The association between change and stability in subjective health ratings and dementia risk: a multi‐cohort study in Swedish and Danish twins

Abstract

AbstractBackgroundPoorer self‐rated health (SRH) is associated with increased dementia risk (Stephan et al., 2021). However, it is unclear whether genetic and/or environmental variance accounts for this risk. Although SRH declines with age (Finkel et al. 2020), it is unknown whether rates of change in SRH are independently associated with dementia outcomes. Using a validated latent dementia index (LDI) (Beam et al., 2022), we tested whether stability and change components of SRH predicted LDI via genetic and environmental pathways in two international cohorts of twins.MethodOur data consisted of two samples of monozygotic (MZ) and dizygotic (DZ) twins. We used data from the Swedish Adoption Twin Study of Aging (SATSA; Ntwins = 548, Npairs = 339) and the Longitudinal Study of Aging Danish Twins (LSADT; Ntwins = 3238, Npairs = 2561). Bivariate Cholesky decompositions and three‐level growth‐curve models with predictive intercepts and slopes were used for our cross‐sectional and longitudinal analyses respectively. All models were estimated in Mplus (Muthen & Muthen, 2017) using maximum likelihood with robust standard errors.ResultsCross‐sectional model results indicate that those with poorer SRH have lower LDI scores (b = ‐0.01, p = .008), even after adjusting for shared genetic covariance, age, sex, education, and depressive symptoms. Predictive growth models indicate that rates of change in SRH are negatively associated with LDI after adjusting for age, sex, education, and depressive symptoms (b = ‐0.17, p = .032). The degree of genetic confounding in the association between rates of change and LDI was inconclusive.ConclusionOur study provides novel insights into the relationship between individual ratings of health and dementia risk. SRH predicted dementia likelihood after controlling for genetic confounding, opening up the possibility that subjective ratings of health account for unique variance in dementia risk. Additionally, changes in SRH also predicted dementia likelihood beyond their mean score. Taken together these results highlight the importance of subjective health measures in understanding dementia risk.