Author + information
- Received March 4, 2016
- Revision received May 3, 2016
- Accepted June 29, 2016
- Published online February 20, 2017.
- Ambarish Pandey, MDa,
- Akhil Parashar, MDb,
- Curtiss Moore, MDa,
- Christian Ngo, MD, MSa,
- Usman Salahuddin, MDa,
- Mandeep Bhargava, MDc,
- Dharam J. Kumbhani, MD, SMa,
- Jonathan P. Piccini, MD, MHScd,
- Gregg C. Fonarow, MDe and
- Jarett D. Berry, MD, MSa,f,∗ ()
- aDivision of Cardiology, Department of Internal Medicine, University of Texas Southwest Medical Center, Dallas, Texas
- bDepartment of Internal Medicine, Cleveland Clinic, Cleveland, Ohio
- cDepartment of Cardiovascular Medicine, Cleveland Clinic, Cleveland, Ohio
- dDivision of Cardiology, Duke Clinical Research Institute, Durham, North Carolina
- eDivision of Cardiology, Ronald Regan-University of California Los Angeles Medical Center, Los Angeles, California
- fDepartment of Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, Texas
- ↵∗Address for correspondence:
Dr. Jarett D. Berry, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9047.
Objectives In this study, the authors performed a meta-analysis of currently available comparative prospective studies to assess the efficacy and safety of exercise training in heart failure (HF) patients with implantable cardioverter-defibrillators (ICD).
Background ICDs have been shown to improve survival in patients with HF. However, many patients with ICDs experience fear of shocks and avoid physical activity. Few data exist for efficacy and safety of exercise training in HF patients with ICDs.
Methods Prospective parallel arm trials with control and exercise training groups that evaluated the efficacy of exercise training in patients with ICDs were included in the meta-analysis. Outcomes of interest were difference in the change in cardiorespiratory fitness (CRF) (ml/kg/min) between exercise and control group and the likelihood of ICD shocks among exercise training compared with that among control participants on follow-up.
Results We included study level data from 6 trials (5 randomized controlled trials and 1 nonrandomized controlled trial). In the pooled analysis, ICD patients undergoing exercise training had significant improvement in CRF (weighted mean difference: 1.98 ml/kg/min; 95% confidence interval [CI]: 0.58 to 3.38). The likelihood of ICD shocks on follow-up was also significantly lower among exercise training than among control participants (pooled odds ratio: 0.47; 95% CI: 0.24 to 0.91).
Conclusions Among patients with HF and ICD implantation, exercise training was associated with significant improvement in CRF and lower likelihood of ICD shocks.
Exercise training or cardiac rehabilitation is associated with improved cardiorespiratory fitness (CRF), quality of life, and lower rates of hospitalization among patients with chronic cardiopulmonary conditions including heart failure (HF) (1–5). The current American College of Cardiology/American Heart Association guidelines recommend cardiac rehabilitation among stable HF patients with reduced ejection fraction with New York Heart Association (NYHA) functional class II to III symptoms (class IIA) (6). However, despite these guideline recommendations, the use of cardiac rehabilitation among HF patients remains low (7).
Patients with HF and implantable cardioverter-defibrillators (ICD) represent a particularly challenging population for cardiac rehabilitation. Previous studies have demonstrated significant behavioral changes among HF patients after ICD implantation, including decreased daily activity levels and avoidance of moderate intensity exercise training (8–10). These behavioral changes are related to patient and provider concerns for ICD shocks during episodes of moderate to high intensity exercise and lack of specific guideline recommendations for exercise training in this patient population (8,11). Small randomized controlled trials (RCTs) have recently demonstrated improvement in CRF with exercise training among HF patients and ICD implantation (12–19). However, these studies were not adequately powered to evaluate the impact of exercise training on risk of ventricular arrhythmias and ICD shocks. Accordingly, we performed a meta-analysis of comparative prospective trials to evaluate the impact of exercise training on CRF and risk of ICD shocks in this patient population.
A comprehensive computerized search of Medline, EMBASE, OVID, Web of Science, and Cochrane database was conducted using MeSH terms and keywords as described in the Online Methods. We initially evaluated all comparative studies including RCTs, nonrandomized parallel group trials, and pre-post within-group design that enrolled patients with ICDs. Both the RCTs and the nonrandomized parallel group trials were included in the analysis because of the paucity of studies addressing this issue. Outcomes of interest were difference in CRF (peak oxygen uptake, ml/kg/min) change from baseline to follow-up between exercise and control group, and the likelihood of ICD shocks on follow-up in exercise versus control participants. Studies failing to report both of the study outcomes were excluded (Figure 1).
