The Association Between Proton Pump Inhibitors and COVID-19 is Confounded by Hyperglycemia in a Population-Based Study

Abstract

Background and Aims: There is conflicting evidence regarding the association between proton pump inhibitors (PPI) and the risk of acquisition and severity of acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection.Aim: To evaluate the association between PPI exposure and infection and development of severe disease in patients infected with SARS-CoV2in a large population-based historical cohort.Methods: Data were extracted from a health maintenance organization database in Israel that insures over 1,200,000 individuals from across the country. All patients who underwent SARS-CoV-2 testing between March and November 2020 were included. Logistic regression and matched analyses were used to compare patients prescribed and exposed to PPIs to those not prescribed PPIs regarding SARS-CoV-2 positivity. In addition, among SARS-CoV-2 positive patients (n = 44,397) the likelihood of developing severe disease, defined by a composite endpoint of death, ICU admission and prolonged hospitalization, was compared in those exposed and not exposed to PPIs.Results: Among 255,355 adult patients who underwent SARS-CoV-2 testing by PCR, 44,397 (17.4%) were positive for SARS-CoV-2 and 12,066 (4.7%) patients were prescribed PPIs in the 3 months before testing. In a multivariable logistic regression model controlling for age, gender, smoking status, BMI, diabetes mellitus, hypertension, COPD, history of ischemic heart disease and fasting blood glucose (FBG) levels, no significant association was found between PPIs and SARS-CoV-2 positivity (p = 0.09 aOR 0.94, 95% CI – 0.88–1.01). Among SARS-CoV-2 positive patients, 910 (2%) had a severe infection. Multivariate logistic regression controlling for the abovementioned confounders, showed no such association between PPIs and severe COVID-19 (p = 0.28). Elevated FBG levels were significantly associated with both PPI exposure (p < 0.001) and severe COVID-19 infection (p < 0.001). These results were reinforced by a matched analysis (n = 655 pairs).Conclusion: PPIs are spuriously associated with severe COVID-19 due to the presence of elevated FBG as a confounder. Our study accounted for the FBG levels of patients and known risk factors for severe COVID-19 infection, which may be the reason for the discrepancy in prior studies. These results may aid in understanding potential confounders when evaluating potential associations of PPIs with other respiratory or viral diseases.