Regional Adipose Distribution and Its Relationship to Exercise Intolerance in Older Obese HFpEF Patients

Abstract

Background

Recent studies indicate that excess total body adipose contributes to exercise intolerance in heart failure with preserved ejection fraction (HFpEF). However, the impact of the pattern of regional (abdominal, cardiac, intermuscular) adipose deposition on exercise intolerance in HFpEF is unknown.

Methods

We measured total body (dual-energy x-ray absorptiometry) and regional adipose (magnetic resonance imaging), peak oxygen uptake (peak VO₂), 6-minute walk distance (6MWD), short physical performance battery (SPPB), and leg press power in 100 older obese HFpEF patients and 61 healthy controls (HC), and adjusted for age, gender, race, and body surface area.

Results

Peak VO₂ (15.7±0.4 vs. 23.0±0.6 ml/kg/min; p<0.001), 6MWD (427±7 vs. 538±10 m; p<0.001), SPPB (10.3±0.2 vs. 10.9±0.2; P<0.05) and leg power (117±5 vs. 152±9 watts, p=0.004) were significantly lower in HFpEF versus HC. Total fat mass, total percent fat, abdominal subcutaneous fat, intra-abdominal fat, and thigh intermuscular fat were significantly higher, while epicardial fat was significantly lower in HFpEF versus HC. After adjusting for total body fat, intra-abdominal fat remained significantly higher while epicardial fat remained significantly lower in HFpEF. Abdominal subcutaneous fat, thigh subcutaneous fat, and thigh intermuscular fat/skeletal muscle ratio were inversely associated, while epicardial fat was directly associated, with peak VO₂, 6MWD, SPPB, and leg power. With multiple stepwise regression, intra-abdominal fat was the strongest independent predictor of peak VO₂ and 6MWD.

Conclusions

In metabolic/obese HFpEF, the pattern of regional adipose deposition may have important adverse consequences beyond total body adipose. Interventions targeting intra-abdominal and intermuscular fat could potentially improve exercise intolerance.

Clinical Trial Registration

www.clinicaltrials.gov Identifier: NCT00959660

Introduction

Heart failure (HF) with preserved ejection fraction (HFpEF) is the fastest growing form of HF and is associated with high morbidity and mortality.(1) Exercise intolerance, manifested as severe exertional dyspnea and fatigue, is a hallmark of chronic HFpEF and is associated with reduced quality-of-life (QoL).(2,3) The mechanisms of exercise intolerance are incompletely understood, however it appears that abnormalities in non-cardiac, systemic factors are important contributors in addition to cardiac function.(3–9)

Obesity is a major independent risk factor for development of HF,(10) and >80% of HFpEF patients are overweight/obese.(11,12) Increased adiposity promotes inflammation, hypertension, dyslipidemia and insulin resistance and impairs cardiac, vascular, pulmonary, and skeletal muscle function, all of which contribute to the pathophysiology of HFpEF.(7,12–15) Multiple lines of evidence suggest that excess body adipose tissue contributes to reduced peak exercise oxygen uptake (peak VO₂) in HFpEF.(12,14–16) Adipose-induced inflammation has wide-ranging adverse effects including coronary and systemic microvascular endothelial dysfunction, capillary rarefaction and impaired skeletal muscle mitochondrial function and protein synthesis that result in reduced skeletal muscle oxygen delivery and extraction.(7,9,14,16) Emerging data suggest that, in addition to the amount of total body adipose tissue, the specific location of adipose tissue may play a role in adverse outcomes, including exercise intolerance.(4,14,17,18) However, the impact of adipose distribution on exercise performance has not been systematically examined in HFpEF.

We aimed to test the hypothesis that older obese HFpEF patients have significantly greater abdominal, cardiac, and intermuscular fat compared to healthy, age-matched controls (HC), out of proportion to total body fat, and that these abnormalities are associated with objective measures of physical function. Therefore, we performed a prospective study in HFpEF patients and HCs using dual-energy x-ray absorptiometry (DEXA) to assess total body adipose mass and magnetic resonance imaging (MRI) for regional adipose mass, and cardiopulmonary exercise testing, 6-minute walk distance (6MWD), and lower extremity muscle power to comprehensively assess physical function.

Methods

Study Participants

As previously described,(13) patients were interviewed and examined by a board certified cardiologist and met these inclusion criteria: ≥60 years of age; body mass index ≥ 30 kg/m²; signs and symptoms of HF defined by National Health and Nutrition Examination Survey score ≥ 3,(19) the criteria by Rich et al.,(20) or both; left ventricular (LV) ejection fraction ≥50%; no segmental wall motion abnormalities; and no significant ischemic or valvular heart disease, pulmonary disease, anemia, or other disorder that could explain the patients’ symptoms.(2,21,22) HCs were recruited from the community and excluded if they had any chronic medical illness, were on any chronic medication, had current complaints or an abnormal physical examination (including blood pressure ≥ 140/90 mmHg), had abnormal results on the screening tests (echocardiogram, electrocardiogram, cardiopulmonary exercise testing), or regularly undertook vigorous exercise.(2,22) The study was approved by the Wake Forest School of Medicine Institutional Review Board. All participants provided written informed consent.

