Comparison of Carotid Plaque Score and Coronary Artery Calcium Score for Predicting Cardiovascular Disease Events: The Multi‐Ethnic Study of Atherosclerosis
Background Coronary artery calcium (CAC) predicts coronary heart disease (CHD) events better than carotid wall plaque presence; however, differences in the utility of CAC burden and carotid plaque burden across the spectrum of cardiovascular disease (CVD) events is unknown.
Methods and Results CVD, CHD and stroke/transient ischemic attack (TIA) events were evaluated prospectively in a multiethnic cohort without CVD at baseline. Carotid plaque score was determined by the number of ultrasound‐detected plaques in the common, bifurcation, and internal carotid artery segments. CAC was detected by computed tomography. Predictive values were compared using Cox proportional hazards models, C‐statistics, and net reclassification, adjusting for traditional CVD risk factors. At baseline, the 4955 participants were mean (SD) 61.6 (10.1) years old and 52.8% female; 48.9% had CAC >0 and 50.8% had at least 1 carotid plaque. After 11.3 (3.0) years of follow‐up, 709 CVD, 498 CHD, and 262 stroke/TIA events occurred. CAC score compared to carotid plaque score was a stronger predictor of CVD (hazard ratio [HR], 1.78; 95% CI, 1.16–1.98; P<0.001 vs HR, 1.27; 95% CI, 1.16–1.40; P<0.001) and CHD events (HR, 2.09; 95% CI, 1.84–2.38; P<0.001 vs HR, 1.35; 95% CI, 1.21–1.51; P<0.001). CAC score and carotid plaque score were weak predictors of stroke/TIA. CAC score had better reclassification statistics than carotid plaque score, except for stroke/TIA, which had similar predictive values.
Conclusions CAC score improved prediction, discrimination, and reclassification of CVD and CHD better than carotid ultrasound measures, although prediction and discrimination were similar for stroke/TIA.
Coronary artery calcium (CAC) and carotid artery plaque have been studied extensively as predictors of cardiovascular disease (CVD) events, to further risk stratify individuals to help identify those who will benefit most from aggressive medical and lifestyle risk reduction strategies.1, 2, 3, 4, 5, 6, 7, 8 Traditionally, CAC has been viewed as a marker primarily for coronary heart disease (CHD) risk given that the location of the calcification is in the coronary arterial bed. Proposed drawbacks of CAC scoring is the potential for it to not detect noncalcified plaques and to underestimate stroke and transient ischemic attack (TIA) risk compared with carotid plaque identification, given that the carotid artery provides cerebral blood flow.4, 9 Larger, population‐based, prospective studies have shown that despite its anatomical location, CAC presence and burden are strong predictors for CVD events, including stroke and TIA.4, 8, 10, 11 CAC presence and CAC score are superior to carotid artery intima‐media thickness (IMT) for prediction of CHD and CVD events, including stroke and TIA6, 8, 12, 13; however, carotid plaque presence is a better predictor of CVD events than carotid IMT.8, 14 Recent studies suggest that a quantitative carotid plaque score may be the best carotid ultrasound predictor of CVD risk.15 The comparative efficacy of CAC scoring and carotid plaque scoring for CVD risk prediction are unclear and have not been investigated in a single, large, multiethnic cohort with long‐term follow‐up.16 The MESA (Multi‐Ethnic Study of Atherosclerosis) is a large, ethnically diverse cohort of individuals without clinically evident CVD at baseline and with over 11 years of follow‐up. Past analyses in the MESA have focused on carotid IMT or the presence or absence of carotid plaque, but the abilities of CAC and carotid plaque scores predict CVD events have not been compared or described in detail.
Study Participants and Design
The MESA is a prospective cohort study of the prevalence, causes, and progression of subclinical CVD. The MESA is a population‐based sample of 6814 men and women aged 45 to 84 years who were free of known CVD at baseline, recruited from 6 US communities. Study objectives and design have been published previously.17 All participants gave informed consent. Protocols were approved by the institutional review boards of the field and reading centers.
