Effectiveness of Exercise-Based Cardiac Rehabilitation for Heart Transplant Recipients: A Systematic Review and Meta-Analysis

Background: Heart Transplant (HTx) is the ultimate chance of life for end stage Heart Failure (HF). Exercise training has consistently shown the potential to improve functional capacity in various chronic heart diseases. Still, the evidence in HTx recipients is scarcer. This study aims to systematically review the literature to evaluate the effectiveness and safety of Exercise-based Cardiac Rehabilitation (EBCR) in HTx recipients and to identify possible moderators of success. Methods: We conducted a systematic review and meta-analysis of randomized controlled trials on the effect and safety of EBCR in adult HTx recipients. The primary outcome was functional capacity, measured by Peak Oxygen Uptake (pVO2). We searched CENTRAL, MEDLINE, Embase, Scopus, and Web of Knowledge databases until December 2020, reviewed references of relevant articles and contacted experts. Usual care (UC), the different dosages of exercise regimens and alternative settings were allowed as comparators. A quantitative synthesis of evidence was performed using random-effects meta-analyses. Results: A total of 11 studies with 404 patients were included. Nine studies comprising 306 patients compared EBCR with usual care. They showed that EBCR improved pVO2 compared to usual care (Mean Difference [MD] 3.03 mL/kg/min, 95% CI [2.28-3.77]; I2 = 32%). In the subgroup analysis, including length of intervention and timing of enrollment after HTx, no significant moderator was found. Two trials, with 98 patients total, compared High Intensity Interval Training (HIIT) and Moderate Intensity Continuous Training (MICT). HIIT attained a significant edge over MICT (MD 2.23 mL/kg/min, 95% CI [1.79-2.67]; I2 = 0%). No major adverse events associated with EBCR were reported. Conclusion: We found moderate quality evidence suggesting EBCR has a significant benefit on functional capacity improvement HTx recipients at the short-term. HIIT showed superiority when compared to MICT. Research focusing long term outcomes and standardized protocols are needed to improve evidence on EBCR effectiveness.


2
Health Services Insights profoundly deconditioned due to many years living with advanced heart failure and thus have poor exercise capacity and cardiac cachexia. 3,5,[8][9][10] Although many patients experience significant early improvement, the extent to which peripheral adaptations secondary to the chronic low output state are fully reversible is unknown. This in part helps to explain the reduced exercise arterial-mixed venous oxygen difference that HTx patients experience during exercise, which reflects tissues difficulties extracting oxygen from the blood. 7 After transplant, patients' health condition is also influenced by chronic immunosuppressive regimens, which affect muscular function and increase the risk of long-term complications such as cancer and renal failure. This, together with cardiac allograft vasculopathy (CAV) are responsible for many longterm complications and deaths. [11][12][13] Exercise-based cardiac rehabilitation (EBCR) showed effectiveness in reverting peripheral alterations and improving oxidative capacity and capillary conductance, especially in the first year after heart transplant. 9,10 At the long term, exercise rehabilitation may play an important role controlling cardiovascular risk factors. 10 Both exercise and reinnervation improve hemodynamics, coronary blood flow and exercise thresholds, lowering resting heart rate and allowing peak heart rate to rise during exercise. 1,6 It has also been hypothesized that exercise itself may accelerate this reinnervation, which could function as an auto-potentiation mechanism. 1,14,15 Functional capacity constitutes a powerful prognostic factor in heart transplant patients as in the heart failure populations. 13,16 Previous experiences have shown that rehabilitation in heart transplant is safe 1,15,[17][18][19] and beneficial. The results of the latest 2017 Cochrane meta-analysis indicated that the inclusion of these patients in cardiac rehabilitation programs was associated with exercise capacity improvement in the short term. 20 Recently published retrospective studies have demonstrated that early rehabilitation after HTx is associated with a reduction in major adverse cardiovascular events in a doseresponse fashion, a reduction in hospitalizations and long-term survival improvement. 9,18,21 Historically speaking, heart transplant recipients have not been exposed to interval-based exercise with higher intensity, as it was considered dangerous due to delayed heart rate response. Nevertheless, high intensity interval training (HIIT) has gained notoriety due to growing evidence showing that this exercise modality may surpass the benefits of more traditional continuous moderate training regimens. 1,[22][23][24] Despite the accumulated knowledge and the recommendations from international societies, 25 up-to-date guidelines with detailed exercise prescription for heart transplant recipients do not exist, and the use of EBCR is sub-optimal in this population. The present study aims to systematically review the literature to evaluate the clinical effectiveness exercise-based cardiac rehabilitation in heart transplant patients and to identify possible moderators of success.

