Objective: To examine the association between diabetes and the prevalence of depression in different sex and age groups by analysing the cross-sectional data from the National. Population Health Survey, conducted in Canada in 1996-1997. Study design: A total of 53 072 people aged 20-64 years were included in the analysis. Depression was defined as depression scale >= 5, based on the Composite International Diagnostic Interview Short-Form (CIDI-SF). Respondents who answered the following question affirmatively were considered to have diabetes: 'do you have diabetes diagnosed by a health professional?' Methods: A multiple logistic regression model was used to adjust for potential confounding effects, and a bootstrap procedure was used to take sampling weights and design effects into account. Results: The prevalence of diabetes was much higher in people aged 40-64 years than in people aged 20-39 years (men: 4.7% vs. 0.5%; women: 3.5% vs. 0.8%, respectively). In contrast, people aged 20-39 years had a slightly higher prevalence of depression than those aged 40-64 years (men: 3.1% vs. 2.9%; women: 6.6% vs. 5.4%, respectively). Diabetes was significantly associated with depression in women aged 20-39 years (odds ratio [OR] = 2.52, 95% confidence interval [CI] = 1.19, 5.32), but not in women aged 40-64 years (OR = 1.62, and 95% CI = 0.65, 4.06). The association was not significant in both age groups in men, but it tended to be stronger in the younger age group. Conclusions: The data suggest that diabetes is significantly associated with depression, particularly in young adults. (c) 2006 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved.