Skeletal muscle adiposity, coronary microvascular dysfunction, and adverse cardiovascular outcomes

Abstract

Background and Aims

Skeletal muscle (SM) fat infiltration, or intermuscular adipose tissue (IMAT), reflects muscle quality and is associated with inflammation, a key determinant in cardiometabolic disease. Coronary flow reserve (CFR), a marker of coronary microvascular dysfunction (CMD), is independently associated with body mass index (BMI), inflammation and risk of heart failure, myocardial infarction, and death. The relationship between SM quality, CMD, and cardiovascular outcomes is not known.

Methods

Consecutive patients (n = 669) undergoing evaluation for coronary artery disease with cardiac stress positron emission tomography demonstrating normal perfusion and preserved left ventricular ejection fraction were followed over a median of 6 years for major adverse cardiovascular events (MACEs), including death and hospitalization for myocardial infarction or heart failure. Coronary flow reserve was calculated as stress/rest myocardial blood flow. Subcutaneous adipose tissue (SAT), SM, and IMAT areas (cm²) were obtained from simultaneous positron emission tomography attenuation correction computed tomography using semi-automated segmentation at the 12th thoracic vertebra level.

Results

Median age was 63 years, 70% were female, and 46% were nonwhite. Nearly half of patients were obese (46%, BMI 30–61 kg/m²), and BMI correlated highly with SAT and IMAT (r = .84 and r = .71, respectively, P < .001) and moderately with SM (r = .52, P < .001). Decreased SM and increased IMAT, but not BMI or SAT, remained independently associated with decreased CFR (adjusted P = .03 and P = .04, respectively). In adjusted analyses, both lower CFR and higher IMAT were associated with increased MACE [hazard ratio 1.78 (95% confidence interval 1.23–2.58) per −1 U CFR and 1.53 (1.30–1.80) per +10 cm² IMAT, adjusted P = .002 and P < .0001, respectively], while higher SM and SAT were protective [hazard ratio .89 (.81–.97) per +10 cm² SM and .94 (.91–.98) per +10 cm² SAT, adjusted P = .01 and .003, respectively]. Every 1% increase in fatty muscle fraction [IMAT/(SM + IMAT)] conferred an independent 2% increased odds of CMD [CFR <2, odds ratio 1.02 (1.01–1.04), adjusted P = .04] and a 7% increased risk of MACE [hazard ratio 1.07 (1.04–1.09), adjusted P < .001]. There was a significant interaction between CFR and IMAT, not BMI, such that patients with both CMD and fatty muscle demonstrated highest MACE risk (adjusted P = .02).

Conclusions

Increased intermuscular fat is associated with CMD and adverse cardiovascular outcomes independently of BMI and conventional risk factors. The presence of CMD and SM fat infiltration identified a novel at-risk cardiometabolic phenotype.

Structured Graphical Abstract

Intermuscular adiposity, a novel ectopic fat depot, is associated with coronary microvascular dysfunction independently of body mass index and other conventional risk factors, and modifies its effect on adverse cardiovascular outcomes in patients with cardiometabolic disease. BMI, body mass index; CMD, coronary microvascular dysfunction; CT, computed tomography; INOCA, ischaemia and no obstructive coronary artery disease; MACE, major adverse cardiovascular event; PET, positron emission tomography.

Open in new tabDownload slide

See the editorial comment for this article ‘Skeletal muscle adiposity in patients with impaired coronary flow reserve: risk marker, treatment target, or bystander?’, by R. de Silva and K. Cheng, https://doi.org/10.1093/eurheartj/ehae909.