Full text articles were retrieved for all titles and abstracts that met the inclusion criteria. Co-primary authors (A.P., A.P.) performed data extraction independently using a pre-standardized questionnaire. The following information was recorded for each study: author, year of publication, nature of study, baseline demographic and clinical characteristics, baseline and follow-up measurements of CRF, mode of CRF assessment, follow-up duration, and ICD shock or antitachycardia therapy events observed on follow-up. All discrepancies concerning study inclusion or outcomes were resolved by the senior author (J.B.). In the event of multiple publications arising from a single trial, only the updated trial with the maximum number of patients was included.
Data synthesis and statistical analysis
Meta-analysis of the outcomes was conducted using “Metan” function, available for Stata version 13 statistical software (Stata Corp., College Station, Texas) (20). The meta-analysis is reported here in accordance with the Preferred Reporting Items for Systematic reviews and Meta Analyses (PRISMA) guidelines (21). We assessed heterogeneity using the I2 test (where I2 ≥50% was assumed to be a result of significant heterogeneity). We found significant heterogeneity across the included studies; hence, we have reported the pooled estimates based on the random effects model. Publication bias was assessed using the funnel plots, and their symmetry was tested by Egger’s regression test, using the “metafunnel” and “metabias” functions, respectively, in Stata software. For the outcome of differences in CRF change from baseline to follow-up between exercise and control group, we have reported the pooled estimate for the weighted mean differences (WMD), which represents the absolute values of mean difference, adjusted for the statistical weight of the model study. For studies such as that by Dougherty et al. (19) with more than 1 follow-up CRF test (n = 1), results from the test performed closest to completion of supervised exercise training protocol was used for analysis. For the outcome of risk of ICD shock, to assess the overall likelihood of ICD shocks, we pooled odds ratios (ORs) of patients with ICD shocks on follow-up in the exercise training versus the control group, using random effects analysis to obtain weighted pooled OR.
Five studies reported the number of participants with ICD shocks in each study group during follow-up. One study, by Berg et al. (13), reported the number of ICD shocks per participant that was used to calculate the total number of ICD shock episodes. The number of participants with ICD shock was imputed assuming 1 ICD shock per person. All p values were 2-tailed with statistical significance specified at 0.05 and confidence intervals (CIs) reported at the 95% level.
We included 6 trials with median exercise training duration of 12 weeks (range: 8 to 24 weeks) and median follow-up of 15 months (range: 2 to 26 months) for ICD-related events. The baseline characteristics of the study participants are summarized in Table 1. All studies included stable HF patients with depressed left ventricular ejection fraction. A total of 98% of the pooled study participants had ICD implants (100% of participants in 5 studies and 68% in 1 study). Most common indication for ICD among the pooled study participants was primary prevention. Inclusion and exclusion criteria for participants in the included studies are detailed in Online Table 1. Among the studies that reported baseline medication use, beta-blockers were used in most of the study participants. In contrast, mineralocorticoid receptor antagonists and antiarrhythmic agents were used in a lower proportion of patients (Table 1). The exercise training protocols used in the intervention arm of these studies are detailed in Table 2. Supervised exercise training was included at some stage of the protocol in 5 of the 6 studies. Baseline and follow-up CRF data were collected in all included studies by using a symptom-limited maximal exercise test.
Effect of exercise training on CRF
All studies reported peak exercise oxygen uptake measurements of CRF at baseline and after exercise training. Pooled analyses using a random effects model showed that exercise-trained participants had a significant improvement in peak oxygen uptake compared with those who received the usual care (WMD: 1.98 ml/kg/min; 95% CI: 0.58 to 3.38; I2 = 68.8%) (Figure 2). Similar improvement in peak oxygen uptake was observed in sensitivity analyses limited to RCTs only (n = 5; WMD: 2.03 ml/kg/min; 95% CI: 0.47 to 3.60; I2 = 74.6%) (Figure 2).
Effect of exercise training on risk of ICD shock events
All studies reported data for ICD shock events during follow-up (Table 3). Most of the ICD shock episodes were observed on long-term follow-up after completion of the exercise training protocol. No exercise-associated ICD shocks were reported in 5 of the 6 included studies. In pooled analyses using a random effects model, exercise training was associated with significantly lower likelihood of ICD shock (pooled OR: 0.47; 95% CI: 0.24 to 0.91; I2 = 57.1%) (Figure 2). Similar findings were also observed on sensitivity analysis limited to RCTs only (n = 5; pooled OR: 0.45; 95% CI: 0.21 to 0.97; I2 = 64.9%) (Figure 3).
