The HEART Camp Exercise Intervention Improves Exercise Adherence, Physical Function, and Patient-Reported Outcomes in Adults with Preserved Ejection Fraction Heart Failure

Abstract

Background:

Despite exercise being one of few strategies to improve outcomes for individuals with heart failure with preserved ejection fraction (HFpEF), exercise clinical trials in HFpEF are plagued by poor interventional adherence. Over the last 2 decades, our research team has developed, tested, and refined Heart failure Exercise And Resistance Training (HEART) Camp, a multi-component behavioral intervention to promote adherence to exercise in HF. We evaluated the effects of this intervention designed to promote adherence to exercise in HF focusing on subgroups of participants with HFpEF and heart failure with reduced ejection fraction (HFrEF).

Methods and Results:

The randomized controlled trial included 204 adults with stable, chronic HF. Of those enrolled, 59 had HFpEF and 145 had HFrEF. We tested adherence to exercise [defined as greater than or equal to 120 minutes of moderate-intensity (40–80% of heart rate reserve) exercise per week validated with a heart rate monitor] at 6, 12, and 18 months. We also tested intervention effects on symptoms (PROMIS-29 and Dyspnea-fatigue index), HF-related health status (Kansas City Cardiomyopathy Questionnaire, KCCQ), and physical function (6-minute walk test, 6MWT). Participants with HFpEF (n=59) were a mean of 64.6 ±9.3 years old, 54% male, 46% non-white with a mean ejection fraction of 55±6. Participants with HFpEF in the HEART Camp intervention group had significantly greater adherence compared to enhanced usual care at both 12 (43% vs. 14%, phi=.32, medium effect) and 18 months (56% vs. 0%, phi=.67, large effect). HEART Camp significantly improved walking distance on the 6MWT (η² = 0.13, large effect) and the KCCQ overall (η²=0.09, medium effect), clinical summary (η²=0.16, large effect), and total symptom (η²=0.14, large effect) scores. In the HFrEF subgroup, only patient-reported anxiety significantly improved in the intervention group.

Conclusions:

A multi-component, behavioral intervention is associated with improvements in long-term adherence to exercise, physical function, and patient-reported outcomes in adults with HFpEF and anxiety in HFrEF. Our results provide strong rationale for a large HFpEF clinical trial to validate these findings and examine interventional mechanisms and delivery modes that may further promote adherence, and improve clinical outcomes, in this population.

Clinical Trial Registration:

URL: https://clinicaltrials.gov/. Unique identifier: NCT01658670

Lay Summary

Heart Failure Exercise and Resistance Training (HEART) Camp is an exercise intervention that was previously successful at improving adherence to exercise in adults with heart failure (HF). In this study, we divided the HEART Camp trial participants based on their ejection fraction (EF) into preserved EF (≥50%) and reduced EF (<50%) subgroups to see how the subgroups responded to the intervention compared to counterparts that were given only paid access to a fitness center. Adults with preserved EF HF significantly improved their exercise adherence with the HEART Camp intervention compared to a group receiving only access to a fitness center.

Graphical Abstract

Introduction

Exercise is one of few successful strategies to improve outcomes in adults with heart failure with preserved ejection fraction (HFpEF). Yet, few adults with HF successfully meet the Heart Failure Society of America (HFSA) exercise guidelines that recommend 30 minutes of supervised moderate intensity exercise five days per week and long-term adherence to exercise interventions remains a challenge.^1,2 Cardiac rehabilitation has been the predominant model of exercise for adults with cardiovascular disease and heart failure; however, Medicare reimbursement is limited to those with an ejection fraction less than 35% (heart failure with reduced ejection fraction – HFrEF), which fails to account for half of the HF population, and cardiac rehabilitation is notoriously underutilized by those adults with HF that are eligible. ^3,4

Several exercise clinical trials have targeted adults with HFpEF and HFrEF.^5–14 The majority of these studies examined short-term effects (4–20 weeks) of exercise on physiologic changes such as peak oxygen uptake (VO₂).^{5–9,11–13} Most of these trials do report efficacy in improving physiologic measures with the exercise interventions, including aerobic, resistance, and/or high intensity interval training, but few report adherence outcomes and no studies to date have resulted in significant improvements in mortality or hospitalizations. Further, few studies included patient-reported outcomes and in those that did, only quality of life^6,10,12 and depression¹³ were reported. Quality of life improvements were reported in some but not all studies and it was unclear if improvements were sustained.

