Cardiac rehabilitation in elderly myocardial infarction survivors: focus on circulatory power

Exercise-based cardiac rehabilitation (EBCR) is paramount after an acute myocardial infarction (AMI). Older individuals have been reported as having a worse prognosis after an AMI, and some series have reported differences in the functional response to EBCR. The peak circulatory power (CP), a non-invasive parameter, has been described as a surrogate for the cardiac power, showing promising results as a comprehensive measure of the cardiovascular response. Whilst this, data concerning the impact of EBCR on CP, particularly among elderly individuals, remains elusive. To address this issue, an observational, retrospective study including all patients admit-ted due to an AMI who completed a phase II EBCR programme be-tween11/2012and4/2017,wasconducted,withCPbeinganalysedby a symptom-limited cardiopulmonary exercise test. A total of 379 patients, 30% aged ≥ 65 years-old, were included. CP significantly improved after the EBCR programme (in all patients, as well as in both subgroups). Older patients presented lower CP than their younger counterparts at the beginning and the end of the programme, while presenting smaller improvements (122 ± 540 vs 293 ± 638 mmHg mL/kg/min, p = 0.013). This was maintained after adjusting for several potential confounding factors. A contemporary ECBR pro-grammewasassociatedwithsignificantimprovementsinCPamong AMI patients. Though those aged ≥ 65 years-old presented smaller improvements in CP than younger individuals, these still presented significant increases in this parameter. These results highlight the importanceofEBCRinthischallenginghigherriskgroupofpatients.


Introduction
Ischaemic heart disease (IHD) is a major cause of morbidity and mortality worldwide [1,2]. Over the years major advances have been made in its overall management, with exercise-based cardiac rehabilitation (EBCR) being central in the current management of individuals with IHD [3][4][5]. Exercise has a broad impact on the cardiovascular (CV) system, on both the heart and the peripheral vasculature, being associated with several beneficial effects in terms of the CV func-tion (such as cardiac contractility and chronotropic reserve) and risk factor control, as well as having several extra-CV actions such as effects on inflammation and on metabolic pathways [6][7][8].
Cardiopulmonary exercise testing (CPET) can have a central role in the assessment of individuals undergoing EBCR, as to provide integrative data on the CV, pulmonary, and musculoskeletal systems, as well as their interactions [9,10]. Indeed, beyond its role in risk stratification and exercise prescription, several studies have highlighted the association between functional parameters and CV events [11][12][13].
While there are currently extensive data on the beneficial effects of EBCR in the setting of IHD, there are still some caveats related to its optimal application [14,15]. Notably, certain subgroups of individuals such as the elderly tend to be underrepresented in different series [7,14,16]. Importantly, this group of patients can have a worse prognosis after an acute myocardial infarction (AMI) [4,17,18]. In this regard, some studies have suggested the beneficial effects of EBCR among elderly individuals, in terms of functional parameters such as the peak oxygen uptake (pVO 2 ) as well as events [14,19,20]. We have previously shown that, in a group of AMI survivors undergoing an EBCR programme, the elderly had significant improvements in functional parameters (as assessed by pVO 2 and exercise duration), though these were smaller than those presented by their younger counterparts [20]. Interestingly, differences in terms of functional parameters between older and younger individuals have also been reported in other studies [21]. As such, the overall effects of EBCR in the elderly have been the focus of heightened interest [14,22].
Beyond classical parameters such as the pVO 2 , others have emerged as to try to provide a broader view of the CV system [9]. Among these, peak circulatory power (CP) has been reported as an interesting CPET parameter, being a non-invasive surrogate of the peak exercise cardiac power [9,23,24]. This parameter, which incorporates both pVO 2 and systolic arterial blood pressure (SBP), provides data on both central and peripheral components of cardiac work, and can thus be of interest as to provide a more comprehensive and integrative view on the CV response to exercise [23,25]. While this parameter has previously been studied in individuals with coronary artery disease (CAD) and in heart failure (HF), data pertaining to the effect of EBCR (particularly in older individuals) remains elusive [10,23,26,27]. Given this background, in the present study we aimed at assessing the effect of an EBCR programme on CP in AMI survivors, and specifically to assess the impact of age on this parameter.

