Effect of Aerobic and Resistance Exercise on Cardiac Adipose Tissues: Secondary Analyses From a Randomized Clinical Trial

Key Points

Question  Are epicardial and pericardial adipose tissues regulated by both endurance and resistance training?

Findings  In this secondary analysis of a randomized clinical trial including 50 individuals with abdominal obesity, endurance and resistance training significantly reduced epicardial adipose tissue mass by 32% and 24%, respectively. While resistance training reduced pericardial adipose tissue mass by 32%, there was no effect of endurance training on pericardial adipose tissue.

Meaning  Endurance and resistance training have the potential to reduce cardiac adipose tissue mass and may have clinical potential given that excessive cardiac adipose tissue is associated with an increased incidence of cardiovascular disease.

Importance  Epicardial and pericardial adipose tissues are emerging as important risk factors for cardiovascular disease, and there is a growing interest in discovering strategies to reduce the accumulation of fat in these depots.

Objective  To investigate whether a 12-week endurance or resistance training intervention regulates epicardial and pericardial adipose tissue mass.

Design, Setting, and Participants  Secondary analysis of a randomized, assessor-blinded clinical trial initiated on August 2016 and completed April 2018. This single-center, community-based study included 50 physically inactive participants with abdominal obesity.

Interventions  Participants were randomized to a supervised high-intensity interval endurance training (3 times a week for 45 minutes), resistance training (3 times a week for 45 minutes), or no exercise (control group).

Main Outcomes and Measures  Change in epicardial and pericardial adipose tissue mass assessed by magnetic resonance imaging, based on a prespecified secondary analysis plan including 3 of 5 parallel groups.

Results  Of 50 participants (mean [SD] age, 41 [14] years, 10 men [26%]; mean [SD] body mass index [calculated as weight in kilograms divided by height in meters squared], 32 [5]), 39 [78%] completed the study. Endurance training and resistance training reduced epicardial adipose tissue mass by 32% (95% CI, 10%-53%) and 24% (95% CI, 1%-46%), respectively, compared with the no exercise control group (56% [95% CI, 24%-88%]; P = .001 and 48% [95% CI, 15%-81%]; P < .001, respectively). While there was a nonsignificant reduction in pericardial adipose tissue mass after endurance training (11% [95% CI, −5% to 27%]; P = .17), resistance training significantly reduced pericardial adipose tissue mass by 31% (95% CI, 16%-47%; P < .001) when compared with the no exercise control group. Compared with the no exercise control group, there was an increase in left ventricular mass by endurance (20 g [95% CI, 11%-30%]; P < .001) and resistance training (18 g [95% CI, 8%-28%]; P < .001). Other cardiometabolic outcomes remained unchanged after the 12-week trial period.

Conclusions and Relevance  In individuals with abdominal obesity, both endurance and resistance training reduced epicardial adipose tissue mass, while only resistance training reduced pericardial adipose tissue mass. These data highlight the potential preventive importance of different exercise modalities as means to reduce cardiac fat in individuals with abdominal obesity.

Trial Registration  ClinicalTrials.gov identifier: NCT02901496

Cardiovascular disease is the leading cause of death globally.¹ Among the risk factors for developing cardiovascular disease is obesity,^2,3 in particular obesity associated with visceral adipose tissue accumulation.⁴ As follows, excessive amounts of cardiac adipose tissue is reported to be associated with cardiovascular disease.^5-9

Cardiac adipose tissue consists of 2 distinct depots, the epicardial depot located immediately next to the myocardium and the pericardial and outermost depot located within and on the pericardium.⁴ Epicardial adipose tissue is proposed to play a cardioprotective role owing to its ability to take up and metabolize large amounts of fatty acids and thereby preventing the formation of atherosclerotic plaques.^10,11 However, increased release of inflammatory adipokines accompanying a pathophysiological expansion of epicardial adipose tissue is proposed to provoke atherosclerosis.⁵ The pericardial adipose depot is less well studied and although its effects on myocardial function are not fully clarified,¹² pericardial adipose tissue is exclusively associated with cardiovascular risk factors, coronary calcification, and incident of coronary heart disease.^7,12,13

