身体质量指数作为肥胖筛查标准的判别准确性评价  

Accuracy of body mass index as a screening standard for obesity

作　　者：李红娟[1] 杨柳[1] 张楠[1] 

机构地区：[1]北京体育大学运动人体科学学院,北京100084

出　　处：《中国预防医学杂志》2014年第6期571-575,共5页Chinese Preventive Medicine

基　　金：国家科技支撑计划课题(2009BAK62B04;2012BAK21B01)

摘　　要：目的以WHO体脂率肥胖判定标准作为＂金标准＂,对中国成年人身体质量指数（简称体质指数,BMI）肥胖判定标准的真实性、可靠性、收益等进行评价,并探讨BMI作为体脂率是否超标的指征的诊断性能及适宜的截断值。方法研究对象为18~79岁成年人689名,测试身高、体重,计算BMI。采用DEXA法对研究对象进行全身身体成分测定。以世界卫生组织推荐的体脂率判定标准（男性F%≥25%,女性F%≥35%）作为肥胖判定的＂金标准＂,以中国肥胖问题工作组研制的中国成年人肥胖判定的BMI标准（BMI≥28kg/m2）作为筛检阳性标准。对BMI标准的真实性、可靠性、收益等进行评价,并采用ROC曲线分析进行截断值判定。结果 BMI作为肥胖筛查标准的真实性评价主要表现为特异度高,灵敏度过低,假阴性率过高;尤其是女性,灵敏度只有0.18,假阴性率高达82%。与体脂率＂金标准＂判断一致性男性尚好（Kappa值为0.32）,女性则较差（Kappa值为0.13）。BMI作为肥胖筛查标准约登指数较低,尤其女性（0.13）;阳性预测值和阳性似然比较高;阴性预测值较低,阴性似然比较高。以体脂率判定的肥胖作为效标时,BMI的诊断性能中等（AUC=0.822,P=0.000）,男女分组诊断性能提高（AUC分别为0.844和0.868,P=0.000）;BMI切点男女分别为25.24和23.98。结论国内现有成年人BMI肥胖判定界值可能偏高,尤其对于女性;在较小样本研究中不宜以BMI作为肥胖分组的依据;应建立以健康指标为参照的、考虑身体脂肪成分和不同性别年龄的各人群肥胖判定标准;以健康干预为目标的BMI筛查截断值确定时应以提高灵敏度、适当降低特异度为原则。Objective To assess the accuracy,reliability and benefit of using body mass index（BMI）of Chinese adults as a screening standard for obesity based on WHO ＂golden standard＂ of BF%for obesity,and to explore the appropriate BMI cut-off for Chinese adults health intervention. Methods 689 adults aged 18-79 years were selected and their BMIs were calculated.Dual energy X-ray absorptiometry（DEXA）was used to estimate body fat percentage（BF%）.Obesity was determined according to BF% which was used as ＂ golden standard＂ by WHO（BF%25%for male and BF 35%for female）,and BMI（28kg/m2）as screening standard for obesity. Results A higher specificity（95%）and a low sensitivity were noticed when BMI was used as screening standard for obesity,especially among women,in which sensitivity was only 0.18 and false negative rate was as high as 82%.The reliability evaluation showed a higher consistency with ＂golden standard＂ in men（Kappa=0.32）than in women（Kappa=0.13）.The Youden index of BMI standard was low especially for women（0.13）.The positive predictive value,positive likelihood and the negative likelihood were high,but the negative predictive value was low.The diagnostic performance of BMI for obesity was fine in general（AUC=0.822,P=0.000）,but was better when used only for men（AUC=0.844）or for women（AUC=0.868,P=0.000）.BMI cut-off was 25.24 for men and 23.98 for women. Conclusions The current BMI threshold for Chinese adult obesity may be high,especially for women,which is not appropriate to be used in smaller sample studies.Therefore,a comprehensive standard for obesity diagnosis should be based on health indicators,BF%as well as gender and age.BMI cut-off value should be chosen to increase the sensitivity and decrease the specificity when used for health intervention.

关 键 词：身体质量指数 体脂率 肥胖 筛检试验评价 ROC分析 

分 类 号：R589.2[医药卫生—内分泌]