A triple lumen catheter was used for the determination of cardiac output during the acute phase of moycardial infarction in 23 patients. The correlation coefficient between thermodilution and the Fick method was 0.91. Precision of the method assessed by multiple repeated determinations was 4.8 or 5.6%. Quadruplicate determinations allowed to detect variations of cardiac output in the range of 0.465 1/min. Variation between two successive determinations of more than 9.4% can probably be attributed to a technical error.