Abstract
INTRODUCTION: Sarcopenia has been proved to be associated with cardiovascular diseases (CVD), chronic kidney disease, and metabolic disorders, but the relationship between sarcopenia and all-cause and cardiovascular mortality risk among middle-aged and older adults across stages 0-3 of cardiovascular-kidney-metabolic (CKM) syndrome remains unclear. This study aimed to investigate the relationship between sarcopenia and all-cause and cardiovascular mortality risk among middle-aged and older adults across stages 0-3 of CKM syndrome based on Nutrition Examination Survey (NHANES) 2011-2018 and the China Health and Retirement Longitudinal Study (CHARLS) 2011-2020. METHODS: Multivariable Cox regression analysis was performed to analyze the association of sarcopenia with all-cause and CVD mortality. Restricted cubic spline (RCS) analysis was conducted to explore the non-linear relationship between body mass index (BMI)-adjusted muscle mass (appendicular skeletal muscle mass divided by BMI, ASMI) and all-cause and CVD mortality, and machine learning (ML) models were developed for mortality risk prediction. The CHARLS database was utilized as validation to enhance the stability of the results. RESULTS: Over an average follow-up of 5.11 years, NHANES recorded 85 all-cause deaths and 12 CVD deaths. After the full adjustment, sarcopenia was significantly associated with all-cause mortality (weighted hazard ratio [HR] = 3.428, 95% confidence interval [CI] 1.484-7.915, p = 0.004) and CVD mortality (weighted HR = 1.070, 95% CI 1.009-1.570, p = 0.049). RCS analysis revealed a nonlinear negative relationship between ASMI and CVD mortality (p for nonlinear = 0.009). ML models demonstrated better predictive performance, with random forest (area under the curve = 0.766) achieving the highest accuracy. In the CHARLS cohort, sarcopenia significantly increases all-cause mortality (HR = 2.519, 95% CI 1.402-4.527, p < 0.001), and subgroup analysis reported consistent results with the main analysis. CONCLUSION: Sarcopenia was significantly associated with all-cause mortality and with CVD mortality in fully adjusted models, though the latter association was modest, and exhibits good predictive performance for mortality among middle-aged and older individuals within CKM stages 0-3.