During the period from March 2010 to April 2017, we retrospectively recruited consecutive patients with chest pain or dyspnea of unknown origin, attended at the cardiology outpatient clinic in 2 University Referral Hospitals in our area, who underwent a cardiac CT performed for clinical reasons according to current guidelines,23 composing a retrospective cohorts study. Subjects were excluded if they had any other cardiomyopathy or if the cardiac CT study was unsatisfactory, owing to low-quality image acquisition. The study protocol was approved by our local ethics committees. The study was exempted from the need of an informed consent by our local ethics committee given the retrospective character of our study and the long-term recruitment. Anonymity of the database was ensured at all times.

A Somaton 64-slice multi-detector CT (Siemens Healthineers, Germany) was used for scanning. Breath holding was utilized to minimize motion artefact. Oral and/or intravenous betablockers were administered as needed to lower heart rate below 65 beats per minute to minimize reconstruction artefacts. Thereafter, a 0.4mg sublingual dose of nitroglicerin was given. A total dose of 1mg/kg intravenous iodine 350mg/g contrast was given to obtain a noninvasive coronary angiography when appropriate. Patients showing high calcium scores (> 400 Agatston Units [AU]) did not receive contrast, and a non-contrast study including only coronary artery calcium score (Agatston score) was performed. The cardiac CT was performed according to the following protocol: tube voltage 100kVp, tube current 55mAs, pitch 0.31, gantry rotation time 0.3s, and collimation 128×0.6mm. The images were reconstructed with a slice thickness of 3 and 0.8mm for calcium scoring and coronary angiography, respectively. The coronary calcium score was calculated according to the method described by Agatston et al.24 Scans were analyzed by an experienced radiologist using a 3-dimensional workstation. We defined CAD as the presence of obstructive coronary stenosis (at least one coronary stenosis > 50%) or a high amount of coronary calcium (Agatston Score≥400AU), based on a definition previously accepted in a few studies.9,13

EAT was identified on non-contrast data sets as a hypodense layer surrounding the myocardium and limited by the pericardium. A volumetric software was used to integrate the whole set of user-defined area traces, from the most cranial at the level of the pulmonary artery to the most caudal at the level of the base of the heart that seemingly included any pericardial content, to come up with an overall EAT volume (cm3).

All subjects were followed for adverse cardiovascular events, including cardiac death, myocardial infarction, revascularization (occurring more than 30 days after the CT), new-onset atrial fibrillation, acute ischemic stroke, transient ischemic attack, or cardiac hospitalizations. Follow-up information was obtained by review of the corresponding medical records in our electronic hospital information system, or telephone contact if needed.

Continuous variables were expressed as means±standard deviations, or medians with [interquartile range], and were compared with Student t test. Categorical variables were reported as frequencies and percentages and compared using Fisher's exact test or the Chi-square test, according to the data size. A Kolmogorov–Smirnov test was firstly used to check the assumption of normal distribution. Taking into consideration the anthropometric variability of EAT volume, we also made the statistical analysis after having indexed EAT to body surface area (EAT-i). A binary logistic regression analysis was performed to determine the relationship between relevant CAD and EAT-i, age, gender, cardiovascular risk factors and cardiovascular comorbidities. Cardiovascular risk factors (hypertension, hyperlipidemia, diabetes mellitus, smoking and obesity) and comorbidities (chronic kidney disease, cerebrovascular disease, peripheral arterial disease, and atrial fibrillation) were grouped in 2 different variables (ie, “cardiovascular risk factors” and “comorbidities”). Odds ratio and 95% confidence interval (95%CI) were calculated for each covariate, as well as the calibration of the model by using Hosmer–Lemeshow test, and C-statistic to assess discrimination. The association of EAT with cardiovascular events during follow up was assessed using Cox regression, in a model which included median EAT-i, age, cardiovascular risk factors, and comorbidities. The first-degree interaction between chronic coronary syndrome and EAT-i was studied using likelihood ratio statistics and the backward elimination method (chunk test). For survival analysis, Kaplan–Meier curves were created to estimate the distribution of cardiovascular events, depending on the presence of elevated EAT-i; the log-rank test was used to test differences in survival between groups. A random subset of 10 patients was selected to evaluate reproducibility of EAT volume measurements, and it was assessed using intra-class correlation coefficient. A 2-tailed probability value of P<.05 was considered statistically significant. Statistical analysis was performed using SPSS software (IBM SPSS statistics, version 22.0 for Mac, Armonk, United States).

