Abstract
BACKGROUND: Empty-nest older adults are considered a high-risk group for frailty due to constrained social support systems, yet the heterogeneity in their frailty progression remains poorly characterized. This study aimed to identify distinct frailty trajectory classes among Chinese empty-nest older adults and explore class-specific predictive factors. METHODS: We analyzed eight years of data from the China Health and Retirement Longitudinal Study. The analysis included 1399 empty-nest older adults after eligibility screening. Frailty was assessed by the frailty index (FI). Growth Mixture Modeling was employed to identify FI trajectory classes, an linear, quadratic, and freely estimated forms were compared. Variable selection was performed via LASSO regression with bootstrap stability verification. Final predictors were analyzed using multinomial logistic regression. RESULTS: A three-class quadratic model best represented the FI trajectories: "Low-increasing", "High-fluctuating", and "Elevated-stable". Common risk factors included older age, rural residence, lower grip strength, death of children, and lower life satisfaction. The "High-fluctuating" trajectory was associated with poorer childhood health and smoking. The "Elevated-stable" trajectory was predicted by worklessness and by drinking. Physiological indicators showed no independent associations. CONCLUSIONS: Frailty among Chinese empty-nest older adults follows heterogeneous pathways shaped by life-course, socioeconomic, and psychophysiological factors. These results support the need for trajectory-specific screening, early risk detection, and tailored interventions for high-risk subgroups.