Diabetes is associated with severe COVID-19 and mortality . The aim of the present study was to determine the effect of age on the association between diabetes and mortality in patients with laboratory-confirmed COVID-19 in Mexico . This retrospective cohort study involved patients aged 20 years or older with symptoms of viral respiratory disease who were screened for SARS-CoV-2 infection across the System of Epidemiological Surveillance of Viral Respiratory Disease in Mexico from January 1 through November 4 , 2020 . Cox proportional-hazard regression was used to calculate the hazard ratio for 28-day mortality and its 95% confidence interval (CI). Among 757,210 patients with COVID-19 (outpatients and inpatients), 120,476 (16 %) had diabetes and 80,616 died . Among 878,840 patients without COVID-19 (those who tested negative for SARS-CoV-2 infection), 88,235 (10.0 %) had diabetes and 20,134 died . Among patients with COVID-19, diabetes was associated with a hazard ratio for death of 1.49 (95% CI 1.47-1.52), adjusting for age, sex, smoking habit, obesity, hypertension, immunodeficiency, and cardiovascular, pulmonary, and chronic renal disease . The strength of the association decreased with age (trend test: P = 0.004). For example, the adjusted hazard ratio for death was 3.12 (95% CI 2.86-3.40) for patients 20-39 years of age; in contrast, the adjusted hazard ratio of death for patients 80 years of age or older was 1.11 (95% CI 1.06-1.16). The adjusted hazard ratios were 1.66 (95% CI 1.58-1.74) in outpatients and 1.14 (95% CI 1.12-1.16) in inpatients . In hospitalized patients 80 years of age or older, no association was observed between diabetes and COVID-19-related mortality (adjusted hazard ratio : 1.03; 95% CI 0.98-1.08). Among patients without COVID-19, the adjusted hazard ratio for death was 1.78 (95% CI 1.73-1.84). In conclusion, in adult patients with COVID-19 in Mexico, the risk of death associated with diabetes decreased with age . No association between diabetes and mortality was observed among inpatients 80 years of age or older . Our findings should be verified in other populations.