Frailty trajectories and hip fracture risk in older Chinese adults: a prospective cohort study with competing risk analysis

Abstract Background Frailty is associated with hip fractures in older adults. However, frailty is a dynamic process, longitudinal trajectories may provide complementary descriptive insights. The exploratory association between frailty trajectories and hip fracture risk remains uncertain, particularly in older Chinese adults. Methods We analyzed longitudinal data from 3,868 community-dwelling Chinese adults aged ≥ 65 years in the China Health and Retirement Longitudinal Study (CHARLS). Frailty was assessed using the CHARLS modified frailty index (CMFI). Latent class mixed models were used descriptively to characterize frailty trajectories. The primary risk analysis used baseline CMFI, measured before follow-up, to evaluate associations with incident hip fractures using cause-specific Cox and Fine-Gray models. Restricted cubic splines were employed to explore nonlinear relationships. Associations between frailty trajectory class and hip fracture were also examined as exploratory analyses. Results During a median follow-up of 7.2 years, 225 incident hip fractures occurred. Two frailty trajectory patterns were identified descriptively: slow-rising and fast-rising. The primary finding was that each 0.1-unit increase in baseline CMFI was associated with a 32% higher risk (HR = 1.32, 95% CI: 1.19–1.46), and this association persisted after accounting for the competing risk of death (SHR = 1.19, 95% CI: 1.07–1.31). A non-linear relationship was observed, suggesting a possible attenuation of the association at higher frailty levels. The associations were generally consistent across the subgroups. In exploratory analyses, the fast-rising trajectory was associated with a higher hazard of hip fracture (HR = 2.39, 95% CI: 1.77–3.23; SHR = 2.05, 95% CI: 1.49–2.84). Conclusions Higher baseline frailty was independently associated with a higher risk of hip fracture. The fast-rising trajectory showed a higher hazard in descriptive analyses but requires time-varying validation. Dynamic frailty monitoring may aid risk stratification, pending future time-varying and external validation.