. On univariate analysis, elevated bilirubin levels, ASA, and ascending cholangitis were significantly associated with inpatient ERCP (p<0.05). Following multivariate adjustment, no factors remained statistically significant (all p>0.05), indicating no independent predictors for the choice of inpatient versus outpatient ERCP. Complications occurred more frequently in outpatients. Multivariate analysis identified age, CFS, ASA, and initial bilirubin as independent predictors of complications. ROC analysis demonstrated limited discriminative performance, with AUC ≤0.65; age and initial bilirubin had the highest predictive accuracy.
Conclusions
No independent clinical variables predicted the choice of inpatient versus outpatient ERCP. Age, frailty, ASA, and initial bilirubin were associated with post-ERCP complications. Higher complications in outpatients highlight the value of structured patient-selection criteria. Larger prospective studies are needed to
validate these findings.
Authors
Annika Parmar, Marium Quaunine, Rebecca Bradley, William King, Karis Harbertson, Reeza Khan
Queen Elizabeth Hospital, London, United Kingdom