Recent findings from the Rehabilitation Therapy in Older Acute Heart Failure Patients (REHAB-HF) trial, which recruited participants during a hospitalization for acute HF to attend 1-on-1 exercise training sessions that began during the hospital admission to 12 weeks, were promsing.¹⁵ Adherence to exercise was reported as high as 76% during the intervention and 78% at 3-month follow-up.¹⁴ However, adherence during the 12-week intervention period was measured by session attendance only and by patient self-report at the 3 month follow-up.¹⁴

Optimizing Exercise Training in Prevention and Treatment of Diastolic Heart Failure (Optimex-Clin)¹⁰ is the only completed trial to date to recruit a HFpEF-only sample and examine long-term (12 month) exercise outcomes. Optimex-Clin randomized participants to high-intensity interval training, moderate intensity continuous training, and a control group that received only a guideline-based recommendation to exercise. Non-significant differences in primary endpoint, peak VO₂ at 3 months were noted.¹⁰ Adherence to Optimex-Clin at 12 months, measured as session attendance, was modest in both the moderate intensity continuous training and high-intensity interval training groups,¹⁰ but proportionally higher when compared to adherence reported in two prior exercise clinical trials in HF: (a) the seminal, multi-center trial, Heart Failure – A Controlled Trial Investigating Outcomes of Exercise Training (HF-ACTION) (34% adherence)¹⁶ and (b) prior work of this research team (42% adherence).¹⁷ The dose-dependent response associated with exercise is difficult to achieve and maintain when participants are non-adherent to exercise.^18–20

Our team developed and tested the multi-component Heart Failure Exercise and Resistance Training (HEART) Camp behavioral intervention to promote long-term adherence to exercise in adults with HF.²¹ We defined adherence as achieving ≥120 minutes of moderate intensity exercise per week validated with a heart rate monitor. The HEART Camp intervention group achieved significantly better long-term adherence at 12 (42%) and 18 months (35%) compared to enhanced usual care (28% and 19% adherence, respectively) in a mixed sample of adults with HFpEF and HFrEF.²² Effects on adherence were moderated by ejection fraction with those with higher ejection fractions at baseline achieving better long-term adherence. HEART Camp also significantly improved physical function, quality of life, and depression.²³

HF-ACTION (n=2,331) reported safety and efficacy of exercise in reduced ejection fraction HF (HFrEF); however, participants with HFpEF were excluded.¹⁶ A similar trial in HFpEF has not been reported. Our knowledge of successful interventional strategies to initiate and sustain exercise in HF, particularly in HFpEF, remains limited. Therefore, the purpose of this secondary data analysis is to evaluate the effects of the multi-component, Heart failure Exercise And Resistance Training (HEART) Camp on adherence to exercise in the HFpEF and HFrEF subgroups as compared to enhanced usual care.

Methods

Design.

The parent HEART Camp study was a two-site, prospective, randomized two-group repeated measures experimental design. Data was collected at 4 timepoints (baseline, 6, 12 and 18 months). Details of the study protocol and the results were previously published.^21–23 In brief, after enrollment and completion of a cardiopulmonary exercise test, all participants took part in a minimum of 6 scheduled, supervised exercise sessions in cardiac rehabilitation prior to randomization. During pre-randomization exercise training sessions, participants learned moderate-intensity continuous training (40%−80% heart rate reserve; HRR)^22,24 and resistance training (10–15 repetitions to volitional fatigue).^25,26 The safety of all participants was assured and plans for participant supervision were clearly outlined in the study protocol. Participants that completed these sessions were then randomized with stratification by study site and gender to 2 groups (HEART Camp or enhanced usual care). After randomization, all participants, including those randomized to enhanced usual care, received standard care for heart failure and were given paid access to a hospital-based exercise training facility where they could exercise independently for 18 months. This paid access was considered an enhancement beyond the standard care of individuals with HF and eliminated a potential financial barrier for participants. Participants randomized to the HEART Camp group received the HEART Camp intervention which is described in more detail below.

Setting and Sample for the Parent Study.

The HEART Camp study team recruited 204 subjects from two, urban medical centers: Bryan Heart Institute in Lincoln, NE and Henry Ford Hospital in Detroit, MI. Inclusion/exclusion criteria are shown in Table 1. All subjects provided written informed consent and the Institutional Review Board at each site approved the study and provided oversight. The CONSORT diagram with screening and randomization details has been published previously.²² Randomization was stratified by site and gender.

Table 1.

