OBJECTIVE

to determine incidence and predictors of late-life depression.

METHODS

this is a 3-year observational cohort study of 3,214 non-demented patients aged 75 and over completing three waves of assessment. The patients were recruited in 138 primary care practices in six urban areas in Germany. Depressive symptoms were measured at baseline, and 18 months and 36 months later using the GDS-15 Geriatric Depression Scale with a cut-off 0-5/6-15. Cox proportional hazard regression models were applied to examine predictors of incident depression, adjusting for sex, age, education, living situation, activities of daily living - and instrumental activities of daily living impairment, somatic comorbidity, alcohol consumption, smoking, mild cognitive impairment and apoE4 status.

RESULTS

the incidence of depression was 36.8 (95% CI: 29.6-45.3) per 1,000 person-years in men and 46.0 (95% CI: 39.9-52.8) in women (sex difference P = 0.069). The incidence increased from 35.4 (95% CI: 29.7-41.9) per 1000 person-years between the ages of 75 and 79 to 75.2 (95% CI: 53.2-103.2) for subjects 85 years and older. After full adjustment for confounding variables, hazard ratios (HR) for incident depression were significantly higher for subjects 85 years and older (HR: 1.83, 95% CI: 1.24-2.70) and those with mobility impairment (HR: 2.53, 95% CI: 1.97-3.25), vision impairment (HR: 1.41, 95% CI: 1.04-1.91), mild cognitive impairment (HR: 1.52, 95% CI: 1.10-2.10), subjective memory impairment (HR: 1.33, 95% CI: 1.01-1.74) and current smoking (HR: 1.69, 95% CI: 1.13-2.53).

CONCLUSIONS

the incidence of depression increased significantly with age. In designing prevention programmes, it is important to call more attention on functional impairment, cognitive impairment and smoking.