BACKGROUND: The timing of mitral valve (MV) surgery to preserve left ventricular (LV) contractility in patients with mitral regurgitation (MR) has been defined by complex cardiac catheterization techniques. Whether noninvasive methods can identify patients with MR, a normal LV ejection fraction, and early LV contractile impairment is unknown. We hypothesized that echocardiographic measures would separate patients with MR and a normal LV ejection fraction into those with and without contractile dysfunction and, thus, prospectively predict the response of LV size and performance to MV surgery. METHODS AND RESULTS: We studied 27 patients with micromanometer LV pressures and radionuclide angiography to obtain a determination of LV volumes and ejection fraction and calculate chamber elastance, a measure of LV contractility, before MV surgery. Echocardiographic studies were performed before MV surgery and repeated at 3 and 12 months after surgery. Age, New York Heart Association class, LV plus maximum pressure per unit change in time, LV systolic and end-diastolic pressures, and echocardiographic posterior wall thickness and radius to wall thickness ratio did not identify preoperative LV contractile dysfunction. However, other echocardiographic measures were related to LV contractility, including LV end-diastolic dimension (r = -0.50, P <.005), LV end-systolic dimension (r = -0.60, P <.0001), and LV fractional shortening (r = 0.50, P =.005). From analysis of receiver operator characteristic curves, an LV end-systolic dimension of >/=40 mm was identified as most predictive for separating patients with MR before surgery into those with and without LV contractile dysfunction (sensitivity of 82% and specificity of 100%). The patients with MR and impaired preoperative LV contractility showed a dramatic deterioration in LV fractional shortening at 3 months after MV surgery (P =.01), which recovered to within the normal range for fractional shortening at 12 months (P =.02) from a progressive reduction in LV end-systolic dimension. This response in LV size and performance temporally differed from that in the patients with MR and normal contractility (2-way analysis of variance P <.0001). However, at 12 months after MV surgery, LV end-diastolic dimension, end-systolic dimension, and fractional shortening were normal in both groups of patients with MR. CONCLUSION: We conclude that echocardiographic measures, particularly an end-systolic dimension of >/=40 mm, may be useful for identifying patients with MR before surgery with early, occult LV contractile dysfunction in whom MV surgery may be recommended to preserve LV systolic performance.