Introduction

Overweight or obesity is prevalent in over 71% of US adults¹ and has become one of the most important threats to public health worldwide. Compared to individuals with normal weight, those with obesity, defined as a body mass index (BMI) ≥ 30 kg/m², experience cardiovascular disease (CVD) events at an earlier age and have a shorter average life span.^2,3 Excess adiposity accelerates atherosclerosis and promotes adverse changes in cardiac structure and function through deleterious effects on the myocardium as well as the vasculature, and through obesity-related comorbidities, including hypertension, dyslipidaemia, and type 2 diabetes mellitus.^4,5 Although increasing BMI is associated with increasing risk of morbidity and mortality across populations, CVD risk is not uniform for individuals with similar BMI and can vary substantially across sex and racial/ethnic groups. Nonetheless, BMI thresholds continue to guide current clinical diagnosis of obesity and candidacy for interventions with potential to improve CVD outcomes.⁶ Besides BMI, other discriminators of cardiovascular risk are needed in individuals at risk for cardiometabolic disease. Imaging technologies such as computed tomography (CT) can be used to assess body composition and distinguish between fat and lean mass directly in vivo via their respective radiodensities, or attenuation, within anatomical compartments.⁷

Recently, intermuscular adipose tissue (IMAT) has emerged as a distinct adipose depot reflecting skeletal muscle (SM) fat infiltration with unique and evolving biological properties. Whereas fatty ‘marbling’ of meat is commercially valued in livestock, IMAT in humans has been associated with insulin resistance and type 2 diabetes.^8,9 Intermuscular adipose tissue can be found in most SM groups, and while IMAT increases with BMI, it can vary considerably between individuals. Early reports suggest that IMAT has a proinflammatory secretome with increased expression of interleukin-6 and tumour necrosis factor, which may affect the metabolic function and insulin sensitivity of neighbouring muscle tissue,^10–12 but the impact of IMAT on CVD events is not well understood.

Coronary microvascular dysfunction (CMD), quantified noninvasively using positron emission tomography (PET) as an impaired global coronary flow reserve (CFR < 2) with normal myocardial perfusion imaging, is independently associated with elevated BMI¹³ and future risk of heart failure (HF), myocardial infarction (MI), and death.^14–19 It is also associated with residual inflammation and myocardial stiffness independently of conventional CVD risk factors in patients with cardiometabolic disease.^20–23 We previously demonstrated an independent inverted J-shaped relationship between BMI and CFR such that in obese patients, CFR decreased linearly with increasing BMI (adjusted P < .0001).¹³ We found that CMD was prevalent in obese patients, worsened with increasing BMI, and was a better discriminator of CVD risk as compared to BMI. Given the limitations of BMI, we sought to investigate the relationship between IMAT, CMD, and cardiovascular outcomes. We hypothesized that measures of both SM quantity and quality are associated with CMD and modify its effect on CVD events independently of obesity.

Methods

Study population

The study population (see Supplementary data online, Figure S1) included consecutive patients undergoing cardiac stress testing with PET/CT at Brigham and Women’s Hospital (Boston, MA) from 2007 to 2014. The most common indication for PET was the evaluation of chest pain, dyspnoea, or their combination. Patients with known coronary artery disease (CAD), including a history of previous MI or revascularization, clinical HF, severe valvular disease, PET evidence of flow-limiting CAD (defined as a summed stress score >2), or abnormal left ventricular ejection fraction (LVEF < 40%) were excluded, as were patients with end-stage disease, including liver, lung, or kidney disease, active malignancy, or planned bariatric surgery at the time of imaging. Medical history, medication use, laboratory findings, height, and weight were ascertained at the time of PET/CT imaging. Body mass index was calculated as the ratio between weight in kilograms and the square of height in metres. The Chronic Kidney Disease Epidemiology Collaboration equation was used to determine estimated glomerular filtration rate (eGFR). The study was approved by the Mass General Brigham Healthcare Institutional Review Board and performed in accordance with institutional guidelines.

Positron emission tomography imaging

Patients were imaged with a whole-body PET/CT scanner (Discovery RX or STE LightSpeed 64, GE Healthcare, Milwaukee, WI) using ⁸²Rb and ¹³N-ammonia as flow radiotracers at rest and following pharmacological stress. Patients were instructed to avoid caffeine or methylxanthine-containing substances 24 h before the scan. Rest LVEF was obtained from gated myocardial perfusion images processed with commercially available software (Corridor4DM, INVIA Medical Imaging Solutions, Ann Arbor, MI). Semi-quantitative visual interpretation was performed using a standard 17-segment, 5-point scoring system to determine summed rest, stress, and difference scores, reflecting scar, ischaemia plus scar, or ischaemia, respectively. Absolute global myocardial blood flow (MBF, mL/min/g) was measured at rest, and peak vasodilator-induced hyperaemia and dynamic images were fitted into a validated two-compartment kinetic model as previously described.²⁴ Coronary flow reserve was obtained as the ratio of global stress to rest MBF.

