To evaluate the hemodynamic impact of coronary stenoses, the fractional (FFR) or coronary flow velocity reserve (CFVR) usually are measured. The combined measurement of instantaneous flow velocity and pressure gradient (v-dp relation) is rarely used in humans. We derived from the v-dp relation a new index, dp_v50 (pressure gradient at flow velocity of 50 cm/sec) and compared the diagnostic performance of dp_v50, CFVR and FFR. Prior to coronary angiography patients underwent non-invasive stress testing. In all coronary vessels with an intermediate or severe stenosis, the flow velocity, aortic and distal coronary pressure were measured simultaneously with a Doppler and pressure guide wire after induction of hyperemia. After regression analysis of all middiastolic flow velocity and pressure gradient data, the dp_v50 was calculated. Using the results of non-invasive stress testing, the dp_v50 cut-off value was established at 22.4 mm Hg. In 77 patients, 124 coronary vessels with a mean 39±19% diameter stenosis were analyzed. In 43 stenoses ischemia was detected. We found a sensitivity, specificity and accuracy of 56%, 86% and 76% for CFVR, 77%, 99% and 91% for FFR and 95%, 95% and 95% for dp_v50. To establish that dp_v50 is not dependent on maximal hyperemia, dp_v50 was recalculated after omission of the highest quartile of flow velocity data, showing a difference of 3%. We found that dp_v50 provided the highest sensitivity and accuracy compared to FFR and CFVR in the assessment of coronary stenoses. In contrast to CFVR and FFR assessment of dp_v50 is not dependent on maximal hyperemia.