This was a registry-based retrospective cohort study using data from our Cardiovascular Diagnosis and Intervention Unit from a tertiary care hospital. The study sample was drawn from all patients who were referred for coronary computed tomography (CT) angiography for evaluation of coronary artery disease (CAD) from January 2008 to December 2013. Most patients (n=584) were referred without a previous ischemia test, and the remainder (n=129) had a previous inconclusive treadmill exercise test or single-photon emission CT scan. Only patients from our primary catchment area were considered for this study. We excluded patients with known cardiovascular disease (previous myocardial infarction, stroke or revascularization procedure, valvular heart disease, previous myocarditis or cardiomyopathy) or with serious life-threatening illness (life expectancy less than 1 year). The final population included patients with full available data on abdominal adipose tissue areas and coronary artery calcium (CAC) score who were followed up for a maximum of 3 years (Figure 1).

Figure 1.

Patient selection flowchart. From the initial 1286 patients referred for cardiac CT due to suspected CAD, the final population included 713 patients. CAD, coronary artery disease; CT, computed tomography; CTA, computed tomography angiography; MACE, major acute cardiovascular events.

(0.15MB).

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the institution's ethics committee. All participants gave their written informed consent to participate in the study.

Arterial hypertension was defined as history of hypertension (according to the European Society of Cardiology Guidelines17) or being on medical treatment with antihypertensive drugs. Smoking history was classified as never, former (participants who smoked for a minimum of 6 months during their lifetime but did not smoke at study entry) and current smoker. Type 2 diabetes mellitus was defined as a history of diabetes (according to current worldwide consensus criteria18) or being on medical treatment for diabetes. Family history of premature coronary heart disease was defined as having a first-degree relative who experienced a fatal or nonfatal myocardial infarction and/or coronary revascularization procedure before the age of 55 years in male relatives and 65 years in female relatives. Dyslipidemia was considered if the patient had a past history of dyslipidemia or was taking lipid-lowering drugs (statins, ezetimibe or fibrates).

All patients underwent a cardiac multidetector CT scan in a 64-slice CT scanner (SOMATOM, Sensation 64, Siemens Medical Solutions, Forchheim, Germany) with 3 different acquisitions: the first for abdominal fat quantification, the second for CAC quantification and the third for coronary angiography. To assess abdominal fat, a single-slice abdominal CT scan was performed between L4 and L5, according to the method described by Borkan et al.19 The scan parameters were 120kV and 216mA with 5-mm thickness. This resulted in an estimated radiation exposure of 0.06 mSv. A blinded expert used the obtained slice to measure abdominal fat distribution: a cursor pointer was used to trace the VAT area by delineating the abdominal wall muscular layer20 and adipose tissue was identified in the areas with attenuation values ranging from−150 to−50 Hounsfield units (HU).21 Total abdominal fat area was measured, and the SAT area was obtained by subtracting VAT from the total abdominal fat area. As an estimate of the relative contribution of the VAT to the total abdominal fat, the VAT to SAT area was calculated (VAT/SAT ratio).

The following scan parameters were used to quantify the CAC: collimation, 24×1.2mm; gantry rotation time, 330ms; pitch, 0.2; tube voltage, 120kV; and tube current, 190 mAs. Image reconstruction of the calcium score acquisition was performed using an effective slice thickness of 3mm. CAC score was reported as the mean Agatston score and was calculated using a detection threshold of 130 HU using semiautomated software (Syngo Calcium Scoring, Siemens Medical Solutions) as described previously.22

Following CAC acquisition, CT angiography was performed (collimation, 64 x 0.6mm; tube current, 850 mAs; all other parameters similar to CAC acquisition scan). Tube current modulation with electrocardiographic pulsing for decreasing radiation dose was used, with full tube current applied at 60% to 65% of the RR interval. In patients with body weight lower than 70kg, tube voltage was reduced to 100kV. A bolus of 80 to 100mL of contrast (Ultravist, iopromide 370mg/mL, Bayer Schering Pharma AG, Berlin, Germany) was injected at 5mL/s via a power injector (Stellant D, Medrad Inc, Warrendale, Pennsylvania, United States) followed by a 40mL saline “chaser”, using a dedicated antecubital vein 18-gauge access catheter. A bolus-tracking technique was used, with a region of interest placed within the ascending aorta, set to detect a predefined threshold of 150 HU. For assessment of CAD, multiphasic sets of the CT reconstructed images were processed on a dedicated workstation (Aquarius Tera Recon Inc, San Mateo, California, United States) and analyzed for detection of at least 1 luminal diameter narrowing higher than 50% in any coronary artery segment by an experienced cardiologist. Severely calcified segments precluding lumen assessment were classified as positive for CAD.

According to our department's policy, only patients from our primary catchment area were followed up in order to collect information on major acute cardiovascular events (MACE). Patient follow-up data were collected by telephone interviews and electronic health record review at 12 and 36 months after cardiac CT. The combined endpoint included death from all causes, myocardial infarction or a revascularization procedure (percutaneous coronary intervention or coronary artery bypass graft surgery) at least 1 month after cardiac CT. Myocardial infarction was defined according to the most recent consensus definition.23 The decision to proceed to myocardial revascularization was made by the attending cardiologist/cardiac surgeon.

Continuous variables are reported as mean±standard deviation or median (interquartile range). Discrete variables are expressed as frequencies and percentages.

