Marital and living status and biological ageing trajectories: a longitudinal cohort study with a 20-year follow-up

Abstract

Biomarkers of ageing (BA) can predict health risks beyond chronological age, but little is known about how marital/living status affects longitudinal changes in BA. We examined the association between marital/living status and BA over time using the-Swedish-Adoption/Twin-Study-of-Aging (SATSA) cohort. Four BAs were analyzed: telomere length (TL) (638 individuals; 1603 measurements), DNAmAge (535 individuals; 1392 measurements), cognition (823 individuals; 3218 measurements), and frailty index (FI) (1828 individuals; 9502 measurements). Individuals were born between 1900 and 1948, and data on marital/living status, BAs, and covariates were collected through nine waves of questionnaires and in-person testing from 1986 to 2014. Mixed linear regression with random effects at twin-pair and individual levels were used to assess BA changes for constant marital/living status. Conditional generalized estimating equation assessed within-individual BA changes for varying marital/living status. Results showed that individuals who were consistently unmarried/non-cohabiting (β = 0.291, 95%CI = 0.189-0.393) or living alone (β = 0.203, 95%CI = 0.090-0.316) were more frail, and experienced accelerated frailty (p-for-interaction with age < 0.001 for marital status; p-for-interaction = 0.002 for living status) and cognitive decline (p-for-interaction < 0.001), compared to those married/cohabiting or living with someone Among individuals whose marital/living status changed, frailty was higher when living alone (β = 0.089, 95%CI = 0.017-0.162) and frailty accelerated when they became unmarried/non-cohabiting or were living alone (p-for-interaction < 0.001). Cognitive decline also accelerated when living alone (p-for-interaction = 0.020). No associations were observed for TL and DNAmAge. In conclusion, being unmarried/non-cohabiting or living alone from mid-to-old age is linked to accelerated cognitive decline and frailty. These findings highlight the potential importance of social support networks and living arrangements for healthy ageing.

Keywords: Ageing trajectories; Biological ageing; Cognition; Cognitive function; Frailty; Living status; Longitudinal; Marital status; Telomere.

Fig. 1

Flow chart. BA Biomarkers of ageing, BMI body mass index, N number of individuals, M number of measurements

Fig. 2

Longitudinal trajectories of BA in people with constant marital and living status. a Longitudinal trajectories of BA by marital status. b Longitudinal trajectories of BA by living status. Note Average changes of BA with chronological age were denoted by smooth lines, with red representing individuals who were married/cohabiting (a) or living with someone (b), and blue for being unmarried/non-cohabiting (a) or living alone (b). Individual BA measurements were presented as orange-connected lines. Mixed linear regression models were used to estimate longitudinal changes of BA, with fixed effects at chronological age, sex, education attainment, smoking status, BMI and calendar year of birth, including an interaction term between marital status and chronological age or living status and chronological age, and with random effects at the twin-pair level and the individual level

Fig. 2

Longitudinal trajectories of BA in people with constant marital and living status. a Longitudinal trajectories of BA by marital status. b Longitudinal trajectories of BA by living status. Note Average changes of BA with chronological age were denoted by smooth lines, with red representing individuals who were married/cohabiting (a) or living with someone (b), and blue for being unmarried/non-cohabiting (a) or living alone (b). Individual BA measurements were presented as orange-connected lines. Mixed linear regression models were used to estimate longitudinal changes of BA, with fixed effects at chronological age, sex, education attainment, smoking status, BMI and calendar year of birth, including an interaction term between marital status and chronological age or living status and chronological age, and with random effects at the twin-pair level and the individual level