Computed tomography-derived body composition analysis

Thoracic body composition metrics, including cross-sectional areas (in cm²) of subcutaneous adipose tissue (SAT), SM, and IMAT,²⁵ were measured at the level of the 12th vertebra (T12) on low-dose, non-contrast axial CT images obtained for attenuation correction during cardiac PET/CT imaging using a standard acquisition protocol (tube voltage 120–140 kV, tube current 20–40 mA, exposure time <2.0 ms, slice thickness 5.0 mm).^26,27 Studies with artefacts related to patient positioning or image acquisition precluding soft tissue assessment at this level were excluded (see Supplementary data online, Figure S1). Subcutaneous adipose tissue, SM, and IMAT were quantified using semi-automated threshold-based segmentation in 3DSlicer (version 4.10.1, https://www.slicer.org/) by a trained reader (A.S.T.) blinded to study details. Tissue-specific Hounsfield units (HU) were used for SM (−29 to +150 HU) and adipose tissue (−190 to −30 HU) as previously described.^27,28 Hepatic attenuation (HA, in HU) was obtained by manual tracing of a circular region of interest within the liver parenchyma at the T12 level. Boundaries of each compartment were verified with excellent intrareader class correlations (≥.997). Randomly selected samples were re-segmented by independent readers (J.P.M., F.J.F.) blinded to patient details with excellent interreader class correlations (≥.972). In addition, epicardial adipose tissue (EAT) volume (in cm³) was quantified using a validated automated deep learning segmentation tool²⁹ and confirmed manually by a blinded reader (B.F.) on a subset of 30 attenuation correction CTs with excellent interclass correlation (.97). All body composition analyses were blinded to clinical, PET, and outcomes data.

Outcomes

Patients were followed over a median 5.8 (Q1–Q3 3.2–7.1) years for the occurrence of a primary endpoint composite of death, hospitalization for nonfatal MI, or HF. Time to first event was analysed. Ascertainment of clinical endpoints was determined by blinded committee adjudication of the longitudinal medical record, the Mass General Brigham Healthcare Research Patient Data Registry, the National Death Index, mail surveys, and telephone calls. To be classified as hospitalization for nonfatal MI or HF, discharge with the corresponding primary diagnosis was required, and only events meeting the universal definition of MI³⁰ or prespecified criteria for clinical signs, symptoms, and escalation of therapy for HF, respectively, were adjudicated as such. All hospitalization events occurred >30 days following PET imaging.

Statistical analyses

Baseline characteristics are presented as rates with percentages for categorical variables and medians with interquartile ranges (Q1–Q3) for continuous variables. We used the Fisher exact test and the Wilcoxon rank-sum test to assess for differences in categorical and continuous baseline characteristics, respectively. Coronary microvascular dysfunction was defined as CFR <2, a prognostically important cutpoint associated with worse cardiovascular outcomes in patients undergoing evaluation for suspected CAD,^31,32 and approximated the median CFR of the clinical cohort. Spearman’s correlation was used to describe the association between the continuous variables of BMI and thoracic body composition metrics SAT, SM, EAT, HA, and IMAT. Fatty muscle fraction (FMF, %) was defined as IMAT/(SM + IMAT) × 100.

Linear regression was used to assess for independent relationships between CFR and body composition metrics. Candidate variables tested included demographic characteristics, medical history and medication use, laboratory findings and noninvasive imaging parameters, with the most clinically important covariates or significant univariable associations included in the multivariable model. To avoid overfitting, demographic and medical history variables (age, sex, anginal symptoms, hypertension, dyslipidaemia, diabetes, tobacco use, family history of premature CAD, BMI >27 kg/m², and oestrogen status) were incorporated into multivariable modelling using a validated pretest clinical risk score for diagnosing CAD (with values 0–8, 9–15, and 16–24 indicating low, intermediate, and high pretest risk, respectively) as previously described.³³ The final multivariable linear regression model for CFR included the pretest clinical score, nonwhite race, BMI, eGFR < 60 mL/min/1.73 m², LVEF, SAT, SM, and IMAT (or EAT, or HA or FMF, modelled separately due to high collinearity). These variables were also incorporated into a logistic regression model to quantify the association between FMF and CMD.

