This study determines the accuracy of Doppler echocardiography (echo) for predicting the presence of pulmonary artery (PA) hypertension from Doppler PA velocity traces. The patient group included 17 patients with congenital cardiac disease who had undergone catheterization. The control group was composed of 15 normal subjects. Doppler traces were analyzed qualitatively and quantitatively. Qualitative assessment included evaluation for a negative presystolic velocity that was the equivalent of the pulmonary a wave detected by M-mode echo. Quantitative assessment included measurement of the following time intervals and ratio of intervals: preejection period (PEP), time to peak velocity (TPV), right ventricular ejection time (RVET), PEP/RVET and TPV/RVET ratios. In the patient group, systolic PA pressure ranged from 22 to 90 mm Hg (mean 50 +/- 23), and mean PA pressure ranged from 12 to 60 mm Hg (mean 32 +/- 17). Five patients had systolic PA pressures of less than or equal to 30 mm Hg and 12 had systolic PA pressures greater than 30 mm Hg. Of 15 control subjects, 14 had a negative presystolic a wave. Of 5 patients with PA pressure less than or equal to 30 mm Hg, 4 had a presystolic negative velocity, and all with higher pressures had no presystolic negative velocity. One patient with pressure less than 30 mm Hg and 2 with PA pressure greater than 30 mm Hg had indeterminate status of presystolic velocity pattern because of turbulence or baseline blanking. The best quantitative indexes for separating patients with normal PA pressure from those with elevated PA pressure were TPV and TPV/RVET, which respectively correlated negatively with systolic PA pressure (r = -0.82, standard error of the estimate [SEE] = 0.02; and r = -0.70, SEE = 0.05). These measurements also correlated negatively with mean PA pressure (r = -0.75, SEE = 0.02; and r = -0.76, SEE = 0.05). Other intervals and ratios had enough individual variability to make them less useful as predictors of PA hypertension.