Methods
This was an observational, retrospective cohort study. The eligible population comprised all patients discharged from the Cardiology Department of the Gaia/Espinho Hospital Centre with the diagnosis of an AMI (according to the International Classification of Diseases, 9th Eition), between November of 2012 and April of 2017. To be included in the study, patients had to have completed a phase II EBCR programme (including at least two assessments in a consultation with a physical medicine specialist and performance of a CPET at the beginning and at the end of the programme) after discharge. The study was approved by the local Institutional Ethics Committee.

EBCR protocol
The EBCR programme has been previously described [15,28,29]. Briefly, this encompassed a predefined 8-week (three sessions per week) outpatient protocol. Before starting, patients were clinically assessed and underwent a CPET [14]. Training intensity was individually prescribed by an expert in EBCR (taking into consideration the heart rate obtained during CPET) [20,28,30].

Cardiopulmonary exercise testing
Patients underwent a symptom limited CPET on a treadmill (Mortara XScribe; Mortara Instruments, Milwaukee, WI, USA) using either a modification of the Bruce protocol or a variation of this protocol (in highly deconditioned patients) [15]. CP (expressed in mmHg mL/kg/min) was defined as the product of pVO 2 (mL/kg/min) and peak SBP (mmHg) [9,23]; the VE/VCO 2 slope (a measure of ventilatory efficiency, incorporating minute ventilation and VCO 2 ) was derived by automatic linear regression from values obtained during the CPET [9,10]. Data related to pVO 2 and the respiratory exchange ratio (RER) have been previously reported [20]. Patients were not asked to discontinue betablockers before the test.

Clinical and analytical variables
As previously detailed, patients were categorized according to age: <65 years-old (younger group) or ≥65 years-old (elderly group) [20]. Data were collected for clinical, analytical, and echocardiographic variables according to the electronic health records (EHR). Arterial hypertension was defined according to the presence of this diagnosis in clinical files. Dyslipidaemia was defined according to previous diagnosis, or the use of antidyslipidaemic medication prior to admission, or by having a low-density lipoprotein cholesterol ≥190 mg/dL [3]. Diabetes mellitus was defined according to previous diagnosis, or the use of antidiabetic agents prior to admission, or by having a glycated haemoglobin ≥6.5% [30]. Arterial hypertension, dyslipidaemia and diabetes mellitus were categorized as present or absent, in accordance with these criteria. Left ventricular ejection fraction was evaluated by the biplane Simpson's method, according to pre-discharge assessment.

Statistical analysis
Continuous variables were presented as mean ± standard deviation or as median [percentile 25-75, interquartile range (IQR)] according to the distribution. Categorical variables were expressed as absolute count (as well as percentage). Continuous variables were compared using unpaired or paired t test for those with normal distribution, or with the Mann-Whitney or Wilcoxon tests (for unmatched and matched data, respectively). The comparison of categorical variables was performed with the χ 2 test. The normality of the distribution was analysed with the Kolmogorov-Smirnov test. The Spearman correlation was used to assess the relationship between CP and the VE/VCO 2 slope. Linear regression analysis was used to assess if age (being ≥65 years-old) was a significant predictor of the change in CP irrespective of potential confounders. Sex, prior history of CAD, number of EBCR sessions, Killip classification, the presence of arterial hypertension, the presence of dyslipidaemia, the presence of diabetes mellitus, smoking status, and the baseline CP were forced into the model. Given that, as previously reported, RER values differed between age groups (as detailed in the Discussion), baseline RER was also included in the model [20]. Results were two-sided, and a p value below 0.05 was considered as significant. Statistical analysis was done using Stata 14 (Stata Corp, College Station, TX, USA).

