Hemodynamic indices of myocardial oxygen consumption (MVO2) were examined in 13 patients with coronary heart disease. The specific aim of this study was to investigate whether clinical results agree with experimental findings in animals. During hemodynamic and respiratory steady-state conditions at rest, the MVO2 (7.6-14.2 ml/min/100 g) was measured directly by myocardial blood flow (argon method) and arterio-coronary venous oxygen content difference. MVO2 was compared with five indirect indices of MVO2. A complex additive parameter consisting of five components of MVO2 had the highest correlation with MVO2 (r = 0.97), as was also demonstrated in a recent experimental study at maximum changes in hemodynamics and MVO2. More easily measurable predictors of MVO2 such as tension-time index (r = 0.923), product of mean systolic aortic pressure and square root of heart rate (r = 0.928), pressure rate product (r = 0.915), and triple product (r = 0.941) were less closely correlated with MVO2. The lower correlations of the readily obtainable indices of MVO2 are probably related to their failure to incorporate factors such as contractility and ventricular dimensions, which are known to exert an important influence on MVO2. The excellent correlation of the hemodynamic additive parameter with MVO2 supports the theoretical concept and the implications of the experimental study. The accurate prediction of MVO2 is based on adequate measurement of MVO2 for the velocity of tension development and maintenance of tension during systolic ejection period, both of which are integrated in the additive index.