Previous studies show that the doubly labeled water method is accurate for measuring energy expenditure in the adult human. To validate this method in infants, carbon dioxide production rate and energy expenditure were measured for 5 to 6 days by doubly labeled water (DLW) and periodic open circuit respiratory gas exchange (RGE) in 10 blinded studies in nine infants following abdominal surgery. Infants were maintained on consistent oral or parenteral nutrition prior to and during study. This avoided diet-related changes in baseline isotopic enrichment of body water which could theoretically contribute to significant errors in calculation of carbon dioxide production rate. For DLW, insensible water loss was assumed to be proportional to respiratory volume and body surface area, where the former was predicted from carbon dioxide production rate. Insensible water loss thus calculated averaged 18% of water turnover. Rates of carbon dioxide production measured by DLW were not significantly different from that of RGE (10.4 +/- 1.1 and 10.5 +/- 0.9 l/kg/day, mean +/- SD, respectively). Energy expenditure was calculated using respiratory quotients from dietary intake (DLW:DIET) and RGE (DLW:RGE) data. There was no significant difference between energy expenditure determined by DLW (DLW:DIET and DLW:RGE) and that measured by RGE (58.5 +/- 6.1, 56.8 +/- 6.1, and 57.3 +/- 5.1 kcal/kg/day, mean +/- SD, respectively). Rate of carbon dioxide production, DLW:diet, and DLW:RGE calculated by DLW differed from corresponding RGE values by -0.9 +/- 6.2, -1.1 +/- 6.1, and 1.6 +/- 6.2%, mean +/- SD, respectively. These findings demonstrate the validity of the doubly labeled water method for determining energy expenditure in infants without concurrent water balance studies.