Inclusion/exclusion criteria for the HEART Camp clinical trial

Inclusion criteria:	Exclusion criteria:
1. Diagnosis of systolic, diastolic, or combined chronic HF (Stage C confirmed by echocardiography and clinical evaluation) 2. 19 years of age or greater;	1. Clinical evidence of decompensated HF 2. Unstable angina pectoris 3. Myocardial infarction, coronary artery bypass surgery, or biventricular pacemaker less than 6 weeks prior
3. Able to speak and read English 4. Telephone access in home 5. Stable pharmacologic therapy per guidelines for past 30 days.	4. Orthopedic or neuromuscular disorders preventing participation in aerobic exercise a.d strength/resistance training 5. Participation in 3 times per week aerobic exercise in the past 8 weeks
	6. Cardiopulmonary exercise test results that precluded safe exercise training
	7. Plans to move more than 50 miles from the exercise site within the next year
	8. Peak oxygen uptake (pVO2) in females > 21 ml/kg/min and in males > 24 ml/kg/min
	9. Pregnancy planned or current

Sample for the Sub-group Analyses.

Recently, the Heart Failure Society of America, the Heart Failure Association of the European Society of Cardiology, and the Japanese Heart Failure Society published a universal definition of HF.²⁷ The new guideline suggests classification of HF on the basis of ejection fraction; individuals with ejection fraction ≥50% are classified as HFpEF.²⁷ We applied this classification to identify our subgroups for this post hoc analysis: HFpEF (EF≥50%) (n=59) and HFrEF (EF<50%) (n=145).

HEART Camp Intervention.

HEART Camp is a theory-driven, multi-component intervention that focuses on participants’ knowledge, attitudes, self-efficacy, self-management skills and social support. The intervention took place over 3 phases, adoption (baseline – 6 months), transition (months 7–12) and maintenance (months 13–18). During the adoption and transition phases, the intervention included individualized, weekly coaching with an exercise coach where coaches and participants discussed goals, self-monitoring, relapse prevention, and self-management strategies. During the maintenance phase, face-to-face interaction between subjects and coaches was stopped and participants were expected to self-manage their exercise. A detailed interventional protocol was published in 2014.²¹

Data Collection and Measurement Tools

Baseline measures included demographic information (age, race, sex and marital status) and clinical information (body mass index, left ventricular ejection fraction [LVEF], and New York Heart Association classification). Adherence to exercise (primary outcome) was calculated based on confirmed minutes of exercise per week. Participants achieving ≥ 120 minutes of moderate intensity exercise per week or more (80% of the HFSA recommended 150 minutes per week) were considered adherent. Minutes of exercise were collected using self-report diaries (collected weekly in the HEART Camp group and every 2 months in the enhanced usual care group) and validated using a heart rate monitor (Polar Electro Inco, Lake Success, NY). In designing our study, we elected to collect adherence data from participants in the enhanced usual care group every 2 months to avoid interventional effects from data collection. We pre-programmed the heart rate monitor with an individualized, target heart rate range that was determined based on results from each participant’s cardiopulmonary exercise test and corresponded to 40–80% of their heart rate reserve. Participants were instructed to wear the heart rate monitor during all exercise sessions and stay within the prescribed target heart rate range. We downloaded data into the Polar Pro Trainer which provides details on date, time spent in exercise, average HR and amount of exercise time within the target range. Any exercise minutes below the 40% heart rate reserve were not counted as minutes of exercise. Adherence levels for 6, 12, and 18 months were computed as the averages across the 4 weeks prior to each measurement time point. Participants achieving ≥ 120 minutes of moderate intensity exercise per week on average over this 4-week period were considered adherent at the given time point.

Secondary outcomes were tested at baseline, 6, 12, and 18 months and included objective physical function (6-Minute Walk Test) and patient-reported health status (Kansas City Cardiomyopathy Questionnaire), anxiety, depression, role satisfaction, fatigue, pain interference, physical function, and sleep disturbance (Patient-Reported Outcomes Measurement Information System (PROMIS)-29). Each of these is described in brief.

The 6-Minute Walk Test (6MWT)²⁸ is an objective measure of physical function performed to assess exercise capacity of individuals with HF.²⁹ During this self-paced test, which has been standardized through guidelines provided by the American Thoracic Society,²⁸ participants walk on a flat, 100-foot walkway while research personnel measure the distance in meters walked in 6 minutes. In the HEART Camp study, participants were asked at baseline, 6, 12, and 18 months to complete two 6-minute walk tests 7 days apart. The farthest distance walked between these two tests was recorded as the 6MWT measurement for that time point.

