Abstract
AIMS: To examine associations of SARS-CoV-2 infection/COVID-19 with insulin treatment in new-onset diabetes. METHODS: We conducted a retrospective cohort study using Veterans Health Administration data (March 1, 2020-June 1, 2022). Individuals with ≥1 positive nasal swab for SARS-CoV-2 (n = 6,706) comprised the exposed group, and individuals with no positive swab and ≥1 laboratory test of any type (n = 20,518) the unexposed group. For exposed, the index date was the date of first positive swab, and for unexposed a random date during the month of the qualifying laboratory test. Among Veterans with new-onset diabetes after the index date, we modeled associations of SARS-CoV-2 with most recent A1c prior to insulin treatment or end of follow-up and receipt of >1 outpatient insulin prescription starting within 120 days. RESULTS: SARS-CoV-2 was associated with a 40% higher odds of insulin treatment compared to no positive test (95%CI 1.2-1.8) but not with most recent A1c (ß 0.00, 95%CI -0.04-0.04). Among Veterans with SARS-CoV-2, ≥2 vaccine doses prior to the index date was marginally associated with lower odds of insulin treatment (OR 0.6, 95%CI 0.3-1.0). CONCLUSIONS: SARS-CoV-2 is associated with higher odds of insulin treatment but not with higher A1c. Vaccination may be protective.