A Randomized, Controlled Trial of Resistance Training Added to Caloric Restriction Plus Aerobic Exercise Training in Obese Heart Failure with Preserved Ejection Fraction

Abstract

Background:

We’ve shown that combined caloric restriction (CR) and aerobic exercise training (AT) improve peak exercise O₂ consumption (VO₂peak); and quality-of-life (QOL) in older patients with obese heart failure with preserved ejection fraction (HFpEF). However, ~35% of weight lost during CR+AT was skeletal muscle (SM) mass. We examined whether addition of resistance training (RT) to CR+AT would reduce SM loss and further improve outcomes.

Methods:

Randomized, controlled, single-blind, 20-week trial of RT+CR+AT vs. CR+AT in 88 patients with chronic HFpEF and body mass index (BMI) ≥28 kg/m². Outcomes at 20 weeks included: the primary outcome (VO₂peak); MRI and dual X-ray absorptiometry; leg muscle strength and quality (leg strength ÷ leg SM area); Kansas City Cardiomyopathy Questionnaire (KCCQ).

Results:

77 participants completed the trial. RT+CR+AT and CR+AT produced non-significant differences in weight loss: mean (95% CI): −8(−9,−7) versus −9(−11,−8) (p=0.21). RT+CR+AT and CR+AT had non-significantly differences in the reduction of body fat [−6.5(−7.2,−5.8) versus −7.4(−8.1,−6.7) kg] and SM [−2.1(−2.7,−1.5) versus −2.1(−2.7,−1.4) kg] (p=0.20 and 0.23, respectively). RT+CR+AT produced significantly greater increases in leg muscle strength [4.9(0.7,9.0) versus −1.1(−5.5,3.2) Nm, p=0.05] and leg muscle quality [0.07(0.03,0.11) versus 0.02(−0.02,0.06) Nm/cm2, p=0.04]. Both RT+CR+AT and CR+AT produced significant improvements in VO₂peak [108(958,157) versus 80(30,130) ml/min; p=0.001 and 0.002, respectively], and KCCQ score [17(12,22) versus 23(17,28); p=0.001 for both], with no significant between-group differences. Both RT+CR+AT and CR+AT significantly reduced LV mass and arterial stiffness. There were no study-related serious adverse events.

Conclusion:

In older obese HFpEF patients, CR+AT produces large improvements in VO₂peak and QOL. Adding RT to CR+AT increased leg strength and muscle quality without attenuating SM loss or further increasing VO₂peak or QOL.

Registration:

ClincalTrials.gov Identifier: NCT02636439

Introduction

Heart failure with preserved ejection fraction (HFpEF) is the most common form of HF, particularly among older persons and women, and is increasing in prevalence^1–4 Even when stable and well-compensated, patients with chronic HFpEF have severe exercise intolerance, with exertional dyspnea and fatigue, and this is associated with severely reduced quality-of-life (QOL).^5–8 However, other than diet and exercise, there are relatively few proven treatments for exercise intolerance in HFpEF.^9–12

Obesity is one of the strongest risk factors for HFpEF^2,5,13–15 and ~85% of HFpEF patients are overweight/obese, such that ‘obese/metabolic HFpEF’ is the most common HFpEF phenotype.^5,11,16 compared to non-obese HFpEF patients, those with the obese HFpEF phenotype have worse symptoms, exercise capacity, hemodynamic abnormalities, and QOL.¹⁷ We have shown that patients with obese HFpEF have reduced skeletal muscle mass and increased thigh muscle fat infiltration and intra-abdominal fat mass, all of which are related to their reduced exercise capacity.^5,11,16–21

In a prior trial,¹¹ we showed that in patients with obese HFpEF, both caloric restriction (CR) and aerobic exercise training (AT) significantly increased VO₂peak, and their effects were additive. These data supported combined CR+AT as a novel treatment to improve exercise intolerance and QOL in older patients with obese HFpEF. However, ~35% of the weight lost (=2.5 kg) during CR was skeletal muscle mass (SM), and this was not attenuated by the addition of AT.¹¹ The loss of SM is of concern as it may attenuate gains in exercise capacity and has been associated with increased risk of frailty, physical disability, injuries, hospitalizations, and death.^22–26 Thus, strategies to further increase VO₂peak during CR+AT by preserving skeletal muscle mass, strength, and muscle quality in HFpEF warrant further investigation. Multiple lines of evidence indicate that resistance training (RT) could be an ideal addition to CR+AT for HFpEF.^{16–18, 20–}25 In HF with reduced EF (HFrEF), RT improves VO₂peak to a similar degree as AT.^26,27 In older adults and patients with HFrEF, RT increases muscle mass, strength, and muscle quality significantly more than AT.^28–29 However, despite these potential advantages, the addition of RT to AT, CR, or their combination has not been formally evaluated in patients with HFpEF. Furthermore, there have been concerns that RT may have potential deleterious consequences in HFpEF by further increasing blood pressure, LV hypertrophy and arterial stiffening.³⁰

Thus, the purpose of this controlled, randomized, single-blind trial was to determine whether adding RT to CR+AT in obese older HFpEF patients further improves VO₂peak (primary outcome), and leg muscle mass, strength and quality, QOL, and LV structure/function and arterial function (secondary outcomes).

Methods

The trial was conducted at Wake Forest University School of Medicine from 2015–2021, institutional review board approved, and registered (NCT02636439). Participants provided written informed consent. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Study Participants

Potential participants were interviewed and examined by a board-certified cardiologist who verified the diagnosis of HF and participant qualification. As previously described and in accord with the ACC/AHA definition of HFpEF,^2,8,10,11 key inclusion criteria were: symptoms and signs of HFLV EF≥50%; age≥60 years; body mass index (BMI)≥28 kg/m²; and either NHANESHF score ≥3³¹ or the criteria of Rich et al,³² or both. Major exclusion criteria were: prior history of reduced LV EF (<45%); significant ischemic or valvular heart disease; or any other disorder that could explain the participants’ symptoms. Participants were clinically stable and were not undergoing regular exercise or diet programs prior to enrollment.

