The Doppler echocardiography and cardiac catheterization studies of all patients who underwent valvular surgery in a three-year period were reviewed to assess the correlation between the estimated severity of valvular disease by both methods. Two-hundred and thirty-five patients (group I: 140 male, age 58 +/- 12; 95 female, age 60 +/- 13) underwent both studies within 6 months. There was agreement on estimation of severity of valve lesions in 140 of 162 patients with aortic valve disease (93% of stenosis, 82% of regurgitations and 79% of mixed lesions), in 58 of 80 patients with mitral valve disease (83% of stenosis, 76% of regurgitations and 33% of mixed lesions) and in 10 of 16 patients with prosthetic valve disfunction. The correlation between both methods was significantly lower in mixed mitral lesions than in the remaining native valve lesions (p < 0.05). Significant disagreement occurred in 4 cases of aortic valve disease, four of mitral valve disease and five of prosthetic disfunction. When disagreement was present, Doppler often underestimated the severity of the disease. Disagreement was more frequent in patients with combined aortic and mitral disease. According to the surgical conclusions cardiac catheterization provided a diagnostic profit in the assessment of the disease severity in 8, 11 and 22% of cases of aortic and mitral valve disease and prosthetic valve disfunction, respectively. Coronary artery disease was present in 19% of patients who underwent coronary arteriography. One-hundred and two patients (group II: 44 m, 48 +/- 15; 58 f, 53 +/- 11) underwent surgery without previous cardiac catheterization.(ABSTRACT TRUNCATED AT 250 WORDS)