Methods

All details of the methods used are given in the online appendix.

Results

In total, 529 articles were found from the electronic searches. Of these articles, 37 were considered eligible after abstracts had been read and reviewed in detail. Twenty-seven trials31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57 were excluded for one or more of the following reasons: pooled data from multiple studies; low-intensity resistance training; whole-body endurance training without resistance training; unsupervised home-based resistance and endurance training; local endurance training protocol using Thera-Band; local endurance training using knee extensor ergometer; no CHF; training intensity was not derived from one-repetition maximum (1RM); resistance in combination with endurance training versus usual care controls and/or a single-leg resistance training. No trials were selected from reference lists of the remaining articles. Finally, 10 trials studying the effects of moderate-to-high intensity resistance training (alone or in combination with endurance training) were considered eligible58,59,60,61,62,63,64,65,66,67 (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Summary of trials studying the effects of moderate-to-high intensity resistance training (alone or in combination with endurance training). CHF, chronic heart failure; 1RM, one-repetition maximum.

Methodological quality

None of the 10 identified trials had a perfect score ( = 9 × “yes”) on the Delphi list. Pu and colleagues67 had the highest score (8 × “yes”), while four trials (40%) scored ⩽3 × “yes”.61,63,65,66 Consequently, the mean (standard deviation) score of the identified trials of the Delphi list was poor: 4.3 (1.7) points (table 1).

Safety monitoring and drop out

Only one trial defined safety/adverse events (defined as “all-cause mortality and non-elective cardiac-related hospital admission”) as a secondary outcome measure.59 Indeed, drug treatment remained unchanged during the trial; no adverse events were reported; and owing to permanent telemetry, major arrhythmic disturbances were excluded. Beckers and colleagues did report the dropout of two patients (7.1%) owing to lymphoma and ischialgia during combined resistance and endurance training, while no dropout was reported in the endurance training group.59

The remaining nine identified trials did not define safety as a primary or secondary outcome. Nevertheless, serious adverse events and/or minimal changes in cardiac drug treatment were reported to be rare.58,63,65,66,67 Five trials (50%) reported anecdotal serious adverse events: a period of severe hypotension in one patient with CHF, requiring a 2-day hospital admission62; withdrawal due to worsening of CHF (n = 1) or because unable to make the time commitment for the study (n = 2)61; and one patient with CHF was reported to drop out owing to hospitalisation for an episode of rapid rate atrial fibrillation.60 These patients with CHF, however, had all been allocated to the endurance-training group, which did not undergo resistance training. To date, only Cider and colleagues have reported the drop out of one patient with CHF of the resistance training group owing to a severe asthma attack requiring hospitalisation.64 Finally, Feiereisen and colleagues reported 1 week of rest in one patient allocated to resistance training owing to low-back pain that could not be directly related to the intervention.58

Patients were supervised during the resistance training programme by a cardiologist, physiotherapist, physical coach, exercise physiologist, registered nurse, “a research assistant” (without further specifications of their background) or a combination thereof.59,60,61,62,63,67 In addition, several trials included a central monitoring system permitting arrhythmia detection during resistance training and/or endurance training59,60,61,62,63; and before the start or at the end of every training session the individual pulse rate and/or blood pressure were assessed.60,61,62,64 None of these trials reported exercise-induced malignant ventricular arrhythmia during resistance training. Finally, six trials (60%) included a warm-up period (consisting of short-lasting, low-intensity walking or cycling, light calisthenics and/or stretching) and/or a cooling down period (consisting of some stretching).58,59,64,65,66,67

Description of actual resistance training programme

Seven trials (70%) reported a mean compliance rate of between 75 and 100%.58,61,62,64,67 Finally, Beckers and colleagues only reported that adherence and compliance were “excellent”.59

Surprisingly, none of the 10 identified trials provided detailed information about the actual training load during progressive moderate-to-high intensity resistance training in CHF. Only Pu and colleagues reported the mean training load during a 10-week resistance training programme (eg, 82% of the most recent 1RM, which was retested every tenth resistance training session). Moreover, these authors did report that four patients with CHF (44.4%) required minor modifications in their resistance regimen because of intermittent mild musculoskeletal symptoms.67

Effects of resistance training

Resistance training does not appear to affect left ventricular contractile function or enhance myocardial deterioration compared with usual care or low-intensity stretching (tables 2 and 3).

