Higginbotham, Sullivan, and Cobb performed a number of exercise tests with invasive monitoring while measuring both oxygen consumption (VO₂) and ejection fraction in both normal patients and those with heart failure [21–23]. At baseline, heart failure patients have an increased resting heart rate, but a reduction in heart rate at peak exercise [21]. These studies were performed prior to heart failure patients being treated with beta-adrenergic blockers, and one can assume that this chronotropic incompetence should be more pronounced in the current era. Heart failure patients were also found to have a reduction in stroke volume at both rest and peak exercise. Of interest, the central arteriovenous (A-V) oxygen difference was increased at rest and increased to a similar level at peak exercise to the A-V oxygen difference in normal controls. However, it is not clear that these central cardiac changes are solely responsible for the exercise limitations of these patients. Various interventions such as inotropes and vasodilators that improve cardiac output and leg blood flow do not improve exercise tolerance [24–26].

Due to these findings, the peripheral muscle response to exercise has also been evaluated. Leg blood flow is reduced due to low cardiac output but also increased leg vascular resistance and abnormal arteriolar constriction [21, 22, 27]. There are also many intrinsic skeletal muscle changes that limit exercise. These include reduced skeletal muscle capillary density, change in skeletal muscle fiber type with a shift from oxidative type I fiber to glycolytic type II muscle fiber, a reduction in oxidative enzymes in the Krebs cycle, a decrease in mitochondrial size and number, an increase in inflammatory markers, and a reduction in muscle mass and strength [27]. One could certainly hypothesize that some of these adaptations that occur could be improved by exercise training. Figure 9.1 outlines the multiple hypothesized and proven benefits of exercise training [28]. In addition to reversing some of the detrimental effects outlined above, exercise also decreases the production of neurohormones, reduces inflammation, and improves endothelial function. However, until the HF-ACTION trial, the effects of exercise training in a large number of patients had never been evaluated.

The HF-ACTION trial was the first large trial designed to evaluate the survival benefit of exercise training in patients with heart failure. The initial trial design was to randomize 3,000 patients to either exercise training or usual care [29]. At the onset of the trial, there was an emphasis placed on ensuring that patients were on excellent medical therapy and 94 % of the patients were on an ACE inhibitor or angiotensin receptor blocker, 94 % were on a beta-blocker, 45 % were on an aldosterone antagonist, and about 40 % had an implantable cardioverter-defibrillator. From a symptom standpoint, the patients were about two thirds New York Heart Association functional class II and 33 % class III. Only 23 patients in the trial had NYHA class IV symptoms at baseline and would caution one that the effects of exercise training in this population are less well understood.

Patients were randomized to either phase II cardiac rehabilitation with 36 monitored sessions or usual care. They were then given either a cycle ergometer or a treadmill to continue exercising as an outpatient. Patients were also given a heart rate monitor to assess the quantity of exercise that patients were performing at home after the 36 sessions were completed. Prior to beginning the trial, each patient performed a symptom-limited metabolic stress test to evaluate for safety of exercise, to rule out ischemia, to determine the initial heart rate training goal, and to determine the peak oxygen consumption. For the patients that were randomized to exercise, after they completed their monitored sessions, they were then given either a treadmill or a cycle ergometer with encouragement to continue to exercise at home. To monitor how much exercise they did, they were also given a heart rate monitor that was downloaded to follow the amount of exercise that they were completing. The initial prescription during the cardiac rehab phase was for walking, treadmill, or stationary cycling for 15–30 min at a heart rate of 60 % of their heart rate reserve assuming that heart rate was below the ventilatory threshold of the patient. The duration was increased to 30 min and then to a total of 40 min over time. One should note that there was no strength component to the exercise prescription, and the training consisted solely of aerobic training. After 18 sessions, the patients began home-based exercise as well. At the conclusion of all 36 sessions, patients were given a goal of exercising 40 min five times a week at a heart rate of 60–70 % of their heart rate reserve.

Of the 1,159 patients randomized to exercise training, only 736 actually completed all 36 sessions of exercise. For the exercise training patients during the first 3 months, the median time of exercise was 76 min per week when the goal was 90. Exercise time increased to 95 min per week by 6 months but then dropped off to 74 min per week at year 1 when the training goal was actually 120 min a week. Although the patients were encouraged to exercise to the goal time, like most humans, the drive to continue to exercise decreased over time.

Despite the demonstration of benefit of exercise training in patients with heart failure outlined in Table 9.2, there still remains a concern about the safety of exercise in this patient population. Additionally, there is little data about the effects of exercise training in patients with NYHA class IV heart failure. Before prescribing exercise for these and all patients, it is important to appropriately evaluate the patients to determine the safety of exercise for them. Patients should be evaluated by an exercise test to evaluate for exercise-induced ischemia and exercise-induced arrhythmias and to help determine a safe exercise training heart rate range for the individual patient. In the HF-ACTION trial, all patients underwent a symptom-limited metabolic gas exchange exercise test on a treadmill using the modified Naughton protocol [33]. This test was performed to exclude patients with noncardiac limitations to exercise and those with ischemia. Additionally, the heart rate training range was made at 60–70 % of the heart rate reserve assuming as long as that target rate was below the ventilatory threshold.

The HF-ACTION trial evaluated a number of safety endpoints between the two arms. They found no difference in the incidence of serious adverse events such as progressive heart failure, unstable angina, arrhythmias, or neurologic events [19]. They also examined the incidence of hospitalization and death rate in the first 3 h after exercise and did see a small increase in hospitalizations with 37 (3.2 %) in the exercise arm compared to 22 (1.9 %) in the control arm. There was no difference in death after exercise between the two arms, a finding confirmed by the Belardinelli trial [30]. Although there was no increase in the incidence of ICD (implantable cardioverter-defibrillator) firing after exercise, the investigators did examine predictors for ICD firing during the trial [34]. Of 1,053 patients that had a defibrillator prior to enrollment in the trial, 20 % of the patients in the exercise arm and 22 % of those in the control arm experienced a shock. Risk factors for ICD firing included those with a previous history of ventricular or atrial arrhythmias or those with exercise-induced dysrhythmias. Additionally, those with a lower diastolic blood pressure and nonwhite race also were associated with an increased incidence of firing. This increase incidence of arrhythmias emphasizes the importance of determining the safety of exercise training for the individual patient before embarking on an exercise training regimen.