Pharmaceutical drugs, such as the glucagon-like peptide 1 analogs,¹⁴ sodium-glucose transporter 2 inhibitors,^15,16 and lipid-lowering drugs,¹⁷ successfully reduce epicardial adipose tissue mass and the risk of developing cardiovascular disease.¹⁰ Also, bariatric surgery promotes loss of epicardial adipose tissue.^18,19 In contrast to these invasive approaches, exercise training may serve as a noninvasive strategy to reduce excessive cardiac adipose tissue. Until now, few studies have investigated the effect of exercise on epicardial and pericardial adipose tissue mass and the existing literature is inconsistent, to our knowledge.^20-24 For instance, it is unclear whether exercise training, both endurance and resistance training, facilitates changes in both epicardial and pericardial adipose tissue.

Based on the evidence that exercise targets visceral adipose tissue and lowers cardiovascular risk,^3,25-27 we set out to investigate the effects of a 12-week endurance exercise and resistance exercise intervention on epicardial and pericardial adipose tissue mass in physically inactive individuals with abdominal obesity. We hypothesized that both endurance and resistance training would reduce epicardial and pericardial fat compared with an inactive control group. This study is explorative as the analysis is based on data from a larger study exploring a different objective.^26,28

This study was part of a larger randomized exercise intervention clinical trial^26,28 performed between August 2016 to April 2018 at the Centre for Physical Activity Research in Copenhagen, Denmark. The study was approved by the ethical committee of the Capital Region of Denmark (H-16018062) and reported to the Danish Data Protection Agency (2012-58-0004). The study was conducted in accordance with the Declaration of Helsinki.²⁹ All participants provided written informed consent.

The complete study protocol has been previously published.²⁸ Briefly, participants were eligible if they were inactive and had abdominal obesity (waist to height ratio was ≥0.5 and/or waist circumference was ≥88 cm for women and ≥102 cm for men) and older than 18 years. Exclusion criteria were ischemic heart disease, diabetes, atrial fibrillation, pregnancy, treatment with immunotherapy or other biological rheumatic drugs, and health conditions that prevented magnetic resonance imaging (MRI) scans and exercise to be performed. Eligible participants were block randomized (1:1:1:1:1) into 5 groups. The randomized sequence was generated by 1 of us (K.K.) using computer-generated block randomization and concealed from all other researchers. To maintain blinding, the randomized sequence number was delivered concealed to a training instructor. The training instructor informed the participant and if assigned to the training intervention, the instructor organized the subsequent training sessions with the participant. The principal investigators thus remained blinded to the training modality. The 5 randomized groups were (1) no exercise plus placebo, (2) no exercise plus tocilizumab, (3) exercise plus placebo, (4) exercise plus tocilizumab, or (5) resistance exercise plus placebo. As described in the protocol,²⁸ participants allocated to no exercise plus placebo, exercise plus placebo, and resistance training plus placebo groups were included in this secondary analysis of this particular study. The no exercise plus placebo and exercise plus placebo groups were also included in the primary study.²⁶

To determine the maximal oxygen consumption rate (Vo₂ max), a graded exercise test on a bicycle ergometer was performed at baseline and after the 12-week intervention as previously described.²⁸ At baseline, 2 Vo₂ max tests were performed to allow familiarization. A 1–repetition maximum (RM) test was used to assess muscle strength during knee extensions (right and left leg separately) and chest press.

