| Objective:The present study aims to develop a simple,reliable and suitable prediction model for appendicular skeletal muscle mass of the elderly in the communities,so as to provide a theoretical basis for the early prevention and control of sarcopenia as an alternative method of instrumental measurement for the daily muscle health monitoring of the elderly in communities.Methods:1.Development of prediction model for appendicular skeletal muscle mass:In this stage,from August to October 2021,data of 1043 elderly people over 60 years old were collected from 15 communities in Hunan Province by multistage sampling method,and the prediction models were developed.(1)Questionnaire survey:General characteristics of the elderly such as gender,age,ethnicity,marital status,smoking and drinking history,disease history were collected through questionnaires.(2)Anthropometric methods:Anthropometric indicators such as height,weight,upper arm circumference,waist circumference,calf circumference,grip strength,gait speed were collected.(3)Bioimpedance Analysis:measuring the skeletal muscle mass of the elderly,and the appendicular skeletal muscle mass data were collected.(4)Statistical methods:The indicators moderately or highly correlated with appendicular skeletal muscle mass were screened out through correlation analysis,then the appendicular skeletal muscle mass was taken as the dependent variable and the indicators screened out by correlation analysis as the independent variables,and the prediction model of appendicular skeletal muscle mass was developed by stepwise linear regression.Meanwhile,Intraclass Correlation Coefficients(ICC)and Bland-Altman test were used in the internal evaluation of prediction model.2.Validation of the prediction model for appendicular skeletal muscle mass:In this stage,from November to December 2021,data of 399 elderly people in Changsha communities of Hunan Province were collected by multistage sampling method and the prediction models were evaluated.(1)Data collection:Questionnaire survey,anthropometric methods and Bioimpedance Analysis were used to collect general characteristics,muscle mass prediction related indicators and appendicular skeletal muscle mass data of the elderly.(2)Validation of model:Bland-Altman test,paired t test and regression analysis were used to compare the difference of appendicular skeletal muscle mass measured by instrument and predicted by model.Bland-Altman test results were expressed as the consistency rate(the number of scattered points falling within the 95%consistency interval/the total number of scattered points),and the higher the consistency rate,the better the prediction effect.Paired t test results were expressed as t value and P value,P<0.05 indicates that the difference was statistically significant.Regression analysis results were expressed as R~2value and P value.The R~2 value is closer to 1 and P<0.05 indicates that the fitting degree of prediction model is better.(3)Application of model:Using the 2019 consensus of the Asian Working Group on Sarcopenia as the diagnostic criteria,the Sensitivity,Specificity and AUC as the evaluation indexes,the diagnostic value of the prediction models in sarcopenia was evaluated.The value range of the above three indexes is 0～1,and the closer it is to 1,the better the diagnostic value of the prediction model is.Results:1.The results of development of prediction models:In the present study,there were gender differences in appendicular skeletal muscle mass(t=27.639,P<0.001).Therefore,three prediction models of appendicular skeletal muscle mass were developed:Model 1:ASM(total)=-31.133-1.573~*Gender(male=0,female=1)+0.213~*height+0.095~*weight+0.253~*calf circumference+0.035~*grip strength;Model 2:ASM(male)=-46.674+0.290~*height+0.064~*weight+0.089~*upper-armcircumference+0.348~*calf circumference;Model 3:ASM(female)=-30.334+0.204~*height+0.093~*weight+0.242~*calf circumference.2.The results of validation of prediction models:Model 1:the consistency rate was 96.99%,R~2=0.9088(SEE=1.078,P<0.001),and having great prediction ability;The AUC was 0.849(95%CI:0.796～0.903,P<0.001),and the specificity and sensitivity were 95.35%and 76.53%,showing great specificity and good sensitivity in the diagnosis of sarcopenia;Model 2:the consistency rate was 93.80%,R~2=0.8435(SEE=1.230,P<0.001),having good prediction ability;The AUC was0.849(95%CI:0.782～0.915,P<0.001),and the specificity and sensitivity were 94.68%and 82.85%,showing great specificity and sensitivity in the diagnosis of sarcopenia;Model 3:the consistency rate was 98.52%,R~2=0.8022(SEE=0.953,P<0.001),having good prediction ability;The AUC was 0.895(95%CI:0.820～0.970,P<0.001),and the specificity and sensitivity were 95.17%and 74.60%,showing great specificity and good sensitivity in the diagnosis of sarcopenia.Conclusions:skeletal muscle mass can be reflected by anthropometric indicators,and the prediction models of appendicular skeletal muscle mass developed in our study can be used as an alternative method to instruments for the early screening of sarcopenia among older adults in community,and is a simple,accurate and low-cost measurement method,at the same time,also can be used for daily monitoring of muscle health in the community older adults. |
PDF Full Text Request