This study investigated the accuracy of a new technique for measuring cardiac output using the derivative Fick principle based on the ratio of change in the partial pressures of end-tidal and mixed expired carbon dioxide produced by short periods of partial rebreathing. A prospective clinical study involving 24 patients following cardiopulmonary bypass for coronary artery bypass grafting or valvular surgery was undertaken in the intensive care unit of a university-affiliated hospital. Haemodynamic measurements were performed after admission to the intensive care unit. Cardiac output was measured simultaneously by bolus pulmonary artery thermodilution and by a non-invasive carbon dioxide partial rebreathing technique. Cardiac output measurement using the new technique demonstrated a significant but consistent underestimate, with a bias of -0.60 +/- 0.87 l/min. This new adaptation of the partial rebreathing technique is reliable in measuring cardiac output in postoperative patients. Reasons for the consistent discrepancy between thermodilution and partial rebreathing techniques are discussed.