Outcome Measures

Outcomes were assessed and images analyzed by individuals blinded to participant group.

LV Morphology and Function

As previously described, LV mass and volumes were assessed by cardiac MRI (1.5 T Siemens Avanto scanner) from a series of multi-slice, multi-phase gradient-echo sequences positioned perpendicular to the LV long axis, spanning apex to base.(13) The epi- and endocardial borders of each slice were traced manually at end-diastole and end-systole, and volumes were calculated by Simpson’s rule. LV stroke volume and EF were calculated from standard formulae.(13)

As previously described, LV filling patterns, mitral annulus velocity, and pulse-wave velocity were assessed by Doppler echocardiography (iE33, Philips Ultrasound).(13)

Body Composition

Total body fat and lean mass were measured by DEXA (Hologic Delphi QDR) using standardized protocols as previously described.(3,4,13)

Regional scans of the thigh, abdomen, and heart were performed by MRI as previously described.(4,13,23) Images were transferred to a dedicated workstation and analyzed using commercially available software (Sliceomatic, Tomovision, Montreal, Quebec) as previously described.(4,13,23) Cross-sectional thigh areas of skeletal muscle (SM), subcutaneous fat (SCF), intermuscular fat (IMF), and bone were measured with images taken at a constant location of 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. Abdominal images were taken from the tenth thoracic to the second sacral vertebrae positioned every 3 cm to cover the entire abdomen. A single slice at the level of the second lumbar vertebra was used for determination of abdominal fat measurements including SCF, intra-peritoneal and retroperitoneal fat. Intra-abdominal fat was calculated as the sum of intra-peritoneal and retroperitoneal fat. For epicardial (within the pericardium) and paracardial (outside the pericardium) fat volumes, axial images were acquired from the diaphragm to the aortic arch using a prospectively ECG-gated, T1-weighted, breath-held, black-blood, single-shot, turbo-spin echo sequence. Pericardial fat was calculated by summing epicardial and paracardial fat.(23)

Physical Function

As previously described, cardiopulmonary exercise testing was performed on a treadmill using the modified Naughton (HFpEF) or the Modified Bruce (HC) protocol using standardized instructions and encouragement to achieve a peak symptom-limited exhaustive effort.(13,24) Continuous expired gas analysis was performed during exercise and averaged over 15-second intervals (Medgraphics Ultima, Medical Graphics Corp.).(2,13,24) Peak VO₂ was calculated as the average of the last 30 seconds during peak exercise.(2,13,24) Ventilatory anaerobic threshold (VAT) and ventilation/carbon dioxide output (VE/VCO₂) slope were assessed as previously described.(21,24)

Six-minute walk distance was assessed as described by Guyatt et al.(25) The short physical performance battery (SPPB) is a validated, reliable measure of physical function in older populations and is strongly predictive of disability, hospitalization, nursing home readmission, and death.(26,27) It consists of three subtasks: standing balance, 4-meter walking speed at usual pace, and time to rise from a chair five times.(26) Leg press power (Watts) was assessed using the Nottingham power rig.(13) Leg power was chosen due to its strong correlation with disability and other adverse outcomes with aging.(28) Muscle quality was calculated as leg power divided by thigh muscle area from the MRI scans.(13)

Statistical Analysis

All analyses were performed with SAS Enterprise Guide (SAS Institute), version 7.1. All outcomes were tested at the 5% 2-sided level of significance. Patient characteristics are presented as mean and standard deviation or frequency and percent. Comparisons of patient characteristics were made by independent samples t-tests for continuous variables and Fisher’s exact or Chi Square tests for categorical variables. Outcome measures are presented as unadjusted (mean ± standard deviation) and adjusted (least squares mean±standard error), with a p-value for each. Unadjusted values were compared using independent samples t-tests, while adjusted values were compared using analysis of covariance with age, gender, race, and body surface area as covariates; a supplemental analysis adjusted for total body fat. Associations between regional fat and physical function measures were made by Spearman correlations and adjusted for body surface area. Multiple linear regression models were constructed to predict physical function using stepwise selection with age, gender, and race forced into the model since these are potential confounders in predicting physical function. The following regional fat measures showing consistent significance in bivariate analyses were included as candidate variables in the stepwise selection process: abdominal SCF, intra-abdominal fat, thigh IMF, and epicardial fat. Effect sizes were reported as partial R-square.