Outcomes

Primary Outcomes

Outcomes were assessed by personnel who were blinded to participant group assignment at baseline and after the 20-week intervention. VO₂peak (in ml/min) was determined by expired gas analysis (Ultima, Medical Graphics) and obtained during the last 30 seconds of exercise on a treadmill (Modified Naughton Protocol) to exhaustion.¹¹

Secondary Outcomes

Total body fat and lean mass were measured by using dual x-ray absorptiometry (DXA, Hologic Inc.).¹¹ Scans of the thigh, abdomen, and heart were performed by magnetic resonance imaging (MRI) as previously described.¹¹ Cross-sectional areas of skeletal muscle (SM), subcutaneous fat (SCF), intramuscular fat (IMF), and bone were measured at the mid-thigh of the left leg. Total thigh area was calculated as the sum of SCF, IMF, SM and bone, and thigh compartment (TC) area was calculated as the sum of SM, IMF, and bone. A slice at the second lumbar vertebra was used to determine abdominal fat measurements, including SCF, visceral and intraperitoneal and retroperitoneal fat. Epicardial and paracardial fat volumes, were assessed by MRI as previously described.¹¹

Knee extensor strength (Nm) was assessed by isokinetic dynamometry (Biodex Medical Systems, Inc., Shirley, NY) at 60° per second with participant seated and hips and knees flexed at 90° as previously described.^11,33 Participants performed four repetitions; two trials on each leg. Peak torque (Nm) for the dominant leg were used in analyses. SM quality was calculated as the ratio of knee extensor strength to thigh muscle area assessed by MRI (Nm/cm²).^11,34 Six-Minute Walk Distance (6MWD) was measured according to the American Thoracic Society and the Short Physical Performance Battery (SPPB) was conducted as previously described.^11,35

Heart failure–specific QOL was assessed with the Kansas City Cardiomyopathy Questionnaire (KCCQ and general QOL was assessed with EuroQOL as well as the 36-item Short-Form Health Survey (SF-36) physical component score.³⁶ Doppler echocardiograms were performed and analyzed per American Society of Echocardiography recommendations. Carotid-femoral pulse-wave velocity were assessed as previously described.¹¹ LV mass and volumes were assessed by cardiac MRI as previously described.¹¹

Randomization

After baseline testing, participants were randomly assigned to either RT+CR+AT or CR+AT, stratified by sex.

Interventions

Caloric Restriction

Participants in both groups were prescribed a hypocaloric diet using meals (lunch, dinner, and snacks) prepared by the Wake Forest Clinical Research Unit Metabolic Kitchen under the direction of a registered dietitian, as previously described and as detailed in the Online Supplement.^11,37,38

Aerobic and Resistance Exercise Training

The exercise prescription was based on ACSM and AHA guidelines for older persons and for HFrEF.^39,40 and were individualized and progressive. Both groups exercised three times per week for 20 weeks. Warm-up and cool-down phases were ~5–10 min of stretching, flexibility, and light walking. The total stimulus phase of the RT+CR+AT was ~60 minutes (40 min for AT and 20 min for RT). To maintain a similar duration between groups, for the CR+AT group, ~20 min of light chair-based range-of-motion, stretching, and flexibility exercises was included at the end of the 40-minute aerobic exercise sessions; this 20 minute added ‘control period’ provided no resistance or aerobic exercise. Sessions were conducted with medical supervision to ensure safety and a uniform, robust stimulus. To eliminate potential for cross-group contamination, treatment groups exercised at different times.

The AT component was identical for both groups and included walking on an indoor track and/or stationary cycling and was identical to that previously described¹¹ and as detailed in the Online Supplement.

The RT component (for the RT+CR+AT group only) was performed immediately after the AT session with heart rate, blood pressure, and cardiac rhythm monitored before and afterwards.^39–40 The RT program was designed to allow a gradual progression so that participants could become familiar with the equipment, minimize muscle soreness, and reduce potential for injury. In accordance with published recommendations,^39,40 and to optimize the RT stimulus for maximal functional gains, the exercise prescription was based on a relative intensity level and progressed at a rate specific to the individual’s strength gains. The RT consisted of two upper body (chest press and compound row) and four lower body (leg extension and flexion, seated leg press, and calf plantar flexion) on Nautilus RT equipment. For the initial three weeks, one set of 8–12 repetitions was performed for each RT exercise which then progressed to two sets of each after four weeks. Using a 2:1 recovery- exercise ratio (1 min recovery after 30 sec RT), the RT (12 total sets of RT exercises) required a total of ~20 minutes. The initial resistance setting on each piece of equipment began at 20–30% of participants’ one repetition maximum (1-RM) but increased to 40–50% of 1-RM after four weeks. During the study, if >12 repetitions were completed on the 2nd set in 2 consecutive sessions, the resistance for that exercise was increased by ~5% for the next session. The 1-RM testing was repeated every 4 weeks to assess gains in strength and to ensure optimal resistance settings.