Separate pooled analysis was not performed for outcomes of appropriate versus inappropriate shock due to the lack of reporting in 3 of the 6 studies, including the largest study by Piccini et al. (17). Among the 3 studies that reported these outcomes, 85% of ICD shock events were appropriate. Furthermore, numerically fewer appropriate shocks were observed among exercise training than among usual care participants (0.10 vs. 0.18 episode per patient, respectively), with no differences in the number of inappropriate shocks.
Effect of exercise training on peak exercise heart rate
Three studies reported data for peak exercise heart rate among study participants (Figure 4). In pooled analyses, a significant statistical heterogeneity was observed across these studies (I2 = 96.5%; p < 0.0001). Furthermore, the peak exercise heart rate did not change significantly among the exercise-trained participants compared to that among the usual care participants (WMD: 3.15 beats/min; 95% CI: −4.25 to 10.54).
Cochrane risk of bias assessment tool was used to perform quality assessment as shown in Online Figure 1. Among the RCTs, random sequence generation was reported in 2 studies while blinding of outcome assessment was performed in 2 studies. Selective or incomplete reporting of outcome was not observed in any of the included studies. A significant publication bias was not observed for peak Vo2 and ICD therapy outcomes among the included studies, as shown in Online Figure 2 (peak Vo2: Egger’s bias coefficient: 1.98; p = 0.08; ICD therapy: Egger’s bias coefficient: −0.72; p = 0.23).
To our knowledge, the present study represents the most comprehensive evaluation of the efficacy and safety of supervised exercise training among this high-risk patient population with HF and an ICD. The principal findings of this meta-analysis indicated that exercise training was associated with a significant improvement in CRF in these patients. Furthermore, exercise training was associated with a lower likelihood of ICD shocks on follow-up.
Our study findings have important clinical significance. Up to 35% of patients with HF have an ICD implanted for primary or secondary prevention (22). HF patients with an ICD have been reported to avoid physical exertion due to fear of shocks (9,23). Furthermore, health care providers have concerns regarding the negative impacts of higher rate of sudden cardiac death or inappropriate shocks with increased heart rate during exercise (8,14,24). These concerns may contribute to barriers to exercise counseling among HF patients as well as a lower participation in regular physical activity and exercise (9,23).
Findings from the present study provide strong evidence supporting the efficacy and safety of supervised exercise training in this patient population. We observed a significant improvement in CRF among the training participants, similar to that observed among patients with moderate HF without an ICD and in patients with HF and preserved ejection fraction (3,4). We also observed a significantly lower likelihood of ICD shocks among exercise-trained participants on long-term follow-up. This finding is particularly significant given the adverse implications of ICD shocks on the natural history of left ventricular dysfunction. From the safety standpoint, contrary to the prevalent concern of increased risk of shocks with physical exertion, exercise-associated ICD shock was not reported during the training phase in 5 of the 6 included studies. One patient from the HF-ACTION study has been previously reported to get ICD shocks during supervised exercise training (1). Taken together, our study findings highlight the clinical benefit and safety of cardiac rehabilitation as an effective adjunct management strategy for patients with HF and ICD implantation.
Findings from the present meta-analysis are supported in an observational study by Davids et al. (25) that demonstrated an increased risk of ICD shock among patients who did not participate in cardiac rehabilitation. However, clinical trials have not been powered adequately to detect a beneficial effect of exercise training against ICD shocks. The present pooled analysis adds significantly to the available literature by demonstrating a substantial reduction in ICD shock risk among participants in exercise training.
Mechanisms by which exercise training may lower risk of ICD shocks are not well understood. Previous studies have demonstrated that higher levels of CRF are associated with a lower risk of ventricular and supraventricular arrhythmia (26,27). The protective effect of exercise training against ICD shocks could be related to the improvement in CRF and lower burden of ventricular arrhythmia. Furthermore, exercise training is associated with favorable remodeling of the left ventricle and reduces the sympathetic excitability, which may also lower the arrythmogenicity of the myocardium (28).
Several points should be considered before exercise training in ICD patients. First, although supervised exercise training may be preferred for these patients, particularly at the start of the training regimen, many studies had a component of home-based exercise as part of their training protocol. Home-based exercise training with heart rate monitoring may be considered an alternative to supervised training in clinical practice, if warranted due to cost reasons or patient preference. Second, consistent with inclusion criteria of the included studies, exercise training should be considered only in stable patients receiving optimal medical therapy. Third, exercise testing may be useful before initiating a structured training program in patients with ICD. This would help identify patients with exercise-induced dysrhythmia who may have a higher risk of ICD shocks during training sessions. Fourth, a majority (two-thirds) of the pooled study participants with reported indications underwent ICD implantation for primary prevention. Risk of ICD shocks with exercise may be higher among patients with ICD implantation for secondary prevention. Finally, target heart rate during exercise training must be tailored to fall well below the ICD threshold for tachycardia detection.
