Original Research  n  Cardiac

Imaging

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Abdominal Aortic Atherosclerosis at MR Imaging Is Associated with Cardiovascular Events: The Dallas Heart Study1 Christopher D. Maroules, MD Eric Rosero, MD Colby Ayers, MS Ronald M. Peshock, MD Amit Khera, MD, MSc

Purpose:

To determine the value of two abdominal aortic atherosclerosis measurements at magnetic resonance (MR) imaging for predicting future cardiovascular events.

Materials and Methods:

This study was approved by the institutional review board and complied with HIPAA regulations. The study consisted of 2122 participants from the multiethnic, populationbased Dallas Heart Study who underwent abdominal aortic MR imaging at 1.5 T. Aortic atherosclerosis was measured by quantifying mean aortic wall thickness (MAWT) and aortic plaque burden. Participants were monitored for cardiovascular death, nonfatal cardiac events, and nonfatal extracardiac vascular events over a mean period of 7.8 years 6 1.5 (standard deviation [SD]). Cox proportional hazards regression was used to assess independent associations of aortic atherosclerosis and cardiovascular events.

Results:

Increasing MAWT was positively associated with male sex (odds ratio, 3.66; P , .0001), current smoking (odds ratio, 2.53; P , .0001), 10-year increase in age (odds ratio, 2.24; P , .0001), and hypertension (odds ratio, 1.66; P = .0001). A total of 143 participants (6.7%) experienced a cardiovascular event. MAWT conferred an increased risk for composite events (hazard ratio, 1.28 per 1 SD; P = .001). Aortic plaque was not associated with increased risk for composite events. Increasing MAWT and aortic plaque burden both conferred an increased risk for nonfatal extracardiac events (hazard ratio of 1.52 per 1 SD [P , .001] and hazard ratio of 1.46 per 1 SD [P = .03], respectively).

Conclusion:

MR imaging measures of aortic atherosclerosis are predictive of future adverse cardiovascular events.  RSNA, 2013

q

1

 From the Departments of Radiology (C.D.M., R.M.P.), Anesthesiology (E.R.), Clinical Sciences (C.A.), and Internal Medicine, Division of Cardiology (R.M.P., A.K.) and the Donald W. Reynolds Cardiovascular Clinical Research Center (R.M.P., A.K.), University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8896. Received December 8, 2012; revision requested January 17, 2013; final revision received February 1; accepted February 20; final version accepted April 1. Address correspondence to C.D.M. (e-mail: [email protected]).  RSNA, 2013

q

84

radiology.rsna.org  n  Radiology: Volume 269: Number 1—October 2013

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

T

he burden of coronary and cerebrovascular atherosclerotic disease has been closely linked to adverse cardiovascular outcomes, but less is known about asymptomatic aortic atherosclerosis. Aortic atherosclerosis has a long subclinical phase and can be assessed by measuring aortic plaque burden and aortic wall thickness (1,2). Both aortic plaque burden and aortic wall thickness can be directly and noninvasively quantified with good reproducibility by using high-spatial-resolution, black-blood magnetic resonance (MR) imaging (3–5). Noninvasive screening tests for subclinical aortic atherosclerosis may offer new insights into atherogenesis within different vascular beds and may help refine traditional cardiovascular risk stratification guidelines. Prior studies in which aortic atherosclerosis has been related to coronary artery disease, cerebrovascular disease, and peripheral arterial disease have provided conflicting results and are limited by the examination of only calcified or discrete raised plaques (6–9). To our knowledge, no large-scale study has been conducted to date to evaluate the independent predictive value of MR imaging measures of aortic atherosclerosis, including mean aortic wall thickness (MAWT) and aortic plaque burden, on future cardiovascular events. These two aortic phenotypes may represent different phases or manifestations of atherosclerosis, with differing clinical implications. The Dallas Heart Study is a multiethnic, population-based probability sample of Dallas County (10). Participants in the Dallas Heart Study underwent extensive phenotypic characterization with MR imaging, including assessment for abdominal aortic wall thickness and

Advances in Knowledge nn MR imaging measures of aortic atherosclerosis are predictive of future adverse cardiovascular events. nn Both aortic plaque burden and mean aortic wall thickness are significant predictors of nonfatal, extracardiac vascular events.

plaque prevalence. We hypothesize that increasing MAWT and aortic plaque burden confer increased risk for future adverse cardiovascular events. The purpose of this study was to determine the value of two abdominal aortic atherosclerosis measurements with MR imaging for prediction of future adverse cardiovascular events.

