Chronic heart failure (CHF) is a complex clinical syndrome of symptoms and signs due to poor cardiac function. Patients with CHF are at risk for premature mortality, a lower quality of life (QOL) and more frequent hospital admissions (1). Because of a better survival and aging of the general population, prevalence of CHF and associated health care costs are expected to rise (2). The CHF patients suffer from low QOL because of their inability to perform daily activities, mainly due to excess dyspnea and fatigue (3). Treatment consists of exercise therapy in combination with optimal titration of medication and optionally cardiac resynchronization therapy. Although exercise therapy has the potential to improve QOL and reduce hospital admissions in patients with CHF (1,4), a substantial proportion of CHF patients does not improve cardiorespiratory fitness (CRF) after cardiac rehabilitation (CR) (5), whereas less than 50% of patients seem to adhere to prescribed physical activity guidelines 1 yr after CR (6,7).
Median total exercise training duration for CR is estimated to be merely 4.9 h (range, 1–17.5 h) in CHF patients in the Netherlands (8). Although a meta-analysis concluded that the effects of cardiac training are independent of the dose of the exercise intervention (1), a median of less than 5 h of training seems insufficient to achieve adequate lasting improvement in exercise capacity or behavioral changes. Indeed, a decline in CRF after an initial improvement with commonly prescribed CR methods is not infrequent (9,10). Long-term CR programs may be needed as a support strategy for long-lasting lifestyle changes (11). Such CR programs should include a gradual transition from supervised exercise training at a hospital facility to unsupervised training in a home-based setting using activities that match individual preferences for better adherence to exercise recommendations (12–16).
We hypothesized that a high training frequency with gradual increases in training time in combination with individual coaching of leisure time physical activities that match individual preferences and a gradual transition from facility to home-based training is more likely to produce long-lasting improvements of CRF in patients with CHF. Therefore, we assessed the effects of a 12-month exercise program with a high-frequency, low-intensity, gradually incremented training time on changes in CRF and long-term outcomes in Dutch CHF patients (Moving Forward Project).
Design and patient recruitment
This is a prospective cohort study among Dutch CHF patients with a propensity score matched control group sampled from the same registry. Since March 2005, all patients with CHF visiting the outpatient clinic of Isala (Zwolle, the Netherlands) were registered in the HArtfalen Registratie Project (HARP) database. All patients starting CR from October 2009 to November 2010 were, after signing a written informed consent, included in the analysis. Baseline characteristics of our study population are described in Table 1. The study complies with the declaration of Helsinki and was approved by the medical ethics committee. Patients received optimal medication treatment and were hemodynamic stable at least 3 wk before CR enrolment.
Moving forward project
At the start of the study, patients were hospitalized for 5 d to perform all baseline measurements. First, a maximal cardiopulmonary exercise test was performed to determine maximal exercise capacity and the corresponding training protocol. Patients were informed about their physical fitness and educated about the advantages of regular exercise. The exercise training program was explained by a sport physician and patients started with training under supervision of a physiotherapist specialized in CR. A flowchart of the Moving Forward Project (MFP) is shown in Figure 1. Training consisted of two main goals: daily graded exercise therapy (GET) for improvement of exercise capacity and resistance exercise training (RT) to improve muscular strength. In the GET protocol patients could choose between either 1) three sessions per day with 1 min of exercise at 40% to 50% of their peak oxygen uptake (V˙O2peak) or 2) two sessions per day for 2 min at 40% to 50% of V˙O2peak or 3) one session per day for 3 min at 40% to 50% of V˙O2peak. The low training volume is advocated by the “SMART” framework of setting goals (17), with an exercise volume that is feasible for all patients and known to induce risk reduction (18). The initial training volume was equivalent to 0.5 MET·h·wk−1. Hence, recommended training duration increased every second week to reach a total training time of 15 to 20 min·d−1 after 12 wk (approximately 4.5 to 6 MET·h·wk−1) and, if possible, 45 to 60 min·d−1 after 24 wk (12–18 MET·h·wk−1). Resistance exercise training consisted of a mix of 8 different exercises distributed over 6 d·wk−1 with and without a Dyna-Band durable resistance band (The Hygenic Corporation, Akron, OH). Resistance exercise training was explained and executed during the supervised training sessions and patients were stimulated to train at home with two series of 10 to 15 repetitions. An overview of our training protocol was presented in Supplemental Figures 1 and 2 (Supplemental Digital Content 1, Graded Exercise Protocols with advised total minutes per day of physical activity during follow-up, http://links.lww.com/MSS/B309, and Supplemental Digital Content 2, Training log graded exercise protocol and resistance training exercises of week 1, http://links.lww.com/MSS/B310).