Statistical Analysis

This 2-arm, randomized, parallel design trial was designed to test the effect of adding RT to a previously tested regimen of CR+AT, using an intention-to-treat analysis, in older patients with obese HFpEF. Analysis on the main outcome of absolute VO₂peak (ml/min) collected at week 20 was conducted using analysis of covariance (ANCOVA) with baseline measure, age, and gender as predefined covariates. Similar analyses were performed for all secondary/exploratory outcomes to evaluate treatment effect. A two-sided p-value of 0.05 was used to determine significance. To assess within-group changes, we used constrained linear mixed-effects models (cLMM) adjusted for age and sex. Baseline and 20-week measures were analyzed together as a dependent variable. The models included treatment, visit, and treatment-by-visit interaction under the constraint that the baseline measures are equal across treatment arms to reflect the design of the randomized trial. Within-group differences were estimated using linear contrasts.

Sample size calculations were based on data from our previous trial of CR+AT.¹¹ The square root of the mean square error from an ANCOVA model was estimated to be 107.5 ml/min. In order to have 90% power to detect a relative effect due to RT of 5.6% (absolute difference of 82.0 ml/min) in VO₂peak at the 0.05 two-sided level of significance, the study required 38 evaluable participants per group. To allow for 15% loss to follow-up, 44 subjects per group (88 total) were randomized. The sample size also had power to test the secondary outcomes.

Results

As shown in Figure 1, 424 participants were screened by telephone and 132 were scheduled for a screening visit. Ultimately, 88 participants (mean±SD: age, 68±5 years; BMI, 40±6 kg/m²) were enrolled and randomized: RT+AT+RT (n=44) and CR+AT (n=44) (Figure 1). Of these, 77 participants (RT+CR+AT n=39; CR+AT n=38) completed the intervention and had evaluable testing for the primary outcome (Figure 1). Two participants (both RT+CR+AT) were excluded from analysis of the primary outcome: one because exercise effort judged by the blinded tester was clearly submaximal; and one due to malfunctioning expired gas collection. Exercise attendance was 82±8% and diet adherence was 99±1%. There were no study-related serious adverse events. Participant characteristics were in accord with those in population based HFpEF studies (Table1).

Figure 1:

Flow of Participants Through the Study

GFR = Glomerular Filtration Rate

CPET = Cardiopulmonary Exercise Testing

Table 1.

Baseline Characteristics of Treatment Groups at Randomization

Characteristic	CR + AT (n=44)	RT + CR + AT (n=44)	p-value
Age, years	67.9±5.4	69.7±5.8	0.13
Women	37 (84%)	38 (86%)	0.76
White	19 (46%)	21 (48%)	0.67
Weight, kg	106.1±19.6	104.3±17.9	0.57
Body surface area, m2	2.08±0.22	2.07±0.20	0.80
Body mass index, kg/m2	40.0±5.9	39.2±5.6	0.71
Body fat, % (by DXA)	50±4	49±4	0.33
Cardiovascular Measures
NYHA Class
II	15 (34%)	16 (36%)	0.82
III	29 (66%)	28 (64%)
Ejection fraction, %	61±5	61±6	0.63
Relative wall thickness	0.51±0.09	0.52±0.08	0.69
Diastolic filling pattern*
Normal	1 (2%)	3 (7%)	0.47
Impaired relaxation	38 (86%)	37 (84%)
Pseudonormal	5 (11%)	3 (7%)
Restrictive	0 (0%)	1 (2%)
e’, cm/s	6.4±1.6	6.5±2.1	0.77
E/e’, ratio	13.3±6.4	13.1±4.5	0.89
N-terminal proBNP, pg/mL, median (IQR)†	55 (42,89)	100 (55,154)	0.048
Systolic blood pressure, mmHg	137±14	135±15	0.59
Diastolic blood pressure, mmHg	79±9	77±10	0.35
Comorbidities
History of Atrial fibrillation	0 (0%)	4 (9%)	0.041
Diabetes mellitus	14 (32%)	20 (45%)	0.19
Hypertension	42 (95%)	42 (95%)	1.00
Current Medications
ACE Inhibitor	8 (18%)	12 (27%)	0.31
Angiotensin receptor blocker	20 (45%)	15 (35%)	0.31
Diuretic	36 (82%)	36 (84%)	0.81
B-blocker	18 (41%)	17 (40%)	0.90
Calcium antagonist	16 (36%)	19 (44%)	0.46
Aldosterone antagonist	2 (5%)	5 (12%)	0.22
Nitrate	2 (5%)	4 (9%)	0.38
Insulin	4 (9%)	10 (23%)	0.07
Oral diabetic agents	11 (25%)	19 (44%)	0.06
Metformin	7 (16%)	13 (30%)	0.13
Sulfonylureas	5 (11%)	6 (14%)	0.75
GLP-1 receptor agonist	0 (0%)	1 (2%)	0.31
D-PP4 inhibitor	1 (2%)	2 (5%)	0.56
Antidepressant	15 (35%)	11 (26%)	0.35

Data presented as mean±SD or N (%), unless otherwise noted. Abbreviations: ACE: angiotensin-converting enzyme; E, E-wave velocity; e’: early mitral annulus velocity (septal); IQR: interquartile range; NYHA, New York Heart Association; RER: respiratory exchange ratio.

^*Diastolic filling pattern determined according to American Society of Echocardiography criteria.

^†Sample size for NT-proBNP collection=27 (N=13 for CR+AT and N=14 for RT+CR+AT)

Primary Outcome

VO₂peak increased significantly in both the RT+CR+AT and CR+AT groups (108 vs. 80 mL/min, respectively, which equates to 7% and 5% improvement from baseline) with no significant between-group difference (44 (−25,112) mL/min; p=0.21). Effect sizes for the primary (and key secondary outcomes) are presented in Figure 2 (Central Figure).

Figure 2 Central Figure”:

Effect Sizes for the Primary (peak VO₂) and Key Secondary Outcomes (strength and body/muscle composition) from Baseline to 20 Week Follow up.

