Body cell mass by anthropometric measures

Abstract

The depletion of body cell mass (BCM), the metabolically active tissue of the body, is an important feature of protein energy malnutrition AIDS and and aging. The gold standard method for BCM estimation, total body potassium counting (TBK), has limited availability and is a very expensive method. Hence we derived predictive equations for BCM, using anthropometric measures. We carried out a retrospective, cross-sectional analysis in a sample of 1379 adult subjects, with 968 healthy and 411 HIV-infected. The entire sample was divided 2:1 into model-development and cross-validation group, respectively. TBK was measured by 4 whole-body liquid scintillation counting. Anthropometric measurements were performed by trained observers and included 10 skinfolds, 2 body lengths, 8 body circumferences, height and weight. A backward-stepwise multiple regression analysis was performed to predict Log10 BCM. The model presented two specific equations, one for males (2= 0.805, SEE=1.08 kg (3.29%), p<0.001) and other for females (2= 0.618, SEE=1.10 kg (4.98%), p <0.001). Validation tests displayed an 2= 0.724, PE=1.09 kg (3.22%) and 2= 0.642, PE=1.10 kg (5.06%), for males and females respectively. Males equation had a p = 0.814 in the Student`s paired t-test, while females equation had a p= 0.715. The test of agreement between different health status and the difference of residuals between all races, had both a p >0.05 in cross-validation group and in the entire sample. These results results express a great precision, performance, and accuracy of the model and don`t show any bias.