After the first week, patients visited our outpatient clinic three times a week to train under supervision of a physiotherapist for 12 wk. In addition, they started on the other days of the week with their personal daily GET protocol and RT exercises at home. After 12 wk, patients were stimulated to train twice a week for an additional 12 wk with a physiotherapist nearby their home to make the transition from supervised training in an outpatient clinic to unsupervised training at home as smooth as possible and to continue their GET protocol and RT exercises at home. After 24 wk, patients continued to train unsupervised at home for another 6 months according to the GET protocol and the RT exercises (Supplemental Figure 1, Supplemental Digital Content 1, http://links.lww.com/MSS/B309). Patients were able to choose their own activity for the GET protocol, whereas exercise intensity (40%–50% V˙O2peak) did not change in this final phase. Exercise volume was increased every other week if possible. RT exercises maintained a mix of 8 different exercises distributed over 6 d·wk−1 with and without a Dyna-Band durable resistance band.
A maximal cardiopulmonary exercise test was performed at baseline and after 3, 6, and 12 months. Every 3 months, patients were seen by a sport physician to discuss the results of the exercise test and possible problems during training.
Maximal cardiopulmonary exercise test
A maximal cardiopulmonary exercise test with indirect calorimetry measurement was performed on an electromagnetically braked cycle ergometer (Lode Corrival, Lode, Groningen, the Netherlands). After a 4-min warm-up at 0 W workload was increased every minute until exhaustion. The graded increase in workload per minute was individualized for every patient (based estimated fitness level) to reach the maximal workload between 8 and 12 min. The subjects breathed through a mouthpiece connected to a turbine flowmeter for continuous measurement of inspired and expired volume. Gas was continuously sampled by a mass spectrometer (MetaMax IIIb, Cortex, Leipzig, Germany) for continuous measurement of O2 and CO2 partial pressures. The gas exchange quantities V˙ E, V˙O2, and V˙CO2 were computed breath-by-breath (19). The system was calibrated with both ambient air and a fixed known gas mixture.
Control patients were recruited from the HARP database. The information recorded in the database includes patient demographics (year of birth and sex) and characteristics at discharge from the hospital being diagnosed with CHF (height, weight, blood pressure, New York Heart Association (NYHA) class, clinical diagnosis, smoking status, heart rate and rhythm, laboratory results (glomerular filtration rate, sodium, potassium, urea, creatinine, and medication). In addition, clinical symptoms and exercise capacity (available for patients participating in CR) were registered at outpatient visits after hospitalization. Left ventricular ejection fraction was obtained by resting echocardiography during or after hospitalization.
For both groups follow-up started from the moment of registration in the HARP database. Incidence of all-cause mortality was collected from the electronic patient files and the Dutch Municipal Personal Records Database.
For our primary aim differences in peak oxygen uptake were compared between start and all follow-up moments as well as between all follow-up moments separately with a paired t test.
Subsequently, patients in the GET-RT program were propensity score matched with controls from the HARP database based on known risk factors associated with increased mortality in CHF patients (20). A propensity score analysis was performed by a logistic regression model for the GET-RT group versus standard care. The following factors were included in the propensity score model; age, left ventricular ejection fraction, NYHA class, serum creatinine, diabetes mellitus, use of beta blocker, systolic blood pressure, body mass index, time since diagnosis, current smoking, gender, use of angiotensin converting enzyme inhibitor and type of CHF (heart failure with reduced ejection fraction (HFrEF)/heart failure with preserved ejection fraction (HFpEF). To minimize the effects of new insights in CHF treatment the year of admission in the HARP database was likewise added to the propensity score model.