Materials and Methods Study Participants and Baseline Examination The Dallas Heart Study is a longitudinal, multiethnic, population-based probability sample of Dallas County residents. Details of the study design have been described previously (10). Briefly, from an initial cohort of 6101 subjects aged 18– 65 years who participated in a detailed survey in regard to medical and family history, 3399 subjects aged 30–65 years returned for a clinical visit involving blood and urine sampling, and 2971 subjects subsequently underwent abdominal aortic MR imaging. Participants with baseline cardiovascular disease (defined as prior stroke, myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting) were excluded. Of the remaining participants, 210 were lost to follow-up, as they were unable to be contacted after their imaging visit and were thus excluded from the study. The current study consisted of 2122 Dallas Heart Study participants who successfully underwent abdominal aortic MR imaging for both phenotypes and were followed throughout the surveillance period. Written informed consent was obtained from all subjects, and the study was approved by the institutional review board of the University of Texas Southwestern Medical Center. The study also conformed to regulations of the Health Insurance Portability and Accountability Act. Implication for Patient Care nn The presence of a thicker aorta at MR imaging indicates increased risk for future adverse cardiovascular events.

Radiology: Volume 269: Number 1—October 2013  n  radiology.rsna.org

Maroules et al

The variable definitions have been described previously (11). Baseline demographics, medical and family history, anthropometric measurements, and laboratory data were ascertained from the initial clinical encounter of the Dallas Heart Study cohort. Hypertension was defined as systolic blood pressure of 140 mm Hg or higher, diastolic blood pressure of 90 mm Hg or higher, or use of baseline blood pressure–lowering medication. Hypercholesterolemia was defined as fasting low-density lipoprotein cholesterol level of 100 mg/dL (2.59 mmol/L) or higher or fasting total cholesterol level of 200 mg/dL (5.18 mmol/L) or higher. Diabetes mellitus was defined as a fasting glucose level of 125 mg/dL (6.9 mmol/L) or higher or use of hypoglycemic medications. Body mass index was calculated by dividing weight in kilograms by height in square meters. Subject race or ethnicity was determined by self-reporting.

MR Imaging Protocol Participants underwent abdominal MR imaging with a 1.5-T whole-body system (Intera; Philips Medical Systems, Best, the Netherlands) (12). Six transverse sections of the infrarenal abdominal aorta were obtained by using a freebreathing, electrocardiogram-gated, T2weighted turbo spin-echo (black-blood) sequence. A single presaturation pulse was used to minimize periadventitial fat.

Published online before print 10.1148/radiol.13122707  Content codes: Radiology 2013; 269:84–91 Abbreviations: HDL = high-density lipoprotein MAWT = mean aortic wall thickness SD = standard deviation Author contributions: Guarantors of integrity of entire study, C.D.M., R.M.P., A.K.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, C.D.M., E.R., A.K.; clinical studies, C.D.M., E.R., R.M.P., A.K.; statistical analysis, C.D.M., C.A., R.M.P., A.K.; and manuscript editing, C.D.M., E.R., R.M.P., A.K. Conflicts of interest are listed at the end of this article.

85

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

Additional parameters included a repetition time of 3 heartbeats, echo time of 45 msec, turbo spin-echo factor of 14, field of view of 264 3 330 mm, and matrix size of 256 3 512 (in-plane resolution of 1.03 3 0.64 mm). The most superior axial section was positioned perpendicular to the infrarenal abdominal aorta, just inferior to the origin of the renal arteries. Each section was centered on the abdominal aorta, and signal from anatomic structures at the periphery of the section was minimized by using saturation bands. Section thickness was 5 mm, and intersection gap was 10 mm. Image acquisition involved the use of a commercial four-element abdominal phased-array receiver coil. Mean imaging time was 6 minutes.