Results

Participant Characteristics

Clinical characteristics of the HFpEF patients and HC are shown in Table 1. HFpEF patients were clinically stable New York Heart Association Class II/III and had typical characteristics of HFpEF, including female preponderance, increased body mass index, increased LV mass and concentric remodeling, increased left atrial size, and Doppler diastolic dysfunction with increased E/e′ ratio.

Table 1.

Patient characteristics

Variable	HFpEF (n=100)	HC (n=61)	p-Value
Age (years)	66.5 ± 5.2	69.3 ± 7.4	0.011
Women	81 (81)	38 (62)	0.010
White	55 (55)	58 (95)	<0.001
Weight (kg)	105.5 ± 17.9	74.5 ± 16.4	<0.001
Body mass index (kg/m2)	39.3 ± 6.1	25.9 ± 4.9	<0.001
Body surface area (m2)	2.1 ± 0.2	1.8 ± 0.2	<0.001
NYHA class
II	60 (60)	--	--
III	40 (40)	--	--
Ejection Fraction (%)	61.1 ± 6.0	59.0 ± 4.8	0.030
LV mass (g)	214 ± 60	129 ± 35	<0.001
LV mass index (g/m2) *	102 ± 25	70 ± 14	<0.001
Relative wall thickness	0.57 ± 0.12	0.38 ± 0.05	<0.001
Left atrial diameter (cm)	4.0 ± 0.5	3.4 ± 0.6	<0.001
Diastolic Filling Pattern †
Normal	2 (2)	48 (80)
Impaired relaxation	87 (87)	12 (20)
Pseudonormal	9 (9)	0 (0)	<0.001
Restrictive	1 (1)	0 (0)
Indeterminate	0 (0)	0 (0)
e′ (cm/s)	6.2 ± 1.5	7.9 ± 1.6	<0.001
E/e′ ratio	13.1 ± 3.7	9.3 ± 2.2	<0.001
History of atrial fibrillation	3 (3%)	--	--
History of established CAD	11 (11%)	--	--
History of hypertension	95 (95)	--	--
History of diabetes mellitus	35 (35)	--	--
Blood Pressure (mmHg)
Systolic	135 ± 14	123 ± 11	<0.001
Diastolic	78 ± 8	75 ± 7	0.018
Medications
Diuretics	76 (76)	--	--
ARBs	35 (35)	--	--
ACE inhibitors	37 (37)	--	--
Beta blockers	40 (40)	--	--
Ca2+ channel blockers	35 (35)	--	--
Nitrates	9 (9)	--	--

Values are mean ± SD, or n (%);

^*LV mass indexed to body surface area.

Diastolic filling pattern determined according to American Society of Echocardiography criteria; Abbreviations: NYHA, New York Heart Association; LV, left ventricle; e′, early mitral annulus velocity (septal); E, E-wave velocity; CAD, coronary artery disease; ARB, angiotensin receptor blocker; ACE, angiotensin-converting enzyme.

Exercise Performance Measures

Peak exercise VO₂ (in ml/min, and indexed to either body weight, lean body mass or leg lean mass), carbon dioxide production, and heart rate were significantly lower in HFpEF versus HC, while ventilation/carbon dioxide slope was significantly higher (Table 2). Peak exercise respiratory exchange ratio (an objective, reliable measure of participant effort) was similar in HFpEF versus HC. There were no significant differences in peak exercise systolic or diastolic blood pressure (Table 2). The 6MWD, SPPB total score, and leg press power (absolute or indexed to thigh muscle area) were also significantly lower in HFpEF patients compared to HC (Table 3).

Table 2.

