Noninvasive Assessment of Plaque Characteristics With Multislice Computed Tomography Coronary Angiography in Symptomatic Diabetic Patients

The study population consisted of 86 (40%) patients with known type 2 diabetes and 129 (60%) patients without diabetes who underwent examination with MSCT coronary angiography for recurrent chest pain complaints. Fifty-one (24%) patients were examined with a 16-slice MSCT scanner, whereas the majority (164 [76%]) underwent examination with 64-slice MSCT. Diabetes was diagnosed according to the American Diabetes Association criteria (17): 1) symptoms of diabetes and a casual plasma glucose level of ≥11.1 mmol/l or 2) a fasting plasma glucose level of ≥7.0 mmol/l. Only patients in sinus rhythm, without contraindications to MSCT, were included (18). All patients gave written informed consent to the study protocol, which was approved by the local ethics committee.

All examinations were performed using a Toshiba Multislice Aquilion system (Toshiba Medical Systems, Tokyo, Japan). First, a prospective coronary calcium scan without contrast enhancement was performed, followed by 16- or 64-slice MSCT coronary angiography performed according to the protocols described elsewhere (19,20). If the heart rate was ≥65 beats/min, additional oral β-blockers (metoprolol, 50 or 100 mg, single dose, 1 h before the examination) were provided if tolerated.

Coronary artery calcium was identified as a dense area in the coronary artery exceeding the threshold of 130 Hounsfield units. A total Agatston score was recorded for each patient.

MSCT angiograms were evaluated by two experienced observers who were unaware of the clinical history of the patients. Coronary arteries were divided into 17 segments according to the modified American Heart Association classification (21). The presence of coronary plaques was evaluated visually using axial images and curved multiplanar reconstructions. One coronary plaque was assigned per coronary segment. Plaques were classified as obstructive and nonobstructive using a 50% threshold of luminal narrowing. As shown in Fig. 1, three types of plaques were identified: 1) noncalcified plaques are plaques having a lower density compared with the contrast-enhanced vessel lumen, 2) calcified plaques are plaques with high density, and 3) mixed plaques are plaques with noncalcified and calcified elements within a single plaque. The presence of coronary plaques on MSCT, the presence of obstructive CAD in general and if located in the left main/left anterior descending (LAD) coronary artery, and the presence of obstructive CAD in one vessel (single-vessel disease) or two or three vessels (multivessel disease) were evaluated. For each patient, the number of diseased coronary segments (segments containing plaques or previously implanted stents) and the number of coronary segments with nonobstructive as well as obstructive plaques were determined. Also, the numbers of segments with, respectively, noncalcified, mixed, and calcified plaques were determined.

Categorical variables were expressed as numbers (percentages) and compared between groups with a χ² test. Continuous variables were expressed as means ± SD and compared with the two-tailed t test for independent samples. When not normally distributed, continuous variables were expressed as medians (interquartile range) and compared using nonparametric Mann-Whitney tests.

To determine the relationship between plaque characteristics and the presence of diabetes, linear regression analysis was performed when the dependent variable was continuous, and logistic regression analysis was performed when the dependent variable was categorical. First, univariate analysis was performed, followed by multivariate analysis with correction for the following variables: age, sex, risk factors for CAD, clinical presentation (typical angina pectoris or atypical chest pain together with the presence of multiple CAD risk factors), and use of statins.

P < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS software (version 12.0; SPSS, Chicago, IL) and SAS software (release 6.12; SAS Institute, Cary, NC).

Baseline characteristics of patients with diabetes and without diabetes are provided in Table 1. In total, 215 patients (136 [63%] men; age [mean ± SD] 58 ± 11 years) were included, of whom 86 (40%) were patients with known type 2 diabetes. Ninety-six (45%) patients used statins (41 [48%] patients with diabetes, 55 [43%] without diabetes, P = 0.47), 91 (42%) used aspirin, 78 (36%) used β-blockers, and 69 (32%) used ACE inhibitors. Patients with diabetes were significantly younger compared with patients without diabetes and had a higher mean BMI and lower prevalence of previous CAD.

