The main findings of our study showed that EFV and PFV were significantly increased among diabetic patients, who in turn had significantly higher SS and multivessel CAD. Both EFV ≥ 119.55 ml and PFV ≥ 125.05 ml showed significantly strong predictability for high-SS. This was consistently found at different cutoffs among diabetics and non-diabetics with respective reasonable sensitivity and specificity. Furthermore, on multivariable regression analysis models corrected for HbA1c level, and other covariables, increased EFV was a significant independent predictor of high-SS along with age.

CMR provides an ideal gold standard for accurately measuring true epi- and pericardial fat volumes. Multiple CMR-based studies have demonstrated significantly larger EAT among diabetic patients in a variety of clinical settings [15], however with deficient data regarding its relation to CAD complexity. To our knowledge, this is the first study to comprehensively assess both EFV and PFV, as well as their impact on CAD severity and complexity among diabetics versus non-diabetics.

Association between epicardial adipose tissue and severity of CAD among diabetics

Epicardial adipose tissue is an active endocrine organ that secretes different adipocytokines via endocrine and paracrine routes to exert various cardiovascular effects. It was linked not only to CAD progression but also cardiomyopathy, particularly in diabetics, arrhythmias e.g. AF, and other cardiovascular and metabolic abnormalities [16]. The relation between EAT and the heart is bidirectional, as EAT exerts paracrine effects on the latter (outside-in) but is also influenced by inflammatory mediators generated in the vascular wall (inside-out), causing changes in the secreted adipokines [17].

Results of our study demonstrated significantly larger EFV and PFV in association with both diabetes and increased complexity of CAD. It was previously demonstrated that both diabetes and increased epicardial fat volume mutually co-exist and synergistically enhance atherosclerosis [2]. This was evident among our study population; whereby diabetic patients had significantly larger EFV, larger BMI, altered plasma cholesterol levels with significantly higher AIP, in association with significantly higher rates of multivessel CAD and higher SS. Presence of either diabetes or increased EFV was demonstrated to interchangeably enhance the development of the other, inducing a state of insulin resistance, systemic inflammation, and excess secretion of proinflammatory adipokines with deleterious effects, which explain the potential biological mechanisms beyond their role in the development of CAD with further increased severity and complexity [3, 4].

Our findings were consistent with the results of a meta-analysis of 21 studies (including 2377 patients with variable degrees of CAD and 2598 participants with no CAD), that showed significantly larger EAT (either echo-measured thickness or CT-measured volume) among the CAD group, and in turn among those having significant stenosis (≥ 50%), however, it did not report about differences in relation to DM and was limited by marked heterogeneity of the included studies [18]. Another meta-analysis including 13 studies (11 echo-measured thickness and 2 CT-measured EFV) with 1102 diabetic patients, demonstrated significantly increased EAT among diabetics compared to control [19].

A previous CMR study demonstrated that increased epicardial fat thickness was significantly associated with a high risk of the composite outcome of myocardial infarction, ischemic stroke, heart failure, and cardiac death among 1554 participants over a median follow-up of 12.7years [20]. On the other hand, it has been suggested that treatment with new antidiabetic drugs as SGLT2 inhibitors [21] and GLP-1 analogues [22] reduced cardiovascular risk and mortality possibly induced by the associated reduction in EAT.

Contradictory results were presented by vanMeijeren et al., demonstrating that CMR-measured EFV was not independently associated with stages of CAD [13]. However, limited by the relatively small sample of CAD patients, the severity of atherosclerosis was not assessed, strict inclusion criteria (excluding diabetics and BMI > 35Kg/m2), and partial confounding due to use of intense statin therapy reflecting inclusion of an already high cardiovascular risk population.

The advantages of CMR extend beyond the mere precise assessment of true epi- and pericardial fat volumes to the exploration of the structural and functional consequences of increased fat volume on adjacent myocardial tissue with possible fatty infiltration. This might provide better demonstration of the impact of increased EFV among diabetics in association with complex CAD. In this respect, results of a recently published study by Bialobroda et al. demonstrated that diabetic patients had significantly increased EFV measured at the atrio-ventricular groove and altered epicardial fat structure evident by decreased T1 relaxation times, that was associated with decreased atrial strain reflecting impaired atrial myocardial function [23]. In our study, myocardial fibrosis was detected grossly using the phase-sensitive inversion recovery technique on late gadolinium enhancement images for assessment of viability based on the percentage of subendocardial enhancement. Unfortunately, our study was limited to the volumetric fat assessment due to unavailability of the respective software analysis tools for T1 and T2 mapping sequences.

On the other hand, previous studies demonstrated an association between EFV and heart failure, whereby increased EFV had deleterious effects in patients having HF with mid-range and preserved EF [24], yet paradoxically improved LV structural and functional consequences among patients with dilated cardiomyopathy [25]. Our study comprised an entity of patients with established CAD who presented for viability assessment, expecting that most of them had severe LV dysfunction secondary to significant CAD. The interplay of diabetes associated with increased inflammatory state, increased EFV, and presence of significant CAD might have impacted the degree of LV dysfunction. Our results demonstrated comparably low LV EF among diabetics and non-diabetics, yet significantly lower LV EF among patients with high SS compared to the low/intermediate SS. On the other hand, EFV showed weak negative correlation with LV EF, yet significantly predicted impaired LV EF ≤ 40%. This was consistent with previous data showing that increased EFV was implicated in the development of obstructive CAD as well as in non-ischemic diabetic cardiomyopathy [15]. Moreover, studies demonstrated that using SGLT2i was associated with reduced EFV and improved LV function regardless of the presence or absence of diabetes [26,27,28].