Exercise performance measures

Variable	Unadjusted Mean ± SD			Adjusted* LSMean ± SE
	HFpEF	HC	p-Value	HFpEF	HC	p-Value
Peak VO2
ml/min	1509 ± 313	1852 ± 608	<0.001	1464 ± 35	1927 ± 48	<0.001
ml/kg/min	14.5 ± 2.6	25.0 ± 7.0	<0.001	15.7 ± 0.4	23.0 ± 0.6	<0.001
ml/kg lean/min†	27.7 ± 5.2	37.4 ± 7.6	<0.001	28.3 ± 0.7	36.6 ± 0.9	<0.001
ml/kg leg lean/min†	87.4 ± 17.9	117.8 ± 24.1	<0.001	90.9 ± 2.3	112.3 ± 3.0	<0.001
Peak carbon dioxide production (ml/min)	1701 ± 403	2123 ± 712	<0.001	1650 ± 43	2207 ± 60	<0.001
Peak respiratory exchange ratio	1.12 ± 0.08	1.15 ± 0.09	0.073	1.12 ± 0.01	1.15 ± 0.01	0.18
Ventilatory anaerobic threshold (ml/min)	1014 ± 244	1115 ± 373	0.063	956 ± 25	1210 ± 35	<0.001
Ventilation/carbon dioxide slope	29.45 ± 3.99	28.56 ± 3.57	0.16	30.11 ± 0.41	27.46 ± 0.57	0.001
Peak heart rate (beats/min)	139 ± 18	153 ± 13	<0.001	139 ± 2	153 ± 3	<0.001
Peak systolic blood pressure (mmHg)	177 ± 18	179 ± 18	0.39	176 ± 2	182 ± 3	0.13
Peak diastolic blood pressure (mmHg)	78 ± 9	76 ± 8	0.13	77 ± 1	78 ± 1	0.62
6-minute walk distance (m)	412 ± 73	563 ± 70	<0.001	427 ± 7	538 ± 10	<0.001
SPPB total score	10.0 ± 1.7	11.3 ± 0.8	<0.001	10.3 ± 0.2	10.9 ± 0.2	0.044
SPPB Balance score	3.8 ± 0.6	4.0 ± 0.2	0.010	3.8 ± 0.1	4.0 ± 0.1	0.15
SPPB Gait speed score	3.8 ± 0.5	4.0 ± 0.0	<0.001	3.8 ± 0.0	3.9 ± 0.1	0.30
SPPB Chair rise score	2.5 ± 1.2	3.3 ± 0.7	<0.001	2.6 ± 0.1	3.0 ± 0.2	0.12
Leg press power (W)	112 ± 53	163 ± 67	<0.001	117 ± 5	152 ± 9	0.004
Leg muscle quality (W/cm2)‡	0.91 ± 0.33	1.45 ± 0.52	<0.001	0.97 ± 0.05	1.28 ± 0.08	0.004

Raw data are presented as mean ± SD;

^*Adjusted for age, race, gender, and body surface area and presented as least square means ± standard error;

^†kg of lean body mass and kg of leg lean mass were measured by dual energy x-ray absorptiometry;

^‡Leg muscle quality is leg power divided by thigh muscle area. SPPB = Short Physical Performance Battery.

Table 3.

Total and regional body composition measures

Variable	Unadjusted Mean ± SD			Adjusted* LSMean ± SE
	HFpEF	HC	p-Value	HFpEF	HC	p-Value
DEXA
Total Mass (kg)	102.4 ± 14.2	75.7 ± 16.3	<0.001	94.4 ± 0.6	88.0 ± 0.8	<0.001
Total Fat Mass (kg)	46.7 ± 10.2	24.1 ± 9.4	<0.001	41.0 ± 0.7	32.9 ± 0.9	<0.001
Total Percent Fat	45.5 ± 6.4	31.5 ± 8.5	<0.001	42.8 ± 0.6	35.6 ± 0.8	<0.001
Total Lean Mass (kg)	53.3 ± 9.1	49.3 ± 11.0	0.016	51.1 ± 0.4	52.8 ± 0.6	0.041
Total Percent Lean	52.2 ± 6.3	65.5 ± 8.2	<0.001	54.7 ± 0.6	61.6 ± 0.8	<0.001
MRI
Abdominal SC fat (cm2)	377.8 ± 145.2	147.7 ± 84.5	<0.001	312.9 ± 12.8	236.9 ± 15.9	0.002
Intra-peritoneal fat (cm2)	116.4 ± 61.5	68.8 ± 60.4	<0.001	110.0 ± 6.4	77.7 ± 8.0	0.007
Retroperitoneal fat (cm2)	94.4 ± 51.3	47.3 ± 35.0	<0.001	95.8 ± 4.1	45.4 ± 5.1	<0.001
Intra-abdominal fat (cm2)	210.8 ± 107.9	116.1 ± 90.0	<0.001	205.7 ± 9.7	123.1 ± 12.1	<0.001
Abdominal SC fat/Intra-abdominal fat ratio	2.34 ± 1.73	1.91 ± 1.31	0.10	1.94 ± 0.18	2.46 ± 0.22	0.11
Total thigh area (cm2)	316.8 ± 80.4	205.3 ± 47.3	<0.001	277.0 ± 5.6	270.7 ± 7.7	0.56
Thigh SC fat (cm2)	165.2 ± 74.8	73.8 ± 41.9	<0.001	132.9 ± 5.4	126.7 ± 7.5	0.55
TC area (cm2)	151.6 ± 28.7	131.5 ± 32.1	<0.001	144.1 ± 2.1	143.9 ± 3.0	0.96
Femur (cm2)	4.3 ± 0.7	5.9 ± 1.0	<0.001	4.2 ± 0.1	6.1 ± 0.1	<0.001
Thigh IMF (cm2)	25.9 ± 9.2	14.4 ± 5.9	<0.001	23.2 ± 0.9	18.9 ± 1.2	0.014
Thigh IMF/TC (%)	17.1 ± 5.2	11.3 ± 4.8	<0.001	16.2 ± 0.6	12.9 ± 0.8	0.003
Thigh SM (cm2)	121.5 ± 25.7	111.2 ± 30.5	0.029	116.8 ± 2.0	118.9 ± 2.8	0.58
Thigh SM/TC (%)	80.0 ± 5.1	84.0 ± 5.0	<0.001	80.9 ± 0.6	82.6 ± 0.8	0.11
Thigh IMF/SM ratio	0.22 ± 0.08	0.14 ± 0.07	<0.001	0.21 ± 0.01	0.16 ± 0.01	0.009
Paracardial fat (cm3)	63.1 ± 39.6	53.9 ± 36.7	0.19	63.6 ± 4.1	53.1 ± 5.5	0.18
Epicardial fat (cm3)	36.0 ± 17.7	54.9 ± 20.6	<0.001	36.3 ± 2.3	54.5 ± 3.0	<0.001
Pericardial fat (cm3)	99.0 ± 53.0	108.8 ± 52.9	0.31	99.8 ± 5.6	107.6 ± 7.6	0.48