Results
The overall study population and baseline characteristics have been previously described [20]. Briefly, a total of 379 patients (81% male gender, mean age 58.8 ± 10.6 years-old, 67% after a ST-segment elevation MI) were included in this study, of whom 30% were ≥65 years-old upon discharge (Table 1).
Patients completed a median of 22 (IQR 16-25) ECBR sessions, with older individuals attending an inferior number [20 (16-24) vs 24 (17)(18)(19)(20)(21)(22)(23)(24)(25)(26), p = 0.057]. As previously reported, pVO 2 differed between groups at both the beginning and the end of the programme (19.68 ± 5.63 vs 24.15 ± 5.72 mL/kg/min and 20.47 ± 5.61 vs 25.75 ± 5.93 mL/kg/min, respectively, p < 0.001 for both comparisons) [20]. Peak SBP did not differ between groups at the beginning [150 (140-  Overall, CP significantly improved after the EBCR programme (all patients: 3676 ± 1120 vs 3434 ± 1066 mmHg mL/kg/min, p < 0.001; younger group: 3911 ± 1084 vs 3619 ± 1034 mmHg mL/kg/min, p < 0.001; older group: 3120 ± 1006 vs 2999 ± 1018 mmHg mL/kg/min, p = 0.018). Older individuals, however, presented lower levels of CP at both the beginning and the end of the programme (Table 2, Fig. 1). In addition, this subgroup had a smaller improvement (delta) in CP when compared to younger individuals ( Table 2). The VE/VCO 2 slope differed between groups at both the beginning Being aged ≥65 years-old was associated with a smaller improvement in CP irrespective of sex, prior history of CAD, Killip classification, ejection fraction, presence of arterial hypertension, dyslipidaemia, diabetes mellitus, smoking status (current or former smoking), number of EBCR sessions and Fig. 1. A contemporary EBCR programme was associated with significant improvements in circulatory power (CP) among a group of AMI survivors. While the elderly had lower CP levels at the beginning and the end of the program, and had a smaller increase than younger patients, these still presented significant improvements. These results highlight the relevance of these programmes among these individuals. baseline CP (Table 3). This result was maintained after inclusion of the baseline RER (p < 0.001). This was also maintained after inclusion in the model of differences in medications at discharge (with an effect on blood pressure) between groups, and after including all medications potentially influencing blood pressure (i.e., angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers, beta-blockers, spironolactone, diuretics, calcium-channel blockers, nitrates, nicorandil, ivabradine; p < 0.001 for both models).

