OBJECTIVE We performed this study to validate the single-beat estimation of end-systolic elastance (E(es)) in an animal model of chronic mitral regurgitation (MR) and to use E(es) to predict postoperative left ventricular (LV) systolic function with the noninvasively derived single-beat E(es) (E(es-s)) in patients with MR. METHODS Eight sheep with MR were studied under 4 different loading conditions. E(es) was measured as the slope of the end-systolic pressure-volume relationship with a conductance catheter during inferior vena cava occlusion. E(es-s) was calculated using a bilinearly approximated time-varying elastance curve. We also studied 105 patients with MR who had undergone mitral valve repair. All measurements in these patients were performed before surgery and repeated at about 1 week after surgery. RESULTS In the animals in a total of 23 different stages, E(es-s) showed a good correlation and agreement with invasively determined E(es) (y = 0.84 x + 0.38; r = 0.8; P < .01, mean difference = 0.1 +/- 0.6 mm Hg/mL). In the clinical study, preoperative E(es-s) (r = 0.77; P < .001) and end-systolic volume index correlated well with postoperative ejection fraction (r = -0.69; P < .001). Multivariate analyses revealed that only E(es-s) was an independent predictor of postoperative LV function. Preoperative E(es-s) less than or equal to 1.0 mm Hg/mL was most predictive for identifying patients with LV dysfunction after surgery (sensitivity 87%; specificity 76%). CONCLUSIONS LV end-systolic elastance can be estimated noninvasively and is useful to detect latent LV dysfunction in patients with MR before surgery.