The peak instantaneous aortic valve gradient derived from Doppler echocardiography is commonly used to predict the severity of aortic stenosis. Peak instantaneous gradient should not be equated with the mean gradient or "peak to peak" gradient measured at cardiac catheterization. The primary purpose of this study is to assess the relationship between the aortic valve gradients, using a two-catheter transseptal technique in 102 patients with aortic stenosis, mixed aortic stenosis and regurgitation, and following aortic valve replacement. These cases were drawn from a series of 111 consecutive transseptal procedures for patients with isolated aortic valve disease. No major complications occurred, and the most common reason for technical failure was inability to engage the atrial septum in postoperative patients. Although the peak instantaneous gradient correlates well with the mean gradient in aortic stenosis (r = .94, P less than .001), mixed stenosis and regurgitation (r = .95, P less than .001), and after aortic valve replacement (r = .86, P less than .001), it systematically overestimates both the mean gradient and the peak to peak gradient. Neither the peak instantaneous nor the mean gradient correlates highly with aortic valve area in aortic stenosis (r = -.48, P less than .01 peak; r = -.58, P less than .001 mean gradient), mixed aortic stenosis and regurgitation (r = -.39, P NS peak; r = -.42, P NS mean gradient) or following aortic valve replacement (r = -.26, P NS peak; r = -.53, P less than .01 mean gradient). Systolic time intervals also were analyzed from the simultaneous left ventricular and ascending aortic pressure tracings.(ABSTRACT TRUNCATED AT 250 WORDS)