Oxygen uptake and carbon dioxide excretion during aorto-coronary bypass surgery were studied in seven patients by indirect calorimetry and compared to blood-gas based measurements. Medium-high dose fentanyl, droperidol and midazolam were used for maintaining anaesthesia. During the period of extracorporeal circulation no external oxygenator was used. Circulation was maintained by two pumps by-passing the left and right heart respectively and the patient's lungs were ventilated with O2/N2 using a Servo 900C ventilator. For indirect calorimetric measurements gas concentrations were analysed by Beckman instruments and gas volumes were measured by the Servo 900C ventilator. Oxygen uptake and carbon dioxide excretion decreased by 31% and 39%, respectively. For invasive measurements during extracorporeal circulation, arterial and venous blood gases and pump flow were used. Using pump flow instead of cardiac output when calculating oxygen uptake circumvented errors in thermodilution measurements. There was a good correlation (r = 0.88) between the invasive and the indirect calorimetric measurements. Further, there was a good correlation between naso-pharyngeal temperature and indirect calorimetric measurements of oxygen uptake (r = 0.87).