Exercise training improves left ventricular diastolic filling in patients with dilated cardiomyopathy : clinical and prognostic implications

Patients with dilated cardiomyopathy (DCM) often have left ventricular (LV) diastolic dysfunction that can precede the development of systolic dysfunction. Recent reports showed that exercise training (ET) improves the exercise capacity of these patients. Although this improvement is primarily due to peripheral adaptations, the contribution of LV diastolic filling has not been well defined. The purpose of this study was to determine whether ET can induce changes in LV diastolic filling that can account for an increase in exercise capacity and whether these changes can influence prognosis. We prospectively studied 55 consecutive patients (mean age, 55 +/- 7 years) with DCM. Patients were randomized into a training group (36 patients) or a control untrained group (19 patients) and matched for clinical and functional characteristics. All patients underwent a pulsed Doppler echocardiographic study, a radionuclide angiographic study, and a cardiopulmonary exercise test before and after a 2-month ET program. On the basis of the Doppler LV diastolic filling pattern at the beginning of the study, patients were prospectively divided into three subgroups: A (restrictive pattern), B ("normal" pattern), and C (abnormal relaxation pattern). In the trained group, peak VO2 (+12%; P < .0001), peak workload (+8.5%; P < .005), and lactic acidosis threshold (+12%; P < .0001) were significantly increased after training without changes in LV ejection fraction. However, only subgroup C demonstrated significant improvement in peak VO2 (+15%; P < .005). No changes were observed in the untrained group. In the trained subgroups a significant increase in rapid filling fraction (RFF), peak filling rate (PFR), peak early filling velocity (E), and E/A ratio was noted. A significant decrease in atrial filling fraction (AFF), peak atrial filling velocity (A), deceleration time of early filling velocity (EDT), and isovolumic relaxation time (IVRT) was observed only in subgroup C. No changes were found in untrained subgroups. A good correlation was found between Doppler and radionuclide LV diastolic filling parameters before and after training (P < .0001). Multiple stepwise regression analysis demonstrated that pretraining E/A ratio (P < .0001) and peak heart rate (P < .0002) were positive predictors of pretraining peak VO2. Posttraining increase in exercise tolerance (P < .0001) and increase in E/A ratio (P < .0001) were the strongest predictors of an increase in peak VO2. The independent