There were conflicting results on the mean changes in peripheral muscle strength and endurance, submaximal exercise capacity, peak Vo₂ and disease-specific QoL after resistance training compared with usual care controls, while the effects on New York Heart Association (NYHA) classification and body composition have never been studied (tables 2 and 3).

The only randomised placebo-controlled trial did show significant improvement in strength, endurance, and functional exercise capacity following a 10-week resistance training programme compared with a 10-week low-intensity stretching programme.67 However, body composition, vastus lateralis muscle enzyme activity and peak exercise capacity were not significantly different between the two groups. Moreover, effects of resistance training on cardiac function, NYHA classification, disease-specific QoL, body composition, peripheral muscle strength and peak Vo₂ seem to be similar to endurance training (alone or in combination with resistance training).58

Effects of resistance training in combination with endurance training

Resistance training in combination with endurance training does not appear to affect left ventricular contractile function or enhance myocardial deterioration compared with endurance training (tables 3 and 4).

Resistance training in combination with endurance training does appear to improve peripheral muscle strength compared with endurance training. Mean changes in NYHA classification, body composition, peripheral muscle endurance and peak VO₂ were similar (tables 3 and 4). Effects on 6 min walking distance and treadmill endurance time have never been studied.

Discussion

Moderate-to-high intensity resistance training has been shown to be a valuable exercise modality in healthy subjects and in patients with (complex) chronic organ failure.19,68,69 Indeed, resistance training should be routinely incorporated in respiratory rehabilitation.70 Based on the striking similarities in systemic factors contributing to exercise intolerance in patients with chronic obstructive pulmonary disease and CHF,4 we reasoned that it might be worthwhile to add moderate-to-high intensity resistance training to exercise-based cardiac rehabilitation programmes.

Even though resistance training at moderate-to-high intensity appears to be safe, its effects in patients with CHF are inconclusive. Indeed, this systematic review of the peer-reviewed English literature identified only one randomised placebo-controlled trial assessing the effects of moderate-to-high intensity resistance training in patients with stable CHF.67 Additionally, moderate-to-high intensity resistance training does not appear to augment the positive effects of aerobic training in CHF, except for peripheral muscle strength. Finally, the methodological quality of the identified trials is generally poor, which limits the validity of most findings.

Methodological considerations

Internal and external validity

None of the 10 identified trials had a perfect score on the Delphi list (table 1). These methodological shortcomings may seriously diminish the internal validity of the reviewed trials.71,72,73 Moreover, we feel it is safe to conclude that the external validity of the reviewed trials is rather limited. Indeed, when it was reported, all trials were performed in a secondary or tertiary care outpatient setting; 72.7% of the patients were men; 60.0% of the patients had an ischaemic aetiology of their disease; and mean left ventricular ejection fraction was 27.9%. Moreover, patients with CHF with coexisting morbidities, like diabetes, intermittent claudication, chronic lung disease, status after stroke, orthopaedic problems, current smokers, pulmonary hypertension, renal impairment, hypercholesterolaemia, arterial hypertension, actively listed on the transplant list or a combination thereof, were mostly ineligible to enrol.58,59,60,64,65,66 Then again, the only randomised placebo-controlled trial comparing progressive resistance training with a low-intensity stretching programme did include patients with CHF with coexisting morbidities, like arthritis, arterial hypertension, coronary artery disease, diabetes and osteoporosis.67 Unfortunately, Pu and colleagues enrolled only older (⩾65 years) community-dwelling female patients with CHF without specification of current cardiac drug treatment.67