Our study is not without limitations. First, we only included 6 studies in this meta-analysis owing to the relative paucity of studies of this issue. Second, possibility of selection bias cannot be ruled out as only published trials were included in the meta-analysis. Third, we are unable to assess impact of exercise training and associated improvement in CRF and reduction in ICD shock rates on long-term clinical outcomes such as HF hospitalization and mortality. Larger trials with extended follow-up are needed to determine the long-term clinical efficacy of exercise training in this patient population. Fourth, we could not determine the specific pooled risk estimates for inappropriate and appropriate shocks among patients taking exercise training versus usual care participants as it was not reported in 3 of the 6 studies. This is particularly relevant considering the significant differences in the pathophysiology of appropriate versus inappropriate shocks. Although appropriate shocks occur mostly in the setting of ventricular arrhythmia, inappropriate shocks are more commonly precipitated by supraventricular tachycardia. However, among the studies that reported these outcomes, up to 85% of ICD shocks observed on follow-up were appropriate. Fifth, our study findings should not be applied to certain arrhythmogenic cardiomyopathies (or inherited arrhythmia syndromes) where exercise is contraindicated. Sixth, as most studies measured follow-up CRF levels immediately after completion of exercise training, we cannot determine the sustainability of the observed CRF improvement after discontinuation of exercise training.
Our study findings suggest that exercise training in patients with ICDs is associated with improved CRF and reduced likelihood of ICD-shocks. Future studies are needed to evaluate whether these beneficial effects of exercise training may translate into an improvement in long-term clinical outcomes among these patients.
COMPETENCY IN MEDICAL KNOWLEDGE: Exercise training under proper supervision in patients with heart failure and ICD leads to improvement in their cardiorespiratory fitness and less likelihood of ICD discharges.
TRANSLATIONAL OUTLOOK: Our study argues for the need for well-powered randomized clinical trials with extended follow-up to determine the long-term clinical efficacy of exercise training in this patient population with heart failure and ICDs. Longer follow-up will also provide answers to the question of maintenance of the observed CRF improvement after discontinuation of exercise training.
Dr. Berry is supported by Dedman Family Scholar in Clinical Care endowment, University of Texas Southwestern Medical Center, and American Heart Association award 14SFRN20740000. Dr. Fonarow has received research support from Agency for Healthcare Research and Quality; and is a consultant for Janssen, Medtronic, and Boston Scientific. Dr. Piccini has received research support from Johnson & Johnson/Janssen Pharmaceuticals and Boston Scientific; and is a consultant for and compensated advisory board member of Forest Laboratories, Medtronic, and Johnson & Johnson/Janssen Pharmaceuticals. Dr. Kumbhani has received honoraria and research support from the American College of Cardiology. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Drs. Pandey and Parashar contributed equally to this work.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- cardiorespiratory fitness
- heart failure
- implantable cardioverter-defibrillator
- New York Heart Association
- randomized controlled trial
- Received March 4, 2016.
- Revision received May 3, 2016.
- Accepted June 29, 2016.
- 2017 American College of Cardiology Foundation
- Pandey A.,
- Parashar A.,
- Kumbhani D.J.,
- et al.
- Pandey A.,
- Garg S.,
- Khunger M.,
- et al.
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Golwala H.,
- Pandey A.,
- Ju C.,
- et al.
- Belardinelli R.,
- Capestro F.,
- Misiani A.,
- Scipione P.,
- Georgiou D.
- Fitchet A.,
- Doherty P.J.,
- Bundy C.,
- Bell W.,
- Fitzpatrick A.P.,
- Garratt C.J.
- Piccini J.P.,
- Hellkamp A.S.,
- Whellan D.J.,
- et al.
- Piotrowicz E.,
- Zielinski T.,
- Bodalski R.,
- et al.
- Dougherty C.M.,
- Glenny R.W.,
- Burr R.L.,
- Flo G.L.,
- Kudenchuk P.J.
- Harris R.J.,
- Bradburn M.J.,
- Deeks J.J.,
- Harbord R.M.,
- Altman D.G.,
- Sterne J.A.C.
- Moher D.,
- Liberati A.,
- Tetzlaff J.,
- Altman D.G.,
- Group P.
- Al-Khatib S.M.,
- Hellkamp A.S.,
- Hernandez A.F.,
- et al.
- Laukkanen J.A.,
- Makikallio T.H.,
- Rauramaa R.,
- Kiviniemi V.,
- Ronkainen K.,
- Kurl S.
- Pathak R.K.,
- Elliott A.,
- Middeldorp M.E.,
- et al.
- Billman G.E.