Optimal Variables and Corresponding Cut-off Points for Identification of Cardiovascular Risk Factors Clustering among Elderly Males

范宇慧,蔡威,陶晔璇,汤庆娅

2009-01-01

Abstract:Objective To explore the optimal cut-off points of body mass index(BMI),percentage of body fat(PBF) and body fat mass index(BFMI) for identification of cardiovascular risk factors clustering among elderly males. MethodsThe data of physical examinations from 1 052 Shanghai elderly males in 2007 were collected.The relationship between cardiovascular risk factors clustering and different strata of BMI,PBF and BFMI was analyzed.Receiver Operator Characteristic(ROC) curve analysis was employed to determine the optimal cut-points for identification of cardiovascular risk factors clustering,and area under curve(AUC) was worked out.The population attributable risk proportion(PARP) of risk factors clustering was calculated. Results Odds ratios of risk factors clustering tended to increase with BMI,PBF and BFMI.BMI≥24 kg/m2,PBF≥21% and BFMI≥5 kg/m2 were the cut-off points that had approximate sensitivity and specificity,and/or had the shortest distance in ROC curve.AUC of all the three indexes was larger than 0.5.Analysis of PARP indicated that BMI under 24 kg/m2,PBF under 21% and BFMI under 5 kg/m2 could prevent 27.1%,37.44% and 36.63% risk factors clustering,respectively. Conclusion BMI≥24 kg/m2,PBF≥21% and BFMI≥5 kg/m2 can well reflect the cardiovascular risk factors clustering among elderly males.