Aortic valve replacement in severe aortic stenosis with left ventricular dysfunction: determinants of cardiac mortality and ventricular function recovery

Objective: The influence of left ventricular (LV) dysfunction on survival of patients with severe aortic stenosis is poorly characterized. Few data are available about preoperative predictors of cardiac mortality and LV function recovery after aortic valve replacement of such patients. The aim of our study was to examine the outcome and the preoperative predictors of postoperative cardiac death and of LV function recovery in these patients. Methods: We evaluated 85 consecutive patients with severe aortic stenosis (aortic valve area < 1cm(2)) and severe depression of LV ejection fraction (EF) < 35% at cardiac catheterization. Among them, 52 underwent aortic valve replacement and they were compared to patients who were not operated on. All patients had a mean clinical follow-up of 53 months and 94% of them had a mean echocardiographic follow-up of 14 months after aortic valve replacement. Results: The mean baseline characteristics included: LVEF 28 +/- 6%, peak-to-peak transvalvular gradient 51 +/- 29 mmHg, aortic valve area 0.63 +/- 0.25 cm(2). Thirty-three patients did not undergo aortic valve replacement: 32 of them died within 3 years. Fifty-two patients underwent aortic valve replacement and 16 had a concomitant coronary bypass surgery. In-hospital mortality was 8%. Postoperative NYHA functional class changed from 2.84 +/- 0.67 to 1.43 +/- 0.44 (P < 0.001) and LVEF from 29 6% to 43 +/- 10% (P < 0.001). At follow-up 10 patients died of heart disease. By multivariate analysis, preoperative LV end-systolic volume index (ESVI) was the only covariate of cardiac death (LVESVI/10 ml/m(2), OR 1.3, Cl 1.1-1.8, P < 0.028). By using a receiver operating characteristic curve, LVESVI less than or equal to 90 ml/m(2) was the best cut-off value (sensitivity and specificity 78%) to fit with a better survival (93% vs. 63%, P < 0.01) and with LVEF recovery after aortic valve replacement (EF improved by 15 +/- 10% vs. 8 5%, P < 0.001). Conclusions: Despite LV dysfunction, aortic valve replacement appears to change drastically the natural history of severe aortic stenosis. Preoperative LV levels predict different postoperative survival rate and LVEF recovery. (C) 2003 Elsevier B.V. All rights reserved.