We performed a 1-to-2 propensity score caliper matched analysis without replacement based on the estimated propensity score of each patient as recommended (21). The log odds of the probability of the patients receiving MFP were modeled as a function of confounders that we identified and included in our dataset. Using the estimated logits, we first randomly selected a patient in the GET-RT program group and then matched that patient with patients in the group receiving standard care with the closest estimated logit value. Patients within the standard care group with an estimated logit of 0.2 SD of the selected patients in the GET-RT program group were eligible for matching.
After the propensity score matching was performed, we compared the baseline covariates between the two groups. Continuous variables were compared using the paired t test or the Wilcoxon signed rank test, as appropriate, and categorical variables were compared using conditional logistic regression. The statistical significance and the effect on outcomes were estimated using appropriate statistical methods for matched data. Cumulative survival curves for time-to-death analyses were made with an extended Cox model, with the start of the GET-RT program as time dependent covariate. Dropouts were analyzed according to the intention to treat principle. Subanalyses were performed between patients that successfully completed GET-RT, those that dropped out, and matched controls, to compare the long-term health outcomes of sustainable exercise training. Statistical analyses were performed in SPSS 17.0. A P < 0.05 was considered statistically significant.
A total of 60 CHF patients were enrolled from October 2009 to November 2010. In this period, 308 patients with CHF were registered at the outpatient clinic, resulting in a CR participation rate of 19.5%. Baseline characteristics of our study populations are described in Table 1. Matching was successful in achieving groups that were comparable on known risk factors for CHF and registration date into the HARP database. In total, 57 patients were able to be matched with 2 controls from the HARP databases, whereas 3 patients were only matched with 1 control patient, resulting in a control group of 117 patients. Baseline characteristics did not differ between the intervention and control group (Table 1). The V˙O2peak measurements during follow-up were measured in 49 patients at 3 months, 45 patients at 6 months, and 38 patients at 12 months. Dropout was mainly related due to long traveling distance to the CR center, too many training sessions per week and fatigue. Patient and disease characteristics did not differ at baseline between dropouts and patients with a complete follow-up (Table 2).
Figure 2 and Supplemental Table 1 (Supplemental Digital Content 3, Timewise comparison of fitness and maximal workload during the GET-RT program, http://links.lww.com/MSS/B311) report the change in V˙O2peak and maximal workload during follow-up. Highest increase in V˙O2peak was observed during the first 3 months (+1.1 mL·min−1·kg−1; 95% confidence interval [CI], 0.4–1.8), and maximal benefits of exercise training were observed after 1 yr (+2.6 mL·min−1·kg−1; 95% CI, 1.4–3.8). Similar results were observed in the change of maximal workload during cardiopulmonary exercise testing, with a significant increase from start to 3, 6, and 12 months (+10 W; 95% CI, 3–18; +17 W; 95% CI, 8–25 and +21 W; 95% CI, 12–30, respectively).
Median total follow-up time was similar in both groups with 100 and 108 months for the GET-RT and control group respectively (P = 0.593). Total follow-up showed 23 (38.3%) deaths in the GET-RT group and 63 (53.8%) deaths in the control group. Survival analysis showed a strong but nonsignificant trend for all-cause mortality in the GET-RT versus control group (hazard ratio [HR], 0.63; 95% CI, 0.39–1.03). Subanalyses revealed that patients that successfully completed 6 months (n = 44) and 12 months (n = 38) of the GET-RT program had a lower risk for all-cause mortality compared to matched controls (HR, 0.45; 95% CI, 0.24–0.84; HR, 0.32; 95% CI, 0.16–0.67, respectively). Furthermore, CHF patients completing the GET-RT program (n = 38) had a lower risk for all-cause mortality (HR, 0.30; 95% CI, 0.13–0.70) compared with CHF patients that dropped out from the GET-RT program s (n = 22), whereas mortality risks did not differ between dropouts (n = 22) and matched controls (n = 42) (HR, 1.48; 95% CI, 0.74–3.00).
