Risk prediction model for sarcopenia in elderly:a systematic review

LU Chengqian; LI Ya; LU Qinyi; JIN Xueqin

doi:10.12102/j.issn.2095-8668.2025.23.008

Abstract：Objective:To systematically review the risk prediction models of sarcopenia in the elderly,and to improve the selection and development of sarcopenia prediction models for healthcare professionals.Methods:Systematic searches of the literature on risk prediction models of sarcopenia in the elderly in CNKI,WanFang Data,VIP,CBM,PubMed,Web of Science,EMbase,Cochrane Library was conducted.The retrieval time was from the inception to October 31,2024.Two researchers independently screened the literature,extracted data,and evaluated the risk of bias and applicability of the included literature according to the PROBAST.Results:A total of 13 articles involving 18 models were included,with a total sample size ranging from 43 to 108 304,and outcome event rates ranging from 4.97%to 37.60%.There were 18 models reported AUC ranging from 0.757 to 0.955.Nine studies underwent internal validation and 1 study underwent both internal and external validation.The 13 studies had good applicability,but there was a high risk of bias.Predictors reported repeatedly by multiple models were age,weight,BMI,sex,and exercise.Conclusion:Current evidence shows that the risk prediction models for sarcopenia in elderly has good predictive performance,but has a high risk of bias.Future studies should strictly follow the specifications of risk prediction model reporting to develop and evaluate the risk prediction model of sarcopenia in elderly to improve their scientific validity,and while constructing the prediction model based on longitudinal data,large-sample,multi-centre external validation should be conducted to assess the feasibility and generalizability of the model.