ANTHROPOMETRIC ESTIMATION OF NEONATAL BODY-COMPOSITION

OBJECTIVE: Estimation of neonatal body composition can be useful in th e understanding of fetal growth. However, body composition methods suc h as total body water and total body electric conductivity are expensi ve and not readily available. Our primary purpose was to develop an an thropometric model to estimate neonatal body composition and prospecti vely validate the model against total body electric conductivity and s econdarily to compare our anthropometric model and a previously publis hed anthropometric formula with total body electric conductivity. STUD Y DESIGN: A total of 194 neonates had estimates of body composition ac cording to total body electric conductivity (group 1). Parental morpho metrics, gestational age, race, sex, parity, and neonatal measurements including birth weight, length, head circumference, and skinfolds (tr iceps, subscapular, flank, acid thigh) were correlated with body fat b y use of stepwise regression analysis. The model was validated in a se cond group of 65 neonates (group 2). RESULTS: There were no significan t differences in any of the parental or neonatal measurements between groups 1 and 2. In group 1, 78% of the variance in body fat with the u se of total body electric conductivity was explained by birth weight, length, and flank skinfold (R(2) = 0.78, p = 0.0001). When prospective ly validated by the subjects in group 2, the model had significant and stronger correlation (R(2) = 0.84, p = 0.0001) with body fat estimate d by total body electric conductivity as compared with the other anthr opometric model (R(2) = 0.54, p = 0.0001). There was no significant (p = 0.11) difference between our anthropometric estimate of body fat an d total body electric conductivity. CONCLUSIONS: The anthropometric mo del developed can be used to reasonably predict neonatal body fat mass at birth.