The Kansas City Cardiomyopathy Questionnaire (KCCQ) is a 23-item, HF-specific tool that measures 5 domains (physical limitation, symptoms, self-efficacy, quality of life, and social limitation) and generates 3 summary scores (overall summary score, clinical summary score, and total symptom score).³⁰ The KCCQ is commonly used in HF clinical trials and was recently recommended as an important quality indicator by the 2020 American College of Cardiology/American Heart Association Task Force on Performance Measures.³¹ Overall summary scores that change by 5 points or greater over time are considered clinically meaningful. ³² The KCCQ has been validated in adults with HFpEF.³³

The Patient-Reported Outcomes Measurement Information System (PROMIS)-29 uses 29 items to measure 7 domains including: anxiety, depression, role satisfaction, fatigue, pain interference, physical function, and sleep disturbance. PROMIS tools were developed by Cella et al to capture patient-reported outcomes.³⁴ These measures have been validated in the HF population.^35,36

Statistical Analyses

All statistical methods were applied across both the HFpEF and HFrEF subgroups. Descriptive statistics were calculated on all demographic and dependent variables. Data were compared between HEART Camp and EUC groups as well as HFpEF and HFrEF groups using t-tests and Chi-square analyses as appropriate. To assess adherence between groups at 18 months (primary aim), a chi-square test was used to test group differences in the proportion of the sample that adhered to at least 80% of the 150 minutes per week recommendation. Tests were carried out at 6, 12, and 18 months, the primary data collection points in the study. A Fisher’s exact test was completed for any time point with less than 5 participants in either group. The analysis included all participants according to the condition to which they were assigned and was based on all available data. (No data imputation was performed.) Missing data was assessed prior to analyses and our missing data rate is 18%.

Analysis of variance tests were completed to calculate effect sizes, reported as η², for secondary outcomes. In order to assess group differences in change over time on continuous outcomes, linear mixed models were performed for each dependent variable. Significant time by group interactions indicate significant differences in change over time between the HC and EUC groups. Variables were examined for normality by examining skew and kurtosis values. While the PROMIS measures were normally distributed, the 6-minute walk test and KCCQ variables were negatively skewed. Square transforms (or 2^nd power transforms) were effective in normalizing the distributions. SPSS version 25 was used for data management and descriptive statistics and SAS 9.4 was used for linear mixed models. Time was treated as a categorical variable since all intervals between time points and across participants were equal. Prior to assessing fixed effects, the most appropriate covariance structure was determined by performing models with various structures (e.g., variance components, unstructured, Toeplitz, autoregressive) and comparing fit statistics. In the even that all fit statistics did not agree on the best error structure, preference was given to the more parsimonious structure. Additional models were performed to examine the effect of covariates (age, gender, body mass index, and race (white vs. non-white)) on intervention effects in both subgroups. Statistical significance was set at the alpha .05 level.

Results

The parent study enrolled 204 adults with stable chronic HF. Fifty-nine (29%) of these participants had an ejection fraction ≥50% (HFpEf subgroup) and 145 had an EF <50% (HFrEF subgroup. Characteristics of the HFpEF and HFrEF subgroups are presented in Table 2. Participant characteristics were balanced between intervention and enhanced usual care groups in both the HFpEF and HFrEF subgroups. Demographic and clinical variables were compared between the HFpEF and HFrEF subgroups. We found HFpEF participants were significantly older (p=.001) and less likely to be taking a beta blocker (p=0.025) compared to HFrEF counterparts. There were no significant differences between the HFpEF and HFrEF subgroups on gender, race, marital status, BMI, New York Heart Association class or peak oxygen consumption (pVO₂).

Table 2.

Participant Demographic and Clinical Variables

	HFpEF Subgroup			HFrEF Subgroup
	HEART Camp n=25	Enhanced Usual Care n=34	p value	HEART Camp n=77	Enhanced Usual Care n=68	p value
Demographic variables
Age, mean (SD)	63.3 (9.4)	65.6 (9.3)	.367	58.6 (13.3)	58.6 (10.1)	.987
Gender, female, n (%)	11 (44%)	16 (47.1%)	.816	34 44.2%)	30 (44.1%)	.996
Married, n (%)	18 (72%)	20 (58.8%)	.296	3 4 (4 4.2%)	41 (60.3%)	.052
Non-Caucasian, n (%)	11 (44%)	16 (47.1%)	.816	4 0 (51.9%)	28 (41.2%)	.195
Clinical variables
Ejection Fraction, mean (SD)	54.2 (5.7)	55.7 (6.4)	.365	34.4 (9.3)	32.9 (10)	.344
NYHA Class, n (%)
I	3 (12%)	3 (8.8%)		2 (12%)	8 (8.8%)
II	14 (56%)	18 (52.9%)	.850	40 (56%)	41 (52.9%)	.046
III	8 (32%)	13 (38.2%)		34 (32%)	18 (38.2%)
IV	0	0		1 (1.3%)	1 (1.5%)
Beta blocker medication, n (%)	23 (92%)	32 (94.1%)	.749	76 (98.7%)	68 (100%)	.346
ACE/ARB medication, n (%)	20 (80%)	28 (82.4%)	.819	72 (93.5%)	55 (80.9%)	.021
BMI, mean (SD)	35.6 (7.2)	36.4 (8)	.708	34.7 (9.0)	33.9 (7.6)	.532
pVO2 from Cardiopulmonary Exercise Test	15.6 (4.1)	15 (4.5)	.662	14.6 (4.0)	15.7 (3.6)	.099

