| Background:Sarcopenia is an age-related disease,characterized by loss of muscle mass,decreased muscle strength and declined physical performance.Sarcopenia affects the quality of life of the elderly significantly.Until now,there is lack of epidemiological data on prevalence and risk factors of sarcopeniain this region.Furthermore,the genetic studies of sarcopenia are relatively limited,which are mainly focusing on the correlation between gene polymorphism and muscle phenotype.However,due to the complexity of the disease,ethnic differences and lack of verification of large sample size,so the results of different studies in this field are inconsistent.In addition,it is still necessary to explore how to use a simple and efficient method to evaluate the occurrence of sarcopenia in the early stage.Objective:(1)To deeply know the prevalence and influence factors of sarcopenia in this region through epidemiological investigation;(2)To explore the polymorphism loci and genotypes of susceptibility genes related to muscle mass loss and sarcopenia;(3)To build models of evaluating the risk of sarcopenia occurrence through machine learning,which facilitate the early screening of sarcopenia in basic medical institutions.Methods:(1)The first part was a cross-sectional study,and a sample of 900 permanent residents aged over 50 from three communities in Changchun City was randomly selected.All subjects received the questionnaire,physical examination,laboratory tests and body composition examination.Sarcopenia was diagnosed according to2019 AWGS.Statistical analysis was performed by SPSS software,to clearly confirm the prevelance of possible sarcopenia and sarcopenia,and influence factors of sarcopenia;(2)In the second part,the correlation between gene polymorphism and sarcopenia was studied.Single nucleotide polymorphism(SNP)was genotyped in 126 subjects with decreased muscle mass and 126 controls from epidemiological survey using Massarray technique,and finally 28 SNP loci of seven candidate genes were included in regression analysis to find out SNP loci and susceptible genotypes associated with decreased muscle mass and sarcopenia,and explore the pathogenesis of sarcopenia from a genetic point of view.(3)In the third part,we used the clinical indicators to build risk assessment models of loss of muscle mass based on machine learning.Furthermore,we assessed the models through sensitivity,specificity,positive predictive value,negative predictive value and area under the ROC curve.In this way,we aim to provide a simple and efficient tool to screen sarcopenia in the early stage.Results:1.In the first part,totally 741 subjects completed the invetigation,and 122 subjects were in the probable sarcopenia group,accounting for 16.5%,and 55 subjects were in the sarcopenia group,accounting for 7.4%.Of these,men of possible sarcopenia and sarcopenia accounted for 8.9% and 9.6%,respectively.Women of possible sarcopenia and sarcopenia accounted for 21.1% and 6.1%,respectively.2.The influence factors of decreased muscle mass and muscle strength were different in men and women.(1)In men,aged 60-69(?=-0.098,95%CI: [-0.294,-0.028],p=0.018),70-79(?=-0.195,95% CI: [-0.501,-0.207],p<0.001)and 80 years and over(?=-0.121,95%CI: [-0.966,-0.305],p<0.001)had a lower muscle mass than those aged 50-59,and more appendicular fat mass(?=-0.284,p<0.001)and greater waist-to-hip ratio(?=-0.142,95%CI: [-3.153,-1.153],p<0.001)were associated with lower muscle mass.And higher BMI(?=1.078,95%CI: [0.263,0.315],p<0.001)was associated with higher muscle mass.(2)In women,higher appendicular fat mass(?=-0.318,p<0.001),waist-hip ratio(?=-0.112,95%CI: [-1.767,-0.606],p<0.001),PHQ-9 score(?=-0.058,95%CI: [-0.022,-0.002],p=0.025)and FT4 level(?=-0.077,95%CI: [-0.037,-0.007],p=0.003)were correlated with lower muscle mass.And higher BMI(?=1.059,95%CI: [0.220,0.259],p<0.001),estradiol level(?=0.056,95%CI: [0.000,0.005],p=0.035),osteocalcin level(?=0.077,95%CI: [0.003,0.016],p=0.007)and bone mineral density(?=0.080,95% CI: [0.019,0.117],p=0.007)were associated with higher muscle mass.(3)In men,aged 60-69(?=-0.329,95%CI:[-8.276,-2.916],p<0.001)70-79(?=-0.377,95% CI: [-10.116,-4.004],p<0.001)and80 and over(?=-0.211,95%CI: [-17.620,-5.320],p<0.001)had lower muscle strength compared with those aged 50 to 59,and diabetes(?=-0.163,95%CI: [-4.519,-1.128],p=0.001)and stroke(?=-0.140,95%CI: [-6.085,-1.045],p=0.006)were correlated with lower grip strength.And faster walking speed(?=0.188,95%CI:[3.473,12.134],p<0.001),greater calf circumference(?=0.139,95%CI: [0.122,0.758],p=0.007),higher albumin level(?=0.154,95%CI: [0.175,0.808],p=0.002)and insulin-like growth factor-1 level(?=0.105,95%CI: [0.001,0.031],p=0.037)were associates with higher grip strength.(4)In women,aged 60-69(?=-0.168,95%CI: [-2.862,-0.587],p=0.003)and aged 80 and over(?=-0.126,95% CI: [-7.584,-1.175],p=0.008)had lower grip strength compared with those aged 50-59,and more full-term births(?=-0.197,95%CI: [-1.909,-0.658],p<0.001),longer duration of diabetes(?=-0.143,95%CI: [-0.165,-0.043],p=0.001)and higher alkaline phosphatase level(?=-0.102,95%CI: [-0.035,-0.004],p=0.015)were associated with lower grip strength.Higher ASMI(?=0.201,95% CI: [0.812,1.934],p<0.001)and physiological score(?=0.119,95%CI: [0.043,0.254],p=0.006)and regular moderate intensity exercise(?