VO_2peak m/min = peak exercise oxygen consumption

Secondary Outcomes

Significant within-group improvements, without between-group differences, were also observed for other expressions of exercise capacity or physical function (VO₂peak mL/kg/min, VO₂peak mL/kg lean body mass/min, VO₂peak mL/kg leg lean mass/min, VO₂peak mL/cm² thigh muscle/min, exercise time, peak workload, peak METs, and 6-minute walk distance and SPPB), as well as multiple HF-specific and general QOL measures, NYHA class, and depression scores. (Table 2).

Table 2.

Exercise Performance and Quality of Life Outcomes by Treatment Group

	CR + AT (n=38)			RT + CR + AT (n=39)			Treatment Effect Size	p-value
Exercise Performance	Baseline	Follow-up	Within group Change (95% CI)	Baseline	Follow-up	Within group Change (95% CI)	Between Group Difference (95%CI)
VO2peak mL/min	1529±375	1609 (1561,1657)	80 (30,130)*	1530±366	1653 (1605,1700)	108 (58,157)**	44 (−25,112)	0.21
VO2peak mL/kg/min	14.6±2.9	16.8 (16.3,17.4)	2.2 (1.7,2.7)**	14.8±2.9	17.1 (16.6,17.6)	2.4 (1.9,2.9)**	0.3 (−0.5,1.0)	0.48
VO2peak mL/kg lean body mass/min	30.0±5.5	32.5 (31.5,33.5)	2.9 (1.9,3.8)**	29.7±5.1	33.3 (32.4,34.3)	3.4 (2.5,4.4)**	0.7 (−0.7, 2.1)	0.30
VO2peak mL/kg leg lean mass/min	94.1±17.8	98.4 (96.7,100.1)	4.9 (3.2,6.6)**	92.6±18.0	98.8 (97.1,100.4)	5.2 (3.5,6.9)**	0.4 (−2.0, 2.8)	0.74
VO2peak mL/cm2 thigh muscle/min	14.1±2.9	15.0 (14.5,15.5)	1.1 (0.6,1.6)**	13.6±2.3	15.3 (14.8,15.8)	1.3 (0.8,1.8)**	0.3 (−0.4,1.0)	0.40
Exercise time, min	9.2±2.4	11.3 (10.8,11.8)	2.2 (1.7,2.6)**	9.2±2.4	11.4 (10.9,11.9)	2.3 (1.8, 2.6)**	0.1 (−0.6, 0.8)	0.83
METs	5.5±1.2	6.5 (6.2,6.7)	1.1 (0.8,1.3)**	5.4±1.2	6.6 (6.4,6.8)	1.2 (1.0,1.4)**	0.1 (−0.2, 0.5)	0.49
Peak heart rate, beats/min	142±21	136 (133,140)	−1 (−4,3)	130±22	139 (136,142)	3 (−1,7)	2 (−2,7)	0.32
Peak systolic blood pressure, mmHg	181±18	179 (173,186)	−6 (−12,1)	187±21	177 (171,183)	−8 (−14,−1)*	−2 (−11,7)	0.63
Peak diastolic blood pressure, mmHg	81±10	72 (69,75)	−7 (−10, −4)**	77±30	73 (71,76)	−6 (−9,−3)**	1 (−3,5)	0.54
Peak RER	1.08±0.10	1.11 (1.08,1.14)	0.01 (−0.02,0.05)	1.10±0.16	1.11 (1.08,1.14)	0.02 (−0.02,0.06)	0.00 (−0.04,0.04)	0.94
VE/VCO2 slope	28.3±4.4	28.2 (27.4,29.0)	−0.9 (−1.7, −0.1)*	29.6±3.8	28.5 (27.8,29.4)	−0.4 (−1.2,0.4)	0.3 (−0.8,1.4)	0.62
6-minute walk distance, m	364±67	419 (405,434)**	61 (47,76)	353±74	405 (391,419)**	47 (33,60)	−14 (−34,6)	0.16
SPPB Score	9.8±1.3	10.2 (9.8,10.7)**	0.8 (0.3,1.2)	9.1±2.0	10.2 (9.8,10.6)**	0.8 (0.4,1.3)**	0.0 (−0.6,0.6)	0.99
Leg strength, Nm	105.8±29.8	97.5 (92.8,102.1)	−1.1 (−5.5,3.2)	89.5±27.0	103.9 (99.5,108.3)*	4.9 (0.7,9.0)*	6.4 (−0.1,12.9)	0.053
Leg muscle quality Nm/cm2	0.96±0.18	0.90 (0.86,0.95)	0.02 (−0.02,0.06)	0.80±0.23	0.97 (0.93,1.01)*	0.07 (0.03,0.11)*	0.07 (0.00,0.13)	0.043
Quality of Life
KCCQ Score
Overall	68±19	91 (88,95)	23 (17,28)**	70±19	87 (83,90)**	17 (12,22)**	−5 (−10,0)	0.07
Clinical	71±17	91 (87,95)	20 (15,25)**	71±19	86 (83,90)**	15 (10,20)**	−5 (−10,1)	0.09
Physical	72±19	91 (88,96)	20 (14,26)**	72±20	87 (83,91)**	15 (9,20)**	−5 (−11,1)	0.08
Social	70±26	95 (90,100)	25 (17,33)**	72±25	89 (84,93)**	18 (11,26)**	−6 (−13,0)	0.07
QOL	60±25	88 (84,93)	26 (19,33)**	65±25	84 (79,88)**	20 (13, 27)**	−5 (−11,2)	0.16
Symptoms	70±19	90 (86,95)	20 (14,26)**	70±21	86 (82,90)**	16 (10,21)**	−4 (−10,2)	0.16
SF-36
Physical composite score	37±9	47 (45,50)	10 (7,13)**	38±9	47 (44,50)**	9 (7,12)**	−1 (−4,3)	0.79
Mental composite score	53±9	57 (55,60)	4 (1,7)*	53±12	57 (55,59)*	3 (0,6)*	0 (−4,3)	0.95
EuroQol Thermometer	72±15	84 (80,88)	14 (9,18)**	70±17	82 (78,85)**	11 (7,16)**	−2 (−7,3)	0.42
CES-D Score	10±6	6 (5,8)	−3 (−5, −1)*	9±7	7 (5,9)*	−2 (−4,−0)*	1 (−2,3)	0.66
NYHA Class	2.7±0.5	1.7 (1.6,1.9)	−1.0 (−1.1, −0.8)**	2.6±0.5	1.8 (1.6,1.9)**	−0.9 (−1.0,−0.7)**	0.1 (−0.1,0.3)	0.53