Cumulative event-free survival curves for the composite endpoint of death or hospitalization for nonfatal MI or HF were compared across categories of CMD and obesity or IMAT median using the log-rank test. Cox proportional hazards models were used to examine for multivariable-adjusted associations of thoracic body composition, CFR, and events. Univariate associations were tested, and sequential Cox models controlled for effects of clinically important covariates: Model 1 was adjusted for traditional risk factors, including pretest clinical score, race, eGFR, LVEF, and BMI, with sequential addition of body composition metrics and CFR to Models 2 and 3, respectively. Nested models were compared with the likelihood ratio test, and the Akaike information criterion was assessed to avoid overfitting. A linear interaction term for CFR and IMAT or FMF was tested for significance in the final adjusted model. The proportional hazards assumption was confirmed for each model using cumulative martingale residuals. In addition, to understand the relative independent contribution of IMAT compared to alternate thoracic ectopic fat depots on clinical outcomes, we sequentially added HA, EAT, and IMAT or FMF to separate Cox models (A, B, and C, respectively) controlling for the effects of pretest clinical score, LVEF, CFR, and BMI at baseline.

To further investigate the presence of effect modification between CMD and IMAT, Poisson regression was performed to compute adjusted annualized rates of events across categories of CMD and IMAT median after adjustment for pretest clinical score, SM and SAT areas. Model fit was assessed with the goodness-of-fit χ² test, with a nonsignificant result indicating adequate fit. A P-value of <.05 was considered to indicate statistical significance, and all tests were two sided. Statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., Cary, NC) and SPSS (version 28.0, IBM Corp, Armonk, NY).

Results

Baseline characteristics

The distribution of baseline characteristics is shown in Table 1. The median age of patients in the overall cohort was 62.6 (53.7–71.6) years, 69.8% were women, and 46.2% were nonwhite. Approximately 45.9% of patients were obese (BMI 30–61 kg/m²), with 21.7% of the cohort classified as having class I obesity (BMI 30–34.9 kg/m²), 13.1% as class II obesity (BMI 35–39.9 kg/m²), and 11.1% as class III obesity (BMI ≥40 kg/m²). Hypertension and dyslipidaemia were prevalent comorbidities present in 72.6% and 58.4% of patients, respectively, and 25.6% of patients had diabetes. Median LVEF was 63% (57–69) and 43.0% of patients had CMD, with median CFR of 2.1 (1.7–2.6) reflecting median peak stress and rest MBF values of 2.3 (1.8–2.9) and 1.1 (.8–1.4) mL/min/g, respectively. Compared with the nonobese, obese patients had higher median indices of SAT (248.7 vs. 101.5 cm²), SM (95.6 vs. 78.9 cm²), IMAT (20.4 vs. 6.5 cm², P < .001 for all), and similar distributions of CFR. They also had higher median EAT (91.7 vs. 60.9 cm³) and lower HA (51.5 vs. 62.3 HU) consistent with higher liver fat content (P < .001 for both). BMI correlated highly with SAT and IMAT (r = .84 and .71, respectively, P < .001) and moderately with SM (r = .52, P < .001) (Figure 1; Supplementary data online, Figure S2). Representative patients of similar demographics and BMI with variable thoracic body composition metrics and CFR values are shown in Figure 2.

