Parenterally fed preterm neonates are known to be at risk for carnitine deficiency. We studied substrate utilization in low-birth-weight infants receiving total parenteral nutrition (TPN) with (A) and without (B) supplementation of 48 mg carnitine ·kg-1 ·d-1 on days 4-7 (birth weights 1334 ± 282 vs 1318 ± 248 g, gestational age 32 ± 2 vs 32 ± 2 wk, A vs B, respectively). TPN consisted of 11 g glucose · kg-1 · d-1 and 2.4 g · kg-1 · d-1 of both protein and fat. Plasma carnitine concentrations at day 7 were for free carnitine 11.8 ± 5.0 vs 164 ± 56 μmol/L and for acyl carnitine 3.8 ± 2.0 vs 33.9 ± 15.4 μmol/L, respectively. Indirect calorimetry at day 7 showed a higher fat oxidation (0.21, -0.31, to +0.60 vs 1.18, 0.70 to 1.95 g · kg-1 · d-1, respectively, P < 0.02, median and interquartile range) in group B and a higher protein oxidation (0.37, 0.30-0.43 vs 0.63, 0.53-0.88 g · kg-1 · d-1, P < 0.001). The time to regain birth weight was also higher in group B (7, 5.5-9 vs 9, 7-14 d, P < 0.05). Carnitine supplementation and calorie intake were the best explanatory variables for metabolic rate (R2 = 0.45, P < 0.002). We conclude that carnitine supplementation of TPN in this dosage does not seem advisable.