To assess late (4 to 5 years) gradient reduction after pulmonary balloon valvuloplasty in childhood, and to compare the effectiveness of valvuloplasty with that of surgical valvotomy, 20 valvuloplasty-treated children and their age- and gradient-matched surgical control patients underwent prospective, noninvasive evaluation. The average age at intervention was 4.3 +/- 1 years for the valvuloplasty group versus 4.7 +/- 0.8 years for the surgical control group (p = NS). Before intervention the peak systolic pulmonary stenosis gradient was 76 +/- 5 and 74 +/- 4.4 mm Hg for the valvuloplasty and surgery groups, respectively (p = NS). Late evaluation consisted of clinical examination, two-dimensional echocardiogram and Doppler study, 24-hour Holter monitoring, 12-lead electrocardiogram, exercise treadmill study and chest radiograph performed an average of 5.3 +/- 0.3 years after valvuloplasty and 11.7 +/- 0.5 years after surgery (p less than 0.01). The patients treated with balloon valvuloplasty had no evidence of restenosis; the residual pulmonary stenosis gradient at follow-up was 24 +/- 2.7 mm Hg (range 8 to 48) versus 35 +/- 3.6 mm Hg (range 19 to 70) immediately after valvuloplasty (p = NS). Comparison of the late residual gradients between treatment groups showed no hemodynamically significant difference (24 +/- 2.7 versus 16 +/- 1.5 mm Hg, balloon versus surgery; p less than 0.01). However, there was, a significant difference in the degree and severity of pulmonary valve insufficiency and ventricular ectopic activity between groups. In the balloon valvuloplasty group 11 patients had no pulmonary insufficiency, and the remaining 9 had mild insufficiency.(ABSTRACT TRUNCATED AT 250 WORDS)