Research Square · Oct 19, 2026
Joint Sarcopenia-Frailty Phenotypes and Subsequent Falls and Hip Fracture in Older Chinese Adults: A Seven-Year CHARLS Study
By Feng X, Chen S, Huang Z et al.
Objective To determine whether joint sarcopenia-frailty phenotypes differentially predict first observed falls and hip fracture in older Chinese adults and whether coexistence confers risk beyond either condition alone. Methods We analyzed the China Health and Retirement Longitudinal Study using 2011 as baseline and 2013, 2015, and 2018 as follow-up visits. Adults aged ≥ 60 years with a classifiable joint phenotype, no baseline report of the corresponding outcome, and at least one subsequent outcome assessment were included. Four phenotypes were defined: neither condition, sarcopenia only, frailty only, and sarcopenia plus frailty. Discrete-time complementary log-log models estimated hazard ratios (HRs) with 20-fold multiple imputation. Standardized 7-year risks, risk differences, and multiplicative and additive interactions were also estimated. Results The falls cohort included 3,987 participants with 1,415 events; the hip-fracture cohort included 4,884 participants with 197 events. Compared with neither condition, adjusted HRs (95% CIs) for falls were 1.02 (0.89–1.16) for sarcopenia only, 1.45 (1.20–1.75) for frailty only, and 1.40 (1.15–1.69) for coexisting sarcopenia and frailty. Corresponding HRs for hip fracture were 1.23 (0.87–1.75), 1.97 (1.29–3.01), and 1.91 (1.13–3.21). Standardized risks were 37.8%, 38.3%, 49.5%, and 48.3% for falls and 3.8%, 4.6%, 7.3%, and 7.0% for hip fracture, respectively. Multiplicative interactions were not significant (P = 0.709 and P = 0.470), and additive-interaction confidence intervals included zero for both outcomes. Key conclusions were unchanged after additional BMI adjustment. Conclusions Frailty, with or without sarcopenia, identified older adults at increased risk of falls and hip fracture, whereas sarcopenia alone did not show a clear association in the joint-phenotype framework. Coexistence did not demonstrate excess risk beyond that expected from the two component states. These findings support joint risk stratification but not statistical synergy.