Incident diabetes within the first two years after SARS-CoV-2 infection: a population-based retrospective cohort study of the Agency for Health Protection of Milan, Italy

Collaborative publication as member of the PASCNET study group.

[Download Paper]

Background Postacute sequelae of SARS-CoV-2 infection (PASC), including persistent symptoms and acute and chronic diagnoses, have become a major research focus. Diabetes mellitus, beyond its established link to COVID-19 severity, is increasingly recognized as a potential long-term outcome. This study investigated the association between SARS-CoV-2 infection and incident diabetes using population-level health administrative data (HAD) from the Agency for Health Protection of Milan, where the epicenter of the pandemic in Italy took place.

Methods This retrospective cohort study included adult residents without a history of diabetes who underwent SARS-CoV-2 testing between 1 March and 31 December 2020. Test-positive individuals were matched 1:1 to test-negative individuals based on sex, age, and testing week. The cohort was followed through 31 December 2021. The incidence of diabetes, identified using an HAD-based case-detection algorithm, was compared between the two groups, and in stratified analyses by sex and age, using weighted Cox models adjusted for chronic comorbidities, area-level deprivation, influenza and pneumococcal vaccinations. Weights were calculated via the inverse probability weighting approach. Effect estimates are presented as hazard ratios (HRs).

Results Our final cohort included 248,176 residents (124,026 test-negative, 124,150 test-positive). Over a median follow-up time of 415 days, 739 positive (0.60%) and 657 negative (0.53%) individuals were newly identified with diabetes. The incidence among positive individuals was 572.82 per 100,000 person-years (CI 531.52–614.12), and that among negative individuals was 509.50 per 100,000 person-years (CI 470.54–548.46). The overall HR was 1.13 (CI 1.02–1.25). In stratified analyses, this effect was prominent in women aged 41–60 years (HR 1.31; CI 1.02–1.68).

Conclusions This study provides population-based evidence supporting an association between SARS-CoV-2 infection and newly detected diabetes. These findings contribute to understanding the long-term health impact of COVID-19 and may inform public health strategies for PASC prevention and management.