Protective effect of physical activity on mortality in older adults with advanced chronic heart failure: A prospective observational study

Abstract

Objectives

The objective of this study was to evaluate the effect on mortality of self-reported physical activity evaluated by the physical activity scale for the elderly (PASE) in elderly patients with advanced heart failure enrolled in a cardiac rehabilitation unit after heart failure decompensation (NYHA class IIIB).

Methods

The study prospectively enrolled 314 elderly patients (≥65 years) with heart failure in NYHA class IIIB (symptomatic with a recent history of dyspnoea at rest) consecutively admitted to cardiac rehabilitation between January 2010 and July 2011. Comprehensive geriatric assessment was performed. Physical activity was evaluated by PASE and stratified in tertiles (0–15, 16–75 and >75). Mortality was collected from September to October 2015 in 300 patients.

Results

The mean age was 74.5 ± 6.1 (range 65–89); 74.7% were men, 132 patients (44.0%) died during the follow-up (44.1 ± 20.7 months). Univariate analysis shows that physical activity level conducted before heart failure decompensation was inversely related to mortality (from 76.0% to 8.2%, P = 0.000). Multivariate analysis confirms that the PASE score predicts mortality independently of several demographic and clinical variables (hazard rate 0.987, 95% confidence interval (CI) 0.980–0.994, P = 0.000). Notably, when considering PASE 0–15 versus 16–75 score and PASE 0–15 versus > 75 score, the hazard rate is 4.06 (95% CI 1.67–9.84, P < 0.001) and 7.25 (95% CI 2.7–19.5, P < 0.001), respectively.

Conclusions

Physical activity level evaluated by the PASE score is inversely related to mortality in elderly patients with advanced heart failure confirming the reduction of mortality exerted by moderate physical activity in such patients.

Introduction

Heart failure (HF) is one of the most frequent and life-threatening diseases in the elderly population.¹ People with HF have a short life expectancy, even in this era of modern medical therapies.^1,2 Regrettably, its frequency is bound to rise due to the aging population and the growing number of people with high risk factors such as obesity and diabetes. HF is a dynamic clinical scenario characterised, especially in the elderly, by frequent decompensation.^1–4

Physical activity (PA) plays an important role in determining HF prognosis,^5,6 and the pathogenic mechanisms by which PA exerts a protective role in the natural history of the disease are still debated.⁷ Actually, it is not known at which intensity PA is capable of modifying HF prognosis in the elderly population.^5–8

The Framingham Heart Study demonstrated in 1142 elderly participants without prior myocardial infarction a protective effect of PA on the risk of HF.⁹ More recently, in the Cardiovascular Health Study conducted on a sample of 5503 subjects aged 65 years and older, free of HF at enrollment, the incidence of HF was inversely related to the weekly metabolic equivalent (MET) consumed but only a high level of PA showed a reduced risk of HF.⁵ It is also known that exercise training reduces mortality, hospital admissions, cardiac events and improves the quality of life in patients with HF.^10,11 Thus, exercising is to be proposed as an essential treatment for the treatment of HF also in the elderly in the advanced phase of disease.¹²

The Heart Failure – A Controlled Trial Investigating Outcomes of Exercise TraiNing (HF-ACTION) study investigated the relation between the intensity of exercise and clinical outcomes in patients with HF. The study demonstrated a reverse J-shaped association between exercise intensity and adjusted clinical risk. Moderate exercise amounts of 3 to less than 5 MET-hours and 5 to less than 7 MET-hours per week were associated with reductions in subsequent risk that exceeded 30%.¹³

Another important item is the tool to measure the PA, especially in older adults.¹⁴ In elderly people, PAs usually performed in everyday life could be assessed with a specific assessment tool (the physical activity scale for the elderly (PASE)) that combines information on leisure, household and occupational activity.¹⁵

The aim of the present study is to evaluate whether PA, evaluated by the PASE score, affects mortality in elderly patients with advanced HF enrolled in a cardiac rehabilitation unit after decompensation (New York Heart Association (NYHA) class IIIB).

Methods

The study prospectively enrolled a cohort of 314 elderly patients (≥65 years) with systolic HF in NYHA class IIIB (symptomatic patients with a recent history of dyspnoea at rest), who were consecutively admitted to the cardiac rehabilitation of the Maugeri Institute of Telese-Terme, Italy, transferred from acute medical units soon after clinical stabilisation of HF decompensation, between January 2010 and July 2012. The mortality was collected in 300 patients between September 2015 and October 2015. On admission, all patients were evaluated with an extensive comprehensive geriatric assessment collecting the following data: age, sex, presence of coronary artery disease, previous stroke, chronic obstructive pulmonary disease (COPD), diabetes, renal insufficiency (creatinine ≥2.0 mg/dl). The ejection fraction (EF) was assessed by cardiac ultrasound (SONOS 5500; Hewlett Packard).