Impact of the glycemic state on measured epicardial and pericardial fat volumes

Furthermore, results of our study demonstrated that on sub-analysis of the diabetic group, both EFV and PFV tended to have a descending pattern of smaller volumes in relation to the glycemic-control categories (uncontrolled (HbA1c ≥ 8.5%), relax-control (HbA1c 7.0-8.4%), and tight-control (HbA1c 6.5–6.9%), respectively), however with borderline statistical significance. Our relatively small sample size might have limited achievement of evident conclusion in this regard. To our knowledge, there is no available data in the literature regarding the impact of the degree of glycemic control on cardiac adipose tissue volume. However, it was previously demonstrated that weight-reduction interventions such as diet, exercise, pharmacological interventions (including antidiabetic drugs), and bariatric surgery significantly reduced EFV [29, 30].

In contrary, Iacobellis et al. demonstrated that echo-measured epicardial fat thickness predicted the development of CAD at 1-year follow-up among well-controlled asymptomatic diabetic obese patients with baseline HbA1c = 6.7% and BMI = 34.9Kg/m2 better than other traditional risk factors including age, hypertension, BMI, and duration of DM [31]. Thus, eliminating the effect of the glycemic state in favor of epicardial fat thickness.

Predictability of EFV for severity of CAD among diabetics

Results of our study demonstrated a significant strong positive correlation between each of EFV and PFV and the syntax score, and significant association of each with high SS. Moreover, our results showed that adjusted for other covariates (including elevated LDL, high atherogenic plasma index, elevated HbA1c, and lowered LV EF), age, and EFV independently predicted high SS on different multivariable regression models.

Previous data showed that echo-measured epicardial fat thickness was positively associated with increased severity of CAD assessed by the Gensini score among diabetic patients [32]. Few CT-volumetric studies demonstrated a significant association of EFV with significant CAD but did not report on their association among diabetics versus non-diabetics. It was shown that CT-measured EFV significantly correlated with and independently predicted the presence and severity of hemodynamically significant CAD that was automatically assessed using quantitative flow ratio [33]. Yu et al., demonstrated that CT-measured EFV ≥ 134.43cm3 was associated with hemodynamically significant CAD (≥ 50% luminal stenosis and with reversible corresponding perfusion defects on single-photon emission computerized tomography-myocardial perfusion imaging (SPECT/MPI) [34]. Moreover, increased EFV was predictive of major adverse cardiovascular events with a follow-up of 76 months [35].

Importantly, our study demonstrated strong predictability of CMR-measured EFV and PFV cutoff values for high-SS among the whole study population (119.55 ml and 125.05 ml, respectively) as well as among both diabetics (126.10 ml and 131.25 ml, respectively) and non-diabetics (109.05 ml and 106.65 ml, respectively). Presenting these cutoffs as strong predictors for increased complexity of CAD is more clinically meaningful than the mere unit increase in volume. This underscores the importance of testing these proposed thresholds among larger studies or developing new ones to be incorporated among traditional risk scores and to be used as therapeutic targets.

Possible role of pericardial adipose tissue in coronary atherosclerosis

Despite the embryological, anatomical, and functional differences between epicardial and pericardial adipose tissues, the latter may be indirectly implicated in the process of atherosclerosis being involved in the chronic inflammatory state seen among diabetics. Few studies showed that a combined increase in both would impact CAD risk and outcomes, particularly among diabetics, yet the underlying mechanism remains unclear [30, 36]. Echo-measured total cardiac adipose tissue above the median value (8.75 mm) was shown to be associated with high cardiovascular risk and all-cause mortality after a follow-up of 6 years among 200 diabetic patients [37]. A previous CT study that assessed pericardial fat enhancement, as a marker of inflammation, demonstrated a significant independent association between PFV and obstructive CAD (OR = 1.26, p = 0.005), as well as CAD burden (OR = 1.25, p = 0.05) in those with greater pericardial fat enhancement [38]. In this respect, our study demonstrated significantly higher levels of EFV, PFV, and their sum among diabetics versus non-diabetics. However, the ratio of epicardial to pericardial fat volumes was comparable among the study groups, denoting maintained relation with possible interaction of both in the pathogenesis of CAD. Moreover, our study demonstrated that increased PFV was significantly associated with higher odds of severe CAD (high-SS) and proposed respective cutoff values among the general population and selectively among diabetics as well as non-diabetics. To our knowledge, this was not assessed in other comparative CMR-volumetric studies, necessitating further investigation.

Limitations

The cross-sectional study design and the relatively small sample size might have limited our results. Our study focused on CMR-measured epi- and pericardial fat volumes, however, did not assess their corresponding echo- and CMR-measured thickness. This might have limited the ability to translate our findings into a simple, easily accessible risk assessment tool for daily clinical practice, albeit provided reliable precise assessment of both epi- and pericardial fat volumes with proposed cutoff thresholds in association with severe CAD among the general population and selectively among diabetics and non-diabetics.