Patient baseline characteristics are shown in Table 1. A total of 179 subjects were included in our analysis, with a mean age of 56±12 years, 103 (57.5%) were men, with a median of 2 [range 1–3] cardiovascular risk factors per individual. In our study ninety-one subjects showed no plaque (50.8%), 35 (19.6%) non-obstructive plaques, and 16 (9%) subjects showed at least one coronary stenosis > 50% (obstructive plaques). A total of 53 patients (29.6%) were diagnosed of having CAD. One-vessel disease was demonstrated in 10 subjects (5.6%), 2-vessel disease in 5 subjects, and three-vessel disease in 1 subject. Left main disease was demonstrated in one patient. Median Agatston Score was 10UA [range 0–380UA]. Absence of coronary calcium was reported in 79 (44.4%) individuals, while an Agatston Score above 400 was found in 42 (23.5%).

Median EAT volume was 96.67cm3 [range 72.36–131.24cm3], and median EAT-i volume 52.51cm3/m2 [range 40.78–68.15cm3/m2]. Patients with type 2 diabetes mellitus (67.56cm3/m2 vs 53.01cm3/m2; P<.001), hyperlipidemia (59.12cm3/m2 vs 50.98cm3/m2; P=.011) and hypertension (60.81cm3/m2 vs 49.13cm3/m2; P<.001) had a significantly higher EAT-i volume. EAT-i volume was also greater in patients with CAD (68.29cm3/m2 vs 50.62cm3/m2; P<.001).

Compared with subjects without CAD, patients with CAD were significantly older (P<.001), were men in a greater proportion (P<.001) and more likely had cardiovascular risk factors: namely hypertension (P<.001), diabetes (P<.001), and a greater mean body mass index (P=.002).

Logistic regression analysis showed that EAT-i (OR, 2.6 per increase in 1cm3/m2 (95%CI, 1.13–6), P=.003), age (P<.001) and male sex (P<.001) were significantly and independently associated with CAD, in a well calibrated model that further showed a good discrimination (Hosmer–Lemeshow test Chi-square=2.039, P=.980; C-Statistic AUC 0.845, 95%CI, 0.784–0.907, P<.001), which included also cardiovascular risk factors and comorbidities.

During a median of 5.53 years follow-up period [range 2.37–6.37 years], 37 cardiovascular events occurred: 9 acute coronary syndromes, 13 late revascularizations, 10 cardiovascular hospitalizations (5 due to acute decompensated heart failure, 2 myopericarditis, 1 third degree heart block, and 2 chest pains of unknown origin), 4 strokes or transient ischemic attack, and one cardiovascular death. Patients with cardiovascular events (Table 1) showed higher EAT-i volume than patients without cardiovascular events (63.79±19.22 vs 53.78±20.93cm3/m2, P=.009). Moreover, they were significantly older (61.35±12.35 years vs 54.18±11.58 years, P<.001), with a higher proportion of hypertension (73.0% vs 53.5%, P=.033) and type 2 diabetes mellitus (35.1% vs 15.5%, P=.007).

A Kaplan–Meier survival analysis is shown in Fig. 1. Those patients with elevated EAT-i showed a significantly higher risk of cardiovascular events during follow-up (log-rank test P<.001, Fig. 1). The selected cutoff value was the median EAT-i (52.51cm3/m2, negative predictive value 90%, positive predictive value 31%). Notably, using multivariable adjusted Cox regression analysis, median EAT-i showed an independent association with adverse cardiovascular events (hazard ratio, 2.44; 95%CI, 1.07–5.56; P=.033) (Table 2). No significant interaction was found between previous chronic coronary syndrome and EAT-i (P=.961).

Fig. 1.

Survival analysis. Kaplan–Meier survival analysis reveals a significantly higher risk of cardiovascular events for patients with elevated EAT-i (log-rank test P<.001). EAT-i, indexed-epicardial adipose tissue.

(0.15MB).

The intra-observer reproducibility calculated from a subset of 10 random patients was excellent with an intra-class correlation coefficient value of 0.976 (95%CI, 0.908–0.994).

Our study has some limitations that should be acknowledged, mainly related to the retrospective nature of the study as well as its cross-sectional design, which precludes obtaining conclusions that could imply a cause–effect relationship. Besides, the sample size and the scarce number events observed make our observations difficult to interpret, requiring additional large-scale studies to expand and understand the results. Nevertheless, despite the relatively small sample size, the baseline characteristics of our population were in line with bigger studies.5

Our results contribute to an expanding body of evidence for the independent role of EAT volume in the development of CAD and cardiovascular events, and it supports the idea of EAT as a tool which could enhance cardiovascular risk stratification.