Exercise Training in Heart Failure: A Long Way to go Yet

This article refers to ‘Impact of exercise-based cardiac rehabilitation in patients with heart failure (ExTraMATCH II) on mortality and hospitalisation: an individual patient data meta-analysis of randomised trials’ by R.S. Taylor et al., published in this issue on pages xxx. Exercise intolerance is one of the cardinal symptoms experienced by patients with heart failure (HF); however, the precise pathophysiology has not been fully elucidated.1 Traditionally, it was thought that cardiac dysfunction was the major determinant of exercise capacity, particularly as peak oxygen consumption is strongly correlated with cardiac output. However, the demonstration of persistent exercise limitation in HF patients following heart transplantation or left ventricular assist device implantation points the contribution of other, non-cardiac, mechanisms.2 As such, we now know that maladaptive responses in the peripheral and pulmonary vasculature, respiratory system, skeletal muscles and autonomic nervous system triggered in the setting of HF are key factors.3 Exercise training can target both the peripheral and cardiac disturbances in HF and hence is an attractive treatment option.4 Preclinical and clinical data have shown exercise to: (i) improve vascular function, by improving endothelial function, decreasing peripheral vasoconstriction, neoangiogenesis, and improving myocardial blood flow; (ii) favourably alter sympathovagal balance together with a reduction in angiotensin II and aldosterone levels; (iii) reverse adverse changes in skeletal muscle morphology, histochemistry and metabolic function as well as restore muscle ergoreflexes; and (iv) improve cardiac function (normalizing myocyte calcium handling, increasing myocyte contractility and left ventricular ejection fraction, reducing end-diastolic volumes).5 These studies promoted a series of randomised controlled trials evaluating the role of exercise training in HF. In this issue of the Journal, Taylor and colleagues performed a meta-analysis upon individual data from 3912 patients with HF and reduced ejection fraction from 18 randomised controlled trials.6 Compared to control (standard care), exercise cardiac rehabilitation (involving both exercise training and HF-related disease management counselling) had no effect on all-cause mortality or hospitalisation and had no effect on HF-specific mortality or HF hospitalisation. Individual patient-level data allowed the authors to perform subgroup analyses to answer the question of which HF patient subgroups (including factors such as age, gender, ejection fraction, New York Heart Association class, HF aetiology, ethnicity, and baseline peak oxygen uptake) might benefit most from exercise training. In this further analysis, no significant interactions were found with respect to mortality or hospitalisation. Several caveats limit the impact upon clinical practice of this meta-analysis and these are related to the design of the constituent studies. The data evaluated were not entirely contemporary, given the publication dates for the included trials ranged from 1999 to 2012. Consequently, patients across the trials were not receiving comparable care, nor were they receiving optimal treatment seen in current practice.7 The three most recent trials contributed 330 patients of whom 87% were on an angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker, and 59% were on a beta-blocker. Furthermore, these trials were conducted before any practical guidelines on exercise training were available, and hence frequency (number of sessions/week), duration of session, intensity of training and degree of supervision varied on a trial by trial basis.4 Only one trial (Heart Failure – A Controlled Trial Investigating Outcomes of exercise TraiNing, HF-ACTION) was powered to detect a mortality effect and this contributed 60% of the patients in the present meta-analysis.8 The remaining studies were phase II and much smaller (< 200 patients). This limits statistical credibility of any positive findings with respect to survival or hospitalisation (though one was not seen). In a small, well-conducted study of 123 patients with HF, supervised exercise which was sustained for a period of 10 years led to improvements in HF symptoms, quality of life and exercise capacity within the first year, left ventricular ejection fraction after 5 years and cardiovascular mortality and hospitalisation at 10 years.9 These findings lend support to a dose–response effect but suggest that more than 5 years of sustained exercise are required to effect structural remodelling and prognostic improvement. Only two further studies with an active treatment arm lasting longer than 12 months are available; one that contributed 99 patients and was published in 199910 and the HF-ACTION study from 2009.8 With the exception of these three trials, the average duration of exercise cardiac rehabilitation in the remainder of the aggregate data was 5.5 months and the average duration of follow-up was 11.5 months; both treatment and follow-up duration are too short to identify a survival or reduction in hospitalisation signal, which is a major limitation of this meta-analysis. Adherence to exercise by patients with HF is frequently poor. At 12 months, only 30% of patients in the active arm of HF-ACTION were adherent, far less than that seen with pharmacological treatments.11 In those patients who managed to sustain a moderate work load regimen, a reduction in cardiovascular mortality and hospitalisation was noted as well as improvement in symptoms and peak oxygen uptake.12 Strategies to improve adherence to an exercise programme have since been published and involve identifying relevant patient-, physician- and healthcare-level hurdles.13, 14 Current HF guidelines provide class I, level of evidence A support for regular aerobic exercise to improve symptoms and functional capacity in patients with HF.15 Taken together, the current study does highlight the need for future studies aimed at maximizing the potential benefits of exercise training in HF. Tailored prescriptions focused on patient-specific factors (e.g. sarcopenia, respiratory muscle weakness, chronotropic incompetence, impaired endothelial function), which are iteratively modified according to response should be developed, much akin to pharmacological dose titration.4 Encouraging exercise prescription compliance should be promoted both in day-to-day clinical practice as well as in trials and this might be objectively undertaken with wearable technologies such as pedometers and accelerometers together with smart-technology.13 Finally, greater effort should be made to recruit patients that reflect those seen in usual clinical practice but those who are under-represented in clinical trials: patients above the age of 70, females, and those with co-morbidities.16, 17 Conflict of interest: none declared.