Usual PA performed before HF decompensation was measured by using the PASE,¹⁶ usually on the second day after hospital admission in all patients and, in any case, within the first week. The PASE is a brief (5 minute) and easily scored questionnaire designed specifically to assess PA in epidemiological studies of persons aged 65 years and older. The PASE assesses PA over a 1-week time frame. The questionnaire combines information on leisure, household and occupational activity. It records participation in leisure activities (including walking outside the home), sport and recreation (light, moderate and strenuous) and muscle strengthening; the frequency is classified according to four categories: never, seldom (1–2 days/week), sometimes (3–4 days/week), and often (5–7 days/week). Duration is also classified according to four categories: less than 1 hour, between 1 and 2 hours, 2–4 hours, or more than 4 hours. Paid or unpaid work, other than work that involves mostly sitting activity, is recorded as total hours per week. Housework is classified as light and heavy. The total score is calculated by multiplying the amount of time spent in each activity (hours/week) or participation (yes/no) in an activity.¹⁶ Cognitive function was assessed with the mini-mental state examination (MMSE)¹⁷ while depressive symptoms were assessed with the geriatric depression scale (GDS) short form 15 items.¹⁸ Disability was evaluated by the Barthel index, which measures the patient's autonomy in 10 domains, including bowel and bladder control. For each category, the score ranges from 0 to 15 points depending on the need for help. The total score ranges from 0 to 100. Higher scores indicate a greater degree of independence.¹⁹ The 6-minute walk test (6MWT) was performed in all patients, in all cases within the first week after admission. Patients were instructed to walk as far as possible along a corridor of 40 metres, for 6 minuteds.²⁰ The test was interrupted whenever the patient displayed symptoms such as angina, severe dyspnoea, dizziness and musculoskeletal pain; the patient was asked to stop walking and resume when possible. Of the 300 patients, 24 (8.0%) were not able to perform the 6MWT in the first week of hospitalisation. These patients were not excluded from the sample, but their score was considered equal to 0.

The severity of comorbid conditions was evaluated using the cumulative illness rating scale (CIRS).²¹ The CIRS is considered to be a valid and reliable measure for the measuring of comorbidity. The index quantifies the burden of chronic disease in a patient by taking into account the number and severity of different diseases among the 14 different conditions analysed. Each condition is evaluated on a scale of zero to four: a score of 0 indicates that there are no problems that affect that system; a score of 1 indicates a condition that does not require therapy; 2 indicates the need for treatment; 3 indicates a serious problem that can be either associated with significant disability or is difficult to control; 4 indicates an extremely serious disease, organ failure, or severe functional impairment.

The ethics committee of Maugeri Foundation approved the study and approved all parts of the project. All the participants provided written informed consent.

Statistical analysis

The data were entered into a specific database and analysed with SPSS 13.0. The Kolmogorov–Smirnov test was used to evaluate the linearity of the data. Analysis of variance was used for comparison between patients who died and survived and among PASE tertiles. The χ² test was employed for discrete variables comparison. Multivariate linear regression analysis was performed with PASE as the dependent variable to assess the independent relationship with age, sex, GDS, MMSE, CIRS, 6MWT, Barthel index and EF. The Cox regression analysis was used to evaluate the predictive effect on mortality. The latest model was built considering only the variables statistically significant in the univariate analysis, namely: age, sex, CIRS, GDS, MMSE, PASE, Barthel index and 6MWT. For analytical purposes, we reorganised the PASE score into tertiles (0–15, 16–75, 76–236). Multivariate Cox regression analysis considering PASE as the dummy variable was also used to estimate the effect on mortality of a PASE score of 16–75 and 76–236 to 0–15 after correction for age, sex, CIRS, MMSE, GDS, PASE, 6MWT and Barthel index. A P value less than 0.05 was considered statistically significant.

Results

The mean age of the study population was 74.5 ± 6.1 years (range 65–89). Among them, 51% were over 75 years; 74.7% were men. All patients were referred to rehabilitation in NYHA IIIB class soon after decompensation treated in an acute medical setting. The ischaemic aetiology affected 242 patients (80.7%). Previous myocardial infarction and angina were present in 52.0% of patients, and 13.7% had previously undergone successful revascularisation by percutaneous transluminal coronary angioplasty and 10.7% by coronary artery bypass grafting. Previous valve surgery was performed in 4%. Cardiac resynchronisation treatment was present in 12.3% while 19.2% had implantable cardioverter defibrillators. Diuretics were used in 287 (95.8%) patients with the exclusion of eight patients in dialysis and five patients in whom the diuretics were temporarily withheld because of a rapid increase in creatinine levels. Beta-blockers were used in 196 patients (65.3%), angiotensin-converting enzyme inhibitors in 216 (72%), angiotensin II receptor blockers in 56 (18.7%), in 32 (10.7%) they were used together. Mineralocorticoid receptor antagonists were used in 138 (46%), digoxin in 13.9%, oral anticoagulants in 68 (22.7%), antiplatelet agents in 208 (69.3%) and lipid-lowering agents in 164 (54.7%); 52.0% were suffering from COPD, 2.7% were on oxygen therapy, 21.5% had had a previous stroke (11.0%) or transient ischaemic attack (10.5%), 45.2% had serum creatinine levels of 2 mg/dl or greater and 2.7% were on dialysis.

