Impaired relaxation is frequently masked by elevated filling pressures, resulting in a pseudonormal flow pattern (E/A >1.0). Because the E/A wave ratio increases as filling pressures rise, it is generally assumed that patients with an E/A ratio of <1.0 (impaired relaxation pattern) have relatively low filling pressures. Nevertheless, patients with an E/A ratio of <1.0 can have as profoundly elevated filling pressures as patients with a pseudonormal or restrictive filling pattern. Because left ventricular (LV) pressure during end-diastole essentially determines atrial afterload, the response of the A-wave velocity to a reduction of atrial afterload by a standardized Valsalva maneuver should allow estimation of LV end-diastolic pressure (LVEDP) regardless of the baseline Doppler flow pattern. This was tested in 20 consecutive patients who were studied by pulse-wave Doppler echocardiography during cardiac catheterization. There was a close correlation between LVEDP and the change in A-wave velocity during the Valsalva maneuver (r = 0.85, SEE 6.7 mm Hg) regardless of the baseline E/A ratio. In patients with a LVEDP of <15 mm Hg the A wave decreased by 21 +/- 15 cm/s. In patients with a LVEDP of >25 mm Hg the A wave increased by 18 +/- 13 cm/s. The change in the E/A ratio during Valsalva correlated fairly with LVEDP (r = -0.72, SEE 8.8 mm Hg), the baseline E/A ratio correlated poorly, and scatter was substantial (r = 0.46, SEE 11.2 mm Hg). Just as elevated filling pressures can mask impaired relaxation, the impaired relaxation pattern can mask the presence of elevated filling pressures. This can be revealed by testing the response of the A wave to the Valsalva maneuver, allowing estimation of LVEDP independent of the baseline E/A ratio.