Estimating Weekly Exercise Duration Requirements to Reduce Possible Sarcopenia Risk in High-Risk Older Adults: A Longitudinal Survival Analysis with Counterfactual Explanation

This study examined the longitudinal association between weekly exercise duration and incident possible sarcopenia and estimated sex-specific exercise increments required to achieve predefined hazard targets among high-risk older adults. METHODS:We analyzed longitudinal data from 4,360 adults aged ≥65 years without possible sarcopenia at baseline in the Korean Longitudinal Study of Aging (KLoSA).Participants were followed for up to 8 years.Possible sarcopenia was defined by handgrip strength (<28 kg for men, <18 kg for women).Sex-stratified Cox proportional hazards models estimated associations between exercise and incident possible sarcopenia.High-risk individuals were defined as those in the upper 20% of the sex-specific Cox linear predictor distribution, and particle swarm optimization-based counterfactual analysis estimated the minimum exercise-duration increase required to achieve target relative hazard reductions. RESULTS:During a mean follow-up of 4.88±2.41years, corresponding to 21,270 person-years, incident possible sarcopenia occurred in 914 of 2,010 men (45.5%; 92.6 per 1,000 person-years) and 1,102 of 2,350 women (46.9%; 96.7 per 1,000 person-years).Using never exercise as the reference, exercise ≥150 min/week was associated with lower risk in men (HR, 0.742; 95% CI, 0.637-0.863)and women (HR, 0.824; 95% CI, 0.709-0.958).In the high-risk group, a 10% relative hazard reduction in men was achieved at a representative exercise increase of approximately 330 min/week, whereas a 20% reduction in women required approximately 510 min/week. CONCLUSIONS:These estimates should not be interpreted as prescriptive targets but as population-level reference points under constrained optimization.The findings support gradual, individualized increases in weekly exercise duration among high-risk older adults.