Erratum. The Incidence of Diabetes Among 2,777,768 Veterans With and Without Recent SARS-CoV-2 Infection. Diabetes Care 2022;45:782-788.

For the article cited above, an issue in the way the outcome (diabetes status) was coded was identified during a follow-up analysis. Participants with ICD-10 codes for diabetes at any time were excluded from the analysis, and incident diabetes was identified based only on laboratory values and glucose-lowering medications. This resulted in a substantial underestimate of the number of cases of incident diabetes in the cohort. The coding issue was corrected, the analysis was repeated, and the corrected code was validated by an independent analyst. After these steps, in agreement with the original article, there remained a statistically significant and positive association of SARS-CoV-2 infection with incident diabetes in the corrected analysis. The largest differences between the original and corrected analysis were seen in the analyses of all male participants in whom the association of SARS-CoV-2 infection with incident diabetes was attenuated in the corrected analysis compared with the original results (odds ratio 120 days: 2.56 [2.32-2.83] original vs. 1.75 [1.63-1.88] corrected; odds ratio all-time: 1.95 [1.80-2.12] original vs. 1.44 [1.36-1.52] corrected). For hospitalized male participants, the differences were smaller. In agreement with the original article, there was no association of SARS-CoV-2 infection with incident diabetes in women in the corrected analysis. Finally, in the corrected models, the P values for the sex * SARS-CoV-2 infection interaction terms were statistically significant except for participants hospitalized in the first 30 days (all available follow-up time). The authors apologize for the error. The online version of the article (https://doi.org/10.2337/dc21-1686) has been updated with the corrected data.