Associations of physical activity and sleep duration with frailty risk and trajectories in middle-aged and older Chinese adults: a longitudinal cohort study from CHARLS

The levels of physical activity (PA) and sleep duration associated with lower frailty risk and more favorable frailty trajectories among middle-aged and older Chinese adults remain to be further clarified. This study aimed to identify levels of physical activity and sleep duration associated with lower frailty risk and to further examine their longitudinal associations with incident frailty and frailty trajectories. Data were derived from the 2011 wave of the China Health and Retirement Longitudinal Study (CHARLS), including 5,181 middle-aged and older adults. Physical activity was assessed using the modified International Physical Activity Questionnaire (IPAQ). Total sleep duration was calculated as the sum of nighttime sleep and daytime napping. Based on the frailty index (FI), participants were categorized as frailty, pre-frailty, or robust. First, Restricted cubic spline (RCS) models were used to examine dose–response associations of moderate-to-vigorous physical activity (MVPAV) and total sleep duration with frailty risk. Using 2011–2018 data, group-based trajectory modeling was applied to identify FI trajectories, and multinomial logistic regression was used to examine their associations with RCS-derived physical activity and sleep duration criteria, with odds ratio (OR) and 95% confidence intervals (CIs). For longitudinal analysis, 2,947 non-frail participants at baseline were followed. Cumulative incidence functions, accounting for death as a competing risk, were used to estimate incident frailty, and Gray’s test compared between-group differences. Multivariable Cox models assessed the associations of being within the RCS-derived lower-risk physical activity and sleep ranges with frailty risk and transitions, using those outside these ranges as the reference. Additional analyses examined the combined number of study-specific criteria met. Hazard ratios (HRs) and 95% CIs were reported. Subgroup and stratified analyses were conducted to assess robustness. RCS analysis showed that a weekly MVPA volume of approximately 561 MET-min/week and a daily sleep duration of 7.00–9.33 h were associated with lower frailty risk. Three distinct FI trajectories were identified over time: a “Frail worsening to more frailty” trajectory, a “Robust worsening to pre-frailty” trajectory, and a “Pre-frail worsening to frailty” trajectory. After adjusting for covariates, reaching the RCS-derived lower-risk physical activity level (OR = 0.05, 95% CI: 0.01–0.14) and being within the identified lower-risk sleep duration range (OR = 0.03, 95% CI: 0.01–0.06) were both associated with a lower likelihood of belonging to the “Frail worsening to more frailty” trajectory. Cumulative incidence function analysis, accounting for death as a competing risk, showed significantly lower frailty incidence among participants reaching the identified physical activity level or being within the identified sleep duration range, as well as among those meeting both study-specific criteria, compared with their respective reference groups. Cox proportional hazards models further showed that reaching the identified physical activity level was associated with a lower risk of incident frailty (HR = 0.87, 95% CI: 0.75–0.99), while being within the identified sleep duration range was associated with a lower risk of incident frailty (HR = 0.73, 95% CI: 0.64–0.84) and a lower risk of transition from robust to pre-frailty (HR = 0.76, 95% CI: 0.66–0.89). The combined analysis showed that participants meeting both study-specific criteria had a lower risk of frailty. In this study, a weekly PA level of approximately 561 MET-min/week, corresponding to about 141 min of moderate-intensity or 71 min of vigorous-intensity physical activity per week, and a daily sleep duration of 7.00–9.33 h were identified as study-specific levels/ranges associated with lower frailty risk.