Coronary Artery Diameter Related to Calcium Scores and Coronary Risk Factors as Measured with Multidetector Computed Tomography

Abstract

Arterial remodeling, an early change of atherosclerosis, can cause dilated arterial diameter. We measured coronary artery diameter with use of noncontrast 64-slice multidetector computed tomography (MDCT), and studied its association with coronary artery calcium levels and traditional coronary risk factors.

We included 140 patients from the ACCURACY trial whose noncontrast MDCT images showed measurable coronary arteries. Using 3 measurements of left main coronary artery (LMCA) and right coronary artery (RCA) diameters within 3 mm of the ostium, we associated the results with traditional coronary risk factors and calcium scores.

The prevalence of LMCA and RCA calcium was 22% and 51%, respectively. Mean arterial diameters were 5.67 ± 1.18 mm (LMCA) and 4.66 ± 1.08 mm (RCA). Correlations for LMCA and RCA diameters in 50 randomly chosen patients were 0.91 and 0.93 (interobserver) and 0.98 and 0.93 (intraobserver). Adjusted odds ratios for the relationship of LMCA and RCA diameters to calcium in male versus female patients were 5.65 (95% confidence interval [CI], 2.78–11.5) and 4.35 (95% CI, 2.24–8.47), respectively. Adjusted ratios and 95% CIs for the association of larger RCA diameter with age, hypertension, and body mass index were 1.36 (1.00–1.86), 3.13 (1.26–7.78), and 1.60 (1.16–2.22), respectively.

Arterial diameters were larger in women and patients with higher calcium levels, and body mass index and hypertension were predictors of larger RCA diameters. These findings suggest a link between arterial remodeling and the severity of atherosclerosis.

An early change of atherosclerosis is arterial remodeling, which can result in dilated arterial diameter. In positive (expansive) arterial remodeling, luminal size is often preserved.^1–4 Conventional coronary angiography and myocardial perfusion imaging detect anatomically significant and hemodynamically relevant luminal stenosis, respectively, but perform less well in depicting atherosclerotic disease in its earlier stages when luminal integrity has not yet been compromised by positive vascular remodeling. Imaging studies of early culprit lesions in patients with acute coronary syndrome have revealed an association of echolucent plaque and positive remodeling with unstable clinical presentation.^5–7 Many studies of normal human coronary artery size have been conducted during postmortem examinations of the heart.^8–17 Investigators using intravascular ultrasonography, cardiac magnetic resonance, and cardiovascular computed tomography (CT) have suggested that coronary plaque rupture could occur in positively remodeled lesions.^18,19 Large positively remodeled lesions as predictors of plaque rupture are the subject of active research. Intravascular ultrasonography yields good views of coronary artery lumina and arterial walls and can help to reveal disease that is not angiographically evident; however, it is an invasive technique that is unsuitable as a screening procedure.^20–22

Coronary artery calcium (CAC), as evaluated by means of cardiovascular CT, is currently used as a surrogate marker of atherosclerosis.^23,24 In this study, we used noncontrast 64-slice multidetector computed tomography (MDCT) to measure coronary artery diameters in patients who had been referred for CAC scoring. We then used these measurements to study the relationship of increased coronary artery diameter to CAC and to traditional cardiovascular (CV) risk factors.

Patients and Methods

We included 140 of 230 patients from the Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography (ACCURACY) trial²⁵ whose noncontrast MDCT images showed measurable coronary arteries. The details of that trial are published elsewhere.²⁵ The subjects—all at least 18 years of age—presented with anginal or atypical chest pain and were being referred for nonemergent invasive coronary angiography. We excluded individuals who had baseline renal insufficiency (creatinine level, ≥1.7 mg/dL), irregular cardiac rhythms, resting heart rates faster than 100 beats/min, resting systolic blood pressures below 100 mmHg, or contraindications to β-blockers, calcium-channel blockers, or nitroglycerin. We also excluded pregnant women and persons with known histories of coronary artery disease (prior myocardial infarction, percutaneous transluminal coronary angioplasty or intracoronary stents, or coronary artery bypass grafting).