Discussion
In the current study, AMI survivors submitted to contemporary EBCR (encompassing aerobic and resistance training, in a background of optimal therapy) presented significant improvements in CP. Though older patients had lower levels of CP and had smaller increases across the programme, these still presented significant improvements.
We have previously assessed the differential impact of an EBCR programme among elderly individuals in terms of classical CPET parameters such as the pVO 2 , exercise test duration and the RER [20]. The current results on CP concur with our previous observations, by depicting an overall gap be-tween age strata, while also reinforcing the beneficial effects of a contemporary EBCR programme among both subgroups of patients [20]. Briefly, our prior data showed that although older individuals had lower functional capacity (namely a lower pVO 2 and exercise duration), these also derived significant benefits from this intervention [20]. In this analysis, we have focused on CP, a parameter which could reflect the overall performance of the cardiac pump, and thus provide a noninvasive estimate of the cardiac power [23,31,32]. Given the data reporting the relationship between CP and CV events, we believe these results are of relevance to the current literature on this topic. This parameter, first described in 2002 by Cohen-Solal et al., has been mainly explored in the setting of HF, where it has emerged as a potentially important component of the overall CPET prognostic assessment [9,23,[31][32][33]. Indeed, while the central role of pVO 2 in the assessment of CV disease is consensual [11,14], CP has been reported as being able to provide additional ancillary data, namely among individuals with HF under beta-blocker therapy [34]. Interestingly, a recent report by Lala et al. [32] on advanced HF patients found that CP was a strong predictor of a composite outcome encompassing death, durable mechanical circu-latory support implantation or cardiac transplantation (at 1 year). Notably, whilst its assessment in the setting of HF has expanded over the years, data on the impact of EBCR on CP remains to be further ascertained, specifically in terms of data among elderly individuals [27,35,36].
Ageing is associated with several changes in both the CV system and other sites, with decreases in overall functional capacity being described in different clinical settings [37,38]. This concept has been particularly studied in terms of pVO 2 , being also described in CP (which incorporates the former in its calculation) [37,39]. The fact that similarly to the results concerning pVO 2 [20], the differences in CP were maintained even after adjusting for several potential confounders highlights the concept that these could (at least partially) be attributed to differences in terms of the overall physiologic response between groups. Furthermore, a higher CV burden in the background of imbalances in terms of CV risk factors (such as arterial hypertension, dyslipidaemia, and diabetes, significantly more prevalent in older individuals) as well as of prior CAD could also be related to these findings. As reviewed by Fleg et al. a plethora of CV changes (potentially modulated by risk factor exposure) could be associated with pathological imbalances in the elderly, whereas extra-CV adaptations should also be kept in mind [40,41]. As elegantly reviewed by Giallauria et al., ageing is a risk factor for frailty, a multifactorial condition involving different mechanistic pathways and having numerous manifestations, being associated with impaired quality of life, lower functional capacity, and worse outcomes [18,[41][42][43][44]. In this regard, data from the SILVER-AMI study, assessing 3041 individuals aged ≥75 years-old with an AMI, reinforced this notion by reporting that mobility impairment was able to modulate the association between age and outcomes [18]. One analysis from this study also showed that in-hospital mobility was an important predictor of functional decline in this population [43]. Interestingly, a prior study on elderly individuals, addressing the potential role of pre-infarction angina, also showed that among patients with this clinical manifestation prior physical activity could affect in-hospital CV outcomes [45]. Notably, and in accordance with the current results, EBCR has been shown to be able to lead to significant improvements in functional capacity among elderly individuals, reinforcing its pivotal role in this higher risk subset of patients [20,21,36].
The current study assessed CP by a CPET, as a product of pVO 2 and SBP, as described in the current recommendations [9,23]. Though the use of mean arterial blood pressure has also been described, given the concordance between measurements described by Hulkkonen et al. [24], we believe this should not limit the interpretation of the current data. Differences in the RER should also be taken into consideration [20]. However, while this differed between groups, the mean values obtained (1.09 ± 0.10 vs 1.09 ± 0.13 in the younger group; 1.05 ± 0.11 vs 1.03 ± 0.11 in the older group) as well as the incorporation of this parameter in the regression model should not hinder the overall assessment of the data on CP. Another point which should be acknowledged relates to the lack of data on musculoskeletal parameters, which could have influenced test termination and as such overall assessment. Given the differences in terms of muscle mass and function associated with ageing, as well as the potential role of exercise in mitigating these changes, further studies should focus on these parameters, as to provide a broader overview on the adaptations leading to functional impairment [14,41,46,47]. In addition, though the programme comprised 24 sessions, elderly patients completed a shorter number (as shown in Table 1). This is in accordance with previous reports and highlights the need for further strategies to promote both referral and overall uptake of EBCR programmes in the elderly [16,48,49]. Though this point should be acknowledged, given the lack of significant differences between groups and the inclusion of this parameter in the regression model, this should not preclude the interpretation of the overall results.
Interestingly, a seminal study reported on a correlation between CP and the VE/VCO 2 slope [50]. While differences in terms of overall patient characteristics (particularly when considering functional capacity and ejection fraction) should be taken into consideration, the current findings concerning the VE/VCO 2 slope reinforce the complexity of responses, while also reinforcing the need for further data on the overall determinants of functional capacity in this patient population. Finally, given the recently reported findings by Anand et al. [51] on peak stress cardiac power (as assessed by stress echocardiography) in terms of mortality and HF development among individuals with an ejection fraction above 50%, it would be interesting to have data on stress echocardiography. Given the morphological and functional changes associated with ageing, additional echocardiographic parameters (encompassing both left ventricular size and potential hypertrophy as well as diastolic function) could also be of interest [28,38,40]. While these points should be acknowledged and pondered with, the current results provide a pragmatic assessment on the impact of a contemporary EBCR programme on CP among elderly individuals, showcasing the relevance of this intervention among this challenging group of patients.

