Desarrollo y validación de ecuaciones para estimar composición corporal en niños de 4 a 6 años de Uruguay
Development and validation of equations for estimating body composition in 4-6 year old children from Uruguay
Rev. chil. nutr; 44 (1), 2017
Publication year: 2017
A correct diagnosis of obesity requires estimating body composition since it is linked to excess fat. Therefore it is necessary to have practical and sensitive methods of measurement. The aim of this study was to develop and validate bioelectrical impedance (BIA) and anthropometric equations to estimate fat mass (FM) and fat-free mass (FFM). Anthropometric measurements, BIA and deuterium dilution technique as the reference standard were performed in 191 children (4-6 years). Two equations were developed using multiple regression models and were validated in a subsample. Bland and Altman was applied to evaluate the bias differences between the means of both methods. The BIA equation was FFM (kg)= 2,089 + (0,346 * Fieight²/Resistance) + (0,312 * Weight) + (0,660 * Sex); with R2= 0.92, SRMSE 0,67 kg, bias 0,140 kg and precision 0,73 kg. The anthropometric equation was FM (kg)= 0.841 + (0,408 * Weight) + (0,137 * Subscapular-Skinfolds-Thickness) + (0,104 * Tricipital-Skinfolds-Thickness) - (0,764 * Sex) - (0,054 * Feight); with R2= 0.89, SRMSE 0,68 kg, bias -0,01 kg and precision 0,76 kg. The equations validated had a good predictive capacity and can be especially useful in epidemiological studies.
Un correcto diagnóstico de obesidad requiere estimar la composición corporal pues ésta se vincula al exceso de grasa, siendo necesario disponer de métodos prácticos y sensibles para su medición. El objetivo fue desarrollar y validar ecuaciones antropométricas y de bio-impedancia para estimar masa grasa (MG) y masa libre de grasa (MLG). Se realizaron mediciones antropométricas, de bio-impedancia y se aplicó la técnica de dilución con óxido de deuterio como patrón oro a 191 preescolares. Se desarrollaron dos ecuaciones utilizando modelos de regresión múltiple y se validaron en una submuestra. Se aplicó Bland-Altman para evaluar el sesgo entre las diferencias de las medias de ambos métodos.