Abbreviations: NYHA, New York Heart Association Classification; ACE, Ace-Inhibitor; ARB, Angiotensin-receptor blocker; BMI, Body mass index.

In the HFpEF subgroup (n=59), 54% identified as men. Ages ranged from 37–85 (mean 64.6 years, SD=9.3) and ejection fraction ranged from 50 to 76 in this subgroup with a mean of 55 (SD=6). Of the 27 non-Caucasian participants (45.8% of the sample), 25 were African-American. In the HFrEF subgroup (n=145), 81 (56%) identified as men and 64 (44%) as women. HFrEF participants were a mean age of 58.6 (SD=11.9). Ejection fraction ranged from 12 to 49 in this subgroup with a mean of 33.7 (SD=9.6). Of the 68 non-Caucasian participants, 66 were African-American. Adherence data are shown in Figures 1a and 1b. Among patients with HFpEF in the Heart Camp group, adherence at both 12 (42%) and 18 months (56%) was significantly greater compared to the EUC group (14% and 0%, respectively). At the 6-month time point, adherence was 42% in the HEART Camp group and 19% in the EUC group; this difference was not statistically significant. Among patients with HFrEF in the Heart Camp group, adherence at 6 (36%), 12 (42%), and 18 months (29%) was not significantly greater compared to the EUC group (38%, 34% and 29%, respectively).

Figure 1a. Longitudinal adherence to exercise in adults with HFpEF from the HEART Camp clinical trial using all available data.

This figure demonstrates adherence for participants that completed data collection at each time point in the HFpEF subgroup. In the HEART Camp group, at 6 months, 9/21 (42%) were adherent to exercise, at 12 months, 9/21 (42%) were adherent, and at 18 months 9/16 (56%) were adherent to exercise. In the EUC group, at 6 months, 6/31 (19%) were adherent to exercise, at 12 months, 4/28 (14%) were adherent, and at 18 months 0/28 (0%) were adherent to exercise.

* statistically significant, α=.05

Figure 1b. Longitudinal adherence to exercise in adults with HFrEF from the HEART Camp clinical trial using all available data.

This figure demonstrates adherence for participants that completed data collection at each time point in the HFrEF subgroup. In the HEART Camp group, at 6 months, 26/73(36%) were adherent to exercise, at 12 months, 28/67 (42%) were adherent, and at 18 months 16/55 (29%) were adherent to exercise. In the EUC group, at 6 months, 24/64 (38%) were adherent to exercise, at 12 months, 21/62 (34%) were adherent, and at 18 months 17/60 (29%) were adherent to exercise.

Note: Adherence based on 120 minutes of moderate intensity or above exercise per week EUC = Enhanced usual care – received paid access to exercise facility

HFpEF participants in the HEART Camp group walked further during the 6MWT than EUC participants at all times points, including baseline, with the difference reaching significance at 18 months (see Table 3). Distance in the HEART Camp group increased by 63.25 meters compared to 13.16 meters in the enhanced usual care group between baseline and 18 months (F(2.7,73.1)=4.14,p=0.011; η² = 0.13, large effect). In a hierarchical linear model, the time*group interaction for 6MWT distance was significant as shown in Table 4, indicating that the groups significantly differed in change over time.

Table 3.

Six-Minute Walk Test Distance Over Time for Participants with HFpEF

Group	Baseline	6 months	12 months	18 months
HEART Camp	383.61	411.21	406.52	446.86
mean (SD)	(117.63)	(120.89)	(140.94)	(112.71)
Enhanced Usual Care	359.45	370.98	368.86	372.61
mean (SD)	(101.46)	(109.94)	(118.11)	(95.55)
t-test p-value	.402	.241	.346	.048

Table 4.

