Predictors of mortality in people with late-life depression: A retrospective cohort study

Objectives: Late-life depression (LLD) is associated with an increased mortality risk in the general older population. It remains however unclear which signs or symptoms are predictive of mortality in those suffering from LLD. Setting and Participants: Patients aged 65 years or older with depressive disorder diagnosed in Southeast London between January 2008 and December 2017. Methods: We assembled patients diagnosed with late-life depression from the Maudsley Biomedical Research Centre Case Register, which is linked to national mortality data. Using depression diagnosis as index date, we followed patients until death or censoring point. Sociodemographic data, scores of Health of the Nation Outcome Scales (HoNOS65 +), which include a physical illness scale, profiles of depressive symptoms, and psychotropic medications were extracted and modeled in multivariable survival analyses to determine predictors of mortality. Results: Of 4,243 patients with LLD (mean age 77.0 years; 61.2% female), 2,327 (54.8%) died over a median follow-up time of 3.5 years. In multivariable Cox regression models, an increased risk of all-cause mortality was associated with older age, cognitive problems, physical illness/disability, impaired activities of daily living, apathy, lack of appetite and mirtazapine prescription; conversely, female gender, non-white ethnicity, guilt feelings, tearfulness, impaired concentration, disturbed sleep and delusions were associated with lower mortality risk. Conclusions: Besides demographic factors, physical health, functioning and cognition, different depressive symptoms were significantly associated with the prognosis of LLD. Elderly patients presenting with depressive symptoms predicting higher mortality risk should be examined and followed more closely.