Joint effects of social isolation and depressive symptoms on incident frailty among middle-aged and older adults: evidence from CHARLS and SHARE

Frailty is a common geriatric syndrome associated with disability, hospitalization, and mortality. Social isolation and depressive symptoms are prevalent in ageing populations and often co-occur, but their independent and joint effects on incident frailty, particularly across different cultural contexts, remain unclear. We used data from two longitudinal cohorts: the China Health and Retirement Longitudinal Study (CHARLS) and the Survey of Health, Ageing and Retirement in Europe (SHARE). Social isolation and depressive symptoms were assessed at baseline. Among participants who were non-frail at baseline, incident frailty was defined as frailty observed at follow-up using a frailty index. Cox proportional hazards models were used to estimate hazard ratios and 95% confidence intervals, with follow-up time calculated from baseline to the follow-up frailty assessment. Additive interaction was evaluated using the relative excess risk due to interaction, attributable proportion due to interaction, and synergy index. Subgroup analyses examined potential effect modification. Across 69,334 person-years in CHARLS and 152,523 person-years in SHARE, depressive symptoms were consistently associated with a higher risk of incident frailty in fully adjusted models (CHARLS: HR 1.83, 95% CI 1.69–1.98; SHARE: HR 1.63, 95% CI 1.52–1.76). Social isolation was independently associated with incident frailty in SHARE (HR 1.30, 95% CI 1.14–1.48) but not in CHARLS. Compared with participants with neither exposure, those with both social isolation and depressive symptoms had an elevated risk of incident frailty in both cohorts. Additive-interaction analyses indicated a negative/subadditive interaction in CHARLS, whereas no clear evidence of additive interaction was observed in SHARE. Depressive symptoms were consistently associated with incident frailty in China and Europe, whereas social isolation showed cohort-specific associations. Their co-occurrence identified individuals at elevated risk. These findings may help identify vulnerable populations and inform future intervention studies targeting social connection and mental health.