MAWT and Aortic Plaque Burden Measurements Images were transferred to a commercial workstation computer and analyzed by using MR analytic software (Magnetic Resonance Analytical Software Systems, MASS, version 4.2; Medis Medical Imaging Systems, Leiden, the Netherlands). After magnification to 400%, luminal and adventitial aortic contours were traced manually by observers who had more than 3 years of experience with advanced aortic imaging (C.D.M. and R.M.P.). Observers were blinded to participant identifiers. Images were excluded from analysis if the signal-to-noise ratio was insufficient to allow full delineation of the aortic contours (n = 42). Aortic plaque was identified by the presence of focal high signal intensity within the aortic wall, associated with luminal protrusion and/or focal wall thickening, on at least one image (3,4). As described previously, aortic plaque burden was calculated with the following equation: [Atot/Atotadv] · 100, where Atot is total aortic plaque area and Atotadv is total adventitial aortic area from all analyzable sections (4). MAWT was calculated as the mean distance between the adventitial and luminal contours. The areas enclosed by the adventitial boundary and the luminal boundary were calculated with the analytical software for each section. Area wall thickness for each section was calculated as the difference between 86

Maroules et al

the radius of a circle with an area equal to that enclosed by the adventitial boundary and the radius of a circle with an area equal to that enclosed by the luminal boundary, as described previously (5). The MAWT for each participant was calculated as the sum of area wall thickness measurements for each section, divided by the number of sections. Prior studies demonstrate good reproducibility of aortic plaque burden and MAWT measurements (4,5).

Coronary Artery Calcium Measurements Coronary artery calcium was measured by using an electron-beam computed tomography (CT) scanner (C-150XP; Imatron, South San Francisco, Calif), with results expressed in Agatston units (13). Duplicate scans were performed 1–2 minutes apart, and the results were averaged. A score of more than 10 Agatston units was selected as a data-derived threshold to define the presence of coronary artery calcium to maximize signalto-noise ratio and reproducibility (13). Cardiovascular Events and Surveillance Death events were ascertained through December 31, 2009, for all subjects in the Dallas Heart Study by using the National Death Index (14). Deaths were classified as cardiovascular in nature if they included International Statistical Classification of Diseases, 10th Revision, codes I00–I99 (n = 53) (15). Subjects were contacted annually to participate in a detailed health survey in regard to interval nonfatal cardiovascular events. In addition, all subjects were tracked quarterly for hospital admissions by using the Dallas–Fort Worth Hospital Council Data Initiative database, which includes hospital claims data for 77 hospitals across 28 counties in North Texas and represents 90% of the health care market volume in this region (16). This database is updated quarterly and consists of more than 23 million patient encounters and more than 7 million identified patients, starting in 1998. Up to 25 International Statistical Classification of Diseases, 9th Revision, diagnosis codes and 25 procedure codes were provided for each inpatient hospitalization. More than 90% of subjects from the initial

imaging visit were followed for nonfatal events with these data sources. Primary records were requested for all suspected cardiovascular events, and these events were each adjudicated separately by two cardiologists. Nonfatal cardiovascular events assessed included (a) nonfatal cardiac events (nonfatal myocardial infarction; hospitalization for unstable angina; coronary revascularization, including percutaneous revascularization and coronary artery bypass graft surgery; hospitalization owing to congestive heart failure; and hospitalization for atrial fibrillation) and (b) nonfatal extracardiac vascular events (nonfatal stroke, transient ischemic attack, cerebrovascular revascularization, peripheral arterial revascularization, and abdominal aortic aneurysm repair).