=0.092,95%CI: [0.153,2.799],p=0.029)were correlated with higher grip strength.3.Influence factors for sarcopenia: Men who aged 80 years and older(OR=363.697,95%CI:3.769-35098.453),had diabetes(OR=13.622,95%CI:1.979-93.770)and became shorter more than 3cm after 40 years old(OR=6.053,95%CI: 1.341-27.323)had an increased risk of sarcopenia.Women with osteoporosis(OR=4.868,95%CI: 1.327-17.865),became shorter more than 3cm after 40 years old(OR=4.766,95%CI:1.167-19.468),had more full-term births(OR=4.544,95%CI:2.015-10.246)and higher trunk fat mass(OR=1.001,95%CI:1.000-1.001)had an increased risk of sarcopenia.In addition,insulin growth factor-1>201.0ng/m L(OR=0.038,95%CI:0.002-0.750),greater calf circumference(OR=0.642,95%CI:0.436-0.945)and thigh circumference(OR=0.68,95%CI: 0.539-0.873),higher MNA initial screening score(OR=0.661,95%CI:0.439-0.996)and physiological score(OR=0.772,95%CI:0.628-0.949)were protective factors of suffering sarcopenia in men.Testosterone>0.34 ng/m L,high BMI(OR=0.172,95%CI:0.086-0.347)and high physiological score(OR=0.831,95%CI:0.715-0.965)were protective factors of suffering sarcopenia in women.4.In the second part of the study,a total of 252 people were genotyped,and the detectable rate of 28 SNP loci of 7 genes was more than 90% and they were included in the final analysis.(1)The risk of decline of ASMI was lower for heterozygous CT genotypes than homozygous TT/CC genotypes at rs2305534 of ACTN3 gene(OR=0.44,95%CI: 0.21-0.96).(2)At rs6218 of IGF1 gene,heterozygous GA genotype had a higher risk of decline of ASMI than homozygous AA/GG genotype(OR=2.98,95%CI:1.31-6.78).The GA/GG genotype had a higher risk of decline of ASMI than the wild-type AA genotype(OR=2.79,95%CI:1.25-6.23).The GG genotype had a higher risk of decline of ASMI than the wild-type AA genotype(OR=2.02,95%CI:1.03-3.95).(3)At rs2288377 of the IGF1 gene,the TT genotype had a higher risk of sarcopenia than the AA/AT genotype(OR=4.51,95%CI:1.15-17.69).(4)At the rs978458 of the IGF1 gene,and TT genotype had a higher risk of sarcopenia than the CC/CT genotype(OR=2.78,95%CI: 1.03-7.48);TT genotypes were associated with an increased risk of sarcopenia compared with wild-type CC genotypes(OR=1.98,95%CI:1.12-3.49).(5)In men,AA genotypes had lower ASMI than the wild-type GG genotypes at rs540874 of ACTN3(?=-0.1319,P=0.0488).The ASMI of GG genotype at rs6218 was higher than that of AA/GA genotype(?=0.1547,P =0.0218).5.In the third part,we built the risk assessment models of loss of muscle mass.When age,gender,BMI and calf circumference were applied to build a model,the sensitivity and specificity of SVM classification model were 0.701±0.074 and0.963±0.018,respectively.The positive predictive value(PPV)was 0.794±0.101;the negative predictive value(NPV)was 0.939±0.019 and the average area under the ROC curve(AUC)was 0.94±0.04.Furthermore,the sensitivity and specificity of ANN classification model were 0.710±0.077 and 0.955±0.020,respectively.The PPV was 0.757±0.102;the NPV was 0.940±0.025 and the AUC was 0.93±0.02.Compared with other models,the results of this model were better and easier.In addition,after our preliminary verification of the established model,the sensitivity,specificity,PPV,NPV and AUC of the SVM classification model were 0.586,0.993,0.895,0.910 and 0.96,respectively.The sensitivity,specificity,PPV and NPV of the regression model were 0.483,0.997,0.993 and 0.951,respectively.The mean error was 0.374kg/m~2,and the mean error percentage was 5.4%.Furthermore,the sensitivity,specificity,PPV,NPV and AUC of ANN classification model were 0.517,0.986,0.789,0.953 and 0.95,respectively.The sensitivity,specificity,PPV and NPV of the regression model were 0.448,0.993,0.867 and 0.948,respectively.The mean error was 0.365kg/m~2,and the mean error percentage was 5.3%.It is suggested that the model has clinical application value,but it still needs to be improved.Conclusion:1.The prevalence of sarcopenia in men was higher than that in women in Changchun City.The prevalence of possible sarcopenia in women was higher than that in men.2.The influence factors for loss of muscle mass and grip strength were different between men and women.3.Aging and history of diabetes had a greater impact on sarcopenia in men;Fertility and history of osteoporosis were associated with a greater risk of sarcopenia in women.Nutrition,physical activity and hormone levels were also associated with sarcopenia.4.Genetic factors might play a role in the individual differences in loss of muscle mass and sarcopenia.The polymorphism at rs2305534 of ACTN3 gene and rs6218 of IGF1 gene were related to the individual differences in decrease of muscle mass.Polymorphism at rs978458 of IGF1 gene was associated with individual differences in sarcopenia.Polymorphism at rs540874 of ACTN3 gene was associated with individual differences in muscle mass in men.5.In primary medical institutions or when DXA can not be accessed to measure muscle mass,age,gender,BMI and calf circumference could be used to build a model and preliminarily evaluate the risk of decline of muscle mass by using SVM or ANN algorithm.Meanwhile,combined with the measurement of grip strength and gait speed,we could preliminarily determine whether there is a possibility of sarcopenia. |
PDF Full Text Request