Introduction and Objective: Studies assessing diabetes mellitus (DM) risk after COVID-19 infection have typically examined narrow periods of the pandemic or did not incorporate advanced causal inference methods. This study assessed the association of COVID-19 infection with incident DM 1-year post-infection, stratified by viral variant periods, among adults aged 20-80y.Methods: We analyzed electronic health record data on 926,385 patients (1,124,735 patient-variant periods) from 34 health systems participating in PCORnet, a national research network. Patients were established in the system at baseline and did not have preexisting DM (defined as A1c ≥6.5; DM diagnosis or medication). COVID-19 variant periods were: Early pandemic (Apr-Sept 2020), Winter 2020/21 (Oct 2020-Feb 2021), Alpha Variant (Mar-Jun 2021), Delta (Jul-Dec 2021), and Omicron (Jan-Dec 2022). Exposure was a positive vs. negative COVID-19 test in an emergency department or inpatient setting. Outcome was a new diagnosis of DM 30-360 days post index COVID-19 test. We used inverse probability (IP) weighting to adjust for confounding (demographics, underlying conditions, baseline and time-varying medications) and informative censoring (for a positive test post-index). IP weighted estimates of the cause-specific cumulative DM incidence by 12 months under COVID-19 positivity versus negativity were calculated via IP weighted estimates of the cause-specific hazards for DM and death.Results: The mean (SD) age was 46.2y (17.1), and 116,205 (10%) tested positive vs. 1,008,530 (90%) negative for COVID-19 on the index date. Positivity vs. negativity was not associated with 1-year DM incidence across the first 4 variant periods (absolute difference in incidence 0.11%-0.3%). Positivity was associated with higher DM incidence in the Omicron variant period (2.92% vs. 2.31%; absolute difference positive vs. negative 0.61%, 95% CI 0.44, 0.80).Conclusion: In a hospital/ED cohort, diabetes incidence was higher among those with vs. without COVID during the Omicron variant period.
S.L. Rifas-Shiman: None. J. Young: None. S. McGrath: None. J.S. Schildcrout: None. N. De Abrew Rajapakse: None. M. Hivert: None. J. Lyons: None. A.C. Powers: None. B. Rasouli: None. R.L. Rothman: Research Support; Cardiohealth Alliance. J.P. Block: None.
National Institutes of Health, NIDDK (U01DK137533)
Source link
Leave a Reply