Association Between Post-ERCP Pancreatitis and New-Onset Diabetes Mellitus: A Retrospective Cohort Study

Background/Objectives: Acute pancreatitis is increasingly recognized as a risk factor for disturbances in glucose metabolism and the development of diabetes mellitus (DM). However, the long-term endocrine consequences of post-endoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP) remain poorly characterized. This study aimed to evaluate the association between post-ERCP pancreatitis and the risk of new-onset diabetes mellitus (NODM). Methods: This retrospective cohort study included patients who underwent ERCP between 2019 and 2024 at a tertiary referral center. New-onset diabetes mellitus was defined using laboratory data and International Classification of Diseases (ICD) diagnostic codes within one year after ERCP. Multivariable logistic regression adjusting for age, sex, hypertension, and coronary artery disease was performed. Results: A total of 2695 patients were included. Post-ERCP pancreatitis occurred in 165 patients (6.1%). New-onset diabetes developed in 9/165 patients (5.5%) in the PEP group and in 27/2530 patients (1.1%) in the non-PEP group. An increased incidence of new-onset diabetes was observed among patients who developed post-ERCP pancreatitis (crude OR 5.35, 95% CI 2.47-11.57; p < 0.001). In multivariable analysis adjusting for age, sex, hypertension, and coronary artery disease, post-ERCP pancreatitis remained significantly associated with new-onset diabetes in the fully adjusted model (adjusted OR 5.33, 95% CI 2.42-11.77; p < 0.001). The absolute risk increase was 4.39%, corresponding to a number needed to harm of 23. Conclusions: An increased incidence of new-onset diabetes was observed among patients who developed post-ERCP pancreatitis. This association remained significant after adjustment for baseline cardiovascular comorbidities. Although the absolute risk increase was modest, these findings may be clinically relevant. Because this was a retrospective study with a limited number of diabetes cases, the findings should be considered hypothesis-generating and require confirmation in prospective studies.