Raw data are presented as mean ± SD;

^*Adjusted for age, race, gender, and body surface area and presented as least square means ± standard error; Intra-abdominal fat = Intra-peritoneal + Retroperitoneal fat; SC fat = Subcutaneous Fat; TC Area = Thigh Compartment Area; SM = Skeletal Muscle; IMF = Intermuscular Fat; TC area = SM + IMF + Femur. Total thigh area = TC Area + Thigh SC fat. Pericardial fat = Paracardial fat + Epicardial fat.

Total and Regional Body Composition Measures

By DEXA, total body mass, total body fat mass, and percent fat were significantly higher, while total lean (muscle) mass and percent lean mass were significantly lower in HFpEF compared to HC (Table 3). By MRI, the following patterns of regional fat deposition were observed in HFpEF versus HC (Table 3): abdominal SCF and intra-abdominal fat were significantly higher; thigh IMF, thigh IMF/thigh compartment (TC,%), thigh IMF/SM ratio were significantly higher while thigh SCF was similar; epicardial fat was 33% lower (p<0.001) while pericardial and paracardial fat were similar. Even when adjusted for total body fat, intra-abdominal fat remained significantly higher while epicardial fat remained significantly lower in HFpEF versus HC; additionally, abdominal SCF/intra-abdominal fat ratio became significantly lower in HFpEF versus HC. Furthermore, there was a significant association between intra-abdominal fat and epicardial fat in the HFpEF patients (r=0.37, p=0.001) but not in the HC (r=0.16, p=0.26).

Relationships between body composition and exercise tolerance and physical function

All measures of physical function were inversely related to abdominal and thigh SCF, thigh IMF, thigh IMF/TC (%), and thigh IMF/SM ratio, and positively related to thigh SM/TC (%) and epicardial fat (Tables 4a–4b, Figure 1). Peak VO₂ showed trends (p<0.10) for inverse relationships with intra-peritoneal, retroperitoneal, and intra-abdominal fat. When adjusted for total body fat, peak VO₂ (ml/min and ml/kg/min), ventilatory anaerobic threshold, and leg press power remained positively related to epicardial fat.

Table 4a.

Relationships of cardiac fat, abdominal fat and thigh composition with Peak VO₂

Variable	Peak VO2 (ml/min)		Peak VO2 (ml/kg/min)		Peak VO2 (ml/kg lean/min)		Peak VO2 (ml/kg leg lean/min)
	Corr.	P-Value	Corr.	P-Value	Corr.	P-Value	Corr.	P-Value
Abdominal SCF (cm2)	−0.60	<0.001	−0.73	<0.001	−0.36	<0.001	−0.35	<0.001
Intra-peritoneal fat (cm2)	−0.12	0.19	−0.13	0.17	−0.16	0.073	−0.08	0.38
Retroperitoneal fat (cm2)	−0.15	0.10	−0.21	0.024	−0.30	<0.001	−0.19	0.036
Intra-abdominal fat (cm2)	−0.13	0.15	−0.15	0.091	−0.24	0.010	−0.13	0.15
Thigh SCF (cm2)	−0.63	<0.001	−0.71	<0.001	−0.34	<0.001	−0.39	<0.001
Thigh IMF (cm2)	−0.46	<0.001	−0.59	<0.001	−0.39	<0.001	−0.39	<0.001
Thigh IMF/TC (%)	−0.57	<0.001	−0.66	<0.001	−0.40	<0.001	−0.40	<0.001
Thigh SM (cm2)	0.51	<0.001	0.46	<0.001	0.19	0.034	0.20	0.026
Thigh SM/TC (%)	0.54	<0.001	0.62	<0.001	0.35	<0.001	0.36	<0.001
Thigh IMF/SM ratio	−0.57	<0.001	−0.66	<0.001	−0.39	<0.001	−0.39	<0.001
Paracardial fat (cm3)	0.00	0.96	0.04	0.65	−0.06	0.50	0.03	0.72
Epicardial fat (cm3)	0.36	<0.001	0.43	<0.001	0.28	0.002	0.32	<0.001
Pericardial fat (cm3)	0.17	0.060	0.23	0.013	0.10	0.28	0.18	0.050

Table 4b.