Statistical Methods Continuous data are presented as mean values with standard deviations (SDs), and categorical data are presented as proportions. Percentiles of MAWT were calculated for the entire population, and this variable was evaluated in quartiles. Multivariate logistic regression was used to assess the associations between individual cardiovascular risk factors and the presence of aortic plaque or increased MAWT. The joint associations of aortic plaque burden and increased MAWT with cardiovascular risk factors were evaluated by using the x2 test for trend. Cumulative incidence curves were constructed for incident cardiovascular events, and the data were analyzed according to the presence or absence of aortic plaque and according to sex- and race-specific MAWT quartiles (5). The primary end point of the present study was incident composite cardiovascular events (nonfatal cardiac events, nonfatal extracardiac vascular events, and death due to cardiovascular disease). Secondary end points were incident nonfatal cardiac events, nonfatal extracardiac vascular events, and death due to cardiovascular disease. All 210 subjects who had no contact after the imaging visit were excluded from the analyses that involved nonfatal end points, as their follow-up time was 0 for these end points. However, these subjects were included for

radiology.rsna.org  n Radiology: Volume 269: Number 1—October 2013

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

the subanalysis of cardiovascular death, as follow-up was complete for this endpoint component. Cox proportional hazards regression was used to assess the independent associations of aortic plaque burden and MAWT for incident cardiovascular events after adjustment for age, sex, ethnicity, body mass index, hypertension, diabetes, cigarette smoking, hypercholesterolemia, and low high-density lipoprotein (HDL) cholesterol level. Aortic plaque burden was analyzed as both a categorical variable (present vs absent) and a continuous variable normalized to aortic area. Since aortic plaque burden was highly skewed with a large point-mass of zero (40% of the cohort), Martingale residuals suggested that a log transformation was appropriate. To take into account the point-mass of zero and to circumvent the fact that the natural log of zero is undefined, we added an offset of 0.009 to every value, which represents the lowest positive aortic plaque percentage in our data set. MAWT was also analyzed as both a continuous variable and a categorical variable (sex- and race-specific quartiles). The contribution of MAWT to models containing coronary artery calcium score fashioned as log (coronary artery calcium score + 1) and traditional risk factors was assessed by using the likelihood ratio x2 statistic (17). Net reclassification index was calculated to assess for the incremental prognostic value of MAWT beyond traditional cardiovascular risk factors. Because there were no established thresholds of risk prediction from the Cox models, we calculated a categoryless net reclassification index from time-dependent models, with confidence limits assessed with bootstrapping (18). Raw percentages of primary and secondary end points were assessed with the Cochran-Armitage trend test. A value of P , .05 was considered to indicate a significant difference. All statistical analyses were performed with software (SAS version 9.0; SAS Institute, Cary, NC).

Results The study population consisted of 2122 Dallas Heart Study participants, with a mean age of 44 years 6 10 (standard

Maroules et al

Table 1 Baseline Characteristics of the Study Population and Those Lost to Follow-up Characteristic Mean age (y)* Male sex Ethnicity   African American  White  Hispanic  Other Hypertension Diabetes Current smoking Hypercholesterolemia Low HDL cholesterol level Family history of myocardial infarction Mean body mass index (kg/m2)* Mean C-reactive protein level (mg/L)† Coronary artery calcium score .10

Current Study Subjects (n = 2122)

Subjects Lost to Follow-up (n = 210)

44 6 10 934 (44.0)

42 6 9 118 (56.2)

.008 .001

997 (47.0) 721 (34.0) 340 (16.0) 42 (2.0) 637 (30.0) 212 (10.0) 552 (26.0) 255 (12.0) 828 (39.0) 658 (31.0) 30 6 7 4.8 6 5.3 424 (20.0)

97 (46.2) 34 (16.2) 76 (36.2) 6 (3.0) 65 (31.0) 32 (15.2) 78 (37.1) 12 (5.7) 92 (44.0) 55 (26.2) 29 6 7 5.1 6 5.4 34 (16.2)

,.0001 ,.0001 ,.0001 ,.0001 .729 .041 .0004 .004 .15 .139 .08 .35 .327

P Value

Note.—Data are numbers of patients, except where otherwise indicated. Numbers in parentheses are percentages. Percentages were rounded. * Data are means 6 SDs, except where otherwise indicated. †