Six months of guidance with a low-intensity, high-frequency, gradually incremented training time program and resistance training showed a significant improvement in peak oxygen uptake and maximal workload among CHF patients, whereas patients were able to maintain these effects during follow-up after 12 months. A strong trend toward improved survival was found for the intervention compared with control group, whereas a statistically significant mortality reduction was found for CHF patients that successfully completed the supervised training period of six months of the GET-RT program or the total training period of twelve months. These findings suggest that a low-intensity, high-frequency, gradually incremented training time program in combination with resistance training and with individual coaching of leisure time physical activities that match individual preferences and a gradual transition from facility to home based training, induces sustainable improvement of CRF in CHF patients and results in a reduced risk for all-cause mortality.
The GET-RT program was associated with continuous improvement in V˙O2peak that extended into the unsupervised period. The amount of increase in CRF in our population aligns with observations of a meta-analysis (22) and the recently published SMARTEX HF study (10). However, we found no decline in CRF after cessation of the supervised training period whereas physical activity patterns typically decline during long-term follow-up (11) as patients are prone to relapse into old habits. We hypothesize that our extended CR program in combination with personal guidance, advice and reassurance may have contributed to the diminished relapse in physical inactivity, resulting in a retained CRF. Previous studies confirmed that intensive and long-term guidance is an effective support strategy for sustainable lifestyle changes (23). Furthermore, we stimulated physical activities that matched the individual patient preference, which may also have contributed to our findings. Beckers et al. (9) found that engagement in physical training of their own choice was the optimal training modality for maintaining physical capacity in CHF patients. Finally, our primary focus has been on session frequency instead of exercise intensity. A recent meta-analysis by Vromen et al. (24) concluded that the increase in exercise capacity with aerobic exercise training for patients with CHF is primarily determined by total energy expenditure, followed by session frequency, session duration and session intensity. Taken together, our GET-RT program induced sustainable improvements in physical fitness among CHF patients. This observation is clinically relevant as previous studies found substantial risk reductions for death and unplanned hospitalization among CHF patients that improved CRF following CR (5,25).
We found an 8-yr mortality rate of 53.8% in our control group, which aligns with national data of Dutch CHF patients (26). More importantly, patients enrolled in the GET-RT program tended to demonstrate a lower mortality rate (38.3%) compared with controls. A plausible explanation for the lack of statistical significance is the high dropout rate (36.7%) among CHF patients in the GET-RT program. Our intention-to-treat analysis will, therefore, likely underestimate the long-term health benefits of the GET-RT program. Indeed, subanalyses of patients completing 6 or 12 months of the GET-RT program showed a significantly lower risk for all-cause mortality compared to matched controls (HR, 0.45; 95% CI, 0.24–0.84 and HR, 0.32; 95% CI, 0.16–0.67, respectively). Furthermore, the health benefits of the GET-RT program were completely lost in patients that dropped-out as they had a similar mortality rate compared to controls (HR, 1.48; 95% CI, 0.74–3.00). The superior survival rates of CHF patients that successfully completed the GET-RT program are likely due to their increase in CRF as patient and disease characteristics were comparable upon study enrolment with patients that dropped out (Table 2). Furthermore, the main reasons for dropout were long traveling distance to the CR center, too many training sessions per week and fatigue, so it did not include disease characteristics or disease progression. These observations strongly suggest that the GET-RT protocol is superior to standard CR in terms of long-term mortality reduction.
Our results show that a low-intensity, high-frequency exercise therapy with increasing exercise volumes and a graded decrease in guidance could be an alternative approach to contemporary CR programs with moderate-intensity continuous exercise training and/or high-intensity exercise training. In the past decade, numerous studies have compared moderate continuous training with high-intensity training with ambiguous results. Although a meta-analysis showed a superior effect of high-intensity exercise training (27), the recently published SMARTEX HF trial, doubling the number of patients of the meta-analysis, found no effect (10). Our GET-RT program not only showed similar improvement in V˙O2peak compared with SMARTEX after 6 months of supervised exercise training but also demonstrated a lasting affect after 6 months of follow-up. The stepwise tapering of supervision enabled patients to gradually take over their own responsibility for adequate habitual physical activity and exercise training.