Protocol for Noncontrast Multidetector Computed Tomography

We used a GE 64 LightSpeed™ VCT (General Electric Medical Systems; Milwaukee, Wisc) set at 120 kVp, 100 mA, a gantry speed of 350 ms/rotation with 227-ms temporal resolution, and a slice thickness of 2.5 mm. The scan mode was cine step. Electrocardiographic triggering was used so that each image was obtained at the same point in diastole, corresponding to 75% of the RR interval. Complete coronary artery views were obtained without contrast-medium injection, and at least 35 consecutive images were captured at 2.5-mm intervals beginning 1 cm below the carina and progressing caudally to include the proximal coronary arteries. The total radiation exposure by means of this technique was less than 1 mSev per patient.²⁶

Coronary Artery Calcium Measurement

Coronary artery calcium measurements were performed on the noncontrast images by an independent, experienced observer who was blinded to the patients' characteristics. Coronary artery calcium was defined as a plaque of at least 3 contiguous pixels (area, 1.02 mm²) with a density of more than 130 Hounsfield units. The lesion's Agatston score was calculated by multiplying the lesion area by a density factor derived from the maximal Hounsfield unit within this area. Total calcium score was determined by adding individual lesion scores from each of the 4 main coronary arteries—left main coronary artery (LMCA), left anterior descending coronary artery (LAD), left circumflex coronary artery (LCx), and right coronary artery (RCA).

Coronary Artery Diameter Measurement

Coronary artery diameter was measured in axial view within 3 mm of the ostia of the LMCA and RCA. The ostium was identified, and 3 separate measurements were recorded 1 cm distal to the ostium and within 3 mm of it. Measurements were made perpendicular to the vessel and from one outer wall to the opposite outer wall. The mean coronary artery diameter was calculated and used for analysis.

In 50 patients chosen at random, a different observer measured the LMCA and RCA diameters by the same means as the first observer, but independently and within a week of each other, to look for interobserver variability. The measurements were repeated on the same 50 patients by the first observer, who used the same methodology to look for intraobserver variability.

Patients were excluded when a clear measurement was not feasible—for example, when a tortuous takeoff, obliterated vessel, or impinging structure with similar nearby opacification made delineation of the 2 structures impossible. Some patients who had very proximal splitting of the LMCA into the LCx and LAD were also excluded. The final study population comprised 140 patients (mean age, 57 ± 10 yr; 81 men).

Statistical Analysis

Ordinal logistic regression was used to determine the odds ratio (OR) of various CV risk factors, including sex, age by 10-year increment, hypertension, diabetes mellitus, dyslipidemia, smoking, body mass index (BMI) per standard-deviation increase, and family history of coronary heart disease (CHD), with ORs organized into quartiles of increasing coronary artery diameters. An additional OR per risk factor was calculated, with adjustment for all other risk factors by means of ordinal logistic regression analysis. A P value less than 0.05 was considered statistically significant. Statistical analysis was performed with use of Stata® software version 10.0 (StataCorp LP; College Station, Texas).

Results

The prevalence of CV risk factors in our patient cohort was as follows: family history of CHD in 37%, hypertension in 31%, dyslipidemia in 30%, obesity in 21%, current smoking in 19%, and diabetes in 11%. The mean LMCA and RCA diameters were 5.67 ± 1.18 mm and 4.66 ± 1.08 mm, respectively. The interobserver and intraobserver correlations for LMCA measurement were 0.91 and 0.98; the values for the RCA were 0.93 and 0.93.

The relationship of the CV risk factors was evaluated across increasing quartiles of LMCA diameter (Table I) and RCA diameter (Table II). Study participants with larger LMCA and RCA diameters were more likely to be male (P <0.0001). In addition, higher BMI was significantly associated with a larger RCA diameter (P=0.0015), and nonsignificant trends were observed in the relationships of increasing age and hypertension.