To convert to Système International units in nanomoles per liter, multiply by 9.524. Data are means 6 SDs, except where otherwise indicated.

deviation [SD]); 1188 women [56.0%]) who were followed throughout the surveillance period. Baseline characteristics of the study population and those lost to follow-up (n = 210) are shown in Table 1 . Those who were lost to follow-up had a higher prevalence of diabetes and smoking but trended toward a lower prevalence of hypercholesterolemia and family history of myocardial infarction and had a similar prevalence of coronary artery calcium. Thus, it is less likely that these subjects had a higher incidence of the cardiovascular end point, resulting in their being lost to follow-up. Aortic plaque was observed in 38.9% of the study population (825 of 2122 subjects). Multivariate analyses for risk factors associated with prevalent aortic plaque and highestquartile MAWT are displayed in Table 2. Evaluated jointly, 17.0% of subjects (361 of 2122) had both prevalent aortic plaque and highest-quartile MAWT, 21.0% (446 of 2122) had aortic plaque alone, 8.0% (170 of 2122) had highest-quartile MAWT alone, and 55.0% (1167 of 2122) had neither. Age, male sex, current smoking,

Radiology: Volume 269: Number 1—October 2013  n  radiology.rsna.org

and hypertension were positively associated with both measures of aortic atherosclerosis. Hypercholesterolemia was positively associated with aortic plaque, and body mass index of more than 30 kg/ m2 was negatively associated with aortic plaque, but neither was associated with MAWT. Low HDL cholesterol level was positively associated with MAWT but not with aortic plaque burden. A total of 143 subjects (6.7%) experienced a cardiovascular event over a mean period of 7.8 years 6 1.5. Of these, 34 subjects (1.6%) experienced a fatal event, 73 subjects (3.4%) experienced a nonfatal cardiac event, and 46 subjects (2.2%) experienced a nonfatal extracardiac vascular event. Kaplan-Meyer plots demonstrate cumulative incidence of composite cardiovascular events based on MAWT quartile and aortic plaque burden (Fig 1). Highest-quartile MAWT and prevalent aortic plaque were associated with the composite outcome (P , .0001 for both). After multivariate adjustment, increasing MAWT quartile trended toward an increased risk of composite 87

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

Maroules et al

Table 2 Multivariate Regression Analyses of Associations between Cardiac Risk Factors and Aortic Atherosclerosis Aortic Plaque Burden, Present vs Absent Variable Age (10-year increase) Male sex African American ethnicity Hypertension Diabetes Current smoking Hypercholesterolemia Low HDL cholesterol level Family history of myocardial infarction Body mass index . 30 kg/m2 C-reactive protein level .3 mg/L*

MAWT, 4th Quartile vs 1st Quartile

Odds Ratio

95% CI

P Value

Odds Ratio

95% CI

P Value

1.08 1.30 0.91 1.52 1.28 2.19 1.41 1.03 0.87 0.79 1.14

1.06, 1.09 1.06, 1.59 0.73, 1.14 1.20, 1.91 0.92, 1.79 1.76, 2.73 1.05, 1.89 0.84, 1.27 0.70, 1.08 0.63, 0.98 0.92, 1.42

,.0001 .01 .41 .0004 .15 ,.0001 .02 .79 .21 .03 .22

2.24 3.66 1.17 1.66 1.24 2.53 1.26 1.59 1.20 1.19 1.01

1.94, 2.57 2.87, 4.67 0.91, 1.51 1.29, 2.15 0.87, 1.77 1.97, 3.24 0.92, 1.73 1.25, 2.02 0.94, 1.52 0.93, 1.53 0.79, 1.3

,.0001 ,.0001 .23 .0001 .24 ,.0001 .15 .0001 .15 .18 .92

Note.—CI = confidence interval. *To convert to Système International units in nanomoles per liter, multiply by 9.524.

