Greater protein intakes are required than have been commonly used to achieve fetal in utero protein accretion rates in preterm neonates. To study the efficacy and safety of more aggressive amino acid intake, we performed a prospective randomized study in 28 infants [mean wt, 946 +/- 40 g (SEM)] of 1 (low amino acid intake, LAA) versus 3 g.kg(-1).d(-1) (high amino acid intake, HAA) at 52.0 +/- 3.0 h of life. After a minimum of 12 h of parenteral nutrition, efficacy was determined by protein balance and was significantly lower in the LAA versus HAA groups by both nitrogen balance (-0.26 +/- 0.11 versus 1.16 +/- 0.15 g.kg(-1).d(-1), p < 0.00005) and leucine stable isotope (0.184 +/- 0.17 versus 1.63 +/- 0.20 g.kg(-1).d(-1), p < 0.0005) methods. Leucine flux and oxidation and nonoxidative leucine disposal rates were all significantly higher in the HAA versus LAA groups (249 +/- 13 versus 164 +/- 8, 69 +/- 5 versus 32 +/- 3, and 180 +/- 10 versus 132 +/- 8 micro mol.kg(-1).h(-1), respectively, p < 0.005), but leucine appearance from protein breakdown was not (140 +/- 15 in HAA versus 128 +/- 8 micro mol.kg(-1).h(-1)). In terms of possible toxicity with HAA, there were no significant differences between groups in the amount of sodium bicarbonate administered, degree of acidosis as determined by base deficit, or blood urea nitrogen concentration. Parenteral HAA versus LAA intake resulted in increased protein accretion, primarily by increasing protein synthesis versus suppressing protein breakdown, and appeared to be well tolerated by very preterm infants in the first days of life.