Body weight, height, and body mass index (calculated as weight in kilograms divided by height in meters squared) were determined as previously described.²⁶ Dual-energy x-ray absorptiometry (version 8.8; Lunar Prodigy, GE Medical Systems) was performed to evaluate changes in body composition, including total fat mass, android and gynoid fat mass, and lean body mass. Fasting blood samples were analyzed for cardiometabolic and inflammatory markers as previously described.²⁶

The MRI scan was performed on a 1.5-T whole-body MRI scanner (Ingenia; Philips). The cardiac MRI was supervised by a senior cardiologist (J.B.R.).²⁸ Steady-state free precession cine images were obtained during repeated breath holds in 2 long axes and in a stack of short axes covering the left ventricle (LV) to rule out wall motion abnormalities and allow for cardiac chamber and mass quantification. Sagittal, axial, and horizontal planes were obtained from 5 retrospective cardiac cycles with a slice thickness of 8 mm (interslice gap 2 mm for the short axis stack), with temporal resolution 30 to 50 milliseconds depending on heart rate (25 acquired phases). The number of short-axis cines acquired was participant specific, depending on ventricular size using retrospective gating in all participants when possible.

We also performed short-TI Inversion Recovery sequences, which are used to null the signal from fat to distinguish water (pericardial effusion) from cardiac fat. These images were compared with the cine images to rule out water. Semiautomatic calculations using cardiovascular imaging software cvi42, version 5.2, were applied for postprocessing using endocardial and epicardial contours to calculate LV ejection fraction, stroke volume, end-diastolic volume, end-systolic volume, and myocardial mass. Epicardial and pericardial adipose tissue volumes were quantified from cine images in the end-diastolic short axis by drawing contours around the epicardial and pericardial fat layers of the entire right ventricle and LV on each slice. Fat areas were generated automatically by the cvi42 software. Each slice area was multiplied by the slice thickness to yield a volume. The total mass of epicardial and pericardial adipose tissue was obtained by summation of all slice volumes and conversion to mass by the specific weight of fat (0.92 g/cm³).³⁰

The MRI scans were analyzed by the same investigator in a blinded manner. A cardiologist (J.B.R.) reanalyzed a subset of the cardiac MRI slices (n = 10) to assess accuracy and interreader variability. The degree of agreement between the 2 assessors (intraclass correlation) was found to be 0.87 (95% CI, 0.30-0.97) for epicardial adipose tissue and 0.98 (95% CI, 0.86-0.99) for pericardial adipose tissue.³¹

Educated personnel supervised all exercise sessions. Participants performed 3 weekly training sessions of 45 minutes during a 12-week period.²⁸ The endurance exercise was high-intensive interval exercise performed on an ergometer bicycle. The resistance exercise was designed as a 45-minute interval-type, medium-load, high-repetition, time-based training. Participants performed 3 to 5 sets of 10 exercises. The resistance exercise load was 60% of 1 RM and increased throughout the 12 weeks to 80%.²⁸

Participants were instructed to maintain their habitual lifestyle during the study. On a monthly basis, self-reported 3-day dietary intake was recorded and during a 4-day period free-living physical activity levels were monitored using accelerometry (AX3; sampling frequency 100 Hz). Postprocessing was conducted as described previously.^26,32

The study was designed to focus on the per-protocol population, where participants were included in the analysis if they completed the trial with a satisfactory training adherence. Completing a minimum of 29 of 36 training sessions (80%) was satisfactory.

A prespecified sample size of 14 individuals in each group was obtained from a power calculation based on the primary study outcome.^26,28 Because the training effects on cardiac adipose tissue were exploratory, no power calculation was performed related to this outcome.

Statistical analyses were performed using SAS software, version 9.3 (SAS Institute) and GraphPad Prism, version 7.02 (GraphPad Software). A 2-sided P < .05 was considered statistically significant and reported with 2-sided 95% CIs, with no default adjustment for multiplicity. Baseline characteristics with normal or nonnormal distribution were reported as mean and SD or median and interquartile range. Postintervention levels of continuous variables adjusted for baseline levels were compared using analysis of covariance performed with postmeasure or absolute or relative change from baseline measure as the dependent outcome, with a fixed factor for group (3 levels) and using the baseline value as a covariate; for each continuous outcome group, means were adjusted for baseline (least squares means). Independent of the result of the overall analysis of covariance model, contrasts between groups were performed. Intervention-induced changes in epicardial and pericardial adipose within groups was evaluated by Wilcoxon matched pairs signed-rank test.

