The feasibility of the thermodilution technique to measure cardiac output (CO) during positive pressure ventilation was assessed in 12 critically ill patients. An indicator (5 ml of 5% glucose in cold water) was injected at the mid and end of the inspiratory and expiratory phases of mechanical ventilation to see the effects of the respiratory cycle on thermodilution data. The 36 measurements at mid-inspiration yielded the smallest coefficient of variation (CV), 2.7%. The CV of 36 randomly selected measurements was 6.0%. CO at end-inspiration gave the highest value (p less than 0.05). These data suggested that one major disadvantage of the thermodilution method, the need for repeated measurement to minimize the variation in data, could be overcome by timing the injection according a certain phase of mechanical ventilation.