TABLE I. Coronary Disease Risk Factors across Increasing Left Main Coronary Artery Diameter

TABLE II. Coronary Disease Risk Factors across Increasing Right Coronary Artery Diameter

We detected LMCA calcification in 22% of the subjects (n=31) and RCA calcification in 51% (n=72). The population was substratified according to LMCA calcium score (0; 1st tertile, 1–6; 2nd tertile, 7–39; and 3rd tertile, ≥40) and RCA calcium score (0; 1st tertile, 1–36; 2nd tertile, 37–212; and 3rd tertile, ≥213). Individuals in the highest quartiles of LMCA and RCA diameters had the greatest likelihood of high CAC levels in those vessels. As Figure 1 shows, only 2% of individ uals with the smallest LMCA diameter (the 1st quartile) had a high burden of local CAC (calcium score, ≥40), compared with 13% among those with the largest LMCA diameters (the 4th quartile). Only 5% of the patients with RCA diameter in the 1st quartile (2.2–3.8 mm) had a CAC score of ≥213, compared with 32% of the patients with RCA diameter in the 4th quartile (5.7–8 mm) (Fig. 2).

Fig. 1 Prevalence of increasing coronary artery calcium score (CACS) categories (by tertiles) across increasing left main coronary artery (LM) diameter categories (quartiles).

Fig. 2 Prevalence of increasing coronary artery calcium score (CACS) categories (by tertiles) across increasing right coronary artery (RCA) diameter categories (quartiles).

Table III describes the univariate and multivariate association of traditional CV risk factors and increasing CAC scores with increasing levels of LMCA and RCA diameters. Only male sex was significantly associated with increasing LMCA diameters in both unadjusted analyses (OR=4.57; 95% CI, 2.35–8.75) and multivariate adjusted analyses (OR=5.65; 95% CI, 2.78–11.50). On the other hand, after all risk factors were taken into account, male sex (OR=4.35; 95% CI, 2.24–8.47), increasing age per 10 years (OR=1.36; 95% CI, 1.00–1.86), hypertension (OR=3.13; 95% CI, 1.26–7.78), and BMI (OR=1.60; 95% CI, 1.16–2.22) were significantly associated with larger RCA diameters.

TABLE III. Ordinal Logistic Regression Describing Association of Traditional Coronary Risk Factors and Coronary Artery Calcium Score with Larger LMCA and RCA Diameters

Discussion

Diameters of the LMCA and RCA were larger in patients with traditional CV risk factors and in those with higher local atherosclerotic burdens, which suggests a link between vascular remodeling and the severity of calcific atherosclerotic disease. We found no previous publications that described the use of cardiac CT to determine the direct association of CAC with coronary artery diameters and the association of coronary artery diameters with various traditional CV risk factors. In this study of patients presenting with chest pain in an elective setting, after we took into account all the traditional CV risk factors, we found that male sex, higher BMI, hypertension, and increasing age per 10 years were significantly associated with larger RCA diameters. Subjects in the highest quartiles of LMCA and RCA diameter were most likely to have high calcium levels in the respective vessels.

Postmortem studies of patients who died suddenly of acute coronary syndromes led to an initial understanding of vulnerable plaques.^27–30 Autopsy findings have indicated that early atherosclerotic lesions can enlarge human coronary arteries as a compensatory mechanism, to optimize luminal cross-sectional areas. In an intravascular ultrasonographic study of patients with stable and unstable coronary syndromes,³¹ positive remodeling and plaque echolucency were significantly more prevalent in unstable than in stable lesions; conversely, negative (constrictive) remodeling was more prevalent in stable lesions. Varnava and colleagues,⁴ who studied 88 men with coronary artery disease who had died suddenly, found that lesions with positive remodeling had a larger lipid core than did lesions with negative remodeling (mean, 39% ± 21% vs 22.3% ± 23.1%; P <0.0001) and a higher macrophage count (mean, 15.6 ± 12.3 vs 8.9 ± 11.6; P=0.005). The common link between compensatory remodeling and plaque rupture might have to do with macrophage-derived foam-cell expression of metalloproteinases,³² proteins that are involved in cell-matrix breakdown. Plaques with a high lipid content and macrophage count might stimulate metalloproteinase production and enable positive remodeling at sites made vulnerable by the soft lipid core. In comparison, lesions with negative remodeling had low vulnerability to plaque rupture in terms of a low lipid content and scanty macrophage count.

