A cohort study of acute pancreatitis in relation to exenatide use

Methods: This cohort study included patients without claims for prior pancreatic disease who initiated exenatide or other antihyperglycaemic drugs between June 2005 and December 2007. Acute pancreatitis was identified with diagnosis codes and confirmed through review of blinded medical records. Poisson regression models provided estimates of rate ratios (RRs) and 95% confidence intervals (CIs) comparing the rate of acute pancreatitis during periods of current (days supplied + 31 days), recent (current definition + 31 days) and past use (>= 32 days beyond current definition) of exenatide relative to other antihyperglycaemic drugs, adjusted for propensity scores. A prespecified nested case-control analysis provided RR estimates adjusted for patient characteristics abstracted from medical records. Results: Initiators of exenatide (N = 25719) had more baseline claims for obesity and concomitant diabetes drugs than comparators (N = 234536). There were 40 confirmed cases of acute pancreatitis in the exenatide cohort and 254 among other antihyperglycaemic drug initiators. Compared to other antihyperglycaemic drugs, the propensity score-adjusted RR for exenatide was 0.5 (95% CI 0.2-0.9) for current use, 1.1 (95% CI 0.4-3.2) for recent use and 2.8 (95% CI 1.6-4.7) for past use. The case-control analysis resulted in a RR of 0.2 for current use (95% CI 0.0-1.4) and 0.1 for recent use (95% CI 0.0-1.3), but an attenuated RR in the past use association (RR 1.1; 95% CI 0.1-11.0). Conclusions: Exenatide use was not associated with an increased risk of acute pancreatitis.