This study demonstrated that EATT was increased in the RA patients compared to the healthy controls as a predictor of subclinical atherosclerosis. Additionally, EATT was correlated with the disease activity scores of the patients with RA that include DAS28-ESR, DAS28-CRP, SDAI, and CDAI. This study also showed that only the increased disease activity score was an independent risk factor for increased EATT in RA patients. To the best of our knowledge, although, there are a few studies in the literature evaluating EATT in RA patients with contradictory results about its association with disease activity, this is the first study that included SDAI and CDAI as novel disease activity markers of RA and found a significant relationship between disease activity and increased EATT.

Epicardial adipose tissue is a true visceral adipose tissue surrounding the subepicardial coronary arteries [24]. It is a metabolically active paracrine and endocrine organ secreting many proinflammatory and proatherogenic cytokines, chemokines, and adipokines [27]. In addition to systemic effects, due to its direct contact with coronary arteries, it may also directly contribute to coronary atherogenesis [28]. Accordingly, EAT plays a role in the development of subclinical atherosclerosis through vascular inflammation and endothelial dysfunction [29]. Several studies have revealed the association of EAT with adverse cardiovascular events ranging from asymptomatic to overt coronary artery disease independently of other risk factors [30]. In recent years, it was shown that the pathogenesis of atherosclerosis is a dynamic process in which inflammation plays a role in all stages [7]. This study showed that EATT increases in patients with RA, which is a chronic inflammatory disease, as a predictor of subclinical atherosclerosis. Likewise, in the studies conducted by Temiz et al. and Fatma et al., EATT was significantly higher in RA patients in comparison to HC [31, 32]. Lima-Martinez et al. conducted a study with 34 female RA patients and showed that healthy women had the lowest EATT values compared to RA patients, and RA patients receiving biological DMARDs had lower EATT in comparison to patients receiving non-biological DMARDs [33]. In the study by Başpınar et al., the epicardial fat area measured by computed tomography was significantly higher in RA patients in comparison to healthy individuals [34]. On the other hand, in the study by Ormseth et al., EATT was not significantly higher in RA patients in comparison to HC (p = 0.06), although RA patients with metabolic syndrome (n = 56) had significantly higher EATT values compared to RA patients without metabolic syndrome (n = 101) [35]. There are a few studies in the literature that have evaluated EATT in other rheumatic diseases in which increased EATT values have been demonstrated in patients with systemic sclerosis, systemic lupus erythematosus (SLE), ankylosing spondylitis, and inflammatory bowel disease. [36,37,38,39]

Corticosteroids have major effects on adipose tissue metabolism via glucocorticosteroid specific receptors which are at the highest density in the visceral adipose tissue. It has been shown that high doses of steroid increase the visceral adipose tissue [40]. Besides, the long-term steroid use is associated with weight gain, even obesity, which in return causes metabolic syndrome, and are related with the features of metabolic syndrome. Kitterer et al., performed the first study that evaluated the effect of long-term steroid therapy on epicardial and pericardial fat deposition. Their study included 61 patients with various rheumatic diseases (mainly vasculitis), and age-sex-BMI matched healthy controls. Larger epicardial fat was found in patients treated with steroids in comparison to steroid-naϊve patients. In addition, significantly higher epicardial fat was found in patients on high-dose steroids (defined as > 7.5 mg prednisone equivalent per day for at least 6 months) compared to patients on low-dose streoids (ie., < 7.5 mg prednisone equivalent per day for at least 6 months). However, no significant difference was observed in epicardial and pericardial adipose fat between patients on low-dose steroids and steroid-naϊve [40]. In the cross-sectional study that included 162 SLE patients and 86 healthy controls, EAT was significantly higher in SLE patients than in controls, and when adjusted for age, sex, race, and waist circumference, only cumulative corticosteroid dose was remained significantly associated with the EAT volume [37]. Even though, corticosteroid use may have had an influence on higher EATT in RA patients in comparison to healthy controls in our study, the majority of our patients were on low-dose steroid (median 5 mg, [IQR, 5–7.5]). Likewise, in the study of Karpouzas et al., 139 RA patients who underwent coronary angiography for EAT measurement and coronary plaque assessment, the EAT volume was found higher in patients with atherosclerosis compared to patients without atherosclerosis, and it was related with the number of segments with plaque as well as mixed plaque presence in RA patients. Yet, they demonstrated that prednisone use did not alter the association between the EAT volume and the plaque burden [41].