Of the 50 randomized participants, 11 participants discontinued or were excluded from the study. Eight of them discontinued without providing a reason (n = 6) or owing to lack of time (n = 2). Three participants were excluded from the analysis. The reasons for exclusion were initiation of medical anticontraceptive/anti-inflammatory treatment during the study (n = 2) and the lack of cardiac MRI follow-up scan (n = 1) (Figure 1). The final study population comprised 39 participants distributed across 3 groups: no exercise control (n = 12), endurance training (n = 14), and resistance training (n = 13) (Figure 1). No individuals were lost to follow-up. Baseline characteristics appeared to be similar between groups (Table 1).

Changes in epicardial adipose tissue mass differed across the groups after the training intervention (P = .005) (Table 2 and eFigure in the Supplement). Whereas epicardial adipose tissue mass remained unchanged in the no exercise control group (2 g [95% CI, −2 to 6]), endurance training reduced epicardial adipose tissue mass by 8 g (95% CI, 4-11) (Figure 2A). Comparing the changes between groups showed that endurance exercise reduced epicardial adipose tissue more than the no exercise control group (9 g [95% CI, 4-15], P = .002) (Table 2). The resistance training group also reduced epicardial adipose tissue mass more than the control group (8 g [95% CI, 2-13], P = .009) (Table 2). The effect of resistance vs endurance training on epicardial adipose tissue mass were not different (1 g [95% CI, −4 to 7], P = .58).

Changes in the pericardial adipose tissue mass differed between groups (P = .005) (Table 2). Endurance training reduced pericardial adipose tissue by 15 g (95% CI, 1-30) (Figure 2B), but the reduction did not differ significantly from the no exercise control group (6 g [95% CI, −15 to 27], P = .58) (Table 2). In contrast and compared with the no exercise control group, resistance training reduced pericardial adipose tissue mass by 34 g (95% CI, 12-55; P = .003) (Table 2). Moreover, comparing the changes in pericardial adipose tissue induced by endurance vs resistance training revealed a more pronounced effect of resistance training (28 g [95% CI, 7-49], P = .01). Other anthropometric measures changed nonsignificantly in response to training (Table 2). An intention-to-treat analysis using baseline carried forward imputation showed that the per-protocol findings were robust (eTable 1 in the Supplement).

Changes in Vo₂ max differed between groups (P = .003) (Table 2); Vo₂ max increased 219 mL/min (95% CI, 104-334) in the endurance training group and remained unchanged in the no exercise control group (−92 mL/min [95% CI, −229 to 44]). Comparing the changes between these 2 groups showed a significant improvement in the Vo₂ max after endurance training vs no exercise (312 mL/min [95% CI, 134-489], P = .001). Vo₂ max also increased in response to resistance training compared with the no exercise control group (271 mL/min [95% CI, 84-460], P = .006). Improvements in Vo₂ max did not differ between the endurance and resistance training groups (40 mL/min [95% CI, −132 to 212], P = .64).

Training modality affected muscle strength (Table 2). There was no change in the 1-RM chest press in the no training control group (1 kg [95% CI, −4 to 3]); in contrast, resistance training increased 1-RM chest press by 12 kg (95% CI, 8-16), which was significantly different from the no training control group (11 kg [95% CI, 6-17], P < .001) (Table 2). One-RM chest press was unaffected by endurance training (−1 kg [95% CI, −4 to 3]) and not different from the no training control group (−1 kg [95% CI, −7 to 4], P = .64). Changes in left and right quadriceps strength (knee extension) in the 2 training groups were comparable with those observed during chest press (Table 2).