Association between thoracic body composition and coronary flow reserve

In univariable linear regression analysis, increasing BMI was inversely associated with CFR [β (se) = −.07 (.03), P = .04 for +10 kg/m² BMI] as was increasing SAT, IMAT, and EAT, while increasing HA was directly associated with CFR (P < .01 for all) (Table 2). After adjustment for clinical covariates and body composition metrics, decreasing SM and increasing IMAT, but not BMI or SAT, remained independently associated with worse CFR [β (se) = .027 (.013) and −.047 (.023), adjusted P = .03 and .04 for +10 cm² SM and IMAT areas, respectively] (Table 2). Findings were confirmed using FMF [β (se) = −.008 (.004), adjusted P = .02 for +1% FMF]. In univariable- and multivariable-adjusted logistic regression analyses, every 1% increase in FMF was independently associated with a 2% increased odds of CMD [odds ratio (OR) 1.02 (1.01–1.04), unadjusted P = .008 and adjusted P = .04] defined as CFR < 2.

Thoracic body composition, coronary flow reserve, and adverse events

Over a median follow-up of 5.8 (3.2–7.1) years, 98 patients met the composite endpoint of death and hospitalization for MI or HF, including 54 deaths (see Supplementary data online, Table S1). In univariable analysis, CFR and IMAT were significantly associated with major adverse events [hazard ratio (HR) 1.81, 95% confidence interval (CI) 1.28–2.56 per −1 U CFR, P = .001 and HR 1.21, 95% CI 1.09–1.35 per +10 cm² IMAT, P < .001, respectively] (Table 3). The addition of body composition metrics SAT, SM, and IMAT into a multivariable model of clinically important covariates, including pretest clinical score, race, BMI, eGFR, and LVEF led to improvement in model statistics (global χ² 65.4 vs. 26.8 for Model 2 vs. Model 1, respectively, P < .001), as did the sequential addition of CFR (global χ² 74.5 vs. 65.4 for Model 3 vs. Model 2, respectively, P < .001) (Table 3). In the final adjusted model, both lower CFR and higher IMAT were associated with increased events [HR 1.78 (1.23–2.58) per −1 U CFR and 1.53 (1.30–1.80) per +10 cm² IMAT, adjusted P = .002 and P ≤ .0001, respectively], while higher SM and SAT were protective [HR .89 (.81–.97) per +10 cm² SM and .94 (.91–.98) per +10 cm² SAT, adjusted P = .01 and P = .003, respectively]. There was a significant interaction between CFR and IMAT such that patients with both CMD and fatty muscle demonstrated the highest risk of events (adjusted P for interaction = .02). Every 1% increase in FMF conferred an independent 7% increased risk of major adverse events [HR 1.07 (1.04–1.09), adjusted P < .001] and modified the effect of CMD on outcomes (adjusted P for interaction = .04).

Intermuscular adiposity compared with alternate thoracic ectopic fat depots

To understand the relative contribution of IMAT compared to alternate thoracic ectopic fat depots on clinical outcomes, we sequentially added HA, EAT, and IMAT or FMF to multivariable-adjusted models controlling for the effects of pretest clinical score, LVEF, CFR, and BMI at baseline. While EAT was associated with increased risk of events in unadjusted or limited adjusted analysis, this association was no longer significant after adjustment for IMAT and other body composition metrics in the presence of CFR (Table 4). In contrast, the highly significant association between IMAT (or FMF) and adverse outcomes was unchanged after further adjustment for other thoracic ectopic fat depots [HR 1.52 (1.27–1.83) per +10 cm² IMAT, adjusted P < .0001] (Table 4). Our results, including the significant interaction between IMAT (or FMF) and CFR, remained unchanged in expanded multivariable-adjusted modelling (see Supplementary data online, Table S2). These findings support IMAT as a robust marker of cardiovascular risk, even as compared to more conventional measures of visceral adiposity such as fatty liver and epicardial fat.

Obesity, intermuscular adiposity, coronary microvascular dysfunction, and outcomes

To visualize the impact of effect modification of IMAT on CMD and cardiovascular events, we stratified results by CMD and obesity (Figure 3A and B) or IMAT median (Figure 3C and D). In adjusted analysis, patients with CMD with or without obesity experienced the highest cumulative rate of events (adjusted P < .001, Figure 3B). Elevated BMI did not further stratify risk among patients with CMD (Figures 3B and  4A). In contrast, only those patients with both CMD and high IMAT experienced the highest rate of adverse events (adjusted P < .001, Figures 3D and 4B), with an adjusted annualized rate of events of 5.1% (adjusted P = .02). The presence of both SM fat infiltration and CMD identified a novel at-risk cardiometabolic phenotype independent of conventional measures of adiposity or CVD.