Moreover, we showed that EATT was significantly correlated with disease activity scores, and in the logistic regression analysis, only disease activity was demonstrated to be an independent risk factor for increased EATT. Besides EAT’s contribution to systemic inflammation, systemic inflammation in return causes adipogenesis leading to an accumulation of fat in epicardial tissue promoting a positive feed-back on inflammation and accelerated atherosclerosis [28, 41]. Therefore, an increase in EATT is expected to be seen in RA patients with higher disease activity scores for which we found a cut-off value of 6.4 mm, and patients with EATT ≥ 6.4 mm in our study were found to have higher DAS28-ESR, DAS28-CRP, SDAI, and CDAI scores. Data about the correlation of EATT with disease activity in rheumatic diseases are insufficient and conflicting. Similar to our study, in the study by Alpaydın et al., DAS28 scores were demonstrated to be positively correlated with EATT (r = 0.365, p = 0.02) [42]. In a study conducted with 115 psoriasis patients, higher EATT values were found in the patients in comparison to healthy individuals, and EATT was correlated with psoriasis disease activity scores [43]. Psoriasis Area and Severity Index scores were also shown as an independent predictor of EAT [43]. Again, in a study conducted with systemic sclerosis patients (n = 30), EATT was demonstrated to be correlated only with disease activity scores (r = 0.45, p = 0.01) [44]. Contrary to our results, Petra et al. and Ormseth et al. did not find any significant correlations between ESR, CRP, DAS28-CRP scores, and EATT values [35, 45]. Temiz et al. showed significant correlations of EATT with ESR, CRP, and Health Assessment Questionnaire scores, but EATT was not correlated with DAS28 scores [31]. Nonetheless, we believe that, as the strength of our study, we showed the correlation of EATT not only with DAS28 but also with other novel disease activity scores of RA, namely SDAI and CDAI.

Another noteworthy finding of our study was that in the univariate and multivariate regression analyses, we found SDAI as an only predictor of increased EATT, and therefore, an independent risk factor for subclinical atherosclerosis in RA patients. As far as we know, this is the first study that showed higher disease activity as a risk factor for EATT. In our study, we did not observe any correlation between EATT and disease duration. Similar results were also found in other studies which reported that the progression of the disease does not result in a significant increase in EATT [31, 45]. These results may be interpreted as disease activity, rather than disease duration, has a greater influence on increased risk of cardiovascular events, even in the early stages of the disease.

There is copious amount of data regarding the association of EATT with coronary atherosclerosis, yet, the number of prospective large-population based studies on the prognostic value of increased EATT for the prediction of future major adverse cardiac events (MACE) are limited. In a prospective study in which 200 patients with a clinical diagnosis of CAD who underwent left and right coronary angiography were followed-up for 26 months, the incidence of death due to cardiovascular events was significantly higher in patients with EATT > 7 mm [46]. The Heinz-Nixdorf Recall study is a prospective population cohort that evaluated the predicted value of EAT volume for coronary events. Of the 4,093 participants without a baseline coronary heart disease, during a follow-up year of 8.0 ± 1.5 years, 130 of them had fatal or non-fatal coronary events developed. The EAT volume was found higher in subjects who developed any coronary event (121 mL vs 95 mL, p < 0.001). The study also showed that the incident of coronary events increased with the EAT volume, and the subjects in the highest EAT volume quartile had a fivefold higher risk of having a coronary event in comparison to subjects in the lowest EAT volume quartile [47]. Again, in a recent study that included 2068 asymptomatic participants without a prior coronary artery disease from the prospective, randomized EISNER (Early Identification of Subclinical Atherosclerosis by Noninvasive Imaging Research), MACE were reported in 223 (11%) of participants during a mean follow-up of 13.9 ± 3 years. By using deep-learning which is a subset of machine-learning, the EAT volume and attenuation were quantified. The EAT volume was shown to be significantly higher in participants with MACE than without MACE (90.6 [IQR, 67.4–128.7] cm3 vs. 77.0 [54.7–103.3] cm3, p < 0.001). In multivariate analysis, adjusted after atherosclerotic cardiovascular disease risk score, the EAT volume was found to be associated with increased risk of MACE (HR, 1.35 [95% CI, 1.07–1.68], p = 0.009). Again, there was a strong correlation between the EAT volume and the subsequent myocardial infarction and cardiac mortality (HR, 1.49 [95% CI, 1.00–2.21], p = 0.046) [48]. On the contrary, Albuquerque et al. evaluated 194 patients with a known coronary artery disease who entered a phase II cardiac rehabilitation program for a mean follow-up of 3.6 ± 1.3 years, and they showed that the EAT when measured by echocardiography was not a predictor of MACE (HR, 1.32 [95% CI, 0.75–2.31, p = 0.33]); which was defined as diagnosis of acute coronary syndrome, coronary revascularization, stroke, ventricular arrhythmias requiring hospitalization and death of any cause [49]. However, this study had some limitations such as shorter follow-up period when compared to other studies, smaller study population and the presence of higher comorbid diseases of the patients. Furthermore, as the study population included patients with a known previous coronary artery disease, this study may be limited to explain the role of EATT for subclinical atherosclerosis and the future cardiovascular events.

The main limitations of our study were its cross-sectional design and the lack of a measurement of arterial blood pressure and waist circumference. However, as we excluded patients with other cardiovascular risk factors such as diabetes mellitus, hypertension, and hyperlipidemia, we could make a more accurate analysis of predictors influencing EATT. Another limitation of this study may be the use of two-dimensional echocardiography to measure EATT as its sensitivity may be lower especially in obese patients and that two-dimensional measurement may not fully assess the entirety of EAT. Yet, a good correlation of echocardiographic EAT with magnetic resonance imaging measurements was previously shown [50]. Echocardiography is still an easily accessible and inexpensive method for measuring EATT. Again, all echocardiographic measurements in this study were performed by a single cardiologist, which might be another limitation of this study. The last limitation of our study was the lack of assessing plaque burden and plaque types in coronary arteries as well as different arterial sites by imaging methods such as computed tomography angiography or vascular ultrasound which could have improved our results of the association of EATT with subclinical atherosclerosis.