Limitations
Several points should be taken into consideration when interpreting the present results. As previously mentioned, this was a retrospective single-centre study, including only those individuals who completed the EBCR programme, with no control group [20]. While this should be acknowledged, prior data from randomized controlled trials in the field should be considered in the face of this caveat [27,28]. Furthermore, patients were under optimized medical therapy (as shown in Table 1), with most being submitted to revascularization. As such, generalization of these findings to other settings should be done cautiously. Secondly, data concerning parameters such as the presence of atrial fibrillation, renal function, and haematological parameters (namely haemoglobin) were not available for the current analysis [52]. Of note, measures of frailty could also be of interest, as this syndrome could lead to limitations in terms of functional assessments as well as EBCR uptake [42,43]. In this regard, methodologies such as the 5-meter gait speed test and the Timed Up and Go test have been proposed as potential ancillary tools in the evaluation of these individuals [43,44]. Given these points, further studies focusing on these parameters could allow additional tailoring of EBCR programmes among the elderly [14,41,44,47]. Arterial hypertension was defined (as described in the Methods section) according to the presence of this diagnosis in the EHR. As such, some individuals under medication due to left ventricular systolic dysfunction or angina could have been missed in this categorization, whereas the lack of ambulatory blood pressure monitoring could also have been of relevance, as to appraise the potential presence of "white-coat" hypertension [4,53]. Albeit these issues, the maintenance of the present results after inclusion of different medications (which could affect blood pressure) on the model should be considered, as further reinforcing the current findings. Thirdly, we dichotomized patients as below or ≥65 years-old. As previously discussed [20], though several definitions have been used to define elderly individuals, this cut-off was chosen given its utilization in different studies [21,36,45,53], with data reporting on its importance in terms of events [54]. Whilst this, the fact that only 3.7% of individuals were aged ≥80 years-old should be considered, as further studies should specifically focus on addressing different responses according to age substrata. Finally, data on CV outcomes was not present, as our aim was to assess the impact of an EBCR programme on functional capacity, as expressed by CP. This information, along with the incorporation of serial echocardiographic assessments (namely stress echocardiography, as explored in the Discussion) should be the focus of future research, as to refine the role of CP in this setting. We believe, however, that while these hindrances should be pondered with, the current study provides novel and relevant data on the CV response to a contemporary EBCR programme among AMI survivors, with a special focus on elderly individuals, thus providing a useful framework for further studies exploring the relationship between CP improvements, potential EBCR programme design personalization and overall CV events.

Conclusions
A contemporary ECBR programme was associated with significant improvements in CP among AMI survivors. Patients aged ≥65 years-old presented significant improvements in CP, though these were less pronounced than those among younger individuals, even after adjustment for several potential confounding factors.
These results highlight the importance of EBCR among elderly individuals, given its physiological and functional benefits, in this challenging higher risk group of individuals.

Author contributions
EMV, AJ, ST, JR, LC, FM, MT were involved in the study's conceptualization and methodology. EMV, MT were involved in the data collection. EMV, RLL, JPN were involved in the statistical analysis. EMV, RLL, MT were involved in the writing of the first draft. EMV, MT, JPN, RFC were involved in the review and editing of the study. All authors read and approved the final manuscript.

Ethics approval and consent to participate
All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the local Ethics Committee (approval number: 229/2017-1).