Four ultrasound methods of estimating the severity of mitral narrowing (MN) were investigated in a prospective manner in 36 patients who underwent ultrasound-Doppler and catheterization (KT) within a space of 24 hours. The correlation of the mean Doppler and KT gradients was poor: r = 0.56 (n = 36, p less than 0.001). The mitral surface areas (MA), calculated from Gorlin's equation (mean A = 0.98 +/- 0.26 cm2) were successively compared with those obtained using the Hatle equation (A = 220/PHT) and by a planimetric determination. The coefficients of correlation were as follows: r = 0.80 (n = 36) and r = 0.70 (n = 32) (p less than 0.001). There was good correlation between the Hatle equation and the planimetric determination: r = 0.80 (n = 32, p less than 0.001). The continuity equation was carried out 15 times; the second measurement site was aortic in 8 cases and pulmonary in 7 cases. Correlations with the Gorlin equation were scored 0.70 and 0.80 respectively (p less than 0.05). The MA was best estimated by the Hatle equation, which was always technically feasible and not influenced by the presence of atrial fibrillation, mitral valve incompetence of previous commissurotomy. Planimetric determination, which is hampered in the presence of considerable calcification, remains valid after commissurotomy. The ultrasound estimation of the MA was very reliable when the planimetric and Hatle estimations concurred. The results obtained from the continuity equation were reliable; however, this method is slow, unreliable in a context of atrial fibrillation and inapplicable in a context of mitral valve incompetence.(ABSTRACT TRUNCATED AT 250 WORDS)