Background: Depression might be an independent risk factor for cognitive decline, a prodromal dementia symptom or a reaction to cognitive/functional impairment.

Objective: To investigate the association between (1) depressive symptoms and (2) depressive symptom patterns over 13 years with incident mild cognitive impairment (MCI) 5 years later.

Materials and methods: We included 724/823 cognitively unimpaired men/women who participated in the population-based Heinz Nixdorf Recall study (t1: 2005–2008, ø62.9 years; t2: 2010–2015, ø68.1 years). Depressive symptoms were assessed in the study center and during six postal follow-ups using the short form of the Center for Epidemiologic Studies Depression Scale (CES-D). Relative risks (RR; 95% confidence intervals) for MCI at t2 (men/women: 71/76) were estimated for CES-D at t1 (linear and dichotomized at ≥17, cut-off for clinically relevant depressive symptoms) and CES-D fluctuations over 13 years (stable low, large fluctuations, stable high/stable around cut-off) using log-linear regression models with Poisson working likelihood adjusted for age, sex, education, diabetes mellitus, coronary heart disease, and stroke.

Results: Fully adjusted risk for MCI at t2 (per CES-D point increase at t1) was elevated for the total cohort (1.053, 1.031–1.076), men (1.046, 1.012–1.081), and women (1.059, 1.029–1.090). Applying the dichotomized CES-D, risk for MCI was substantially increased for the total cohort [2.22 (1.38–3.58)] and in women [2.59 (1.46–4.58)]. Large CES-D fluctuations and stable high/stable around cut-off were associated with increased RR for MCI in the total cohort and in women compared to stable low symptoms.

Conclusion: Depressive symptoms predicted MCI in cognitively unimpaired participants of our population-based study. Adequate treatment of depression may therefore contribute to the maintenance of normal cognition and delay dementia onset.