Discussion

In this work, we demonstrate a novel relationship between IMAT, CMD, and cardiovascular outcomes by leveraging advances in noninvasive imaging. First, we found that lower SM quantity and quality were associated with CMD, independently of conventional risk factors. Every 1% increase in FMF conferred an independent 2% increased odds of CMD. Second, in adjusted analyses for outcomes, we showed that both higher IMAT and lower CFR were associated with increased risk of CVD events, while higher SM and SAT were protective. Every 1% increase in FMF conferred a 7% increased risk of CVD events. Third, the presence of elevated IMAT, not BMI, modified the effect of CMD on outcomes such that patients with both CMD and fatty muscle, not obesity per se, demonstrated the highest risk of CVD events. Fourth, these findings were independent of other conventional thoracic ectopic fat depots and were notably informative for identifying cardiometabolic risk in patients without traditional visceral adiposity. Our data support that thoracic SM quantity and quality are tied to coronary microvascular function and together identify a novel at-risk cardiometabolic phenotype (Structured Graphical Abstract).

These findings may be especially relevant for the common clinical demographic represented here, namely patients with ischaemia and no obstructive CAD (INOCA).³⁴ Obesity-related CVD risk is not well captured by BMI, especially in women.³⁵ While increased thoracic IMAT was independently associated with worse CVD events, higher thoracic SM and, to a lesser extent, SAT demonstrated a protective effect. This underscores that, especially in INOCA patients, assessments of BMI and traditional adiposity—which largely reflect lower-risk SAT—are likely to be insufficient to appropriately stratify CVD risk. Our findings are clinically relevant and may help to better discern CVD risk among otherwise similar patients, especially younger, female individuals with fewer traditional risk factors and high prevalence of overweight or obesity. Figure 2A/C and B/D demonstrate thoracic body composition results in two middle-aged Hispanic women with class II obesity. Despite similar age and BMI, normal renal function, LVEF, and myocardial perfusion, patient (B/D) has increased IMAT and decreased CFR as compared to patient (A/C) and thus is at significantly increased risk of CVD events. These findings are relevant for ongoing studies investigating cardiovascular effects of fat and lean weight-modifying glucagon-like peptide-1 receptor agonist therapies. For example, it is possible that perceived lean muscle mass loss in response to incretin-based therapies may involve loss of IMAT resulting in improved skeletal muscle quality.³⁶

Recently described, IMAT reflects an ectopic adipose tissue depot that is interspersed among SM fibres residing within the boundary of the muscle fascia.⁹ Considered a marker of overall muscle quality, it has been associated with systemic inflammation, insulin resistance, the metabolic syndrome, and coronary artery calcification in a manner comparable to visceral adipose tissue.^12,37,38 We previously reported on the independent associations between residual inflammation, abnormal CFR, and impaired myocardial strain in a cohort of clinical trial patients with cardiometabolic disease and otherwise well-controlled CVD risk factors and preserved cardiac function.²⁰ We found that individuals with impaired CFR demonstrated the strongest association between markers of HF and inflammation, building on prior work identifying abnormal global CFR and CMD as robust markers of future risk of HF hospitalization in patients with preserved LVEF.^15–17 Heart failure with preserved ejection fraction (HFpEF) is prevalent in obese patients and often preceded by exercise intolerance, which may reflect limitations in SM related to functional and structural impairments also common to the myocardium, including reduced oxidative capacity, microvascular dysfunction, capillary rarefaction, and fibrosis.^39–41 Using a deep learning model for abdominal body composition analysis, we showed that deficient skeletal muscularity but not excess conventional adiposity was independently associated with CMD and future adverse events, especially HF,⁴² in patients with no obstructive CAD. In particular, a sarcopenic CMD phenotype in these female-predominant patients with normal or elevated BMI was associated with the highest CVD risk. Reduced lean body mass, as part of a distinct yet poorly understood SM pathology, has been hypothesized to be a sex-specific contributor to the pathophysiology of HFpEF,⁴³ and CMD and IMAT may help to mediate this effect.