Relationships of cardiac fat, abdominal fat and thigh composition with other exercise performance measures

Variable	Ventilatory Anaerobic Threshold (ml/min)		6-Minute Walk Distance (m)		Total SPPB Score		Leg Press Power (W)
	Corr.	P-Value	Corr.	P-Value	Corr.	P-Value	Corr.	P-Value
Abdominal SCF (cm2)	−0.39	<0.001	−0.56	<0.001	−0.33	<0.001	−0.67	<0.001
Intra-peritoneal fat (cm2)	−0.00	0.97	−0.12	0.23	−0.03	0.78	0.05	0.60
Retroperitoneal fat (cm2)	−0.08	0.40	−0.19	0.054	−0.08	0.43	0.07	0.51
Intra-abdominal fat (cm2)	−0.02	0.80	−0.15	0.13	−0.04	0.65	0.08	0.41
Thigh SCF(cm2)	−0.42	<0.001	−0.58	<0.001	−0.38	<0.001	−0.63	<0.001
Thigh IMF (cm2)	−0.29	0.001	−0.46	<0.001	−0.33	<0.001	−0.36	<0.001
Thigh IMF/TC (%)	−0.37	<0.001	−0.54	<0.001	−0.40	<0.001	−0.49	<0.001
Thigh SM (cm2)	0.35	<0.001	0.39	<0.001	0.41	<0.001	0.57	<0.001
Thigh SM/TC (%)	0.35	<0.001	0.50	<0.001	0.40	<0.001	0.48	<0.001
Thigh IMF/SM ratio	−0.37	<0.001	−0.54	<0.001	−0.40	<0.001	−0.50	<0.001
Paracardial fat (cm3)	−0.00	0.97	−0.02	0.84	0.01	0.94	0.15	0.12
Epicardial fat (cm3)	0.27	0.002	0.36	<0.001	0.21	0.034	0.43	<0.001
Pericardial fat (cm3)	0.13	0.16	0.16	0.12	0.09	0.39	0.30	0.002

Data adjusted for Body Surface Area and presented as Spearman correlation coefficients. P value indicates significance for each correlation for the overall sample (HFpEF and HC). SCF = Subcutaneous fat; Intra-abdominal fat = Intra-peritoneal + Retroperitoneal fat; SM = Skeletal Muscle; IMF = Intermuscular Fat; TC = Thigh Compartment; SPPB = Short Physical Performance Battery. Pericardial fat = Paracardial fat + Epicardial fat.

Figure 1. Peak VO₂ (ml/kg/min) versus Regional Fat Measures.

Regression plots for baseline measures of Peak VO₂ (ml/kg/min) versus (A) Intra-Abdominal Fat, (B) Abdominal SCF, (C) Thigh IMF, (D) Thigh IMF as Percent of Thigh Compartment (TC), (E) Paracardial Fat, (F) Epicardial Fat. Similar results were found for non-indexed Peak VO₂ (ml/min) as well. Regional fat measures obtained using MRI. Thigh compartment calculated as Intermuscular Fat+Skeletal Muscle+Bone. Spearman r and p values additionally included for reference.

Multiple linear regression with age, gender, race forced in showed that intra-abdominal fat was the strongest independent predictor of peak VO₂ and 6MWD with partial R-square of 0.36 for peak VO₂ and 0.2 for 6MWD, respectively (Table 5), while abdominal subcutaneous and epicardial fat remained significant but modest predictors (inverse for epicardial).

Table 5.

Predictors of Physical Function

	Peak VO2 (ml/kg/min)		6-minute Walk Distance (m)		Leg Press Power (W)
Baseline Characteristic	Partial R-Square	p-value	Partial R-Square	p-value	Partial R-Square	p-value
Age	0.140	<0.001	0.156	<0.001	0.053	0.021
Gender	0.355	<0.001	0.093	<0.001	0.293	<0.001
Race	0.086	0.001	0.005	0.43	0.015	0.23
Intra-abdominal Fat	0.358	0.003	0.200	<0.001	--	--
Abdominal SCF	0.077	<0.001	0.182	<0.001	0.075	0.006
Epicardial Fat	0.034	0.046	0.033	0.048	--	--

^*Age, Gender, Race forced into model; Thigh IMF did not reach significance for Peak VO₂ or 6-minute Walk Distance.