Hierarchical Linear Models of Physical Function, Heart failure-related Health Status, and Patient-Reported Symptoms in Participants with HFpEF

		Num DF	Den DF	F	p-value
Six Minute Walk Testa	Time	3	89	3.584	0.017
	Group	1	58	2.928	0.092
	Time*Group	3	89	5.718	0.001
KCCQ Overall Summary Score	Time	3	130	0.595	0.619
	Group	1	57	0.062	0.805
	Time*Group	3	130	3.981	0.009
KCCQ Clinical Summary Score	Time	3	131	0.742	0.529
	Group	1	57	0.014	0.906
	Time*Group	3	131	5.698	0.001
KCCQ Total Symptom Score	Time	3	131	0.647	0.586
	Group	1	57	0.006	0.939
	Time*Group	3	131	4.179	0.007
KCCQ Quality of Life	Time	3	131	1.123	0.342
	Group	1	57	0.354	0.554
	Time*Group	3	131	2.518	0.061
Dyspnea-Fatigue Inventory	Time	3	131	2.063	0.108
	Group	1	57	0.419	0.520
	Time*Group	3	131	0.602	0.615
PROMIS Anxiety	Time	3	131	0.639	0.591
	Group	1	55	2.044	0.158
	Time*Group	3	131	1.019	0.386
PROMIS Depression	Time	3	130	1.371	0.255
	Group	1	57	1.266	0.265
	Time*Grour	3	130	1.347	0.262
PROMIS Social Satisfaction with Roles	Time	3	131	1.304	0.276
	Group	1	57	0.209	0.650
	Time*Group	3	131	1.877	0.137
PROMIS Fatigue	Time	3	132	0.891	0.448
	Group	1	58	0.276	0.602
	Time*Group	3	132	2.267	0.084
PROMIS Pain Interferencea	Time	3	91	1.623	0.190
	Group	1	58	0.270	0.605
	Time*Group	3	91	0.264	0.851
PROMIS Physical Function	Time	3	130	2.133	0.099
	Group	1	57	0.538	0.466
	Time*Group	3	130	0.241	0.868
PROMIS Sleep Disturbancea	Time	3	97	1.433	0.238
	Group	1	59	0.296	0.588
	Time*Group	3	97	1.853	0.143

^a.Indicates an Autoregressive covariance structure, otherwise Variance Components structure.

We identified significant differences between adults with HFpEF in HEART Camp and enhanced usual care on the KCCQ Overall Summary Score (F(3,96)=3.42,p=0.02; η²=0.09, medium effect), the Clinical Summary Score (F(3,95)=6.17,p=0.001; η²=0.16, large effect), and Total Symptom Score (F(2.7,85.3)=5.31,p=0.003; η²=0.14, large effect). Between group comparisons of overall change (time*group interactions) for the Dyspnea-Fatigue Inventory and domains of the PROMIS-29 were non-significant in hierarchical linear models. Results from all hierarchical linear models of data for the 6-minute walk test, KCCQ, Dyspnea-Fatigue Inventory and PROMIS-29 are reported in Table 4. We report means by group over time in Figures 2a, b, & c for those scores with significant time*group interactions.

Figure 2a. Kansas City Cardiomyopathy Questionnaire (KCCQ) Overall Summary Score.

This figure shows mean KCCQ overall summary scores in the HFpEF subgroup from baseline to 18 months. ANOVA indicated a statistically significant change by group over time. (F(3,96)=3.4,p=.02)

Figure 2b. Kansas City Cardiomyopathy Questionnaire (KCCQ) Clinical Summary Score.

This figure shows mean KCCQ clinical summary scores in the HFpEF subgroup from baseline to 18 months. ANOVA indicated a statistically significant change by group over time. (F(3,95)=6.2,p=.01)

Figure 2c. Kansas City Cardiomyopathy Questionnaire (KCCQ) Clinical Summary Score.

This figure shows mean KCCQ clinical summary scores in the HFpEF subgroup from baseline to 18 months. ANOVA indicated a statistically significant change by group over time. (F(3,95)=6.2,p=.01)

In participants with HFpEF, the rate of attrition at the end of the 18-month study was 25.4%, with 44 participants completing heart rate validation and exercise diaries at 18 months. We had slightly greater attrition (32.2%) with study instruments as 40/59 completed all data points. The differences in attrition rates between groups was non-significant at all time points. Interestingly, when non-completers were examined against completers on our measurement tools the only difference we found in baseline measures was on the PROMIS-Sleep Disturbance. Completers had significantly lower scores (47.5) on the domain than non-completers (54.8) (t(57)=2.65, p=.01 indicating non-completers had greater sleep disturbance. A sensitivity analysis completed with all data from the parent HEART Camp study (previously reported) found no differences in findings as a result of excluding participants that did not complete the study.²² In the current study, we completed a sensitivity analysis in the HFpEF subgroup, again with no changes in outcomes. This indicates that attrition did not bias outcomes, or dilute significant findings.

