We evaluated the accuracy of an indirect CO2 Fick method for measuring cardiac output in 30 critically ill mechanically ventilated patients. When the Fick principle was applied to CO2 using estimated PaCO2, the cardiac output obtained underestimated the thermodilution technique showing a lack of accuracy. However, there was a significant correlation between thermodilution and CO2 rebreathing methods using measured (r = 0.92; p less than 0.001) and estimated (r = 0.60; p less than 0.01) arterial PCO2. The regression equation using measured arterial PCO2 was y = 0.59 + 0.91x, and for estimated arterial PCO2 was y = 1.7 + 0.33x. The results suggest that the CO2 rebreathing method using measured arterial PCO2 may be useful to determine cardiac output in those seriously ill patients on artificial ventilation not requiring right heart catheterization.