OBJECTIVE: The throughflow method is a new technique for continuous and minimally invasive measurement of cardiac output by the Fick principle, which uses ventilation of the 2 lungs with unequal inspired gas concentrations by means of a double-lumen endobronchial tube. It exploits steady-state gas exchange and thus permits rapid repetition of measurement. DESIGN: Comparison of paired measurements by the throughflow method using N(2)O exchange with bolus thermodilution. SETTING: Departments of anesthesiology in 2 university teaching hospitals. PARTICIPANTS: Nine patients undergoing cardiac surgery in the precardiopulmonary bypass period. INTERVENTIONS: Patients intubated with a double-lumen endobronchial tube were ventilated with 45% nitrous oxide (N(2)O) to the left lung (zero to the right lung). Arterial blood gas samples were taken to measure alveolar deadspace to allow correction for the alveolar-arterial N(2)O difference and to correct for the presence of unmeasured shunt perfusion. MEASUREMENTS AND MAIN RESULTS: Throughflow measurements correlated with thermodilution (r = 0.719, p < 0.05) with a mean bias of -0.208 L/min (-5.2%). The standard error of the bias was 0.060 L/min, with 95% confidence limits for the bias of -0.088 L/min and -0.328 L/min. The limits of agreement between the 2 methods were +0.960 L/min and -1.376 L/min. CONCLUSIONS: The throughflow method showed good agreement with thermodilution. It permits continuous cardiac output measurement without the need for sampling of mixed venous blood, using techniques of lung isolation, which are readily available in clinical anesthetic practice.