PGI56 EVALUATION OF MORTALITY ASSOCIATED WITH PROTON PUMP INHIBITORS AND H2 BLOCKERS: A REAL WORLD EVIDENCE STUDY

Abstract

The aim of this analysis was to estimate all-cause mortality among patients taking proton pump inhibitors (PPIs) and histamine H2 antagonists (H2 blockers). Patients identified through a federated network of electronic medical records were required to have taken PPIs or H2 blockers. Among these two cohorts, patients were required to have two records of these treatments recorded in their medical history at least three months apart. Patients treated with PPIs were matched 1-to-1 with patients treated with H2 blockers, using a greedy-nearest-neighbor algorithm. The risk of mortality was measured in the one year, five years, and ten years following the index treatment. All criteria were defined using ICD9/10, CPT, and RxNorm terminology. Kaplan-Meier curves and risk ratios (95% CI) were used to compare groups. The mean age was 54.4 ± 18.8 (N=1,455,638) and 46.6 ± 22.7 (N=496,538) among PPI- and H2 blocker-treated patients. In the matched analysis (N=448,446), H2 blocker patients were 0.089 (0.778,0.841), 0.8 (0.785,0.815), and 0.799 (0.786,0.812) more likely to die than PPI patients in the one year, five years, and ten years following treatment. The survival probability was 0.25%, 1.34%, and 2.35% higher in PPI-treated patients than in H2-treated patients in the one year, five years, and ten years following treatment (all with p<0.05). Taking H2 blockers is associated with a small increased risk of mortality. This burden is observed in patients with and without an indication for PPI or H2 blocker use. PPIs were not associated with an increased risk of death, in contrast to other studies. This difference may be due to a broader and less selective patient population, or use of different controls.