This analysis was prespecified and included all MESA participants with Exam 1 CAC evaluation and follow‐up data (N=4955) who also had Exam 1 carotid plaque assessment (N=4955) measured by the University of Wisconsin Ultrasound Reading Center (Madison, WI).8 Because plaque assessment only was available for a subset of MESA participants and may introduce bias into the analysis, multiple imputation was used to account for missing data from the entire MESA cohort (N=6783) and findings were analyzed with and without imputation. Demographic, medical history, and laboratory data were obtained from July 2000 to August 2002 and have been described previously in detail.17
At Exam 1, B‐mode ultrasound was used to image the near and far walls of the right and left distal common carotid artery (CCA), carotid bulb, and proximal internal carotid (ICA) using a Logiq 700 ultrasound system (13 MHz transducer; General Electric Medical Systems, Wauwatosa, WI). The carotid bifurcations and internal carotid arteries were interrogated thoroughly at 9 MHz from both longitudinal and transverse approaches to identify the thickest regions. Images were stored on super‐VHS videotape and digitized at a high resolution and frame rates using a Medical Digital Recording (MDR) device (PACSGEAR, Pleasanton, CA) and converted into DICOM digital records. Carotid plaque presence was defined as a focal abnormal wall thickness (IMT, >1.5 mm) or a focal thickening of >50% of the surrounding IMT.18, 19 A total plaque score (range, 0–12) was calculated to describe carotid plaque burden. Of the 12 segments analyzed for each participant, 1 point per plaque was allocated for the near and far walls of each segment (CCA, bulb, and ICA) of each carotid artery that was interrogated. For example, if a participant had a plaque on the near and far wall of the bulb and the far wall of the ICA, their total plaque score would be 3. The excellent reproducibility of the University of Wisconsin Ultrasound Reading Center's carotid ultrasound measurements using MESA images has been previously described in detail.8 Carotid IMT was not included in this analysis because it has been compared to CAC in the MESA, previously, and the carotid plaque score appears to be superior to it for CVD risk prediction.8
Methods for computed tomographic (CT) scanning and interpretation have been reported previously.20 CAC was assessed at all 6 MESA sites at Exam 1 by using either a cardiac‐gated electron‐beam CT scanner (Chicago, Los Angeles, and New York Field Centers) or a multidetector CT system (Baltimore, Forsyth County, and St. Paul Field Centers). All scans were over‐read by a trained radiologist or cardiologist using an interactive scoring system.21 CAC was determined by the Agatston score and reported as a continuous variable with excellent reproducibility.21, 22
Participants were followed from the baseline examination for a median of 11.3 (3.0) years. They were contacted by telephone every 9 to 12 months to inquire about interim hospital admissions, CVD outpatient diagnoses, and deaths. Events were verified with death certificates and medical records. Two physicians, blinded to study data, independently reviewed and classified CVD events. In cases of disagreement, a mortality and morbidity committee determined the final classification. CVD was defined as CHD (definite or probable myocardial infarction, CHD death, resuscitated cardiac arrest, definite angina, and probable angina—if followed by coronary revascularization), stroke (fatal or nonfatal), or other atherosclerotic CVD death.12 Stroke and TIA were adjudicated by neurologists. Stroke was defined as a focal neurological deficit lasting 24 hours or until death with a clinically relevant lesion on brain imaging, and no nonvascular cause was identified. TIA was defined as a focal neurological deficit lasting 30 seconds to 24 hours, without brain imaging suggesting stroke. In the present analysis, the composite of TIA and stroke are reported because the number of strokes was small. A detailed description of the MESA follow‐up methods is available at http://www.mesa-nhlbi.org.