Methods
This systematic review and meta-analysis was implemented according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 26 The study protocol was registered in PROSPERO and can be consulted at CRD42021239110.

Eligibility criteria
Study designs. This systematic review included randomized controlled trials (RCT) comparing exercise-based cardiac rehabilitation programs to usual care in the management of HTx recipients. It also comprised studies comparing different training modalities (eg, moderate intensity vs high intensity exercise training) and settings (eg, home-based vs centerbased). Narrative reviews, preclinical studies, and editorial or opinion articles were excluded. Previous reviews and metaanalysis were assessed as guide, and reference lists were searched to identify additional RCTs.
Participants. We included studies examining interventions in adult HTx recipients with no restrictions regarding sex, ethnicity, and socioeconomic background at any time after the procedure.
Interventions. We defined EBCR as an intervention including physical exercise prescription by a cardiac rehabilitation specialist, performed at the hospital, cardiac rehabilitation center, at home or under a hybrid format including more than one location. 27 High intensity interval training (HIIT) can be defined as interval durations of up to 4 minutes, with an intensity of ⩾85% of heart rate peak, ⩾80% Peak Oxygen Uptake (pVO2), or a Rating of Perceived Exertion (RPE) ⩾15, interspaced with up to 3 minutes active recovery intervals. 28,29 Moderate-Intensity Continuous Training (MICT) includes continuous aerobic exercise of intensity 60% to 75% of heart rate peak, 50% to 65% of pVO2 or 12 to 15 RPE. 28,29 To be considered for this study the intervention should last at least 4 weeks long and comprise an aerobic exercise modality.
Comparators. Usual care (UC) was defined as the standard management programs for adult heart transplant recipients, as proposed by the International Society for Heart and Lung Transplantation (ISHLT). 25 This involves regular follow-up consultations to ensure the safety and optimal dosing of medicines and detect the development of complications or disease progression that may require a change in management.
Outcomes. The primary outcome of this review was (i) functional capacity and exercise tolerance, measured with pVO2 or 6-minute Walking Test. Secondary outcomes assessed were (ii) Adverse Events (All-cause mortality, cardiovascular mortality, All-cause Hospitalizations and cardiovascular

Search strategy
We searched for studies meeting our eligibility criteria in 5 bibliographic databases: MEDLINE, SCOPUS, ISI -Web of Science, EMBASE, and Cochrane Central Register of Controlled Trials (CENTRAL) -our full queries are displayed in Supplemental Material 1. We have also performed manual searching through gray literature across clinicaltrials.gov to retain efficacy in the identification of additional published, unpublished, or ongoing trials. In addition, we browsed trial registers, contacted study authors, and searched the references of all relevant primary studies, as well as of other relevant systematic reviews. All studies published until December of 2020 were included. No limitation concerning language of publication was applied.

Study selection
Two authors (RC and RP) independently screened the titles and abstracts of all records. Subsequently, relevant full-text articles were obtained and read by 2 independent authors. Inter-reviewer discrepancies were solved by discussion and consensus or by a third reviewer (EM) when agreement was not reached. When needed, authors were contacted regarding additional information for study eligibility assessment, or to request the full text when unavailable.

Data extraction
Data was independently collected by 2 authors, using a standardized form. We retrieved data on study design, country, maximum follow-up length, inclusion and exclusion criteria, number randomized (for intervention and comparator), number for results (intervention and comparator), age, sex, intervention and comparator description and components, setting, frequency, intensity, exercise length, and time after transplant. Data regarding primary and secondary outcomes were also collected. Inter-reviewer disagreements were solved either by consensus or by a third reviewer when agreement could not be reached. Study authors were contacted to provide missing information.

