Joint association of frailty and social isolation with depressive symptoms among middle-aged and older Chinese adults: a nationwide cohort study

Frailty and social isolation are common in later life and each has been linked to poor mental health. However, their joint association with incident depressive symptoms, and whether co-occurrence confers risk beyond either factor alone, remains unclear. Participants aged ≥ 45 years without baseline depression and with complete exposure data were included ( n = 7330). Frailty was assessed using a 30-item frailty index and defined as ≥ 25. Social isolation was measured by a composite scale (range 0–6) capturing marital status, living alone, frequency of contact with children, recent social participation, and residence; scores ≥ 3 denoted isolation. Incident depressive symptoms was defined as CESD-10 ≥ 10 at follow-up. Cox models estimated adjusted hazard ratios (HRs) for frailty, isolation, and their joint categories. Interaction was examined on multiplicative and additive scales using cross-product terms and RERI/AP/SI, respectively. Over follow-up, 2007 incident depressive symptoms cases occurred. Compared with non-frail participants, frailty was associated with depression risk (adjusted HR 1.47, 95% CI 1.24–1.74). Social isolation was also associated with increased risk (adjusted HR 1.14, 95% CI 1.02–1.28). Joint analyses showed a graded pattern versus non-frail and non-isolated: non-frail+isolated HR 1.11 (95% CI 0.98–1.25), frail + non-isolated HR 1.36 (95% CI 1.10–1.67), frail+isolated HR 1.86 (95% CI 1.43–2.42). Additive interaction was suggestive but not definitive (RERI 0.39; 95% CI − 0.14–0.93; AP = 0.21; − 0.03–0.46; SI = 1.85; 0.81–4.21); multiplicative interaction was not evident (interaction HR 1.24, 95% CI 0.89–1.72). Subgroup analyses by gender and cardiometabolic comorbidities showed broadly consistent patterns (all P for interaction >0.05). Frailty and social isolation independently predict incident depressive symptoms among middle-aged and older Chinese adults, with co-occurrence identifying the highest-risk subgroup. While interaction findings were suggestive on the additive scale, they were not statistically definitive.