1. The myocardial perfusion relative to left ventricular (LV) workload may be estimated by the subendocardial viability index (SVI). The SVI is a pressure-time integral ratio: the numerator is the area between aortic and LV pressures during diastolic time (DT) and the denominator is the area under the LV pressure curve during systolic time (ST). New non-invasive tonometric devices allow estimation of SVI but neglect LV end-diastolic pressure (LVEDP) in the calculation. The aim of the present study was to determine the haemodynamic correlates of SVI and to test the effects of neglecting LVEDP on SVI estimation. 2. High-fidelity pressures were recorded at rest at the aortic root and LV level in 38 subjects (33 men/five women; mean (+/-SD) age 47 +/- 14 years; nine controls and 29 patients with various cardiac diseases). The SVI (1.16 +/- 0.28) was positively correlated with the DT/ST ratio (1.71 +/- 0.35; r(2) = 0.81) and was negatively correlated with LVEDP (15 +/- 7 mmHg; multiple r(2) = 0.94). The SVI was not related to aortic pressure (mean, pulse, mean systolic, mean diastolic). In 17 patients with LVEDP > 14 mmHg, the SVI calculated assuming zero LVEDP was 33 +/- 15% higher (range 16-70%) than the actual SVI. 3. The DT/ST ratio was the main determinant of the myocardial perfusion relative to cardiac workload and accounted for 81% of SVI variability, whereas aortic pressure did not contribute. Although LVEDP accounted for only 13% of SVI variability, it should be taken into account in the non-invasive calculation of SVI in patients with known or suspected increases in LV filling pressure.