cardiovascular events (adjusted P for trend = .06) (Table 3). When evaluated as a continuous variable, MAWT was associated with increased risk for composite cardiovascular events (hazard ratio, 1.28 per 1 SD increase, equivalent to 0.33-mm increase; P = .001) and nonfatal extracardiac vascular events (hazard ratio, 1.52 per 1 SD; P , .001) (Fig 2). Sensitivity analyses were performed, which included the 210 subjects lost to follow-up, with censoring of these individuals at the end of the study period. These analyses had no significant effect on the results. Increasing MAWT trended toward an increased risk of cardiovascular death, but this relationship did not reach significance (hazard ratio, 1.28 per 1 SD; P = .07). After further adjustment for coronary artery calcium score in the multivariable model, MAWT remained independently associated with composite cardiovascular events (hazard ratio, 1.21 per 1 SD; P = .02). The likelihood ratio x2 statistic was also modestly improved when MAWT was added to the model, including risk factors and coronary artery calcium score (x2 = 5.42, P = .02). However, the C statistic for the risk factor and coronary artery calcium score model (0.826) was unchanged with the addition of MAWT 88

(P = .9). Similarly, when MAWT was added to a model with traditional risk factors, there was no improvement in the C statistic or in clinical risk categorization by using the category-free net reclassification index metric. Aortic plaque burden, in contrast, was not associated with increased risk for composite events after multivariable adjustment when analyzed as either a categorical (present vs absent) or continuous variable (Table 3 and Fig 2, respectively). However, aortic plaque burden conferred a significantly increased risk for nonfatal extracardiac vascular events (hazard ratio, 1.46 per 1 SD increase, equivalent to 4.2-unit increase; P = .03) (Fig 2). Including the 210 subjects lost to follow-up did not significantly change these findings.

Discussion In a large, multiethnic population free of cardiovascular disease, we demonstrated that MR imaging measures of aortic atherosclerosis are predictive of future adverse cardiovascular events. Our results suggest that MAWT but not aortic plaque burden is a significant independent predictor of composite cardiovascular events, and both MAWT and aortic

plaque burden are significant independent predictors of extracardiac events. These results validate the prognostic implications of these subclinical aortic atherosclerosis measures and clarify the relevance of these measures for specific clinical manifestations of cardiovascular disease. Atherosclerosis imaging constitutes a new paradigm in cardiovascular risk stratification, beyond traditional risk factors and biomarkers (2,19,20). Plaque and arterial wall thickness measurements obtained with MR imaging are promising for the quantification of atherosclerosis (3,20–25). In prior studies, aortic atherosclerosis has been associated with cardiovascular events and mortality. In these studies, however, investigators limited their evaluation to aortic calcification by using spine radiography or CT (6,9,17,19,26). Such analyses require exposure to ionizing radiation. Furthermore, by limiting assessment to calcified plaque, early atherogenic changes, such as vessel wall thickening and noncalcified (soft) plaque burden, may be overlooked. To our knowledge, the present study is the first population-based study on the evaluation of the prognostic value of aortic wall thickness and aortic plaque with MR imaging with respect to fatal and nonfatal cardiovascular events. Prior reports have demonstrated good reproducibility of MAWT and aortic plaque burden assessment with black-blood MR imaging (4,5,27). Results of the present study also highlight independent associations between MR imaging measures of aortic atherosclerosis and traditional cardiac risk factors. MAWT and aortic plaque burden were most strongly associated with hypertension and smoking, which corroborated observations from prior studies (28,29). Although MAWT was positively associated with body mass index, a paradoxical inverse association was observed with aortic plaque burden and body mass index. Similar inverse associations between coronary atherosclerosis and body mass index have been reported. Kovacic et al demonstrated a dichotomous relationship between body size and coronary artery calcium, where body size is positively associated with coronary calcium

radiology.rsna.org  n Radiology: Volume 269: Number 1—October 2013

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

Maroules et al

Figure 1

Figure 1:  Graphs show cumulative incidence curves for incident composite cardiovascular events based on (left) MAWT quartiles and (right) aortic plaque burden (present vs absent). AP = aortic plaque, AWT = aortic wall thickness.

