Details
The U.S. Food and Drug Administration (FDA) initiated this study in the Sentinel System to assess safety signals that emerged from published reports suggesting a potential association between proton pump inhibitor (PPI) use and risk of severe COVID-19. To determine whether use of PPIs increases the risk of COVID-19 hospitalization and severe COVID-19 outcomes, the FDA leveraged the Sentinel System, which offers large sample sizes and advanced methods for adjusting for confounding factors.
An initial comparative safety analysis using data from April to December 2020 produced conflicting results across data sources: an increased risk of COVID-19 hospitalization was observed with PPI use compared to famotidine use among individuals with commercial insurance (hazard ratio 1.55, 95% confidence interval 1.11, 2.18), while a protective effect was observed among individuals with Medicare (hazard ratio 0.83, 95% confidence interval 0.72, 0.95). To explore this inconsistency, the FDA conducted a follow-up analysis using more comprehensive propensity score (PS) methods for confounding adjustment (i.e., including additional covariates in the user-specified PS model and applying a high-dimensional PS model) and extended the study period through January 2024 for commercial data and June 2023 for Medicare data. This follow-up analysis found no significant increase in the risk of COVID-19 hospitalization or severe COVID-19 outcomes with PPI use compared to famotidine use in either population. Hazard ratios for COVID-19 hospitalization ranged from 0.96-1.02 (all 95% confidence intervals included 1.0) in commercial data and 0.93-0.99 (95% confidence intervals included 1.0 or the upper bound was slightly less than 1.0) in Medicare data, across various propensity score methods. For severe COVID-19 outcomes, hazard ratios ranged from 1.02-1.11 in commercial data and 0.93-1.00 in Medicare data across various PS methods, with all confidence intervals including 1.0.
Findings from this cohort study in the Sentinel Distributed Database aligned with results from an FDA-led case-control study conducted using Medicare data, strengthening FDA’s conclusion that there is no association between PPI use and the risk of COVID-19 hospitalization. Since the follow-up analysis employed more extensive control for confounding and produced results that were robust across analytic methods and data sources, the initial conflicting results are due to inadequate adjustment for confounding. Based on the study findings, FDA determined that no regulatory action was required.