All participants had normal cardiac function at baseline with a mean (SD) LV ejection fraction of 69% (7.2%). While measures of systolic function and LV volumes did not change between the 3 groups after the intervention, LV mass changed across groups (P < .001) (Table 3). Compared with the no training control, there was an increase in LV mass by both endurance (20 g [95% CI, 11-30], P < .001) and resistance training (18 g [95% CI, 8-28], P < .001). There was no difference in the effect of resistance vs endurance training on LV mass increase (2 g [95% CI, −7 to 12], P = .63).

There were no changes between the groups in systolic blood pressure, heart rate (Table 3), or plasma levels of total cholesterol, low-density lipoprotein and high-density lipoprotein cholesterol, triglyceride level, glucose, hemoglobin A_1c, C-reactive protein, tumor necrosis factor α, interleukins 6, 1β, 8, and 10, and adiponectin after the intervention (eTable 2 in the Supplement).

The compliance to training was comparable between the 2 groups: 89% vs 93% in the endurance and resistance training groups, respectively (P = .07) (eTable 3 in the Supplement). Average energy intake or free-living physical activity, assessed monthly, did not differ between groups nor did it change during the intervention (eTable 4 in the Supplement).

This study showed that endurance and resistance training reduced epicardial adipose tissue mass in physically inactive people with abdominal obesity and moreover that pericardial adipose tissue mass was only reduced by resistance training. Because epicardial and pericardial adipose tissues are emerging as risk factors for cardiovascular disease, there is a growing interest in preventive strategies to reduce fat accumulation in these depots.¹⁰ Here we demonstrate that 2 training modalities are capable of reducing cardiac adipose tissue mass. Most studies^20,21 but not all³³ report reduced epicardial adipose tissue after endurance training. Our study contributes by demonstrating that not only endurance, but also resistance training, can reduce epicardial adipose tissue.

Three months of resistance and endurance training were associated with reduced epicardial adipose tissue by 24% and 32%, respectively. These effects represent the isolated effect of training because food intake remained unchanged. Dietary restrictions are reported to reduce epicardial adipose tissue by 32%,¹⁹ so it is possible that combining training and dietary restriction may have a greater effect on epicardial adipose tissue mass than exercise alone. Furthermore, in comparison, treatment with sodium-glucose transporter 2 inhibitors for 6 months¹⁶ and treatment with glucagon-like peptide 1 analogs for 3 months¹⁴ resulted in 14% and 29% reductions in epicardial adipose tissue, respectively. Considering that pharmacologic treatment can be associated with adverse effects and is associated with increased economic costs compared with training-based treatment strategies, our study highlights the importance of exercise training as a means to reduce cardiac adipose tissues.

Although pericardial adipose tissue has been suggested to be a stronger cardiovascular risk marker than epicardial adipose tissue,¹² the effect of exercise training on pericardial adipose tissue has largely been neglected. Our study showed that pericardial adipose tissue mass was reduced by resistance training but not endurance training. Overall, our results suggest that resistance training may be superior to endurance training as resistance training reduced pericardial adipose tissue and improved fitness and strength, while endurance training only improved fitness. Nevertheless, both exercise modalities were associated with reduced epicardial adipose tissue, suggesting that people with specific training preferences or requirements can benefit from both training modalities.

Both training modalities increased LV mass. Exercise is known to stimulate cardiac remodeling including mild cardiac hypertrophy; hence, these results were expected.^34,35 Given the healthy study population without heart disease and the relative short training duration (12 weeks), we did not expect to find alterations in systolic cardiac function. Previous studies primarily report increased diastolic function or a small increase in stroke volume after 6 months of exercise training.^36,37 Moreover, previous studies have shown that antidiabetic drugs, such as glucagon-like peptide 1 analogs¹⁴ and sodium-glucose transporter 2 inhibitors,¹⁶ reduced epicardial adipose tissue volume and nonfatal and fatal cardiac events.³⁸ Whether the reduction in epicardial adipose tissue directly improves cardiac function and protects against cardiovascular disease, or is merely a biomarker, remains unclear. Further studies are needed to clarify whether training interventions can improve cardiac function via a reduction in epicardial and pericardial adipose tissue mass in a cardiovascular disease population.

