Mitral regurgitation (MR) imposes a volume work load on the left ventricle (LV) that is proportional to the severity of the regurgitation. However, it is not clear how semiquantitative grading 1–4 + is related to quantitative regurgitant flow. Method 74 patients with MR were investigated using multiplane transesophageal echocardiography. MR was graded semiquantitatively on a 4 point scale by two observers, integrating information on jet size and morphology in multiplane views and pulmonary venous flow patterns. Blinded to the results of the semiquantitative MR grading, regurgitant stroke volume (RSV) was calculated as the difference between thermodilution stroke volume and forward stroke volume through the mitral annulus (SV ma), obtained by the mitral annulus area multiplied by mitral inflow velocity-time integral recorded by pulsed Doppler at the mitral annulus. Assuming elliptical shape, the mitral annulus area (A) was calculated as A = Jrab, where a and b were radius of mitral annulus measured by 2-dimensional echocardiography at four and two chamber views. Results Semiquantitative MR grading correlated best with mean RSV (r = 0.97 P < 0.001 Fig.) using an exponential fit. RSV dramatically increased when MR grading was >2+ with significantly increasing slope. 4+ MR (81 ± 36 ml) involved much more than double RSV as 2+ MR (19 ± 12 ml p < 0.001). Conclusion Because its relationship with regurgitant flow is nonlinear, the semiquantitative MR grading score does not accurately reflect the amount of volume load on the LV, especially above MR grading 2 +.