Thigh IMF, Intra-abdominal Fat, nor Epicardial Fat reached significance for Leg Press Power.

Discussion

Multiple recent studies have shown that excess total body adipose is a key contributor to the severely reduced peak VO₂ in the metabolic/obese HFpEF phenotype.(3,4,8,9,15) Prior studies in other disorders suggest that the location of excess adipose has effects independent of total body adipose, and that regional adipose may be an important determinant of impaired cardiac function and exercise intolerance.(12,14,15,23) However, the impact of regional adipose in HFpEF is unkown. In this study, we examined the distribution of adipose tissue in HFpEF patients compared to HCs and the relationships of regional adipose depots with multiple objective measures of exercise tolerance and physical function. The major new findings of this study include that older, obese HFpEF patients have relatively larger depots of abdominal and thigh adipose, and smaller amounts of epicardial fat compared to HC. There was also a significant association between intra-abdominal fat and epicardial fat in the HFpEF patients but not in the HCs. Additional novel findings are that abdominal SCF, thigh SCF, and thigh IMF/SM ratio were inversely associated with all objective measures of physical function, including peak VO₂, 6MWD, ventilatory anaerobic threshold, SPPB, and leg power, while epicardial fat was directly associated with these physical function measures. In multiple linear regression, intra-abdominal fat yielded the highest partial R-square among all the regional fat measures and was the strongest independent predictor of peak VO₂ and 6MWD. These data suggest that in HFpEF, the pattern of regional adipose distribution may have important adverse consequences beyond those of total body adipose.

Previous studies in other disorders showed that Increased abdominal fat is associated with concentric remodeling, LV diastolic dysfunction and adverse subclinical changes in LV mechanics.(29–33) Obokata et al. recently reported that compared to non-obese HFpEF and controls, obese HFpEF patients had more concentric LV remodeling, greater right ventricular dysfunction, increased right and left heart filling pressures, impaired pulmonary vasodilation and reduced peak VO₂ during maximal supine cycle exercise.(15) We confirm and significantly extend these findings by showing, in a relatively large cohort of older, obese HFpEF patients compared to HC, that obese HFpEF have increased left atrial size, LV mass, relative wall thickness, and impaired Doppler measures of diastolic function, as well as impairments in multiple objective measures of physical function, including peak VO₂, VAT, SPPB, and leg muscle power, and by examining the relationships of physical function with multiple measures of total and regional adipose using both DEXA and MRI.

Increased adiposity promotes inflammation, hypertension, dyslipidemia and insulin resistance, all of which contribute to the pathophysiology of HFpEF.(13,16) Moreover, adipose-induced inflammation has wide-ranging adverse effects including coronary and systemic microvascular endothelial dysfunction, capillary rarefaction and impaired skeletal muscle mitochondrial function and protein synthesis that results in reduced skeletal muscle oxygen delivery and utilization.(7,9,14) A consequence of adipose-mediated inflammatory endothelial dysfunction and capillary rarefaction is that it reduces O₂ delivery to the active muscles resulting in fatigue even during low-level exercise.(7,14,22) Indeed, a novel finding of our study was that excess abdominal adipose was inversely associated with peak VO₂, VAT, 6MWD, and leg muscle power. Taken together, these data suggest that central adiposity may be an important contributor to the impaired exercise tolerance and physical function in older, obese HFpEF patients.

In addition to multiple novel findings, our results confirm, in a much larger group of subjects, our prior reports that increased thigh IMF area and IMF/SM ratio are significantly related to the severely reduced peak VO₂ in older obese HFpEF patients,(3,4) suggesting that abnormal skeletal muscle composition and its adverse consequences contribute to exercise intolerance in HFpEF.(34,35) We extend these findings by demonstrating that increased thigh SCF, IMF, IMF/TC ratio, and thigh IMF/SM ratio were associated with other objective measures of impaired physical function, including lower VAT, SPPB, 6MWD, and leg power. The mechanisms whereby increased thigh IMF is associated with increased anaerobic metabolism during submaximal exercise may be related to the deleterious effect excess adiposity has on skeletal muscle mitochondrial function. Specifically, Bharadwaj et al. reported that thigh fat volume, thigh SCF and thigh IMF were negatively associated with skeletal muscle citrate synthase activity in sedentary older individuals.(36) Moreover, we previously reported that older HFpEF patients with obesity have decreased skeletal muscle oxidative fibers, enzymes, capillarity and mitochondrial function that result in decreased diffusive O₂ transport and impaired peak and reserve oxygen utilization by the active muscles,(4,7–9,22,37) confirming findings reported in an animal model of HFpEF.(37)

