Longitudinal bidirectional associations among sarcopenia risk, social isolation, and frailty in Chinese older adults using an item response theory-derived frailty index.

Introduction: Frailty is a common geriatric syndrome characterized by multisystem decline and increased vulnerability in older adults, and may interact with sarcopenia risk and social isolation through shared behavioral and functional pathways. However, their longitudinal bidirectional associations remain insufficiently understood. In addition, the frailty index, one of the most widely used tools for frailty assessment, is typically constructed within a conventional equal-weight cumulative-deficit framework and may not fully capture heterogeneity in the contribution of individual deficits. Methods: This longitudinal observational study used data from the China Health and Retirement Longitudinal Study from 2011 to 2015. A total of 3,159 adults aged 60 years and older with repeated measures of sarcopenia risk, social isolation, and frailty were included. Frailty was assessed using a 40-item item response theory-derived frailty index. Sarcopenia risk was measured using a composite score based on estimated skeletal muscle mass, grip strength, and gait speed. Social isolation was assessed using a multidimensional social isolation index. Longitudinal associations were examined using cross-lagged panel models and random-intercept cross-lagged panel models. Results: Reciprocal longitudinal associations were observed. At the between-person level, higher sarcopenia risk was associated with higher frailty. At the within-person level, higher sarcopenia risk predicted later social isolation, higher social isolation predicted later frailty, and sarcopenia risk also directly predicted later frailty. Frailty more strongly predicted later social isolation and later sarcopenia risk. Bootstrap-based indirect-effect analyses further suggested that social isolation may represent a modest longitudinal indirect pathway linking sarcopenia risk and frailty, with more consistent evidence for the reverse pathway. No significant subgroup differences were observed by age, sex, education, or residence. Conclusion: Sarcopenia risk, social isolation, and frailty formed a linked longitudinal process in older adults. Social isolation may represent a modest indirect pathway linking sarcopenia risk and frailty in both temporal directions, with stronger evidence for the reverse pathway. Early identification of both muscle-related vulnerability and social disconnection may help delay frailty progression.