Principal Component Analysis of Relationship between Anthropometric Indices and Plasma Glucose in Adolescents

XU Hong,TU Yunming,YANG Yuping,PENG Wei,LIU chang,JIANG Fuqing,XIAO Cailan,XIANG qun,LIANG Shangdong
DOI: https://doi.org/10.3969/j.issn.1007-7146.2013.03.011
2013-01-01
Abstract:Objective:To evaluate the effects of anthropometric indices including body mass index(BMI),waist circumference(WC),hip circumference(HC),waist hip ratio(WHR),and waist-to-height ratio(WHtR) on fasting plasma glucose(FPG) in adolescents in Nanchang region.Methods:Weight,height,BMI,WC,HC,WHR,WHtR and FPG of 731 adolescents at the ages of 12-18 were measured.The results were analyzed with SAS software.Results:Five anthropometric indices including BMI,WC,HC,WHR and WHtR were significantly correlated with FPG with the Pearsons correlation coefficients ranging from 0.14 to 0.38.Principal component analysis(PCA) was performed to form four principal components(PCs).Multiple regression analyses showed that combined analyses of four PCs was better than any one of five anthropometric indices alone in the explanation of the variance of FPG.Conclusion:All the five anthropometric indices including BMI,WC,HC,WHR and WHtR may be correlated with FPG in both male and female adolescents,and the accuracy of predicting the FPG using combined five anthropometric indices is higher than that using one of them alone.
What problem does this paper attempt to address?