In contrast to our pre-specified hypothesis, we found that epicardial fat, even when adjusted for body surface area or for total body fat, was substantially (33%) lower in HFpEF versus HC. However, these results are consistent with multiple prior studies in patients with HF with reduced EF (HFrEF) which showed significantly reduced epicardial fat.(38,39) They differ, though, from Obokata et al. who reported finding greater epicardial fat thickness in non-obese and obese HFpEF participants compared to controls.(15) The discrepancy between studies may be due to the techniques used to quantify epicardial fat. Obokata used echocardiography to measure one-dimensional thickness of epicardial fat at a single point adjacent to the right ventricular free wall. We used MRI to acquire volumetric measures of epicardial and paracardial fat around the entire heart.(40) Adipose tissue provides an extremely high signal-to-noise ratio for MRI whereas it tends to scatter ultrasound waves, degrading signal-to-noise for echocardiography. Furthermore, echocardiography overestimates epicardial fat and underestimates paracardial fat compared to MRI(41), likely due to difficulty in distinguishing the pericardium. Finally, volumetric measurements by MRI are not only more accurate and reproducible than by echocardiography, they preclude systematic errors from measuring epicardial fat thickness at a single point that are due to individual differences in epicardial fat distribution, such as occurs in HF patients from redistribution of epicardial fat due to LV remodeling.(40,42)

While our study does not address the potential mechanism(s) for reduced epicardial fat observed in obese HFpEF, prior studies in HFrEF have highlighted that epicardial fat plays a major role in regulating fatty acids to cardiac muscle as a local energy source, and that the heart can switch fuel sources in response to stress.(43) For example, in the setting of obesity and diabetes, myocardial fatty acid β-oxidation increases at the expense of glucose utilization.(44) Given the high prevalence of impaired glucose tolerance in metabolic/obese HFpEF, it seems possible that increased reliance on fatty acid β-oxidation may, at least partly, explain the observed reduction in epicardial fat in HFpEF (and in prior studies of HFrEF) compared to HC’s.

In contrast to abdominal and thigh fat which were inversely related to physical function, epicardial adipose was directly associated with all physical function measures, and remained a significant, independent, positive predictor of physical function even after adjusting for total body fat. This suggests that, unlike abdominal and thigh fat, epicardial adipose may not adversely impact physical function in HFpEF. This would be consistent with data regarding epicardial fat and myocardial energy utilization discussed above.

The finding of a significant association between intra-abdominal fat and epicardial fat in the HFpEF patients but not in the HCs further supports the validity of our key finding that obese HFpEF may have a unique pattern of regional fat distribution.

Clinical implications

Reducing total body adipose via bariatric surgery can prevent incident HF and improve exercise intolerance in established HFrEF.(12,13) Our data suggest that in obese/metabolic HFpEF, in addition to reducing total fat mass, additional benefit may be realized by targeting specific fat depots, such as intra-abdominal and intermuscular adipose. Indeed, dietary weight loss significantly improved symptoms, exercise capacity, and quality-of-life in older obese HFpEF patients, and these benefits were related to improved body and regional composition, including reduced fat mass and improved thigh muscle mass/intermuscular fat ratio.

Strengths and Limitations

Our study has several strengths, including prospective design, relatively large numbers of HFpEF and HCs, comprehensive measures of body composition by both DEXA and MRI, and multiple objective measures of physical function. Our study also has potential limitations. Although DEXA can potentially overestimate fat and lean mass in presence of edema, our HFpEF patients were well-compensated without overt fluid overload. Moreover, our MRI assessments confirm and extend our DEXA results by demonstrating regional differences in fat distribution between HFpEF and HC. Average BMI in the HC was lower than HFpEF, and even though we adjusted for body size, our data do not definitively prove that the pattern of regional fat distribution we identified is specific to obese HFpEF, which would require a control group matched for BMI and comorbidities.

Conclusions

In patients with the metabolic/obese HFpEF phenotype, in addition to total body adipose, the pattern of regional adipose deposition may have important adverse consequences for HFpEF and its primary manifestation, exercise intolerance. Interventions targeting specific adipose depots, such as intra-abdominal and thigh intermuscular fat, have the potential to improve the clinically important outcome of exercise intolerance.

Abbreviations List

6MWD

six minute walk distance

DEXA

dual energy x-ray absorptiometry

EF

ejection fraction

HC

age-matched healthy control

HF

heart failure

HFpEF

heart failure with preserved ejection fraction

HFrEF

heart failure with reduced ejection fraction

IMF

intermuscular fat

LV

left ventricle

MRI

magnetic resonance imaging

NYHA

New York Heart Association

SCF

subcutaneous fat

SM

skeletal muscle

SPPB

short physical performance battery

TC

thigh compartment

VAT

ventilatory anaerobic threshold

VO₂

oxygen consumption