CAC score was analyzed continuously as natural log transformation (ln) of (CAC+1) because over 50% of participants had zero CAC.12 Carotid plaque was analyzed as ln (carotid plaque score +1) and as the transformed and untransformed score ranging from 0 to 12 to allow for more‐direct comparison in the current analyses and to other studies, respectively. Measures were divided by the SD to present regression coefficients in SD units, allowing for comparison between 2 measures with different units. Survival analysis was performed using a univariate approach with Kaplan–Meier analysis assessing associations between CAC and carotid plaque (predictors) and CHD, CVD, and stroke/TIA events (outcome variables). Cox proportional hazards models were utilized to assess multiple covariates on survival and to account for potential confounders. Each model was adjusted for age, sex, race/ethnicity, education (categories), family income (family income [1 if <$25 000, otherwise 0]), heart rate, body mass index, smoking (never, former, or current), total cholesterol, high‐density lipoprotein (HDL) cholesterol, lipid‐lowering medication use, presence of diabetes mellitus, systolic blood pressure, and antihypertension medication use. The strength of association for plaque score and CAC score was based on the relative size of their hazard ratios (HRs) and the corresponding chi‐square test or Z test of the HRs will be reported, as in previous MESA publications.8, 12
Receiver‐operating characteristic curves for survival models and net reclassification improvement (NRI) analyses were used to assess the improvement in risk stratification generated by plaque score and CAC score for each outcome.23 Categorical NRI analyses with bootstrapped SEs were used to assess the improvement for each outcome based on 10‐year risk of 10%.
Multiple imputation using sequential chained regression was used to account for all missing values.8 Estimates were averaged across 10 imputed data sets and Rubin's rules for combining the SEs were used for all models. The multiple imputation model included traditional CVD risk factors, demographic information, CAC scores, and the known outcomes of the entire cohort over the 11‐year follow‐up period. The percentage of missing data was 0% for the outcome and <1% for all covariates, except low‐density lipoprotein (LDL)‐cholesterol (2% missing), income (4% missing), and carotid plaque (23% missing).
Baseline characteristics of complete cases (those with carotid plaque measurements, n=4955) and all participants using multiple imputation (n=6783) are described in Table 1. At baseline, 4955 participants underwent baseline carotid ultrasound evaluation with plaque assessment and calculated plaque score by the University of Wisconsin. Participants in the complete case analysis were mean (SD) 61.6 (10.1) years old and 52.8% were female. Participants were ethnically diverse with 1924 (38.8%) white, 614 (12.4%), Chinese, 1288 (26.0%) black, and 1129 (22.8%) Hispanic participants. At least 1 carotid plaque was found in 2516 (50.8%) participants who had a mean plaque score of 2.6 (1.8; Table 2). CAC was present in 2424 (48.9%), with an average CAC score of 270.6 (519.9) units among those with detectable CAC. CVD events are shown in Table 2.
Comparison of CAC Score and Carotid Plaque Score for Predicting Events
The unadjusted Kaplan–Meier survival curves for CVD, CHD, stroke, and stroke/TIA are shown in Figure 1 by CAC score and Figure 2 by carotid plaque score. They show progressive increases in risks of these events, with increasing CAC burden and carotid plaque burden. After adjustment for traditional risk factors, CAC score had a higher HR for predicting CVD events (HR, 1.78; 95% CI, 1.16–1.98; P<0.001) compared to carotid plaque score (HR, 1.27; 95% CI, 1.16–1.40; P<0.001), though both provided incremental improvement over traditional risk factors alone (Table 3). As with CVD, similar findings for CHD events were observed for CAC score (HR, 2.09; 95% CI, 1.84–2.38; P<0.001) and carotid plaque score (HR, 1.35; 95% CI, 1.21–1.51; P<0.001). Both CAC score (HR, 1.24; 95% CI, 1.05–1.47; P=0.012) and carotid plaque score (HR, 1.15; 95% CI, 0.98–1.35; P=0.08) have a mild incremental predictive value for stroke/TIA. Neither CAC nor carotid plaque predicted stroke events alone (without TIA), but the data were limited by the low number of isolated stroke events (results not shown).
Table 3 shows that CAC score significantly improved the Harrell's C‐statistic over traditional CVD risk factors for CVD and CHD events (both P≤0.001). The carotid plaque score marginally improved the C‐statistic over CVD traditional risk factors alone for CVD (P=0.034) and CHD (P=0.049) events, but the incremental difference in the area under the curve is small. Neither CAC score nor carotid plaque score significantly improved the C‐statistic over traditional CVD risk factors alone for stroke/TIA (both P>0.10).
