4. Discussion:
Patients with CAD had higher values of EAT. EAT-T by echocardiography correlated with EAT-V by CTA and was an independent predictor of CAD. Individuals with an EAT-T (PLAX) ≥3.9 mm and EAT-T (PSAX) ≥3.8 mm had Odd’s ratio of 2.8 and 2.6 respectively of having CAD. The sensitivity of these parameters to detect CAD was over 80% with specificity in the range of 50%. The marker was an independent risk factor after adjusting for clinical risk score of QRISK 3 which includes traditional risk factors.
Previous studies have shown positive correlation of EAT with indices of obesity like BMI (8). So, to nullify the effect of obesity on EAT and for better refinement of data, its indexed value was also evaluated in our study. Odds ratio of having CAD was higher for indexed EAT-T values as compared to absolute values in our study, which further signified the importance of neutralizing the effect of obesity on EAT measurement by indexing it to body surface area.
Epicardial adipose tissue surrounding the heart has been postulated in contributing to the development of coronary atherosclerosis by various mechanisms (3). One such mechanism is EAT acting as inflammatory organ leading to local secretion of various adipo-cytokines and inflammatory mediators like tumour necrosis factor, interleukin-6 and free fatty acids from epicardial adipocytes which may adversely affect coronary arteries (9, 10). Recent studies have also found that EAT was associated with the presence of high-risk vulnerable plaques (11).
EAT-V measured by CTA in multiple studies has been associated with CAD. Iwasaki et al (12) measured EAT-V in patients who underwent CTA and found that patients with significant coronary artery stenosis had higher EAT-V. K Yamashita et al (13)showed that EAT-V was associated with total coronary plaque burden and vulnerable lipid rich necrotic plaques, and Ueno et al(14) showed independent association of chronic total occlusions with EAT-V. A recent study by Wenji Yu et al in Chinese population also revealed association of EAT volume to obstructive CAD(15). Thus EAT-V has emerged as a reliable risk predictor of CAD, plaque burden and vulnerable plaques, and EAT-T by echocardiography in our study showed good correlation with it.
On the other hand, echocardiographic EAT-T has not been widely studied due to the proposed limitations of measurement in linear dimensions, inter-observer variation and lack of good acoustic windows(16, 17). There are only few studies of EAT-T by echocardiography and have shown its association with CAD severity (assessed by various scores like Gensini score, SYNTAX score, and number of epicardial vessels involved), and with the presence of CAD (cut off being taken from 50% to 70% diameter stenosis) (18, 19). Verma B. et al showed that EAT-T has positive correlation with waist circumference, LDL-C level, Gensini score, and SYNTAX score, and also showed that EAT-T is independent predictor of CAD after adjusting for all conventional risk factors (20). All these studies have adjusted for conventional risk factors in a multivariate model for showing EAT-T as independent predictor of CAD. However, in our study we included QRISK3 score in multivariate model. Since calculation of QRISK3 score require all conventional risk factors and is a better predictor of CAD and future CV events than individual conventional risk factor, therefore putting it in a multivariate model enhanced the independent predictive value of EAT-T.
Erkan et al observed mean EAT values of 4.3±0.9 mm, 5.2±1.5 mm, and 7.5±1.9 mm in patients with normal coronary arteries, minimal CAD, and significant CAD, respectively. In that study, a EAT cut-off value of 5.8 mm predicted significant CAD (AUC: 0.875; P < 0.001, 95% CI: 0.825–0.926) in that study (21). While various other studies found a threshold of 6-7 mm as elevated EAT-T(22, 23). In our study, optimum EAT cut-off value to predict CAD was identified as ≥3.9 mm on PLAX view (AUC: 0.68, 95% CI: 0.58-0.79) with a sensitivity of 84 % and a specificity of 55 %, respectively. These variations in EAT cut-offs in different studies may be explained by ethnic differences in the threshold for association with CAD as described by Moharram et al (24).
Our study had the following limitations. This was a single centre study with moderate number of patients. The above findings and cut-offs need to be validated in larger prospective cohort studies. The measurement of EAT is operator-dependent and inter-observer variations may occur; although we observed good inter-observer correlation. Also measuring EAT-T in obese and those with sub-optimal echo windows is a challenge, these patients were excluded from this study. Also, this was a hospital-based study and may not reflect the normal population. However, a CT scan-based study to show correlation with CAD may not be feasible in normal population.