Abstract
OBJECTIVES: To determine whether diabetes mellitus increases depressive symptoms in older Latinos in a population-based cohort. DESIGN: Prospective cohort study. PARTICIPANTS: Individuals from the Sacramento Latino Study on Aging aged 60 and older in 1998-99 and followed annually until 2008 (N = 1,586). MEASUREMENTS: Diabetes mellitus was defined according to self-report, fasting blood glucose of 126 mg/dL or greater, glycosylated hemoglobin of 6.5% or greater, or diabetic medication use. Depressive symptoms were defined as Center for Epidemiologic Studies Depression Scale (CES-D) score of 16 or greater or use of antidepressant medication. Multistate Markov modeling was used to assess the effects of time-dependent diabetes mellitus on transitions between three states over time: low CES-D score (normal), high CES-D score or treated (depressed), and death. Bivariate analyses identified covariates significantly associated with any transition, including sex and baseline measures of age, education, body mass index, hypertension, and stroke. RESULTS: In a fully adjusted model, participants with diabetes mellitus had a 35% higher rate of developing depressive symptoms or starting treatment with an antidepressant (hazard ratio (HR) = 1.35, 95% confidence interval (CI) = 1.13-1.62). Time-dependent diabetes mellitus was associated with a lower rate of regression from depressed to normal (HR = 0.72, 95% CI = 0.59-0.88) and a 2.3 greater rate pf progression from depressed to death (HR = 2.31, 95% CI = 1.57-3.40). CONCLUSION: Diabetes mellitus increased the risk of developing depressive symptoms in older Mexican Americans. Older Latinos with diabetes mellitus should be screened for depressive symptoms and prioritized for close follow-up, potentially through greater reliance on team-based models of care.