Baseline data presented as mean±SD. Follow-up data presented as LSMeans (95%CI) adjusted for baseline value, age, and sex from ANCOVA. Within-group change estimates derived from cLMM. Abbreviations: CES-D: Centers for Epidemiologic Studies Depression Scale; KCCQ: Kansas City Cardiomyopathy Questionnaire; MET: metabolic equivalent; Nm: Newton-meter; NYHA: New York Heart Association; SF-36: Short Form 36-item Health Survey; SPPB: Short Physical Performance Battery; RER: respiratory exchange ratio; VAT: ventilatory anaerobic threshold; VE/VCO₂: ventilatory equivalent for carbon dioxide; VO_2, volume of oxygen consumption.

^*significant at p<0.05

^**significant at p<0.001

The total weight loss (Table 3) of −8 versus −9 kg (8% vs. 9%) was significant in both the RT+CR+AT and CR+AT groups, respectively. DXA revealed that both the RT+CR+AT and CR+AT groups lost significant amounts of body fat (−6.5 vs. −7.4 kg; a decrease of 12% vs. 13%), percent body fat (−2.2% vs. −2.5%), and SM (−2.1 vs. −2.1 kg), respectively. However, the percentage of SM (relative to total body weight) increased significantly in both groups (2.0% vs. 2.5%, respectively). There were no significant between-group differences for these DXA derived body composition measurements.

Table 3.

Body Composition and Cardiac Function Outcomes by Treatment Group

	CR + AT (n=38)			RT + CR + AT (n=39)			Treatment Effect Size	p-value
Body Composition	Baseline	Follow-up	Within group Change (95% CI)	Baseline	Follow-up	Within group Change (95% CI)	Between Group Difference (95% CI)
Weight, kg	106±20	96 (95,97)	−9 (−11, −8)**	104±18	97 (96,98)	−8 (−9, −7)**	1 (−1,3)	0.21
DXA Measurements
Skeletal muscle, kg	51.6±10.8	49.4 (48.8,50.1)	−2.1 (−2.7, −1.4)**	51.8±9.5	49.4 (48.8,50.0)	−2.1 (−2.7, −1.5)**	−0.1 (−1.0,0.8)	0.86
Fat, kg	53.0±10.5	45.2 (44.3,46.0)	−7.4 (−8.1, −6.7)**	51.3±10.6	46.0 (45.1,46.8)	−6.5 (−7.2, −5.8)**	0.8 (−0.4,2.0)	0.20
Skeletal muscle, %	48.2±4.0	51.0 (50.6,51.4)	2.3 (1.8, 2.7)**	49.2±4.2	50.6 (50.2,51.1)	2.0 (1.6,2.5)**	−0.4 (−1.0,0.2)	0.23
Fat, %	49.6±4.1	46.5 (46.1,47.0)	−2.5 (−3.0, −2.1)**	48.6±4.5	46.9 (46.5,47.4)	−2.2 (−2.7, −1.8)**	0.4 (−0.2,1.1)	0.19
MRI Measurements
Thigh
Subcutaneous fat, cm2	181±64	141 (136,145)	−24 (−29, −19)**	162±71	141 (136,145)	−24 (−29, −18)**	0.2 (−6.2,6.6)	0.95
Skeletal muscle, cm2	111±30	109 (106,111)	−3 (−5, −1)*	113±21	109 (107,111)	−3 (−5, −1)*	0.5 (−2.8,3.8)	0.77
Intramuscular fat, cm2	34±11	29 (28,30)	−4 (−5,−3)**	32±10	30 (29,31)	−4 (−5, −3)**	0.3 (−1.1,1.9)	0.66
SM:IMF ratio	3.5±1.1	4.0 (3.8,4.2)	0.4 (0.2,0.6)**	3.7±1.1	4.0 (3.8,4.1)	0.4 (0.2,0.5)**	0 (−0.3,0.2)	0.85
Abdominal
Subcutaneous fat, cm2	375±121	326 (315,338)	−58 (−71,−45)**	400±110	329 (319,340)	−57 (−6, −45)**	3 (−13,19)	0.72
Visceral fat, cm2	194±106	165 (156,174)	−36 (−48, −25)**	194±89	172 (163,180)	−30 (−40, −20)**	6 (−6,19)	0.29
Epicardial fat, cm2	24.1±13.2	25.9 (22.9,28.9)	0.5 (−2.6,3.6)	25.3±9.8	25.0 (22.5,27.5)	−0.5 (−3.2,2.2)	−0.9 (−4.9,3.0)	0.64
Paracardial fat, cm2	24.0±13.8	24.9 (21.8,27.9)	−0.1 (−3.5,3.3)	24.6±12.9	23.6 (21.1,26.2)	−1.0 (−3.9,1.9)	−1.2 (−5.2,2.8)	0.55
Pericardial fat, cm2	48.1±26.4	50.8 (45.5,56.0)	0.2 (−5.4,5.9)	49.9±21.4	48.6 (44.2,53.1)	−1.8 (−9.1,5.5)	−2.1 (−9.0,4.8)	0.54
Cardiac Function
MRI Measurements
LV mass, g	101±20	100 (98,103)	−3 (−5, −1)*	107±21	102 (100,104)	−1 (−3,1)	1 (−2,5)	0.37
End-diastolic volume, mL	126±25	123 (119,127)	−5 (−9,0)*	135±26	123 (120,127)	−5 (−9—,1)*	0 (−5,6)	0.89
LV mass/end-diastolic volume	0.81±0.13	0.82 (0.79,0.85)	0.00 (−0.03,0.03)	0.80±0.13	0.83 (0.80,0.86)	0.02 (−0.01,0.05)	0.01 (−0.03,0.05)	0.49
Ejection fraction, %	61±5	61 (60,62)	0 (−1,1)	61±6	60 (59,61)	−1 (−2,0)	−1 (−2,1)	0.36
Echo-Doppler Measurements
LV mass, g	204±50	190 (179,200)	−14 (−25,−3)*	202±47	191 (181,202)	−11 (−22, −1)*	2 (−13,17)	0.82
Relative wall thickness	0.51±0.09	0.53 (0.50,0.55)	0.02 (−0.01,0.04)	0.52±0.08	0.52 (0.50,0.55)	0.00 (−0.02,0.03)	−0.01 (−0.04,0.03)	0.66
Left atrium diameter, cm	3.9±0.6	4.0 (3.8,4.1)	0.1 (−0.0,0.2)	3.9±0.6	3.9 (3.8,4.0)	0.0 (−0.1,0.1)	−0.1 (−0.2,0.1)	0.28
E/A ratio	0.89±0.32	0.92 (0.86,0.99)	0.06 (−0.00,0.13)	0.87±0.27	0.91 (0.85,0.98)	0.04 (−0.03,0.10)	−0.01 (−0.10,0.08)	0.83
e’, cm/s	6.4±1.6	6.3 (5.9,6.6)	−0.1 (−0.6,0.3)	6.5±2.1	6.1 (5.7,6.4)	−0.4 (−0.9,0.0)	−0.2 (−0.7,0.3)	0.42
E/e’ ratio	13.3±6.4	13.6 (12.5,14.6)	0.4 (−0.8,1.6)	13.1±4.5	14.1 (13.1,15.2)	1.2 (0.0,2.3)*	0.6 (−0.9,2.1)	0.42
Pulse wave velocity, cm/s†	1011±256	965 (909,1020)	−72 (−144,0)	1057±352	959 (907,1011)	−84 (−152, −16)*	−6 (−83,71)	0.88