Given that excessive epicardial and pericardial adipose tissue are risk factors of cardiovascular disease,^6,7 our results may have clinical potential as a means to reduce cardiac adipose tissue mass. The beneficial cardiovascular-lowering risk of current antidiabetic therapeutics has been hypothesized to be in part driven by a reduction in epicardial and pericardial volume.^14,16,39,40 Hence, this study shows that both endurance and resistance training may have the potential as preventive strategies.

Strengths of this study include the randomized, placebo-controlled clinical study design, the fact that every exercise session was supervised, and that MRI was used to determine epicardial and pericardial adipose tissue mass. Epicardial and pericardial adipose tissues covering the ventricles and not the entire heart were measured, which may underestimate the true amount of these fat depots. The study followed a per-protocol design and replaced dropouts, and this may introduce a bias (attrition bias) and limit the conclusion to only apply to people who performed exactly the prescribed amount of exercise. The small group sizes, which fell below the numbers prespecified by the power calculation, limit the interpretations of the findings and introduce a risk of type 2 errors. However, an intention-to-treat analysis using baseline carried forward imputation showed that the per-protocol findings were robust (eTable 1 in the Supplement). Importantly, the training effect on epicardial and pericardial adipose tissue was not the primary objective of the main study^26,28 rendering this analysis exploratory, and thus the findings in this study need replication. We did not combine resistance and endurance training, which would have been interesting to reveal their potential additive effects.

In this secondary analysis of a 12-week randomized clinical trial, including physically inactive people with abdominal obesity, supervised high-intensity interval endurance training and resistance training reduced epicardial adipose tissue, while only resistance training reduced pericardial adipose tissue mass.

Corresponding Author: Regitse Højgaard Christensen, MD, Rigshospitalet, Department 7642, Blegdamsvej 9, 2100 Copenhagen Ø, Denmark (regitse.hoejgaard.christensen@regionh.dk).

Accepted for Publication: May 1, 2019.

Published Online: July 3, 2019. doi:10.1001/jamacardio.2019.2074

Correction: This article was corrected on August 7, 2019, to fix the corresponding author’s email address.

Author Contributions: Dr R.H. Christensen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs R.H. Christensen, Wedell-Neergaard, and Lehrskov contributed equally.

Concept and design: R.H. Christensen, Wedell-Neergaard, Lehrskov, Legaard, Dorph, Nymand, Karstoft, Krogh-Madsen, Rosenmeier, Pedersen, Ellingsgaard.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: R.H. Christensen.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: R.H. Christensen, Larsen, R. Christensen.

Obtained funding: R.H. Christensen, Pedersen.

Administrative, technical, or material support: R.H. Christensen, Wedell-Neergaard, Lehrskov, Legaard, Dorph, Larsen, Fagerlind, Seide, Nymand, Ried-Larsen, Krogh-Madsen.

Supervision: Karstoft, Krogh-Madsen, Rosenmeier, Pedersen, Ellingsgaard.

Conflict of Interest Disclosures: Dr Ried-Larsen reports personal fees from Novo Nordisk outside the submitted work. No other disclosures were reported.

Funding/Support: This study was funded by TrygFonden. The Centre for Physical Activity Research is supported by a grant from TrygFonden. The Centre of Inflammation and Metabolism is a member of the Danish Center for Strategic Research in Type 2 Diabetes, which is funded by the Danish Council for Strategic Research (grants 09-067009 and 09-075724). Salary to Dr R.H. Christensen was financed by a grant from the Danish Heart foundation (grant 16-R107-A6704-22970). Salary to Dr Lehrskov was partly financed by a grant from Danish Diabetes Academy, which is supported by the Novo Nordisk Foundation. Running costs were financially supported by Aase and Ejnar Danielsen Foundation (grant 10-001271). Dr R. Christensen is employed at the Parker Institute at Bispebjerg and Frederiksberg Hospital, which is supported by a core grant from the Oak Foundation (grant OCAY-13-309).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.