Abstract
BACKGROUND: Observational studies have reported an association between metformin and improved outcomes in COVID-19, but most have been small and with significant limitations. OBJECTIVE: To evaluate the association between preadmission metformin exposure and mortality in patients with diabetes mellitus hospitalized with coronavirus disease 2019 (COVID-19) infection. DESIGN: Retrospective cohort analysis using electronic health records extracted from the American Heart Association COVID-19 Registry. PARTICIPANTS: Adults (n = 11,993) with diabetes mellitus but without chronic kidney disease (CKD) or need for hemodialysis who were hospitalized with COVID-19 between January 25, 2020, and February 9, 2022. MAIN MEASURES: We used propensity score modeling to address differences between metformin and non-metformin users prior to multivariable log-binomial models to examine the association between metformin use at time of hospital admission for COVID-19 infection and in-hospital death; composite of in-hospital death or discharge to hospice; composite of in-hospital death, discharge to hospice, or ICU admission; and composite of in-hospital death, discharge to hospice, ICU admission, or mechanical ventilation. KEY RESULTS: Compared to metformin non-use, pre-admission metformin use was associated with lower risk of in-hospital death (risk ratio (RR) 0.81 [95% CI 0.75-0.90]); composite of in-hospital death or discharge to hospice (RR 0.79 [95% CI 0.74-0.87]); composite of in-hospital death, discharge to hospice, or ICU admission (RR 0.90 [95% CI 0.86-0.95]); and composite of in-hospital death, discharge to hospice, ICU admission, or mechanical ventilation (RR 0.9 [95% CI 0.84-0.98]). Metformin use was also associated with lower risk of death due to respiratory cause (RR 0.86 [95% CI 0.74-0.97]) but not cardiovascular (RR 0.84 [95% CI 0.58-1.2]) or other (RR 0.78 [95% CI 0.60-1.0]) causes. CONCLUSIONS: Pre-admission metformin use was associated with lower risk of in-hospital mortality and markers of disease severity among adults with diabetes mellitus without CKD and not requiring hemodialysis who were hospitalized with COVID-19 infection.