This study included 249 consecutive patients with suspected significant aortic valve stenosis. Using contemporary ultrasound equipment we reassessed the value of out-patient Doppler-derived maximum pressure gradients for the prediction of peak-to-peak pressure gradients measured later at heart catheterization. It was possible to record the pressure gradient in 97% of the patients by Doppler examination and in 86% at heart catheterization. There was a fair, statistically significant correlation between Doppler and invasive gradients (n = 201, r = 0.80, p < 0.05, SEE = 21 mm Hg), independent of coexistent aortic regurgitation, atrial fibrillation, left ventricular function, number of days between the examinations and other variables analysed. A Doppler gradient > 80 mm Hg was 98% (90-100%) predictive of critical aortic stenosis (gradient > 50 mm Hg as determined by heart catheterization). Similarly, a Doppler gradient of < or = 30 mm Hg was 98% (87-100%) predictive of non-critical stenosis. In the remaining patients (53% of the population) the Doppler gradient did not contribute decisively to clinical management.