Noninvasive imaging methods have been used to evaluate early signs of atherosclerosis. However, brightness-mode ultrasonography and cardiac magnetic resonance do not enable complete evaluation of the coronary arteries.^33–36 Conventional coronary angiography and myocardial perfusion imaging imperfectly depict atherosclerosis, whether in its earlier stages or when the disease is mature but has not yet compromised luminal integrity by means of positive remodeling. Coronary artery calcium scoring on cardiovascular CT has been used to evaluate atherosclerosis. In 1,195 patients without histories of coronary artery disease, investigators observed extensive atherosclerosis (calcium score, >400) in 31% of patients in whom the results of myocardial perfusion imaging studies were normal.³⁷

The higher resolution of MDCT can help to detect positive vascular remodeling. Achenbach and associates³⁸ used MDCT for measurement and a “remodeling index” to evaluate increases in coronary artery diameters by dividing the vessel areas in the lesions by the reference segments. Of note, if the whole coronary artery is remodeled positively, all that can be seen is an increase in the overall diameter of the vessel, with no comparison segment available. Montoya and colleagues³⁹ concluded that patients with positively remodeled coronary segments and low-attenuation plaques on CT angiography were at higher risk of developing acute coronary syndrome than were patients who had lesions without those characteristics.

In our study, we found a significant correlation of coronary artery diameter with CAC and with male sex and BMI, and we observed a trend toward larger coronary diameter with increasing age and hypertension. The effects of traditional CV risk factors on coronary atherosclerosis and coronary diameter have been studied by means of postmortem histologic analysis, invasive angiography, and echocardiographic evaluation of proximal coronary arteries. Results varied when quantitative angiography was used to determine the effects of age on normal coronary artery size.^40,41 Some postmortem studies revealed a slight tendency for adult coronary artery size to increase with age^17,42; however, this finding was not always reproducible.¹⁶ Guerra and associates⁴³ found no independent effect of age on the size of coronary arteries. Dyslipidemia had no significant effect on coronary artery diameter in our study group, which was similar to what Dodge and colleagues⁴¹ found in subjects who had normal arteries. Many studies have revealed a relationship between coronary artery size and left ventricular mass. Resting casual blood-pressure measurements are a poor predictor of left ventricular mass.^42,44,45 We did not calculate left ventricular mass in our patients; however, we noted a trend toward increased correlation with hypertension. A postmortem study of 2,403 individuals revealed that hypertension was associated with larger RCA and LMCA diameters.⁴⁶ Body surface area (BSA) was found to be a strong independent predictor of coronary artery diameter in persons without coronary calcium.^47,48 Conversely, Leung and co-authors⁴⁰ detected no significant effect of BSA on coronary size. However, their study documented smaller coronary diameters in female subjects, and this difference was attributed to a typically smaller BSA in women. Few studies have shown that women tend to have smaller arteries than men, even after normalization for BSA.^40,49,50

Limitations of the Study

We measured coronary diameters in axial view rather than obtaining reoriented cross-sectional images. Because remodeling is not radially symmetric, it is possible that we underestimated the coronary diameters.

Conclusion

A simple method for accurately determining coronary artery size would be valuable if differences could be shown between the normal population and patients with early coronary atherosclerosis. Our study shows the feasibility of measuring coronary artery diameters by means of noncontrast MDCT. Coronary artery diameters were associated with CAC and with the presence of traditional CV risk factors. Our finding of larger diameters of the LMCA and the RCA in patients with risk factors and higher local atherosclerotic burden suggests a link between arterial remodeling and the severity of atherosclerosis. This screening method is potentially useful in detecting early coronary artery atherosclerosis and increasing our understanding of subsequent risks of CV disease. However, more studies are needed to confirm this association and to clarify conflicting and limited existing data, with use of different imaging methods, on the relationship between traditional risk factors of coronary atherosclerosis and coronary artery diameters.