Abstract
BACKGROUND: Associations of diabetes (DM) types with severe COVID have not adequately been investigated in clinical databases. METHODS: Analytical cohort of n=5713 patients with diabetes was selected from the electronic medical record-based COVID Registry in the urban academic health care system from March 2020 to March 2021. All cases in the registry tested positive for SARS-CoV-2 infection on the clinical assessments. Severe COVID was defined in cases requiring hospitalization or if fatal outcome. Demographics variables, clinic visits, vaccination status, multi-morbidity, ambulatory medication data, biochemical markers, and other diabetes characteristics were collected on or before the date of COVID diagnosis. After excluding pregnancy-related diabetes, cases were classified into type 1 (DM1), type 2 (DM2) and other/hybrid DM based on disease coding and clinical variables. Univariable and multivariable logistic regression models were used to assess the association of the three diabetes types with severe COVID after adjustment for age, sex, race/ethnicity. FindingsDiverse cohort included 37.6% Latino, 37.5% Black, 19.5% White, and 5.3% other; 53.7% were women. DM1 was identified in n=141 cases, Type 2 diabetes in n=2749, and other/hybrid diabetes in n=2823. The diabetes subtypes could not be reliably explored due to limitations of the clinical database. Compared to DM2, the odds of severe COVID were similar in DM1 (1.11, 95%CI 0.65, 1.81, p=0.700), but higher in other/hybrid DM (1.53, 95%CI 1.34, 1.76), after adjustment for sex, age, and race/ethnicity. SUMMARY: Other and hybrid diabetes types are not rare and are associated with a higher risk of severe COVID-19. The diabetes subtypes could not be reliably explored due to limitations of the clinical database. Presentation: No date and time listed