Effects of depression and cognitive impairment on increased risks of incident dementia: a prospective study from three elderly cohorts

Abstract

Depression is usually accompanied with cognitive impairment and increases risk of incident dementia. However, evidence has been limited on the effect size of depression with cognitive impairment and their synergistic effect on future dementia. To explore this, we examined three large cross-country population-based prospective cohorts. Depressive symptoms were assessed by epidemiologic scale, while cognitive impairment was defined by subjective cognitive tests. Dementia was ascertained by self-reported physician-diagnosed conditions. Cox proportional hazard models were employed to determine the hazard ratio (HR) and 95% confidence interval (95% CI), with adjustments of potential confounding variables. Addictive and multiplicative interactions were calculated to evaluate the synergistic effect. A total of 64,706 participants were included at baseline (mean age: 63.9, female: 55.2%), where 4197 (6.5%) individuals had depressive symptoms only, 28,175 (43.5%) individuals had cognitive impairment only, 11,564 (17.9%) individuals had both, and 20,770 (32.1%) individuals had neither. Compared with the neither group, all the other three groups had higher risks of subsequent dementia (depression only: HR 1.65, 95% CI 1.26-2.17; cognitive impairment only: HR 2.71, 95% CI 2.33-3.14; depression with cognitive impairment: HR 3.51, 95% CI 2.95-4.17). There was insignificant additive (RERI, 0.15, 95% CI -0.45-0.75; AP, 0.042, 95% CI -0.13-0.21; SI, 1.06, 95% CI 0.83-1.37) and multiplicative (0.78, 95% CI 0.58-1.06) interaction between depression and cognitive impairment on subsequent dementia. We found depression with cognitive impairment has higher risks of dementia than either condition alone and no significant synergistic effect exists between these two factors.

Fig. 1. Proportional Venn diagrams of subgroups division by cohorts.

The distribution of sample size across different subgroups in cohort: A Health and Retirement Study (HRS), B English Longitudinal Study of Ageing (ELSA), and C Survey of health, Ageing and Retirement (SHARE). The size of the circles represents the proportion corresponding to each subgroup.

Fig. 2

Cumulative probability and risk table of dementia survival curve across subgroups.

Fig. 3. Effect of depression and cognitive impairment on hazard risks of incident dementia.

Left tabular figures are pairwise comparisons among subgroups. The numbers in each cell represent the HR (Hazard ratio) and its 95% confidence interval for the corresponding row compared to the column, while the colors indicate Z-value after multiple comparison correction. Statistical significance: Z > 1.96, P < 0.05; Z > 2.58, P < 0.01; Z > 3.29, P < 0.001. Right bar graphs represent the point estimates adjusted for covariates in each model, and the error bars represent the 95% confidence intervals.