An independent t test was used to assess differences in abdominal adipose tissue areas and the VAT/SAT ratio between participants according to the presence or absence of the combined endpoint.

Cox proportional hazards regression was used to perform the multivariate analysis using the following models: model 1 included age and sex; model 2 included age, sex, current smoking status, history of hypertension, type 2 diabetes mellitus, dyslipidemia, and familial history of premature CAD. Model 3 included all variables from model 2 plus coronary calcium score. A final model (model 4) was built including all variables from model 3 plus a dichotomic variable for presence/absence of CAD. The proportional hazards assumption of the Cox models was evaluated with Schoenfeld residuals.24 There was no evidence of departure from the assumption of proportionality. VAT, SAT, VAT/SAT ratio and CAC+1 were included in the models as a base-2 logarithm due to their skewed distribution (proportionality was assessed using the Kolmogorov-Smirnov test). One unit variation of the base-2 logarithmic transformation would be equivalent to a doubling of the variable of interest.

All statistical analyses were conducted using Stata 13.1 for Mac (StataCorp, College Station, Texas, United States).

The final population included 713 participants, 437 (61% men), with a mean age of 57.7±10.2 years. The participants’ characteristics according to the incidence of the combined endpoint during the follow-up period are depicted in Table 1.

During a median follow-up of 1.3 years (interquartile range, 1.1-1.9 years) there were 18 deaths, 2 myocardial infarctions, and 54 revascularization procedures, for a total of 66 events of the combined endpoint. These participants showed a significantly increased VAT/SAT ratio (1.06±0.74 vs 0.80±0.52, P=.0001) and a trend to having a higher VAT area (166.9±71.2 vs 150.0±76.2 cm2) and lower SAT area (188.6±94.7 vs 216.5±122.5 cm2), which was statistically significant (Table 1).

The total event rate was 6.0 (95% confidence interval [95%CI], 4.7-7.6) per 100 person-years of follow-up (1.2% in patients without obstructive CAD; 19.1% in patients with obstructive CAD in cardiac CT angiography). There was a significant increase in the event rate across the tertiles of the VAT/SAT ratio (3.0 events in the first tertile, 6.5 events in the second tertile and 8.5 events per 100 person-years for the last tertile, P=.0076). Figure 2 depicts the cumulative hazard curves of the combined endpoint over time according to the tertile of the VAT/SAT ratio.

Figure 2.

Cumulative incidence of the combined endpoint over time according to the tertile of the VAT/SAT ratio. There was an increase in the event rate across the tertiles of the VAT/SAT ratio. MACE, major acute cardiovascular events; SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue.

(0.13MB).

In the age- and sex-adjusted analysis, doubling of the VAT/SAT ratio was associated with a 1.47 (95%CI, 1.05-2.07) increased hazard of the combined endpoint (Table 2). This association remained significant after adjustment for traditional cardiovascular risk factors (model 2, adjusted hazard ratio, 1.50; 95%CI, 1.06-2.13).

Visceral adipose tissue was not associated with the combined endpoint in the multivariate analyses (Table 3). However, higher SAT was independently associated with a lower risk of death or MACE irrespective of age, sex, cardiovascular risk factors, and CAC (HR, 0.70, 95%CI 0.51-.97, P=.033).

In the model including CAC score and traditional risk factors for CAD, a doubling of the VAT/SAT ratio was associated with a 1.43-fold increased hazard of the combined endpoint (95% CI, 1.01-2.01) (Table 2). The CAC and VAT/SAT ratio seemed to have an additive effect in the prediction of the combined endpoint of death or MACE (Figure 3).

Figure 3.

Annual combined endpoint event rates stratified by coronary artery calcium score and tertiles of the VAT/SAT ratio. CAC and the VAT/SAT ratio seemed to have an additive effect in the prediction of the combined endpoint of death or MACE. CAC, coronary artery calcium; MACE, major acute cardiovascular events; SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue.

(0.08MB).

The prevalence of CAD was 29% (208 patients). The VAT/SAT ratio was a significant predictor of the combined endpoint (1.43-fold increased hazard of death or MACE per doubling of the VAT/SAT ratio) after adjustment for cardiovascular risk factors, CAC, and the presence of CAD (model 4, Table 2).

The limitations of this study include the observational design and caution should be exercised when drawing conclusions about causality. All decisions to proceed to myocardial revascularization were made by the attending cardiologist and cardiac surgeon. We excluded revascularization procedures performed during the first month after coronary CT to avoid procedures that were prompted in the short-term by the result of the coronary CT angiography. Therefore, we can assume that most revascularization procedures were performed due to persistent angina despite anti-ischemic therapy or a positive ischemia test. Our cohort included a white population, so its external validity, especially to other ethnic groups, remains to be shown. Only patients from our primary catchment area were followed up to accurately collect information on MACE. Those patients would necessarily come to our center if they had a MACE, and therefore we tried to reduce to a minimum the potential loss of information on MACEs that would be managed in other centers. Only a small proportion of the initial sample underwent the 3-year follow-up (n=107). A sensitivity analysis (Table of the supplementary material) showed no significant differences in age, sex, cardiovascular risk factors, calcium score, areas of abdominal fat, and prevalence of CAD between groups according to the presence or absence of 3-year follow-up information. In addition, we did not systematically collect detailed information about the medical drugs patients were prescribed.