The prevalence of cognitive impairment (MMSE < 24) and depression (GDS ≥ 5) were, respectively, 33.9% and 44.0%. A total of 132 patients died (44.0% after a mean follow-up of 44.1 ± 20.7 months). Univariate analysis showed that patients who died were older and had higher comorbidity and GDS scores, with lower systolic blood pressure values and lower scores on the MMSE, PASE, Barthel index and 6MWT. No significant difference was found for sex, body mass index, diastolic blood pressure, HR and EF (Table 1).

With the rise of PA, stratified among PASE tertiles, we observed an increase in MMSE, EF, 6MWT and Barthel index from the lowest to the highest tertiles, while a decrease in age, GDS and CIRS score was observed. No statistical differences were found for sex (Table 2). Mortality decreased with the increase in PASE scores (P for trend = 0.000) (Figure 1 and Table 2).

Multivariate linear regression analysis on the PASE score demonstrated an inverse relationship with age and GDS score, while a significant positive correlation is present for CIRS, MMSE score, 6MWT, Barthel index and male sex. No statistical difference was found for EF (Table 3).

Cox regression analysis showed that PA usually performed in the week before the hospitalisation predicts mortality independently of the GDS, CIRS, Barthel index, MMSE score and 6MWT (Table 4). Figure 2(a) described the survival curves obtained with Cox regression analysis of PASE stratified in tertiles on mortality. When we analysed PASE tertiles as dummy variables with the lowest PASE tertiles as reference we found a higher hazard rate (HR) versus the highest tertiles (HR 7.25, 95% confidence interval (CI) 2.7–19.5, P < 0.001) and a lower HR (HR 4.06, 95% CI 1.67–9.84, P < 0.001) for a PASE score between 16 and 75 (Figure 2(b)).

Discussion

This study shows that daily PA, whether pursued for its own sake or linked to the activity of daily living is a predictor of mortality regardless of comorbidities, disability and physical function in elderly patients with heart failure in NYHA class IIIB. The risk of death is inversely related to PA levels: the higher the involvement in leisure and in household and occupational activity the lower mortality.

Mortality in elderly patients with decompensated severe HF

HF in the elderly is characterised by high mortality both in the acute and in post-acute phase. For a sample of 279 elderly (>75 years old) hospitalised in cardiac rehabilitation for decompensation of chronic established HF, Passantino et al. recently recorded mortality rates of 10%, 36% and 77% for in-hospital, 12-month and 5-year mortality, respectively.²²

PA and mortality in the elderly

PA designates a global status of general health. The literature is consistent with the hypothesis that physical function in elderly patients is an independent predictor of death.²³ There is evidence that strongly supports the positive association between increased levels of PA and improved health in older adults.²⁴ Muscle repair by means of PA is a possible therapeutic strategy in the field of cardiovascular muscle repair.²⁵ In the elderly, PA is referred to as any activity that has energy cost. Thus, it often refers to housework, shopping, gardening more than structured exercise programmes, as suggested by the World Health Organization and the American College of Sports Medicine guidelines: 150 minutes per week of moderate intensity PA is needed for health benefit in older adults.^26,27

PA and mortality in elderly patients with decompensated severe HF

PA is a safe and valid procedure with a protective effect on general functional capacity and prognosis in the elderly patient with HF.^7,10,15 In a meta-analysis conducted on 81 studies, no exercise-related deaths occurred during more than 60,000 hours of exercise training.²⁷ In the 57 studies in which a cardiopulmonary exercise test was performed there was a mean increase of maximum oxygen uptake of 17%.²⁷ An increasing number of studies has demonstrated that exercise training is effective in reducing HF mortality. The first study, conducted by the ExTraMATCH Collaborative group analysis, found a 35% lower risk of death and a 28% lower risk for the composite endpoint of mortality or hospitalisation for 801 patients enrolled in nine randomised controlled clinical trials.¹² A similar result was established by Smart and Marwick in a meta-analysis conducted on 11 randomised clinical trials (including 729 patients) in which the exercise group displayed a 39% lower risk of mortality.²⁷ The HF-ACTION trial, which was conducted on 2331 patients with stable systolic HF with left ventricular EF of 35% or less, in NYHA functional class II–IV²⁸ randomly assigned to 36 sessions of supervised, moderate-intensity training followed by home-based training or the usual care group, after a median follow-up time of 30 months, revealed a 11% reduction in the training group of all-cause mortality or hospital stay.

