人体测量学
医学
标准差
标准误差
统计
协议限制
横断面研究
生物电阻抗分析
参考值
体质指数
核医学
数学
病理
内科学
作者
Manuela Dittmar,Helmut Reber
标识
DOI:10.1046/j.1532-5415.2002.50015.x
摘要
To cross-validate existing anthropometric and bioimpedance equations to establish their validity and accuracy for estimating intracellular water (ICW) in healthy older Germans and to develop a new equation with improved accuracy and precision for predicting ICW from multifrequency bioimpedance analysis (MFBIA).Cross-validation study.University of Mainz.One hundred fifty-five healthy volunteers aged 60 to 80 years (77 men, 78 women; mean ages +/- standard deviation 67.7 +/- 4.8 and 68.6 +/- 5.5 years, respectively).ICW was measured by whole-body counting of (40)potassium ((40)K) ((40)K method) as the reference method and compared by cross-validation techniques against five existing bioimpedance and three anthropometric prediction equations. A new equation for estimating ICW from MFBIA was developed using the (40)K method as criterion method.Compared with the (40)K method, the existing bioimpedance and anthropometric equations showed large prediction errors in ICW estimates for older men (-32.3% to +37.7%) and women (-34.2% to +26.6%), depending significantly and positively on ICW volume and inversely on weight. A new equation for estimating ICW from MFBIA was developed (R(2)=0.933, standard error of the estimate (SEE)=0.92 L) involving phase angle at 5 kHz, impedance, height, and gender, with data from 100 subjects chosen at random. Cross-validation on an independent group (n=55, R(2)=0.958, SEE=0.68 L) showed no significant bias (0.013 +/- 1.52 L).Published bioimpedance and anthropometric prediction equations are not applicable to older Germans because they might be population-specific. The bioimpedance equation of the manufacturer of the bioimpedance analyzer used in this study provides accurate estimates of ICW for normal weight, but not overweight, older men. The newly developed equation improves accuracy and precision of ICW estimates by MFBIA.
科研通智能强力驱动
Strongly Powered by AbleSci AI