Baseline data presented as mean±SD. Follow-up data presented as LS Means±SE (95%CI) adjusted for baseline value, age, and sex from ANCOVA. Within-group change estimates derived from cLMM. Abbreviations: DXA: dual X-ray absorptiometry; SM: skeletal muscle; IMF: intramuscular fat; LV: left ventricular; E: E-wave velocity; E/A: early to atrial filling velocity; e’; early mitral annulus velocity (septal); MRI: magnetic resonance imaging

^*significant at p<0.05

^**significant at p<0.001

^†determined using carotid-femoral artery by Doppler echocardiography

Thigh MRI (Table 3) revealed that both groups significantly decreased SCF area (−23.5 vs. −24.0 cm²; a decrease of 14% vs. 13%), IMF area (−3.6 vs. −4.2 cm²; a decrease of 11% vs. 12%), and SM area (−2.9 vs. −3.0 cm²; a decrease of 2% vs. 2%), respectively. Consequently, thigh skeletal muscle SM:IMF ratio increased significantly in both groups by 1%. There were no significant differences between groups for changes in thigh tissue composition.

Abdominal MRI (Table 3) revealed that abdominal SCF visceral fat decreased significantly within both groups, with no significant between-group differences.

Despite decreased SM in both groups, leg muscle strength improvement was observed in RT+CR+AT (4.9 Nm [5.4%]) but not CR+AT (−1.1 Nm [−1%]) with a between-group difference of 6.4(−0.1,12.9) Nm, p=0.053. Also, muscle quality improved significantly in the RT+CR+AT group (0.07 Nm/cm²=8.7%) but not in CR+AT (0.02 Nm/cm² = −1%), with a between-group difference of 0.07(0.00,0.13) Nm/cm², p=0.043.

Both RT+CR+AT and CR+AT significantly reduced LV mass and arterial stiffness with no between-group differences. (Table 3). Epicardial, pericardial, and paracardial fat were all unchanged from baseline to follow-up. (Table 3)

Discussion

This study is the first to examine whether addition of RT to CR+AT in older individuals with obese HFpEF further improves exercise capacity and attenuates loss of SM during CR+AT. In these older patients with obese HFpEF, we achieved excellent adherence to both exercise and dietary interventions and considerable weight reduction. This was associated with clinically meaningful improvements in both physical function and QOL, in accord with our prior trial results.¹¹ This trial significantly extends prior knowledge by demonstrating that addition of RT appeared safe, was well tolerated, and significantly improved leg strength and muscle quality, without further increasing VO₂peak or attenuating SM loss. Both RT+CR+AT and CR+AT significantly reduced LV mass and arterial stiffness, potentially important outcomes in HFpEF.