Coronary plaque characteristics on MSCT in the entire population and in patients presenting with or without diabetes are shown in Table 2 and Fig. 2. After exclusion of 64 (2%) segments of nondiagnostic quality (n = 11 small caliber, n = 45 with motion artifacts due to elevated heart rate, and n = 8 with increased signal-to-noise ratio), a total of 2,941 coronary segments were included in the analysis. CAD was completely absent on MSCT in 41 (19%) patients. In the remaining 174 (81%) patients, 917 (31%) segments with plaques (n = 871 [30%]) or previously implanted stents (n = 46 [1%]) were observed. Of segments containing plaques, 675 (77%) showed nonobstructive and 196 (23%) showed obstructive CAD. In general, noncalcified plaques were observed in 204 (23%) segments, mixed plaques in 271 (31%) segments, and calcified plaques in 396 (46%) segments.

As can be derived from Table 2, the average number of diseased segments was higher in patients with diabetes (4.9 ± 3.5) compared with nondiabetic patients (3.9 ± 3.2) (P = 0.03). In particular, nonobstructive coronary plaques were more frequently observed on MSCT in the former population (Fig. 2A). In addition, CAD tended to be more severe in diabetic patients as both left main/LAD coronary artery disease and multivessel disease were more frequently diagnosed, although the difference did not reach statistical significance.

For plaque types, however, significant differences were observed between diabetic and nondiabetic patients as patients with diabetes presented with significantly more segments containing noncalcified plaques (1.3 ± 2.0 vs. 0.7 ± 1.2, P = 0.005) as well as calcified plaques (2.3 ± .8 vs. 1.5 ± 2.1, P = 0.02). Accordingly, the relative distribution of plaque types, as illustrated in Fig. 2B, also differed because plaques in patients with diabetes were more frequently either noncalcified (114 of 406 [28%] vs. 90 of 465 [19%]) or calcified (198 of 406 [49%] vs. 198 of 465 [43%]). In contrast, plaques in patients with diabetes were less frequently mixed (94 of 406 [23%] vs. 177 of 465 [38%]) (P < 0.0001).

The results of uni- and multivariate analyses of the correlation between MSCT plaque characteristics and the presence of diabetes are depicted in Table 3. After correction for baseline characteristics, the correlation of the number of diseased coronary segments as well as the number of segments with nonobstructive plaques and the presence of diabetes remained statistically significant. For plaque type, both the number of coronary segments with noncalcified and calcified plaques remained significantly correlated with diabetes.

In the present study, differences in coronary plaque characteristics between patients with and without diabetes were observed using MSCT coronary angiography. A significant, positive correlation between the presence of diabetes and coronary plaque extent was demonstrated. In particular, diabetes was associated with an increased number of nonobstructive plaques, indicating more diffuse CAD compared with that in patients without diabetes. Also, differences in the distribution of coronary plaque types were observed, with diabetic patients showing more noncalcified and calcified plaques and less mixed plaques.

In the present study, a larger plaque burden was observed in patients with diabetes. Similar observations have been reported in previous invasive as well as postmortem studies (22,23). Nicholls et al. (23) recently reported observations in 654 subjects (including 128 with diabetes) using intravascular ultrasound; the authors demonstrated that diabetes was a strong, independent predictor of percent plaque volume and total atheroma volume, indicating that diabetes appears to be associated with a substantial increase in (diffuse) plaque burden.

In addition, diabetes was associated with more nonobstructive plaques in the current study. This has also been observed in studies using invasive coronary angiography (8,9). The increased total plaque burden may be related to the increased event rate, as observed in diabetic patients. Moreover, it has been suggested that plaque rupture may occur often in nonobstructive lesions, referred to as vulnerable plaques (24–27). Many of these nonobstructive lesions will not be associated with stress-inducible ischemia, resulting in normal results on functional imaging tests, such as nuclear imaging or stress echocardiography (28,29). Whether the larger total plaque burden and the increased prevalence of nonobstructive lesions in diabetic patients translates into a higher event rate remains to be determined in future studies.