CAC score significantly reclassified non‐CVD events (3% [2–5%]) and CHD events (13% [5–18%]). Carotid plaque score did not consistently reclassify CVD or CHD events or nonevents (Table 3). Neither CAC score nor carotid plaque score consistently reclassified events or nonevents for stroke/TIA (Table 3).
In participants with any carotid plaques (ie, carotid plaque score >0; n=2516), the ability of CAC score to predict CVD, CHD, and stroke/TIA events remained similar to or stronger than carotid plaque score (Table 4). In those with a CAC score >0 (n=2424), carotid plaque score also predicted CVD events and CHD events, similar to or weaker than CAC score (Table 4). In analyses where both CAC score and carotid plaque score were included in the same models, CAC score consistently was a stronger predictor than the carotid plaque score for CVD and CHD events, but not for stroke/TIA events (Table 4). Results with and without multiple imputation were very similar (Table S1).
The ability of ultrasound to evaluate carotid IMT and carotid plaque presence have been compared to CAC presence previously, and carotid plaque presence appears to be a better predictor of CVD events than carotid IMT.8 However, the CAC score is a continuous variable that has not been directly compared to the carotid plaque score, a robust carotid ultrasound marker of atherosclerotic burden, in a large, multiethnic cohort with over a decade of follow‐up. Our main finding is that both the CAC score and the carotid plaque score improve prediction of CVD and CHD events compared to traditional CVD risk factors alone; however, the predictive value of the CAC score consistently is superior to the carotid plaque score for all total CVD and CHD. For prediction of stroke and TIA events, CAC and carotid plaque score performed similarly.
Autopsy studies have demonstrated that atherosclerosis is a systemic disease,24 and subclinical markers of atherosclerosis have significant overlap, supporting the notion that atherosclerosis, even at the early stages, is a systemic process. The relationships between the locations of subclinical vascular disease have been thought to be linked to the types of events they may predict and could impact which type of study providers should use to identify arterial injury. Numerous studies have shown that the degree of CAC has additional incremental value for CVD and CHD events,2, 21, 25 but more‐recent studies with longer follow‐up periods also show an association with CAC and cerebrovascular disease events.10, 11 Similarly, the carotid plaque score is associated with future CHD events despite distance between the carotid and coronary arterial beds.26, 27, 28 It has been proposed that the heterogeneity of mechanisms of cerebrovascular events is a plausible explanation for why more‐recent findings contradict the previously emphasized importance of the anatomical location of subclinical disease in regard to the type of events they predict.11, 16, 29, 30
This study highlights that although CAC score was, overall, a better predictor of CVD and CHD events, the carotid plaque score had similar predictive value for stroke/TIA events—as predicted by their anatomical location. Though inferior to CAC score for predicting CVD and CHD events, carotid ultrasound remains a relatively inexpensive, noninvasive, radiation‐free technique to augment CVD risk prediction, so it may be a more‐suitable test in children and young adults. We previously reported the incremental value of increased carotid IMT on risk prediction in the MESA,8 so it was not included in the present analyses. Also, it recently downgraded to a level III recommendation in the most recent guideline statement on the assessment of CVD risk prediction,5 though measuring carotid IMT without considering carotid plaque presence or absence already had not been recommended.18 Whereas detecting carotid plaque is less operator dependent than accurately imaging and measuring carotid IMT, measuring a CAC score is the least operator dependent of these tests.
Though the MESA is a large study designed to asses subclinical determinants of CVD, its observational design has well‐known limitations. There were a significant number of subjects who were missing carotid ultrasound assessments of plaque at Exam 1. We attempted to mitigate this though the use of multiple imputation to account for missing data and observed no significant differences between the complete case analyses and the imputed data analysis. The data on stroke and TIA are presented together because of the low number of these event types within the MESA. Finally, the units of CAC score and carotid plaque score are not equal, with CAC ranging from 0 to infinity and carotid plaque being a discrete 0 to 12 score. Regression coefficients were presented in SD units of the measure, allowing for more‐generalizable comparisons.
The CAC score improves prediction, discrimination, and reclassification of CVD and CHD risk better than carotid ultrasound measures. CAC and carotid plaque scores had similar prediction and weaker discrimination and reclassification for stroke/TIA events.