The protective impact of PA in reducing HF-related disability is due also to the effect exerted on other organs and systems, such as the brain and muscles compromised during the natural history of HF,^3,7,15 with a linear improvement in maximal oxygen consumption and maximal cardiac output following endurance training.²⁹ It has recently been demonstrated that impairments in either gait speed or instrumental activities of daily living, measured within 1 year before the diagnosis of incident HF, were independently associated with mortality.³⁰ There is a strong systemic effect of PA also mediated by an improvement in the endothelial progenitor cell number and function in chronic HF patients, independent of age,³¹ that could explain the efficacy on different systems, i.e. our data also demonstrate an inverse relationship of the PASE score with age and GDS, while a significant positive correlation is found for CIRS, MMSE score, 6MWT, Barthel index and male sex. No statistical effect is found for EF. These data show that PA is related only partially to clinical status, i.e. a positive relationship is observed between CIRS and PASE and no statistical difference was found with EF, a specific determinant of HF severity. This means that the presence of more disease is related to greater PA independently of the status of HF, while an increase in depressive symptoms and reduced physical functioning and cognitive status are related to reduced PA.

What type of PA should be recommended in elderly patients with HF?

The PA in the elderly with advanced HF is conditioned by the general clinical conditions, the presence of comorbidities, disability and frailty. Even when the elderly are able to perform low-intensity resistance exercises, gradually distributed over a period, this ability is often limited by general conditions, attitude to PA and the reduction of autonomy, namely the ability to move independently and perform all the functions of everyday life. We suggest enhancing PA through the activities of daily living by turning leisure, household and occupational activity into structured exercise, also considering the lack of efficacy of innovative procedures such as electrical myostimulation on top of exercise training, which do not demonstrate any significant additional improvement in exercise capacity in cardiac HF patients.³² Daily PA in the elderly population with HF plays a significant role in conditioning the body and in the management of associated diseases such as diabetes, obesity and hypertension,³³ by improving muscle tone and reducing the phenomenon of age-related sarcopenia and cachexia, which modulates HF and is common in patients with advanced HF.^7,34

PA as objective measure to LVAD implant as destination therapy in the elderly

ESC guidelines identify the left ventricular assist device (LVAD) as destination therapy in highly selected patients with end-stage HF despite optimal pharmacological and device therapy. These are patients who are not suitable for heart transplantation, but are expected to survive more than 1 year with good functional status. LVAD is supposed to improve symptoms, and reduce the risk of hospitalisation and of premature death.³⁵

PASE could be a valid and easy instrument better to stratify elderly patients for LVAD implantation as destination therapy. Combined low levels of PASE, disability and 6MWT identify a cohort of HF patients with a non-reversible frailty³⁶ and high mortality (in our study 76%) who should not be suitable for LVAD implant, but rather for palliative care. Advanced HF contributes significantly to the frailty status and is potentially reversible with LVAD (LVAD-responsive frailty). However, many patients with advanced HF may be frail due to conditions unrelated to HF severity, which is not treatable/reversible with LVAD (LVAD-independent frailty) (i.e. cognitive impairment, non-temporary disability, lack of social support). Thus, in our experience a slightly lower PASE score and disability and functioning only slightly reduced identify patients with low mortality (8.2%) and potentially reversible frailty, in which LVAD as destination therapy should be the gold standard to reverse end-stage HF treatment.

Limitations of the study

The main limitation of the study is the incomplete availability of N-terminal pro-b-type natriuretic peptide (NT-pro-BNP) values. Cox regression performed in the 182 patients with NT-pro-BNP values, the PASE score is still protective on mortality (HR 0.987, 95% CI 0.978–0.996) while NT-pro-BNP is an independent predictor of mortality (HR 1.017, 95% CI 1.011–1.023). Moreover, a subscore on different items of the PASE questionnaire is omitted as only the total score of PASE is evaluated.

Conclusions

Our results demonstrate that PA performed before HF decompensation in the elderly with advanced HF is associated with a lower mortality. Our findings also enhance the importance of PA in vulnerable populations such elderly with advanced HF. A small amount and intensity of PA, even less than of that recommended in the general population, is an effective measure to improve survival even in patients with comorbidity, compromised physical functioning and disability.

Author contribution

FC, CA, CM, DB, CN and PA contributed to the conception or design of the work. FC contributed to the acquisition, analysis and interpretation of data for the work. FC, CA, CN and PA drafted the manuscript. FC, CA, IL, CN and PA critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work ensuring integrity and accuracy.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References