Strengths and limitations
The unique assets of this study are the large prospective database with extensive information about patient and treatment characteristics, the repetitive cardiopulmonary exercise testing, and the long-term follow-up period of 8 yr. Limitations of our study are the observational design, the relatively large number of dropouts, and the absence of monitoring of habitual physical activity with activity trackers (e.g., accelerometers). The observational design with propensity matching only ensures balance in measured and not in unmeasured confounders. By matching on the best-known independent predictors of mortality in patients with CHF though, we tried to minimize the effect of a selection bias. Both groups were comparable on all 13 highly significant independent predictors of mortality in patients with CHF (20). The gradual loss of follow-up in V˙O2peak and maximal workload measurements during the first year of follow-up can be taken into account as an attrition bias. The probability of attrition though did not seem to depend on the severity of CHF as baseline characteristics with known predictors of the severity were comparable in patients with complete and incomplete follow-up (Table 2). Furthermore, data loss due to dropout was only relevant for cardiopulmonary outcomes as survival was obtained in the whole study population using the Dutch Municipal Personal Records Database. Finally, accelerometer data would have provided more insight in time-dependent changes in habitual physical activity and the prevalence of patients achieving the exercise guidelines. Although this information was not available in the current study, we found a strong increase in V˙O2peak suggesting that CHF patients adopted a physically active lifestyle during the transition from hospital-based to home-based exercise training.
A high-frequency, low-intensity extended CR program, supporting transition to home environment by exercise training and intensive counseling, results in significant increases in both peak oxygen uptake and maximal workload in patients with CHF. The improvement of fitness characteristics remained present after 6 months of unsupervised exercise training, which is highly unusual in CHF patients. Compared with matched controls, a strong trend toward improved survival was found after a mean follow-up of 8 yr, and moreover, higher mortality rates in patients that failed to complete the study protocol were seen. Further studies with a randomized controlled design are necessary to validate our observations and overcome the limitations of the study.
The authors would like to thank Petra Koopmans for the statistical help, Esmée Bakker and Annemiek Huis in’t Veld for data collection and data entry and Karin Schiphorst, Mineke Vegter, Suzanna de Vries, Henri van de Wetering and Margriet Wijma for their part in setting up the Moving Forward Project.
Financial support: This research received no specific grant from any funding agency in the public, commercial, or non-for-profit sectors. Dr Eijsvogels is financially supported by a grant from the Dutch Heart Foundation (2017T051).
Conflict of interest disclosure: The authors declare that there is no conflict of interest.