Interpretation of current literature

On average, the effects of moderate-to-high intensity resistance training on peak Vo₂ and disease-specific QoL in patients with CHF seem inconclusive. At first sight, this may be somewhat surprising, in particular because of the cross-sectional positive relationship between skeletal muscle function and peak Vo₂ in CHF.4,6,7,8,9,10,11 Nevertheless, in light of the various methodological shortcomings it is rather difficult to speculate whether or not the lack of improvement in peak Vo₂ and disease-specific QoL is due to problems with study design or truly reflects the lack of a causal relationship. Then again, Feiereisen and colleagues did not find significant correlations between changes in peak Vo₂, LVEF and skeletal muscle strength after resistance training, endurance training or a combination thereof in CHF.58

Surprisingly, in this systematic review, only four trials (40%) reported a hypothesis,59,61,64,67 while only two trials (20%) clearly stated their primary outcome(s).59,67 This, together with the lack of statistical comparisons for one or more (baseline) outcome(s) between the experimental group and the control group60,61 makes the process of scientific inference about the effects of resistance training (alone or in combination with endurance training) in patients with CHF less consistent and unclear. Additionally, sometimes interpretations of the results do not seem entirely consistent with the reported results59,61,62,67 (see online appendix for details). Finally, differences in baseline peak Vo₂, differences in the duration and intensity of the intervention(s) and the diversity in exercise outcomes may also partially explain conflicting results between the identified trials (tables 2 and 4).

Safety

Although safety has never been set as a primary outcome, our systematic review does not provide evidence that resistance training at an intensity of 50–80% of the 1RM is hazardous for patients with clinically stable CHF. In fact, the drop-out rates were rather low and serious adverse events related to the intervention appear absent. This may be due to careful screening at baseline and/or during follow up, together with the supervised setting. Moreover, resistance training does not appear to cause a reduction of the left ventricular contractility function or to enhance myocardial deterioration in patients with CHF.

Clinicians who consider resistance training in patients with CHF do have to take into account the above-mentioned methodological limitations. None of the identified trials included patients with CHF with unstable angina pectoris, complex ventricular arrhythmias, uncontrolled atrial fibrillation and/or pulmonary hypertension; and only one trial reported the enrolment of a patient with CHF with an internal cardioverter-defibrillator, without reporting group allocation.62 Finally, endurance training has been shown to partially reverse left ventricular remodelling in patients with clinically stable CHF. This benefit was not confirmed with combined resistance and endurance training.74 It may therefore be premature to generally add resistance training to endurance training in CHF rehabilitation.

Recommendations

It seems appropriate to design new randomised controlled trials to quantify the effects of moderate-to-high intensity resistance training in CHF. These randomised controlled trials have to include a larger and more heterogeneous sample of clinically stable patients with CHF with short-term and long-term follow-up. It may even be worthwhile assessing the effects of a multidisciplinary approach, in which cardiologist and physiotherapist collaborate closely with specialised nurses, psychosocial counsellors, occupational therapists and dieticians.

Left ventricular remodelling,74 submaximal exercise parameters, skeletal muscle endurance and patient-centred outcomes, like daily physical activity level using validated accelerometers may be of clinical relevance75 and, in turn, should be considered as possible future primary or secondary outcomes.

Finally, authors have to report more carefully the safety of resistance training as well as the actual training load over time of individual patients with CHF to make a continuous evaluation and evolution of these programmes possible. Future studies should also consider moderate-to-high intensity single-leg resistance training49 and be more specific about patient selection. Indeed, resistance training may be of particular interest in disabled patients with CHF and skeletal muscle atrophy or weakness, or both.2,4

Conclusions

We believe that there is a clear rationale for studying the effects of moderate-to-high intensity resistance training in clinically stable patients with moderate-to-severe CHF. Nevertheless, the current peer-reviewed literature does not provide good evidence that moderate-to-high intensity resistance training has statistically significant and clinically relevant positive effects on important outcome measures, such as peak Vo₂, skeletal muscle endurance, submaximal exercise capacity and disease-specific QoL. On the contrary, most of the included trials had substantial methodological limitations. Consequently, our current understanding of the effects of moderate-to-high intensity resistance training in CHF is limited. Even though moderate-to-high intensity resistance training does not seem be harmful for patients with CHF, the current peer-reviewed evidence is inadequate to generally recommend its incorporation into exercise-based rehabilitation programmes for these patients.