Evaluation of a mathematical model for predicting the relationship between protein and energy intakes of low-birth-weight infants and the rate and composition of weight gain

A model for predicting the relationship between protein and energy intakes of low-birth-weight (LBW) infants and the rate and composition of weight gain is described. It is based on linear multiple regression equations summarizing the rates of weight gain, nitrogen retention, and energy retention of 101 previously studied LBW infants fed protein intakes ranging from 2.25 to 3.9 g.kg-1.d-1 and concomitant energy intakes ranging from 115 to 147 kcal.kg-1.d-1 plus current theory concerning nutrient retention and body composition. To test the validity of the model, three combinations of protein and energy intake predicted by the model to result in specific rates and compositions of weight gain were fed to 44 LBW infants, and the observed rates of weight gain, protein accretion, and fat accretion were compared with the rates predicted by the model. Differences in these and other outcome variables between two of the groups, the intakes of which differed only in energy, also were compared to provide additional insight into the effect of concomitant energy intake on protein utilization. Across groups, actual outcomes correlated closely with predicted outcomes, supporting the validity of the model for the total population. However, outcomes of individual infants deviated as much as 30% from predicted outcomes; the magnitude of the deviation was independent of birth weight, gestational age, or size for gestational age.