SummaryMyocardial oxygen consumption indices that are frequently applied to man such as tension-time index (TTI), pressure-rate product (P 9 HR) and triple product (TP) have not been fully validated so far. These easily obtainable indices and a modified TTI (P -x/H--R), therefore, were examined in I0 closed-chest dogs with very broad variations of hernodynarnics and oxygen consumption (3-36 rnl/rnin -100 g) analyzing 162 steady states. Myocardial blood flow was directly measured by a differential pressure coronary sinus catheter. I~O 2 was varied by administration of catecholamines and other inotropic drugs, atropine, beta-blocking agents and hypoand hypervolernia. Over a wide range of hernodynarnic states, correlations with directly measured MVO 2 of TTI (r = 0.63), P 9 HR (r = 0.87), TP (r = 0.65) and P 9 %/HR (r = 0.80) are not satisfactory due to neglect of contractility and cardiac volumes by these terms. Better correlations are obtained when relating these indices to IV[Vo2 under different inotropic states. At normal and moderately increased contractility, correlations with M~O2 rose as follows: TTI (r = 0.96), P 9 HR (r = 0.91), TP (r = 0.96) and P 9 ~ (r = 0.94). Significant rises in correlation are due to the close relationship between peak pressure and dP/dtma x at only moderately increased contraction velocity. Correlation differences within this inotropic range must be related to incorporation or neglect of ejection time as a partial determinant of MVO 2. At markedly increased contractility, results for these indices, however, are in part very poor: TTI (r = 0.40), P -HR (r = 0.81), TP (r = 0.38) and P 9 ~/HR (r = 0.76). Within this inotropic state neglect of dP/dtma x as a rna]or determinant of MVO 2 and the inverse relationship between ejection time and dP/dtma x mainly account for these correlation shifts. It is concluded that non-invasively obtainable indices, currently in use, are no reliable predictors of actual overall MVO2 of the left ventricle if the contractile state of the rnyocardium is not checked invasively before. The broad variability of the relation of the energy demand of velocity of tension development to rnaintenance of systolic wall tension is not sufficiently considered by these terms. Appropriate caution, therefore, is necessary when applying those indirect indices of M'~rO2 to humans.