Quality assessment and publication bias
The systematic review and meta-analysis were performed using Review Manager (RevMan) 5.4. To assess the possible risk of bias (RoB) for each study, we collected information using the Revised Cochrane risk-of-bias tool for randomized trials (RoB 2). 30 Each study was evaluated across 7 domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. Studies were categorized as overall "high risk," "low risk," or "unclear risk" of bias with the help of RoB 2. Studies with 3 or more individual domains categorized as unclear risk or 1 high risk domain were classified as "overall high risk." We computed graphic representations of potential bias within and across studies using RevMan. Assessments of overall quality of evidence were performed using the GRADE approach for the primary outcome.

Data synthesis
Random-effects meta-analysis regarding functional capacity (measured with pVO2) was performed comparing EBCR with UC (primary analysis) and HIIT versus MICT (secondary analysis). The Mean Difference (MD) was used as the effect measure. For the studies reporting values of mean and standard deviation (SD) at the beginning and end of intervention both for experimental and control groups, a new mean and SD was calculated using the difference between the final and initial measurements and adjusting for the paired nature of this comparison. A similar adjustment was considered for a crossover trial used in a separate comparison of HIIT versus MICT. Those values were then used to calculate the difference between the experimental and control group and respective standard error, which is presented in the secondary analysis. Correlation between final and initial measurements was calculated from the studies providing SD at baseline, follow-up, and change score. When no information regarding SD was reported, we contacted authors for full description of results. In case of no further information, a mean of the remaining standard deviations was imputed in the analysis. Heterogeneity was analyzed using the Cochran's Q statistic, the Chi-square test and I-square statistic (I 2 ). Moderate or severe heterogeneity was considered if I 2 > 50% and I 2 > 75%, respectively. To assess potential moderators of heterogeneity in the primary analysis, the following subgroup analysis were performed: time after heart transplant, intervention length, resistance training (RT) inclusion, overall risk of bias, setting, and aerobic training intensity.
A qualitative description was performed for other outcomes that could not be included in the meta-analysis.

Study selection
Out of the 39 552 records initially identified, 11 studies were included in the meta-analyses: 9 studies comparing exercisebased cardiac rehabilitation with usual care, 14,15,31-37 2 studies comparing High Intensity Interval Training and Moderate Intensity Continuous Training. 23,38 The selection process and reasons for exclusion are further detailed in Figure 1.

Studies and patients' characteristics
A total of 11 studies with 404 patients (mean age 52 years) were included in 2 separate meta-analysis. For the primary meta-analysis 306 patients from 9 studies comparing EBCR versus UC, 14,15,31-37 and 98 patients from 2 studies comparing HIIT with MICT. 23,38 Only one study with 40 patients compared home-based and center-based settings. 39 This study was naturally not depicted with meta-analytical variables as it was an isolated investigation. In all trials, most patients were male. In all studies functional capacity was reported as pVO2.
One study 40 was excluded from the analysis due to concerns regarding patients matching with Tegtbur et al 36 study. An email was sent to address this issue, but no answer was provided.
Further detailed information regarding each study included in both meta-analysis is outlined in Table 1. More specific information regarding studies only included in qualitative synthesis is available as Supplemental Material 2.

EBCR characteristics
In most studies (n = 8), patients were recruited 6 or more months after transplant. 14,15,[31][32][33][34]36 Overall, 7 trials performed center-based exercise (CB), 15,23,32,33,35,37,38 3 studies performed home-based exercise (HB), 14,34,36 1 was a hybrid regimen (HB and CB). 31 In most of the studies frequency of training was 3 sessions per week. The length of session ranged from 28 to 52 minutes with a median length of 30-minutes. None of the studies had an intervention that lasted longer than a year. One study compared HB and CB 39 and it consisted in 8 weeks of MICT and RT performed 3 times per week.

