People exposed to proton-pump inhibitors shortly preceding COVID-19 diagnosis are not at an increased risk of subsequent hospitalizations and mortality: A nationwide matched cohort study

Abstract

Aims: To assess whether exposure to proton-pump inhibitors (PPIs) shortly preceding COVID-19 diagnosis affected the risk of subsequent hospitalizations and mortality. Methods: This population-based study embraced first COVID-19 episodes in adults diagnosed up to 15 August 2021 in Croatia. Patients were classified based on exposure to PPIs and burden of PPI-requiring morbidities as nonusers (no issued prescriptions, no recorded treatment-requiring conditions between 1 January 2019 and COVID-19 diagnosis), possible users (no issued prescriptions, but morbidities present; self-medication possible) and users (≥1 prescription within 3 months prior to the COVID-19 diagnosis; morbidities present). Subsets were mutually exactly matched for pre-COVID-19 characteristics. The contrast between users and possible users informed about the effect of PPIs that is separate of the effect of PPI-requiring conditions. Results: Among 433 609 patients, users and possible users were matched 41 195 (of 55 098) to 17 334 (of 18 170) in the primary and 33 272 to 16 434 in the sensitivity analysis. There was no relevant difference between them regarding mortality (primary: relative risk [RR] = 0.93 [95% confidence interval 0.85–1.02; absolute risk difference [RD] = −0.34% [−0.73, 0.03]; sensitivity: RR = 0.88 [0.78–0.98]; RD = −0.45% [−0.80, −0.11]) or hospitalizations (primary: RR = 1.04 [0.97–1.13]; RD = 0.29% [−0.16, 0.73]; sensitivity: RR = 1.05 [0.97–1.15]; RD = 0.32% [−0.12, 0.75]). The risks of both were slightly higher in possible users or users than in nonusers (absolutely by ~0.4–1.6%) indicating the effect of PPI-requiring morbidities. Conclusion: Premorbid exposure to PPIs does not affect the risk of death or hospitalization in adult COVID-19 patients, but PPI-requiring morbidities seemingly slightly increase the risk of both.