Electronic frailty index and outcomes after first ERCP in older adults: a retrospective cohort study

The electronic frailty index (eFI) summarizes health deficits recorded in electronic health records. Its association with outcomes after endoscopic retrograde cholangiopancreatography (ERCP) warrants evaluation in defined clinical populations. This retrospective single-center cohort included 645 adults aged ≥65 years undergoing first ERCP for choledocholithiasis, classified as non-frail (n = 466) or frail (n = 179) using an eFI cutoff of ≥0.15. Five perioperative outcomes were assessed using multivariable logistic regression, with procedural adjustment in 600 complete cases and multiple imputation for missing procedure duration. ROC analyses compared binary frailty status with CCI. Kaplan–Meier and Cox analyses evaluated 365-day mortality after ERCP in all 645 patients and 30-day readmission after discharge in 636 discharge survivors. Frailty was associated with postoperative infection (adjusted OR 2.97, 95% CI 1.47–6.02) and intraoperative hypoxemia (3.27, 1.38–7.70). The PEP association attenuated after procedural adjustment (OR 1.83, 95% CI 0.97–3.44), and the pain estimate was sensitive to missing-data handling. Discrimination was modest and varied by outcome. There were 36 deaths within 365 days after ERCP and 61 readmissions within 30 days after discharge. After adjustment for age, sex, and CCI, frailty was associated with readmission (HR 2.02, 95% CI 1.10–3.68; P = 0.022), whereas an independent association with mortality was not established (HR 2.02, 95% CI 0.92–4.44; P = 0.079). Among hospitalized adults aged ≥65 years undergoing first ERCP for choledocholithiasis, frailty was associated with infection, hypoxemia, and 30-day readmission. Associations with PEP, pain, and mortality remained uncertain. Incremental predictive value beyond CCI was not established.