Sources of Funding
This research was supported by contracts N01‐HC‐95159 through N01‐HC‐95169 from the National Heart, Lung, and Blood Institute, grant ES015915 from the National Institute of Environmental Health Sciences, grant R831697 from the US Environmental Protection Agency, and by grants UL1‐RR‐024156 and UL1‐RR‐025005 from the National Center for Research Resources and by a T32 HL 07936 Ruth L. Kirschstein National Research Service Award from the National Heart, Lung, and Blood Institute to the University of Wisconsin–Madison Cardiovascular Research Center.
Stein is a member of the Wisconsin Alumni Research Foundation. The remaining authors have no conflicts to disclose.
Table S1. Risk Factor‐Adjusted Cox Regression Models for Predicting Incident Events With Multiple Imputation
The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. Participating MESA investigators/institutions can be found at http://www.mesa-nhlbi.org.
- ↵Nambi V, Brunner G, Ballantyne CM. Ultrasound in cardiovascular risk prediction: don't forget the plaque!. J Am Heart Assoc. 2013;2:e000180. DOI: 10.1161/jaha.113.000180.
- ↵McClelland RL, Jorgensen NW, Budoff M, Blaha MJ, Post WS, Kronmal RA, Bild DE, Shea S, Liu K, Watson KE, Folsom AR, Khera A, Ayers C, Mahabadi AA, Lehmann N, Jockel KH, Moebus S, Carr JJ, Erbel R, Burke GL. 10‐year coronary heart disease risk prediction using coronary artery calcium and traditional risk factors: derivation in the MESA (Multi‐Ethnic Study of Atherosclerosis) With Validation in the HNR (Heinz Nixdorf Recall) Study and the DHS (Dallas Heart Study). J Am Coll Cardiol. 2015;66:1643–1653.
- ↵Hecht HS. Coronary artery calcium scanning: past, present, and future. JACC Cardiovasc Imaging. 2015;8:579–596.
- ↵Goff DC Jr., Lloyd‐Jones DM, Bennett G, Coady S, D'Agostino RB Sr., Gibbons R, Greenland P, Lackland DT, Levy D, O'Donnell CJ, Robinson JG, Schwartz JS, Shero ST, Smith SC Jr., Sorlie P, Stone NJ, Wilson PW; American College of Cardiology/American Heart Association Task Force on Practice G . 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63:2935–2959.
- ↵Nambi V, Chambless L, Folsom AR, He M, Hu Y, Mosley T, Volcik K, Boerwinkle E, Ballantyne CM. Carotid intima‐media thickness and presence or absence of plaque improves prediction of coronary heart disease risk: the ARIC (Atherosclerosis Risk in Communities) study. J Am Coll Cardiol. 2010;55:1600–1607.
- ↵Nasir K, Bittencourt MS, Blaha MJ, Blankstein R, Agatson AS, Rivera JJ, Miemdema MD, Sibley CT, Shaw LJ, Blumenthal RS, Budoff MJ, Krumholz HM. Implications of coronary artery calcium testing among statin candidates according to American College of Cardiology/American Heart Association cholesterol management guidelines: MESA (Multi‐Ethnic Study of Atherosclerosis). J Am Coll Cardiol. 2015;66:1657–1668.
- ↵Gepner AD, Young R, Delaney JA, Tattersall MC, Blaha MJ, Post WS, Gottesman RF, Kronmal R, Budoff MJ, Burke GL, Folsom AR, Liu K, Kaufman J, Stein JH. Comparison of coronary artery calcium presence, carotid plaque presence, and carotid intima‐media thickness for cardiovascular disease prediction in the Multi‐Ethnic Study of Atherosclerosis. Circ Cardiovasc Imaging. 2015;8:e002262.
- ↵Hermann DM, Gronewold J, Lehmann N, Moebus S, Jockel KH, Bauer M, Erbel R. Coronary artery calcification is an independent stroke predictor in the general population. Stroke. 2013;44:1008–1013.
- ↵Gibson AO, Blaha MJ, Arnan MK, Sacco RL, Szklo M, Herrington DM, Yeboah J. Coronary artery calcium and incident cerebrovascular events in an asymptomatic cohort. The MESA Study. JACC Cardiovasc Imaging. 2014;7:1108–1115.
