During the past years, substantial methodological and interpretational limitations of indirect calorimetry, particularly concerning fuel utilisation, have been discussed. The aim of the present study was to evaluate short-time intraindividual variability of two consecutive gas exchange measurement series and of calculated data on total energy expenditure/24 h and carbohydrate and fat utilisation. 24 healthy volunteers (16 f, 8 m, 34.7 +/- 13.1 yrs) were admitted to the study. Trials were performed supine after an 12 h overnight fast. After a resting period of 30-45 min and following equilibration of respiratory values for at least 10 min prior to the test, indirect calorimetry measurements were performed using the Sensor-Medics 2.900 device (canopy). Two measurements series lasting up to 30 min each were performed 15-20 min apart. Total energy production/24 h as well as that obtained from carbohydrate and fat utilisation were calculated in both measurement series. Protein utilisation was derived from estimated urinary 24 h nitrogen excretion. O2-consumption, CO2-production, the respiratory quotient and total energy production/24 h show acceptable mean coefficients of variation of 3.7%, 4.6%, 3.5% and 3.6%, respectively. In contrast, carbohydrate and fat utilisation values demonstrate a coefficient of variation of 21.2% and 17.4%, respectively, suggesting considerable impression of estimates of fuel utilisation by indirect calorimetry. We conclude that for research purposes, particularly over short-time periods, indirect calorimetry provides sufficient accuracy only in estimating total resting energy production, while considerable uncertainty exists in using this method to assess carbohydrate and fat utilisation.