Risk of bias assessment
Overall bias classification was high risk. Out of the 11 studies, 4 were classified as overall unclear risk 23,31,32,38 and 7 as overall high risk. 14,15,[33][34][35][36][37] Selection bias related to random sequence generation was low risk in almost all trials. Three studies 14,35,36 were classified as unclear risk due to vague description of the randomization process. Two were deemed as high risk due to imbalance of the groups at baseline. 33,34 Allocation concealment was only described in 3 studies. 23,32,38 The remaining 8 studies were classified in the unclear risk category because we were unable to find if concealment of intervention and control groups was initially done.
All studies were classified as unclear risk of bias for performance bias because of the inherent problem of blinding participants and personnel staff in exercise interventions. Considering detection bias the most frequent result was unclear risk, with 8 studies not clarifying this point. 14,15,23,[33][34][35][36][37] Only 3 were classified as low risk of bias since blind assessment was performed. 31,32,38 Regarding attrition bias 8 studies were categorized as low risk 14,15,23,[31][32][33]37,38 and 3 studies 34-36 were classified as high risk because there were patient losses to follow-up sufficient enough to dictate this classification. As for selective reporting all studies were categorized as low risk.
The graphs in Figure 2 schematically represent the bias assessment.
In the sub-analysis regarding the potential moderators of intervention success -time after heart transplant (<6 or ⩾6 months), intervention length (⩽3 or >3 months), resistance training inclusion, overall risk of bias assessment, setting (home-based, center-based or hybrid) and intensity of aerobic exercise (HIIT or MICT) -none of these variables explained significantly (P > .05) the gain in pVO2 of EBCR over UC. Nevertheless, we observed an increase in the mean difference of pVO2 of at least 1 mL/kg/min when considering HIIT studies and unclear risk of bias compared to MICT and high risk of bias, respectively (Supplemental Figures 1-6).    analyzed the maintenance of effect on pVO2 after intervention cessation. 38,41,42 Yardley et al 41  Adverse events. None of the included studies reported on allcause mortality or cardiovascular mortality. Nytrøen et al 15 reported a myocardial infarction in the usual care group which forced this patient to withdraw but none in the intervention group. The remaining studies did not report considerable adverse events. The rate of the events was similar in HIIT and MICT groups and the incidence and nature of the events were similar to the general HTx population.
Cardiac allograft vasculopathy. Only one study 43  Quality of life. The diversity in the evaluation methodologies used to measure QoL, as well as missing data, limited the implementation of meta-analysis.
Studies comparing EBCR versus UC evaluated QoL using Short

Health Services Insights
Hospital Anxiety and Depression Scale (HADS), regarding the patients present at Hermann et al's 32 study. Anxiety score (HADS-A) decreased significantly in the exercise group (4.7 ± 1.8 to 1.8 ± 0.8, P = .01) but not in the control group (3.2 ± 1.6 to 3.7 ± 2.3, P = NS). No impact was reported on depression score (HADS-D).
Yardley et al 41 reported the difference at baseline and at 5-year follow-up for some of those patients enrolled in Nytrøen et al. 15 Anxiety assessed by the HADS-A questionnaire decreased in the HIIT group and increased in the control group with a significant difference at the 5-year follow-up (P = .01). There were no differences between groups when comparing the depression scores at 5-year follow-up.
In Nytroen et al 23 no impact was reported regarding HIIT versus MICT on HADS-A or HADS-D scores.

