LAUSR

OBJECTIVE To examine associations of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection/coronavirus disease 2019 with incident diabetes. RESEARCH DESIGN AND METHODS We conducted a retrospective cohort study using Veterans Health Administration data. We defined all patients without preexisting diabetes with one or more nasal swabs positive for SARS-CoV-2 (1 March 2020–10 March 2021; n = 126,710) as exposed and those with no positive swab and one or more laboratory tests (1 March 2020–31 March 2021; n = 2,651,058) as unexposed. The index date for patients exposed was the date of first positive swab and for patients unexposed a random date during the month of the qualifying laboratory test. We fit sex-stratified logistic regression models examining associations of SARS-CoV-2 with incident diabetes within 120 days and all follow-up time through 1 June 2021. A subgroup analysis was performed among hospitalized subjects only to help equalize laboratory surveillance. RESULTS SARS-CoV-2 was associated with higher risk of incident diabetes, compared with no positive tests, among men (120 days, odds ratio [OR] 2.56 [95% CI 2.32–2.83]; all time, 1.95 [1.80–2.12]) but not women (120 days, 1.21 [0.88–1.68]; all time, 1.04 [0.82–1.31]). Among hospitalized participants, SARS-CoV-2 was associated with higher risk of diabetes at 120 days and at the end of follow-up in men (OR 1.42 [95% CI 1.22–1.65] and 1.32 [1.16–1.50], respectively) but not women (0.72 [0.34–1.52] and 0.80 [0.44–1.45]). Sex ∗ SARS-CoV-2 interaction P values were all <0.1. CONCLUSIONS SARS-CoV-2 is associated with higher risk of incident diabetes in men but not in women even after greater surveillance related to hospitalization is accounted for.