OBJECTIVE--To assess the left ventricular pressure pulse, in particular its time course, reconstructed from the continuous wave Doppler signal of functional mitral regurgitation using the simplified Bernoulli equation. DESIGN--Prospective study with simultaneously recorded high fidelity left ventricular pressure and continuous wave Doppler traces of functional mitral regurgitation, along with indirect left atrial pressure, electrocardiograms, and phonocardiograms. SETTING--Tertiary referral cardiac centre. PATIENTS--9 patients (age 60 (17) years) were studied immediately before or 1-20 h after routine cardiac surgery. RESULTS--104 cardiac cycles were analysed. There were no consistent differences between directly measured and reconstructed pressures in the time intervals from Q to + dP/dt (mean (SD) 125 (35) v 130 (35) ms and from Q to -dP/dt (389 (30) v 387 (28) ms or from Q to maximum pressure (267 (40) v 270 (40) ms, all P = NS). The time from Q to the onset of pressure rise (67 (30) v 64 (30) ms, P < 0.01) and the duration of total left ventricular systole (404 (50) v 408 (50) ms, P < 0.01) measured by the two methods were effectively identical, though the small difference was consistent enough to be statistically significant. The calculated peak pressure drop between the left ventricle and the left atrium (45-100 mm Hg) significantly underestimated left ventricular pressure (72-150 mm Hg; 70 (11) v 105 (15) mm Hg, P < 0.01) even if mean left atrial pressure (14 (4.0) mm Hg) was taken into account. Compared with those directly derived from left ventricular pressure, values of pressure measured at + dP/dt (26 (6.5) v 53 (10) mm Hg, P < 0.01) and -dP/dt (30 (8.0) v 60 (10) mm Hg, P < 0.01), and those of the rates of increase (675 (155) v 815 (155) mm Hg/s, P < 0.01) and fall (610 (145) v 845 (175) mm Hg, P < 0.01) were all significantly underestimated by Doppler. The underestimation in peak rates of pressure change could not entirely be explained by a scaling effect of absolute pressure. To investigate interrelations between the two methods throughout the cardiac cycle, reconstructed left ventricular pressure was plotted against the direct record. The plots confirmed that the reconstructed pressure was always less than directly measured pressure, the relative degree of underestimation falling as the pressure rose. This was not the effect of acceleration but probably reflects changing geometry of the regurgitant orifice. CONCLUSION--The continuous wave Doppler trace of functional mitral regurgitation is suitable for studying the timing of overall mechanical events and normalised rates of change of pressure in the left ventricle. Estimates of atrioventricular pressure drop by this method and particularly its absolute rates of change seem to be less reliable.