The doubly labeled water (2H(2)18O) method used to estimate total energy expenditure (EETotal) is particularly sensitive to analytic error in preterm infants, because of their high percentage of body water and the high ratio of water flux to CO2 production. To evaluate further use of this method, the EE of 12 preterm infants was measured by indirect calorimetry and 2H(2)18O simultaneously and continuously for 5 days. Initial infant weight, age, and postconceptional age were (means +/- SD) 1,674 +/- 173 g, 4.4 +/- 2.6 wk, and 34.6 +/- 1.6 wk, respectively. The indirect calorimeter system included an air-temperature-controlled chamber and heart rate monitor. EE was measured by indirect calorimetry for 85.6 +/- 4.7% of study time and estimated from the linear regression of heart rate on EE for 14.4 +/- 4.7% of study time. The 2H(2)18O method entailed an initial dose of 100 mg 2H2O and 250 mg 18O/kg and a final dose of 75 mg 18O/kg; urine was collected twice daily. 2H and 18O enrichments were measured by gas-isotope-ratio mass spectrometry. EE was calculated from measured 2H and 18O dilution spaces (NH, NO), turnover rates (kH, kO), and measured respiratory quotient. The ratio of 2H to 18O dilution spaces was 1.01 +/- 0.01 and the ratio of kO to kH was 1.16 +/- 0.04. Estimation of EE from 2H(2)18O and indirect calorimetry agreed within 1%, although individual variability in methods was large.