Another important finding of the present study is the difference in distribution of different coronary plaque types between patients with and without diabetes. Relatively more noncalcified and calcified plaques were observed in patients with diabetes. At the same time, the proportion of mixed plaques (possibly regarded as an intermediate phase of coronary plaque development) was significantly lower in patients with diabetes. Accordingly, these observations could suggest a more rapid development of atherosclerosis in the presence of diabetes, with faster progression from noncalcified lesions to completely calcified lesions. A faster progression of atherosclerosis in patients with diabetes had been suggested previously on the basis of event rates in patients undergoing nuclear perfusion imaging (6,30). In the general population, a normal perfusion scan is associated with a low (<1%) hard event rate, which is sustained over long-term follow-up. In patients with diabetes, the hard event rate is equally low in the first 2 years in patients with a normal perfusion scan, but an increased event rate (despite the initial normal myocardial perfusion scan) is observed after 2 years of follow-up. This observation has been considered to be related to a faster progression of CAD in diabetic patients.

The increased prevalence of both noncalcified and calcified plaques also may have implications for calcium scoring. In a recent study by Raggi et al. (31), 10,377 asymptomatic individuals (including 903 patients with diabetes) were followed for a period of 5 ± 3.5 years after coronary calcium scoring with electron- beam computed tomography. Higher mortality was observed in diabetic patients compared with nondiabetic patients despite similar coronary calcium scores, a finding that was observed for every level of coronary calcification. The authors hypothesized that the difference in prognosis in diabetic and nondiabetic patients despite similar calcium load could be attributed to the presence of extensive diffuse noncalcified atherosclerosis, which could not be detected by calcium scoring. In line with these suggestions, the current study indeed demonstrated the presence of diffuse atherosclerosis with a significantly higher amount of noncalcified coronary plaques in the diabetic patients. Accordingly, calcium scores may underestimate total coronary plaque burden to a higher extent in patients with diabetes, and, thus, MSCT coronary angiography may have substantial incremental value over coronary calcium scoring, although this concept needs further study.

This study is a comparative study, describing coronary atherosclerosis in patients with and without diabetes. Examinations were performed at a single time point and were not repeated over time. Also, MSCT angiograms were evaluated visually as no reliable quantitative algorithms are currently available. Two scanner generations (16- and 64-slice MSCT) were used during the study, which could have affected the accuracy of detection of different plaque types. Follow-up data are not yet available, and these data are needed to determine whether the observations on MSCT may provide prognostic information and may potentially be used to identify diabetic patients at increased risk. Finally, patients in the present study were referred for noninvasive cardiac evaluation of chest pain with known or suspected CAD. Accordingly, the findings may not be applicable to asymptomatic diabetic patients.

In addition, several limitations of MSCT in general should be mentioned. MSCT is still associated with an elevated radiation dose, and administration of contrast media is also required. Finally, the presence of ischemia cannot be determined on MSCT, and abnormal MSCT findings should ideally be combined with functional data.

In summary, differences in coronary plaque characteristics on MSCT were observed between patients with diabetes and without diabetes. Diabetes may be associated with a higher coronary plaque burden as determined on MSCT. Also, more noncalcified and calcified plaques in combination with less mixed plaques were observed in patients with diabetes, possibly reflecting faster progression of CAD in the presence of diabetes. MSCT may be used to identify differences in coronary plaque burden, which may eventually be useful for risk stratification of patients with diabetes.

G.P. is supported financially by a training fellowship grant from the European Society of Cardiology, by a Huygens scholarship, and by Toshiba Medical Systems Europe. J.D.S. is supported financially by The Netherlands Heart Foundation (grant 2002B105). J.W.J. is an established investigator of the Netherlands Heart Foundation (grant 2001T032). J.J.B. has received research grants from GE Healthcare and BMS Medical Imaging.