Diabetes incidence associated with depression and antidepressants in the Melbourne Longitudinal Studies on Healthy Ageing (MELSHA)

Objective: Diabetes may be associated with depression and antidepressant medication (ADM) use, but published findings remain equivocal. The authors' aimed to determine the risk of diabetes incidence associated with baseline depression exposures (symptoms and/or ADM use). Methods: A prospective cohort study was conducted in a regionally representative sample of noninstitutionalised older Australian people (N=1000, aged 65+year), who were followed up biennially between 1994 and 2004 (attrition was ≈24%). Analyses excluded participants for prevalent diabetes at baseline, determined by self-report or specific medications. Diabetes incidence was ascertained by first self-report at any follow-up wave. Depression exposures (baseline predictors) were defined by the Psychogeriatric Assessment Scales (PAS) depression scale and ADM use, and classified as: (1) 'symptomatic' (PAS score 5+); (2) 'ADM use'; (3) 'symptomatic or ADM use'; (4) 'symptomatic and no ADM use'; (5) 'asymptomatic (PAS score <5) and ADM use' and (6) 'symptomatic and ADM use'. Covariates were demographic, lifestyle, functional health and chronic disease factors. Cox regressions were used to determined hazard ratios with 95% confidence intervals (HR [95% CI]) for diabetes incidence according to depression exposures, adjusted for significant covariates. Results: Baseline response rate was 70.3%. Depression predictors of diabetes incidence were 'symptomatic' (2.29 [1.28,4.10]), 'symptomatic or ADM use' (2.13 [1.32,3.44]) and 'symptomatic and no ADM use' (2.38 [1.28,4.45]), after adjustment for significant covariates. Being asymptomatic was not a protective factor among those prescribed antidepressants. Conclusions: Older people with depressive symptoms are at least twice more likely to develop diabetes than those without depressive symptoms, regardless of antidepressants.