Background This study evaluated the correlation and variability between noninvasive and invasive measures of mitral stenosis severity before and after balloon mitral commissurotomy (BMC) in a large group of patients with symptomatic mitral stenosis. Factors related to variability between measurements were determined. Methods The Doppler transmitral gradient, Doppler half-time valve area, and 2-dimensional echocardiographic (2D) mitral valve area (MVA) were measured immediately before and 1 day after BMC in 272 consecutive patients with mitral stenosis and compared with their respective measures during cardiac catheterization. Results The correlation coefficient for the comparison of noninvasive and invasive measurements of the transmitral gradient was 0.63 before BMC and 0.60 after the procedure; for 2D versus Gorlin-derived MVA, 0.39 and 0.57, respectively; and for Doppler half-time versus Gorlin-derived MVA, 0.31 and 0.18, respectively. A large degree of variability in the measurement of MVA was present among the 3 techniques before BMC and increased after BMC. Before BMC, for the comparison of 2D and Gorlin-derived MVA, variables predictive of the discrepancy were age, echocardiographic score, transmitral gradient during catheterization, and cardiac index. For the comparison of Doppler half-time versus Gorlin-derived MVA, age, heart rate during cardiac catheterization and echocardiography, cardiac output and left ventricular end-diastolic pressure predicted the difference between the 2 measures. Conclusions In symptomatic patients with mitral stenosis, there is significant variability between noninvasive and invasive measures of mitral stenosis severity despite careful, reproducible measurements. The difference between noninvasive and invasive measures of MVA before BMC is strongly related to cardiac output. (Am Heart J 1999;138:777-84.)