- ↵Folsom AR, Kronmal RA, Detrano RC, O'Leary DH, Bild DE, Bluemke DA, Budoff MJ, Liu K, Shea S, Szklo M, Tracy RP, Watson KE, Burke GL. Coronary artery calcification compared with carotid intima‐media thickness in the prediction of cardiovascular disease incidence: the Multi‐Ethnic Study of Atherosclerosis (MESA). Arch Intern Med. 2008;168:1333–1339.
- ↵Polak JF, Szklo M, Kronmal RA, Burke GL, Shea S, Zavodni AE, O'Leary DH. The value of carotid artery plaque and intima‐media thickness for incident cardiovascular disease: the Multi‐Ethnic Study of Atherosclerosis. J Am Heart Assoc. 2013;2:e000087. DOI: 10.1161/jaha.113.000087.
- ↵Iino R, Yokoyama N, Konno K, Naito K, Isshiki T. Impact of combined assessment of coronary artery calcium score, carotid artery plaque score, and brachial‐ankle pulse wave velocity for early coronary revascularization in patients with suspected coronary artery disease. Int Heart J. 2012;53:154–159.
- ↵Stein JH, Korcarz CE, Hurst RT, Lonn E, Kendall CB, Mohler ER, Najjar SS, Rembold CM, Post WS. Use of carotid ultrasound to identify subclinical vascular disease and evaluate cardiovascular disease risk: a consensus statement from the American Society of Echocardiography Carotid Intima‐Media Thickness Task Force. Endorsed by the Society for Vascular Medicine. J Am Soc Echocardiogr. 2008;21:93–111; quiz 189‐190.
- ↵Touboul PJ, Hennerici MG, Meairs S, Adams H, Amarenco P, Bornstein N, Csiba L, Desvarieux M, Ebrahim S, Fatar M, Hernandez Hernandez R, Jaff M, Kownator S, Prati P, Rundek T, Sitzer M, Schminke U, Tardif JC, Taylor A, Vicaut E, Woo KS, Zannad F, Zureik M. Mannheim carotid intima‐media thickness consensus (2004–2006). An update on behalf of the Advisory Board of the 3rd and 4th Watching the Risk Symposium, 13th and 15th European Stroke Conferences, Mannheim, Germany, 2004, and Brussels, Belgium, 2006. Cerebrovasc Dis. 2007;23:75–80.
- ↵Carr JJ, Nelson JC, Wong ND, McNitt‐Gray M, Arad Y, Jacobs DR Jr., Sidney S, Bild DE, Williams OD, Detrano RC. Calcified coronary artery plaque measurement with cardiac CT in population‐based studies: standardized protocol of Multi‐Ethnic Study of Atherosclerosis (MESA) and Coronary Artery Risk Development in Young Adults (CARDIA) study. Radiology. 2005;234:35–43.
- ↵Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr., Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827–832.
- ↵Mitchell JR, Schwartz CJ. Relationship between arterial disease in different sites. A study of the aorta and coronary, carotid, and iliac arteries. Br Med J. 1962;1:1293–1301.
- ↵Gaibazzi N, Rigo F, Facchetti R, Carerj S, Giannattasio C, Moreo A, Mureddu GF, Salvetti M, Grolla E, Faden G, Cesana F, Faggiano P. Differential incremental value of ultrasound carotid intima‐media thickness, carotid plaque, and cardiac calcium to predict angiographic coronary artery disease across Framingham Risk Score strata in the APRES multicentre study. Eur Heart J Cardiovasc Imaging. 2016;17:991–1000.
- ↵Katano H, Mase M, Nishikawa Y, Yamada K. Calcified carotid plaques show double symptomatic peaks according to agatston calcium score. J Stroke Cerebrovasc Dis. 2015;24:1341–1350.
- ↵Rim JH, Lee HY, Yoo SM, Jung HY, White CS. Carotid Doppler ultrasonography as a surrogate for coronary CT angiography to exclude subclinical coronary atherosclerosis in asymptomatic patients with a negative coronary calcium score. J Clin Ultrasound. 2013;41:164–170.