Effect of Adding RT on Exercise Capacity and QOL

VO₂peak (ml/min) was the pre-planned primary outcome as it assesses total oxygen utilization during maximal exercise and is not directly influenced by changes in bodyweight. Both had significant, clinically meaningful increases in VO₂peak, by 7 and 5%, respectively, without significant between-group difference. When VO₂peak was expressed in ml/kg/min, there were improvements of 2.2 and 2.4 ml/kg/min (15% and 16%), in the RT+CR+AT and CR+AT groups, respectively. This magnitude of improvement in indexed peak (ml/kg/min) VO₂ is large, and substantially exceeds the clinically meaningful differences of 1.0 mL/kg/min and 10% for this clinical population.⁴¹ However, contrary to our hypothesis, addition of RT to CR+AT did not result in further increased VO₂peak.

Other secondary measures of exercise performance, including exercise time to exhaustion (2.3 vs. 2.2 minutes; 1 minute is clinically meaningful) and 6-minute walk distance (47 vs. 61 meters; 30 meters is clinically meaningful) also showed large, clinically meaningful, significant improvements in both groups, supporting improved exercise capacity. The Short Physical Performance Battery (SPPB), also improved significantly, by 0.8 units in both groups; clinically meaningful change is 0.6 units.³⁵

While both the disease-specific (KCCQ) and multiple general measures (SF-36; EuroQOL) of QOL improved significantly with CR+AT, the addition of RT did not produce further improvements in these measures. The KCCQ score at baseline (mean of 69) indicated moderately impaired QOL yet both CR+AT and CR+AT+RT groups produced large improvements of 15–20 points during the intervention, greatly exceeding the threshold for clinically meaningful improvement of 5 points, confirming the strong effect of exercise plus dietary weight loss interventions in these patients.^11,42 NYHA class by a blinded observer improved as well. Depressive symptoms, measured by the CES-D instrument, also significantly improved.

Effect of Adding RT on Body Weight/Composition

Both RT+CR+AT and CR+AT produced large, significant reductions in body weight, body fat, subcutaneous and visceral fat, and thigh intramuscular fat (IMF), with no between-group differences. By MRI, both groups had significant reductions, with no significant between-group difference in abdominal subcutaneous (15% vs. 14%) and visceral fat (16% vs. 18%), respectively. It is known that a reduction of 5–10% total bodyweight and/or body fat, particularly abdominal visceral fat, has many favorable health benefits in clinical populations, including patients with diabetes and with HFpEF.^11,43,44 Despite improved exercise capacity and QOL, there was no significant change in epicardial or pericardial fat by MRI, a robust technique for measuring this outcome.

Most weight loss studies, including our prior trial of CR+AT in obese HFpEF, observed that 30–40% of the weight lost is lean body mass, which is primarily SM tissue.^43,45 Skeletal muscle loss is concerning as it may hinder further gains in exercise capacity, reduce basal metabolic rate/caloric expenditure, and has been associated with disability, hospitalizations, and death.^23,24,46 Despite providing a guideline-based RT program, and optimal protein intake (≥1.2 gm protein/kg body weight/day), the RT+CR+AT group had no significant difference in SM loss (−2.1 kg) and thigh skeletal muscle area loss (~3 cm²) as the CR+AT group. This outcome was contrary to our hypothesis, as prior studies have shown that RT, in the absence of caloric restriction, increases muscle mass in older adults in general and in patients with HFrEF.^27,29 Moreover, RT has been shown to attenuate SM loss during hypocaloric weight loss in healthy adults.⁴⁷ The mechanisms for the lack of attenuation of SM loss in older patients with obesity and HFpEF, despite the addition of RT, could be catabolism of low quality SM, thereby preserving and enhancing quality of the remaining muscle tissue.

Lower Extremity Muscle Mass, Strength and Quality

Although it has been postulated that reduced SM plays a major role in the age-related decline of strength, initial SM and changes SM explain only a small portion (~5%) of variability of age-related strength decline.⁴⁸ In Health ABC,⁴⁹ the loss of SM was associated with the decline in strength in older adults, but the decline in strength was much more rapid than the concomitant loss of SM, suggesting there is a decline in muscle quality during aging. Moreover, maintaining or gaining muscle mass did not appear to prevent aging-associated declines in muscle strength in Health ABC participants.⁴⁹ The present investigation demonstrated, for the first time in this clinical population, that the addition of RT to CR+AT resulted in a significantly greater improvement in both isokinetic lower extremity strength and skeletal muscle quality. Peak torque during leg extension significantly increased by 16% (89.5 to 103.9 Nm) in the RT+CR+AT but decreased by 8% (105.8 vs. 97.5 Nm) in CR+AT group, indicating that addition of RT completely prevented this loss and even improved muscle strength. Although several studies have demonstrated the positive impact of RT on muscle strength in weight stable patients with HFrEF,^26–30 this is the first study to demonstrate that RT can produce improvement in lower limb muscular strength during CR+AT induced weight loss in older patients with obese HFpEF. These positive effects of adding RT may be clinically important given the age of these patients, the concomitant effect of aging, and the additional SM loss due to the HFpEF condition which we previously reported.^16–18,23 Indeed, muscle strength is an important clinical outcome among older persons, independent of muscle mass; reduced skeletal muscle strength is strong predictor of a range of adverse outcomes, including nursing home placement and death.

In addition to the well-recognized age and HF-related reductions in SM quantity, there is substantial evidence that SM quality may decrease as well.^9,20,21,48. SM quality is as the capacity to generate force relative to the mass/volume of contractile tissue. Declines in SM strength with aging are not completely explained by declines in muscle mass.^49,50 While strength alone quantifies the amount of force a muscle can generate, larger muscles are not necessarily stronger. A smaller muscle may be more effective at generating force, due to more contractile proteins, less fat infiltration, or other physiological properties that can alter the quality of the muscle.^9,10,48,49 In the present study, lower extremity (quadriceps) skeletal muscle quality, assessed by the ratio of knee extensor strength to thigh muscle area (expressed in Nm/cm²) as previously described,¹¹ improved significantly in RT+CR+AT (by 21%) but did not improve in the CR+AT group. Thus, the current study indicates that during CR+AT induced weight and fat loss, addition of RT can significantly improve muscle quality in older patients with obese HFpEF. Although the clinical importance of muscle quality is less established than muscle strength and mass, because it takes into account both strength and body composition, muscle quality is a highly relevant outcome in older persons^23,43.