Discussion
In this systematic review we have analyzed the effectiveness of EBCR in the follow-up of HTx patients. Overall, 11 primary studies were identified, including 404 patients, 9 for the EBCR versus UCR and 2 studies for the HIIT and MICT comparison. We found moderate quality evidence that EBCR in HTx recipients leads to improvement in functional capacity in shortterm follow-up (less than a year) compared to UC alone.
We also found moderate quality of evidence that HIIT is feasible, safe, and superior to MICT at improving functional capacity in this population at the short-term. Importantly, this benefit may be lost at the long term 42 reinforcing the uncertainty of HIIT long-term benefits compared with MICT.
Regarding the impact of EBCR at the long term, Yardley et al 41  Only one study 39 compared head-to-head the setting of the intervention. The center-based approach reached meaningful improvement when compared to the home-based intervention. However, this result should be interpreted with caution, since patients in the home-based exercise group utterly failed to raise their functional capacity. Studies by Tegtbur et al 36 and Bernardi et al 14 conducted home-based experiences and shown that patients were able to significantly improve their exercise capacity. In subgroup analysis including only studies with HB, pVO2 still improved meaningfully. As many of HTx recipients live at long distances from the transplant center, HB programs are of particular relevance in this population and the message should be on to put the majority of HTx patients exercising, at home or at a rehabilitation center.
Few studies reported on adverse events. None reported prognostic benefit. This may be related to the short follow-ups. Peak Oxygen Uptake is a well-documented powerful prognosticator in those with chronic cardiovascular conditions but the evidence addressing the relation of pVO2 and survival in heart transplant recipients is scarcer. 46 Retrospective data 16 suggests that pVO2 is associated with long-term survival in HTx recipients. A more recent retrospective study 47 has shown that participation in cardiac rehabilitation was associated with a 29% reduction in 1-year readmissions. So, the finding of 3.03 mL/kg/min improvement in pVO2 is relevant not only for the improved capability of patients, but also for what it can mean prognostic wise. However, this should be confirmed in long-term trials.
Only one study investigated the effects on cardiac allograft vasculopathy (CAV) 43 showing a reduction of 50% in progression as measured by intravascular ultrasound. The pathophysiological process of CAV is complex and involves many factors, including innate and adaptive immune responses as well as traditional risk factors. The way by which exercise affects CAV progression is not fully understood. Theoretically, this may result in part from the reduction in pro-inflammatory state. Nevertheless, it is not yet fully understood to what extent the effect on CAV depends on the HIIT effect, such that this finding needs to be confirmed in future studies.
We did not find any consistent effect of EBCR on QoL. This can relate to different reporting methods, which precludes aggregation of data, the relativity low number of patients, and the inherent problematic quality of the trials. HIIT demonstrated a beneficial impact on anxiety at both short 32,45 and long-term. 41 Caution should be taken since we are discussing isolated reports.

Strengths and limitations
This systematic review is important as it is the first to shed some light on the longer follow-up studies in the heart transplant recipient's field. Its importance also comes from the fact that it compares HIIT and MICT interventions besides the traditional comparison of EBCR and UC. This is also, that we are aware, the first meta-analysis using the mean differences of pVO2 values attained during the rehabilitation. This matters because takes into consideration the baseline fitness levels of the individuals and enables us to see a more precise estimate of the real impact of exercise in HTx populations. Other strong asset of this review is the query designed who allowed us to reach 39 550 references and the rigorous methodology implemented. We believe this study to be the most comprehensive and up to date meta-analysis in HTx rehabilitation.
Major impediments for stronger evidence are the relatively low number of patients, the scarcity of trials (only 9 studies 9 for the primary analysis and 2 studies for the HIIT vs MICT comparison) and the short follow-up length. In fact, this meta-analysis included only one more trial 23 than the latest Cochrane meta-analysis published, 20 which reflects the lack of investigation on the field. Other limitation is the suboptimal quality of the studies retrieved and the lack of standardization in the exercise regimens across studies. Most studies included in the meta-analysis began exercising patients many months which could represent an external validity problem, as the majority of real-life programs commence in the first month post-transplant. So, as of today, and because of these limitations, we can only say that exercise improves short-term functional capacity (less than a year). Longer follow-up studies are needed to see the real impact of a one-time intervention and, perhaps, to arrange and ideal timing for a repeated structured program. Another particularly important limitation that proceeds from the short-term follow-up is the nonexistence of prospective randomized data on mortality and hospitalizations.

Conclusion
We found moderate quality evidence suggesting EBCR has a significant benefit on functional capacity improvement in HTx recipients at the short-term. HIIT showed a slight superiority when compared to MICT. To sustain a prolonged effect, HTx recipients need to continue exercising. This may be facilitated by HB or hybrid programs. Further research with focus on long-term benefits and more standardized protocols are needed to build more robust evidence on EBCR effectiveness in this population.