STUDY OBJECTIVE - The aim was to study the feasibility of measuring heat production by the human left ventricle with a view to using this variable as an index of left ventricular mechanical efficiency. DESIGN - The transcoronary temperature difference was derived from catheter mounted thermistors placed percutaneously in the aortic root and coronary sinus. Left ventricular blood flow was measured by continuous thermodilution in the coronary sinus, and heat removal by coronary venous blood was calculated from blood flow and the transcoronary temperature difference. Diffusional heat loss was measured using temperature/time curves recorded in aorta and coronary sinus after a bolus injection of cold saline into pulmonary artery. The heat loss from the system into the endothermic reactions of haemoglobin was calculated from left ventricular oxygen extraction using an assumed respiratory quotient. The energy released by left ventricular myocardial metabolism (EEO2, calculated from oxygen extraction), was compared to measured left ventricular heat production, and the mechanical efficiency of the left ventricle was calculated by the formula: Efficiency = (EEO2 - HLV)/EEO2. PATIENTS - Fifteen conscious patients with anginal chest pain were studied at the time of cardiac catherisation and coronary arteriography. MAIN RESULTS - The transcoronary temperature difference was in the range 0.10-0.32 (mean 0.21) degrees C. Total left ventricular heat production, equal to the sum of heat removed by the blood stream, diffusional loss and endothermic reactions (HLV) was in the range 1.5-4.6 (mean 2.7) watts. The values of EEO2 obtained were in the range 2.4-6.5 (mean 4.0) watts, and the calculated mechanical efficiency of the left ventricle was 0.24-0.55 (mean 0.34). CONCLUSIONS - The measurement of heat production by the human left ventricle is safe and practical. This technique promises to be of value in the clinical investigation of the relationship between myocardial function and energy utilisation.