The results of the study do not constitute endorsement by ACSM and are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
1. Taylor RS, Sagar VA, Davies EJ, et al. Exercise-based rehabilitation for heart failure
. Cochrane Database Syst Rev
2. Campbell DJ. Heart failure
: how can we prevent the epidemic? Med J Aust
3. Malik FA, Gysels M, Higginson IJ. Living with breathlessness: a survey of caregivers of breathless patients with lung cancer or heart failure
. Palliat Med
4. Abela M. Exercise training in heart failure
. Postgrad Med J
5. Bakker EA, Snoek JA, Meindersma EP, et al. Absence of fitness
improvement is associated with outcomes in heart failure
patients. Med Sci Sports Exerc
6. Ades PA, Maloney A, Savage P, Carhart RL Jr. Determinants of physical functioning in coronary patients: response to cardiac rehabilitation
. Arch Intern Med
7. Conraads VM, Deaton C, Piotrowicz E, et al. Adherence of heart failure
patients to exercise: barriers and possible solutions: a position statement of the study group on exercise training in heart failure
of the heart failure
association of the european society of cardiology. Eur J Heart Fail
8. Vromen T, Spee RF, Kraal JJ, et al. Exercise training programs in Dutch cardiac rehabilitation
centres. Neth Heart J
9. Beckers PJ, Denollet J, Possemiers NM, Wuyts K, Vrints CJ, Conraads VM. Maintaining physical fitness
of patients with chronic heart failure
: a randomized controlled trial. Eur J Cardiovasc Prev Rehabil
10. Ellingsen O, Halle M, Conraads V, et al; SMARTEX Heart Failure
Study (Study of Myocardial Recovery After Exercise Training in Heart Failure
) Group. High-intensity interval training in patients with heart failure
with reduced ejection fraction. Circulation
11. ter Hoeve N, Huisstede BM, Stam HJ, van Domburg RT, Sunamura M, van den Berg-Emons RJ. Does cardiac rehabilitation
after an acute cardiac syndrome lead to changes in physical activity
habits? Systematic review. Phys Ther
12. Deka P, Pozehl B, Williams MA, Yates B. Adherence to recommended exercise guidelines in patients with heart failure
. Heart Fail Rev
13. Blumenthal JA, Babyak MA, O’Connor C, et al. Effects of exercise training on depressive symptoms in patients with chronic heart failure
: the HF-ACTION randomized trial. JAMA
14. Duncan K, Pozehl B, Norman JF, Hertzog M. A self-directed adherence management program for patients with heart failure
completing combined aerobic and resistance exercise training. Appl Nurs Res
15. O’Connor CM, Whellan DJ, Lee KL, et al. Efficacy and safety of exercise training in patients with chronic heart failure
: HF-ACTION randomized controlled trial. JAMA
16. Witham MD, Gray JM, Argo IS, Johnston DW, Struthers AD, McMurdo ME. Effect of a seated exercise program to improve physical function and health status in frail patients > or = 70 years of age with heart failure
. Am J Cardiol
17. Bowman J, Mogensen L, Marsland E, Lannin N. The development, content validity and inter-rater reliability of the SMART-goal evaluation method: a standardised method for evaluating clinical goals. Aust Occup Ther J
18. Eijsvogels TM, Molossi S, Lee DC, Emery MS, Thompson PD. Exercise at the extremes: the amount of exercise to reduce cardiovascular events. J Am Coll Cardiol
19. Beaver WL, Lamarra N, Wasserman K. Breath-by-breath measurement of true alveolar gas exchange. J Appl Physiol Respir Environ Exerc Physiol
20. Pocock SJ, Ariti CA, McMurray JJ, et al. Predicting survival in heart failure
: a risk score based on 39 372 patients from 30 studies. Eur Heart J
21. Austin PC. Statistical criteria for selecting the optimal number of untreated subjects matched to each treated subject when using many-to-one matching on the propensity score. Am J Epidemiol
22. Haykowsky MJ, Timmons MP, Kruger C, McNeely M, Taylor DA, Clark AM. Meta-analysis of aerobic interval training on exercise capacity and systolic function in patients with heart failure
and reduced ejection fractions. Am J Cardiol
23. Gregory S, Bostock Y, Backett-Milburn K. Recovering from a heart attack: a qualitative study into lay experiences and the struggle to make lifestyle changes. Fam Pract
24. Vromen T, Kraal JJ, Kuiper J, Spee RF, Peek N, Kemps HM. The influence of training characteristics on the effect of aerobic exercise training in patients with chronic heart failure
: a meta-regression analysis. Int J Cardiol
25. Martin BJ, Arena R, Haykowsky M, et al. Cardiovascular fitness
and mortality after contemporary cardiac rehabilitation
. Mayo Clin Proc
26. Vaartjes I, Hoes AW, Reitsma JB, et al. Age- and gender-specific risk of death after first hospitalization for heart failure
. BMC Public Health
. 2010;10:637, 2458-10-637.
27. Pattyn N, Coeckelberghs E, Buys R, Cornelissen VA, Vanhees L. Aerobic interval training vs. moderate continuous training in coronary artery disease patients: a systematic review and meta-analysis. Sports Med