All primary analyses were performed with the HFrEF subgroup. As mentioned above, adherence did not differ significantly at any time point. Of the 13 outcome variables, only the PROMIS-Anxiety differed significantly over time (significant time*group interaction). PROMIS Anxiety improved for the HEART Camp group, whereas anxiety increased in the EUC group. Otherwise, no differences were observed in the HFrEF subgroup.

To examine the effect of covariates in the models testing intervention effects in both subgroups, we added age, gender, body mass index, and race (white vs. non-white) as covariates. Although these were significant covariates in some cases, there was no change in significance of the time by group interaction in any of the models. Since the effectiveness of the intervention was not altered by the addition of the covariates, we reported the unadjusted models in the interest of model parsimony.

Discussion

The purpose of this study is to evaluate the effects of the HEART Camp intervention on adherence to exercise, physical function, and patient-reported outcomes in the subgroups of adults with HFpEF and HFrEF as compared to enhanced usual care. In the HFpEF subgroup, adherence was significantly greater at 12 (42.9% vs. 14.3%), and 18 (56% vs 0%) months in the HEART Camp intervention group compared to enhanced usual care. Physical function also improved significantly (based on 6-minute walk test distances) over time in the HEART Camp group compared to enhanced usual care counterparts. Patient-reported outcomes, including heart failure-related health status (KCCQ Overall Summary Score), symptoms (KCCQ, total symptom score), and clinical health status (KCCQ Clinical Summary Score) were also statistically and clinically significant when comparing HEART Camp to enhanced usual care. These findings are critically important given the limited successful treatment approaches in HFpEF. Our HFrEF subgroup did not achieve significant differences in adherence, physical function, or patient-reported outcomes aside from improvements in anxiety in the HEART Camp intervention group compared to EUC over time. This is the first manuscript to report long-term adherence outcomes out to 18 months and the first to report sustained improvement in health status from 12 to 18 months in an exercise trial in adults with HFpEF.

Difficulties with initiation and continued adherence to exercise have plagued adults with HF as well as HF clinicians. HF-ACTION, the largest study of exercise in HF to date, included only subjects with reduced ejection fraction and reported adherence to the target of 120 minutes per week at 12 months to be 35%.^16,37 In our parent study of HEART Camp we enrolled adults with stable chronic HF and adherence was similar at 42% in the intervention group at 12 months compared to 19% in the enhanced usual care group.¹⁷

Interventions, such as financial incentives, have been used to successfully encourage short-term (less than 6 months) participation in exercise or standard cardiac rehabilitation programs.³⁸ However, studies of long-term exercise programs or exercise adherence are limited with little data to inform effective intervention strategies. To date, Optimex-Clin is the only completed longer term (12 month) trial focused on adults with HFpEF. Investigators in the Optimex-Clin trial defined adherence as 70% or greater attendance at sessions. Using this definition, the trial’s moderate continuous training (comparable to HEART Camp moderate intensity aerobic training), achieved 60.4% adherence when considering the full 12-month study, which is similar to the adherence we achieved in our HFpEF subgroup of 56% at 18 months. When Optimex-Clin investigators considered the home-based portion of the intervention (months 4–12) alone, adherence was 58.5%.¹⁰ In the recent REHAB-HF trial, adherence was high; however, adherence was measured by session attendance during the intervention period and by patient self-report in the follow-up period.¹⁴ Each of these surrogates for adherence are difficult to validate. Important to note is that the definition of adherence in HEART Camp was based on minutes of validated, moderate intensity exercise (defined as achieving 40–80% of heart rate reserve). These differences in adherence definitions and measurements contribute to the difficulty of comparing adherence across studies. Establishing a standard definition of adherence is critical to enhancing inter-trial comparisons in the future and should be made a priority by clinicians and exercise scientists, particularly in HFpEF where a large-scale clinical trial has yet to be undertaken. Without standardization, adherence will remain the perpetual “Achille’s heel” of exercise programs in HF.²

Adults with HFpEF in our HEART Camp group responded more favorably than EUC peers in terms of physical function improvements as measured by 6MWT distances at 18 months. The findings were not only statistically significant (p= .048) but also associated with a positive moderate change in clinical status. These results are consistent with short-term studies of exercise interventions with HFpEF.^5,6 Maldonado-Martin et al. reported significant improvements in 6MWT distances in an exercise training group vs. attention-control in a 16 week exercise intervention trial. This intervention also improved ventilatory thresholds and peak oxygen consumption, but these were not significantly correlated to improvements in the 6MWT.⁵ Caloric restriction in combination with exercise have been reported to synergistically work to improve 6MWT distance in HFpEF.⁶ Physical function and health status continued to improve over time for adults with HFpEF in the HEART Camp intervention group. These improvements may have contributed to the observed increase in adherence seen between 12 and 18 months.

