Growth, protein, and energy balances were studied in two groups of very-low-birth-weight premature infants fed pooled pasteurized human milk (HM) or a preterm formula (PF). Each infant was studied at 33 and 36 weeks gestational age with a combined technique of nutrient balance and indirect calorimetry measurement. Weight and length gains were higher with PF than with HM, but head circumference growth was similar with both milks. Although the volume of milk given was lower, energy intake was higher with PF than with HM in both studies (126 to 130 vs 103 to 109 kcal/kg/day). Percentage of energy absorbed was better with PF than with HM (94% vs 84%) at 33 weeks, and similar (95%) with both milks at 36 weeks. Energy expenditure, which had increased from 33 weeks to 36 weeks, was higher with PF than with HM (57 to 63 vs 46 to 52 kcal/kg/day) during both studies. Energy retention accounted for about 50% of energy absorbed with both milks, but was higher with PF than with HM (60 vs 40 to 50 kcal/kg/day) in both studies, and resulted in fat accretion well above that seen during intrauterine growth for both milks in both studies. Protein intake was higher with PF than with HM (3.1 vs 2.4 gm/kg/day) in both studies, giving a protein accretion similar to the intrauterine accretion with PF (2.2 gm/kg/day), but lower with HM. Our results suggest that the nutritional value of pooled pasteurized human milk for VLBW infants should be reconsidered, especially because of its low protein content, and that energy density of preterm formulas must be questioned in view of elevated fat deposition.