BACKGROUND: Controlled exercise training (ET) is a valuable therapeutic addition to pharmacological treatment in most patients with chronic heart failure (CHF), reducing long-term mortality, preventing cardiac remodelling and improving functional capacity. Despite the fact that the mechanism underlying its benefits might be multifactorial, a sustained improvement in autonomic balance is usually attributed as a major effect. Nevertheless, not all eligible subjects show the same response to ET, probably due to several differences in the subpopulations enrolled. We hypothesize that some heart rate variability (HRV) indexes could be valid tools to optimize the selection and follow-up of CHF patients receiving ET intervention. METHODS: Forty patients with CHF and left ventricular ejection fraction (LVEF)</=40% under complete evidence-based pharmacological treatment were included; 20 were assigned to a program of controlled ET on a 3-times/week basis during 24 weeks, training group (TG) and 20 received a standard follow-up program, control group (CG). In each patient, full clinical assessments, echocardiography, HRV analysis and 6-minute-walk test were performed at the beginning and the end of the study. RESULTS: After 24 weeks, patients in the TG showed a significant improvement in LVEF, 6-minute walk test, functional class of symptoms and HRV parasympathetic related indices (HF and rMSSD). Patients in the CG did not exhibit any improvement in the aforementioned indices and experienced more adverse events. Moreover, an initial value of HF<150 ms(2)/Hz or rMSSD<20ms predicted better outcomes of the ET program, including improvements in systolic function, the distance walked in 6minutes, and the functional class of symptoms, along with a reduction in clinical events. CONCLUSIONS: In CHF patients, HRV indexes related to parasympathetic function are valid and clinically useful tools to select and follow-up those candidates that could experience superior functional improvement after ET.