Interestingly, we did not find significant differences in adherence or most patient-reported outcomes when comparing HEART Camp to enhanced usual care in our HFrEF subgroup. This suggests the findings reported in our primary outcome paper were highly influenced by adherence in the HFpEF subgroup.²² It is unclear whether these adherence findings were driven by differing pathophysiological and/or symptom profiles in HFpEF and HFrEF or disparate responses to exercise that led to improved adherence among those with HFpEF. Our findings are consistent with a subgroup analysis of REHAB-HF, in which older adults with HFpEF responded more favorably to the intervention than HFrEF counterparts.³⁹ In their analyses, adults with HFpEF responded favorably with large effect sizes on the short physical performance battery, the KCCQ, and six-minute walk test distance. Adherence was similar across HFpEF and HFrEF subgroups during the intervention and at follow-up. However, as noted above, adherence was defined based on session attendance and patient self-report, which may not correspond to objectively measured minutes of moderate intensity exercise. Taken together, these findings support the potential usefulness of the HEART Camp intervention to promote adherence to exercise in adults with HFpEF and ultimately, improve patient-reported outcomes in this population.

Several prior exercise interventions trials in HFpEF have tested intervention effects on quality of life and health status outcomes.^6,8,13,40 These trials have reported significant improvements in the physical domains of both the Short-form-36 (SF-36) and the Minnesota Living with Heart Failure Questionnaire (MLHFQ) following supervised aerobic and/or resistance training interventions.^40,41 In this subgroup analysis, participants in the intervention group reported significantly greater change over time in their KCCQ overall summary, clinical summary, and total symptom scores compare to EUC. To date, one shorter term (20 weeks) study⁶ and one longer term (12 months) study, Optimex-Clin¹⁰ have used the KCCQ to test exercise effects on health status in HFpEF. Interestingly, Kitzman and colleagues reported that in a trial of a caloric restriction intervention compared to and combined with exercise, caloric restriction alone improved KCCQ scores, but exercise did not.⁶ These findings may have resulted from the short-term intervention of only 20 weeks. Comparatively, our findings and the findings of Optimex-Clin over 12 months showed that although initially scores may be similar across intervention and control groups, over the long-term health status (as measured with the KCCQ) do seem to improve with exercise interventions.¹⁰

We tested interventional effects on symptoms using the PROMIS-29. We did not record significant changes between the groups in our HFpEF subgroup. We are the first reported study to examine change in specific symptoms in a trial of exercise in adults with HFpEF aside from depression. Prior studies have demonstrated that improvements in depression may be achieved with exercise interventions in adults with HFpEF.^13,40 Although we did not find significant changes here, we did note trends in improvement. Future fully powered studies are needed to see if these trends would reach significance. In our HFrEF subgroup, the HEART Camp intervention group reported significantly greater improvements in anxiety compared to EUC counterparts.

We would like to acknowledge the limitations of this study. This analysis is a retrospective, subgroup analysis from a parent study. Therefore, findings should not be generalized beyond this sample. We recruited from 2 sites in the U.S., Detroit, MI and Lincoln, NE. Despite the small number of sites, our sample was representative relative to race and biological sex across groups. Further, we employed a consistent interventional protocol at both sites and closely monitored intervention fidelity as previously published.⁴² We noted no differences in the intervention effects between sites believe our intervention has high potential for translation. We acknowledge that overall, our sample was younger and included more men than typically reported in the general HFpEF population. In the parent study, randomization occurred after a run-in period which further limits the generalizability of our findings. The run-in period allowed for monitored exercise sessions and equipment orientation prior to independent exercise to reduce the risk of participant injury or adverse events. Future studies may incorporate a stratified sampling approach to further diversify the sample.

Conclusion

The findings of these sub-group analyses are highly encouraging, and additional research on alternative, cost-conscious delivery methods and a larger clinical trial of our multi-component intervention to test long-term adherence in adults with HFpEF is warranted. Exercise remains one of few treatment strategies to successfully promote positive outcomes in adults with HFpEF. To further drive these improvements, interventions to support long-term adherence are critical.

Bullet Points.

• Adults with heart failure with preserved ejection fraction (HFpEF) (EF≥50%) responded favorably to the Heart Failure Exercise and Resistance Training (HEART) Camp.

• The HEART Camp intervention promotes adherence to exercise in adults with HFpEF compared to paid access to a fitness center alone.

• The HEART Camp intervention improved physical function, health status, and symptoms in adults with HFpEF over time compared to paid access to a fitness center alone.