Zhiyong Hou, Ming Yu, Keqiang Lu, Yiming Ma, Ran Jia, Guichuan Lai, Yuqi Bai, Weijun Huang
INTRODUCTION: Late-life depression symptoms are implicated in dementia. We examined how cumulative burden, duration, trajectory, and variability of depression symptoms are associated with incident dementia. METHODS: A prospective cohort analysis of 23,305 dementia-free adults from four population studies (English Longitudinal Study of Ageing [ELSA], Health and Retirement Study [HRS], Survey of Health, Ageing and Retirement in Europe [SHARE], China Health and Retirement Longitudinal Study [CHARLS]) with repeated Centre for Epidemiologic Studies of Depression scale (CES-D)/Euro-Depression scale (EURO-D) assessments across three waves and a pooled median follow-up of ≈10.8 years. Exposures included CumSD/cumulative average depression symptom score (CumADS), high-symptom exposure duration, visit-to-visit variability (Standard deviation [SD], coefficient of variation [CV], variation independent of the mean [VIM]), and time-course patterns. Associations were analyzed using multivariable-adjusted Cox regression. RESULTS: Each 1-unit increase in cumulative score was associated with a 3%-8% higher dementia hazard across cohorts. Highest versus lowest cumulative quartiles showed markedly elevated risk. Sustained high exposure for 4 years conferred ≈2.7-3.9× greater risk. Higher variability and worsening trajectories were also linked to higher incidence. Associations were robust across subgroups. DISCUSSION: Persistent and unstable depression symptoms independently predict higher dementia risk, supporting longitudinal mood monitoring and sustained management. HIGHLIGHTS: Multi-cohort study of 23,305 adults (ELSA, HRS, SHARE, CHARLS). Cumulative depression burden shows a dose-response with dementia risk. Highest versus lowest quartile: dementia hazard up to 18× (HRS). Sustained high symptoms (4 years) linked to ≈2.7-3.9× greater risk. Visit-to-visit variability independently associates with higher dementia risk.