Impact of RT on Cardiac Structure and Function and Arterial Stiffness

CT+AT resulted in small but significant decreases in LV and end-diastolic volume as well as a nearly significant (p=0.06) increase in LV E/A ratio consistent, with our previous findings.¹¹ Addition of RT resulted in significant improvements with no significant between-group difference in LV mass (by echo) and end-diastolic volume (by MRI). Despite limited evidence, there are lingering concerns that RT may negatively impact LV systolic and/or diastolic function in HF patients via increased blood pressure and LV afterload. While several studies have examined the effect of RT in HFrEF,^26–29 this is the first trial to formally test the effects of adding RT to traditional AT in HFpEF. A novel finding was the favorable effects on arterial stiffness, measured by carotid-femoral pulse-wave velocity on Doppler echocardiography, which were observed in both groups. This suggests that the concern that RT may adversely HFpEF by increased afterload may be unfounded. These results suggest that supervised RT, when added to CR+AT, appears to have no adverse effect on cardiac or arterial structure and function, is safe, and was not associated with any exercise-related adverse events. Finally, improved LV mass, end-diastolic volume, and arterial stiffness with CR and exercise are highly relevant to HFpEF where abnormalities in each of these contribute to the pathogenesis and severity of the disorder.

Strengths and Limitations

Strengths of the current study include that it was a randomized, controlled, single-blind clinical trial with a well-defined population of older patients with obese HFpEF, extensive phenotyping including with serial CPET and MRI, professionally administered CR diet, and medically supervised AT and RT with careful and frequent monitoring. Thus, efficacy and safety could differ under other conditions. We did not test effects of RT alone, without concomitant CR +AT. The minimum BMI for study participants was 28, which is based on World Health Organization recommendations for dietary intervention for the general population, and which includes most patients with HFpEF. However, our data do not address safety and efficacy of this intervention in patients with lower BMIs. The participants had typical clinical features of obese HFpEF, including severe exercise intolerance, LV hypertrophy, and LV diastolic dysfunction, 93% were on maintenance diuretics, and met predetermined criteria for HFpEF utilized in prior publications^{2,8,10.11,31,32} and in accord with 2013 recommendations of the AHA and ACC which were the most current at the time of trial design. Patients enrolled in the study were medically stable and well-compensated; therefore, the results may not apply to other populations, particularly those who are unstable or have acute decompensated HFpEF, who may require other approaches.⁵⁰ Although the RT intervention was in accord with ACSM guidelines for RT at the time, we cannot exclude the possibility that increased stimulus (time and/or intensity) and larger sample size could have produced significantly increased muscle mass. However, SM strength and muscle quality did improve significantly with RT, and these outcomes are important to patients. Finally, we cannot exclude that greater intensity of protein supplementation or RT may have reduced SM loss.

Conclusion

In older patients with obese HFpEF, combined CR+AT produced robust weight loss and improvements in physical function (including the primary outcome of VO₂peak), QOL, body composition, as well as cardiac and arterial function. The addition of RT to CR+AT was well-tolerated and resulted in increased leg strength and leg muscle quality without attenuating SM loss or further increasing VO₂peak or QOL. These positive findings are particularly notable in a study population, predominantly non-white obese women, that is unduly burdened by obese HFpEF and are often under-represented in clinical trials.

What is new ?

This is the first clinical trial to add resistance exercise training (RT) to an already effective lifestyle intervention of caloric restriction (CR) and aerobic exercise training (AT), in older patients with obese HFpEF. The addition of RT to CR+AT was well-tolerated and resulted in increased leg strength and leg muscle quality without attenuating SM loss or further increasing exercise capacity (VO₂peak) or QOL. These positive findings are particularly notable in a study population, predominantly non-white obese women, that is unduly burdened by obese HFpEF and are often under-represented in clinical trials.

What are the clinical implications?

The results of this study suggest that supervised resistance training (RT), when added to caloric restriction and aerobic exercise training (CR+AT), appears to have no adverse effect on cardiac or arterial structure and function, is safe, and was not associated with any exercise-related adverse events. Finally, improved LV mass, end-diastolic volume, and arterial stiffness with CR and exercise are highly relevant to HFpEF where abnormalities in each of these contribute to the pathogenesis and severity of the disorder. Clinicians should consider adding supervised RT to improve skeletal muscle strength and muscle quality of older patients with obese HFpEF.

Non-standard Abbreviations and Acronyms

ACC

American College of Cardiology

ACSM

American College of Sports Medicine

AHA

American Heart Association

AT

aerobic exercise training

cLMM

constrained linear mixed-effects models

CPET

cardiopulmonary exercise test

CR

caloric restriction diet

DXA

dual x-ray absorptiometry

HFpEF

heart failure with preserved ejection fraction

HFrEF

HF with reduced ejection fraction

KCCQ

Kansas City Cardiomyopathy Questionnaire

LV

left ventricle

QOL

quality-of-life

RT

resistance training

SM

skeletal muscle mass

SPPB

short physical performance battery

TC

thigh compartment area

VO₂peak

peak exercise oxygen consumption

IMF

intramuscular fat

6MWD

Six-Minute Walk Distance

SF 36

36-item Short-Form Health Survey

cLMM

constrained linear mixed-effects models