The expansion of IMAT is thought to promote an adverse muscle phenotype associated with metabolic derangements, including impaired bioenergetics, mitochondrial dysfunction, and increased catabolism with loss of lean muscle mass.⁹ Emerging data suggest that the human IMAT secretome is highly immunogenic and inflammatory relative to that of SAT or even VAT depots in individuals with obesity, and may alter immune cell trafficking to neighbouring muscle in a manner that may be critical to regional glucose homeostasis.¹² Taken together, these findings support the hypothesis that common systemic factors involving inflammation and altered glucose metabolism may promote parallel insults to myocardial and SM beds, which may manifest as microvascular dysfunction and intramuscular fat infiltration, respectively, and ultimately contribute to adverse CVD events, including myocardial injury and HFpEF outcomes.

Our study has important limitations including its observational, single-centre design. Despite the use of multivariable-adjusted models, confirmatory and sensitivity testing showing robust data results, residual confounding may persist. Although we leveraged low-dose CTs routinely obtained with clinical cardiac PET studies for the purpose of attenuation correction, we demonstrate excellent segmentation quality assurance, and the ability to obtain concurrent imaging in these patients represents an important strength of this study. Because these were thoracic studies, we focused our cross-sectional assessments at the level of T12 for thoracic body composition analysis. As such, results may differ when compared to abdominal body composition results obtained from the lumbar spine level,^42,44 and abdominal visceral adipose tissue could not be quantified (although we would expect it to have a less important role in this female-predominant cohort since women have significantly lower visceral and higher subcutaneous adiposity as compared with men³⁵). Since cardiac PET is performed using vasodilator stress, no exercise testing results were available. As only patients with normal visual myocardial perfusion and no evidence of flow-limiting CAD were included, findings are generalizable to an INOCA population. Relative to less accurate or more cumbersome methods for body composition analysis,⁴⁵ opportunistic CT used in this manner may provide personalized prognostic information without added cost or radiation exposure and is poised to grow with rapid developments in artificial intelligence and machine learning. Future studies assessing the impact of therapeutic strategies such as supervised exercise and nutritional, medical, and/or surgical weight loss interventions on SM quality and CMD outcomes are warranted.

Conclusions

In patients without flow-limiting CAD, increased intermuscular adiposity is associated with CMD and adverse cardiovascular outcomes independently of BMI and conventional ectopic fat depots and risk factors. The presence of CMD and SM fat infiltration identified a novel at-risk cardiometabolic phenotype.

Supplementary data

Supplementary data are available at European Heart Journal online.

Declarations

Disclosure of Interest

B.F. reports research grants from the National Institutes of Health, AstraZeneca, and Ionis. R.B. reports research grants from Novartis, Amgen, Heartflow and Nanox AI and consulting fees from Heartflow, Nanox AI and Caristo. S.Do. reports research grants from Pfizer, Attralus, GE Healthcare and Siemens and consulting fees from Pfizer. M.F.D.C. reports research grants from Alnylam Pharmaceuticals, Intellia, Gilead Sciences, Sun Pharma, and Xylocor and consulting fees from Valo Health and MedTrace. M.T.L. reports research grants from the National Academy of Medicine, the National Institutes of Health, the American Heart Association, AstraZeneca, Ionis, Johnson & Johnson Innovation, Kowa Pharmaceuticals, and the Risk Management Foundation of the Harvard Medical Institutions. F.J.F. reports research grants from Pfizer, consulting fees from Boston Scientific and holds a related patent (US11322259B2, WO2019051358A1). All other authors report no relevant disclosures.

Data Availability

Data may be shared in a deidentified and anonymized format for use as approved by the independent review board and according to federal and institutional regulations.

Funding

This research was supported by a Lemann Foundation Cardiovascular Research Postdoctoral Fellowship (A.C.A.H.S.) and the Gilead Sciences Research Scholars Program in Cardiovascular Disease and NIH K23HL135438 (V.R.T.).

Ethical Approval

The study was approved by the Mass General Brigham Healthcare Institutional Review Board and performed in accordance with institutional guidelines.

Pre-registered Clinical Trial Number

None supplied.

References

Author notes