during early stages of the disease but inversely associated with coronary calcium at a later stage (30). Furthermore, the Pathobiological Determinants of Atherosclerosis in Youth investigators reported a numerical decrease in abdominal aortic fatty lesion percentage with increasing body mass index, although these findings were not significant (31). It is possible that this inverse relationship may be due to reverse confounding by the welldescribed lower body mass index seen in smokers, but these results were consistent after adjustment for smoking status. These findings suggest a complex interrelationship between subclinical atherosclerosis and obesity, which may depend on the specific site and manifestation of subclinical disease. Both MAWT and aortic plaque burden were strongly associated with nonfatal extracardiac vascular events. These results are in agreement with prior observations in which aortic calcification have been related to stroke and peripheral arterial disease (6,7,19,32). MAWT, but not aortic plaque burden, was significantly associated with increased risk of composite cardiovascular events and trended toward increased risk of cardiovascular death after multivariate adjustment. Arterial wall thickness has proved useful in predicting cardiovascular disease

Table 3 Hazard Ratios for Composite Cardiovascular Events Associated with MAWT Quartiles and Aortic Plaque Burden Unadjusted Measure and Parameter MAWT   1st quartile†   2nd quartile   3rd quartile   4th quartile Aortic plaque burden   (present vs absent)

Adjusted*

Hazard Ratio

95% CI

P Value

Hazard Ratio

95% CI

P Value

… 1.28 2.57 4.98 2.47

… 0.62, 2.51 1.39, 4.73 2.80, 8.85 1.77, 3.46

… .53 .003 ,.0001 ,.0001

… 1.02 1.35 1.73 1.04

… 0.49, 2.09 0.07, 2.58 0.91, 3.27 0.72, 1.52

… .96 .37 .09 .83

Note.—CI = confidence interval. Unadjusted P for trend for MAWT was less than .0001. Adjusted P for trend for MAWT was .06. * Adjusted for age, sex, ethnicity, body mass index, hypertension, diabetes, cigarette smoking, hypercholesterolemia, and low HDL cholesterol level. †

Data for this row were reference values.

risk, with most attention given to carotid intima-media thickness (1,25). Differences in risk associated with MAWT and aortic plaque burden may be related to a higher sensitivity of MAWT for detection of early subclinical atherosclerosis, as remodeling of the vessel wall precedes development of discrete plaque, and small changes in the aortic wall can be detected with MR imaging (4,33). Another possibility is that MAWT and discrete aortic plaque are measures of fundamentally distinct atherogenic phenomena and thus

Radiology: Volume 269: Number 1—October 2013  n  radiology.rsna.org

may be associated with different atherosclerotic events. Finally, MAWT is a continuous measure and thus may have more power for detection of associations with our limited numbers of cardiac events. However, the null results with aortic plaque and cardiac events were consistent, even when the continuous form of aortic plaque burden was used.

Clinical Implications In a prior report, we provided population normative data for MAWT (34). 89

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

Maroules et al

Figure 2

Figure 2:  Graphs show hazard ratios for cardiovascular events associated with (left) MAWT and (right) log-transformed aortic plaque burden as continuous variables. Hazard ratios are for 1 SD increase in each measure. CVD = cardiovascular death, logAP = log-transformed aortic plaque.

We extend those data by demonstrating that the presence of a thicker aorta at MR imaging indicates increased risk of cardiovascular events. The additive information over coronary artery calcium score appears modest, and there was no increment in the C statistic by including MAWT in models that included coronary artery calcium score and traditional risk factors. There was also no statistical change in C statistic or net reclassification index when compared with traditional risk factors alone. As such, our results do not encourage screening for global cardiovascular risk by using only these aortic measures. However, when a thicker aortic wall is encountered on MR images, this finding does portend an adverse cardiovascular prognosis, and both increased MAWT and the presence of aortic plaque may be particularly informative for extracardiac atherosclerotic events. In addition, these measures 90

can be useful subclinical atherosclerosis measures to investigate the correlates and determinants of different phases and manifestations of atherosclerotic disease that are distinct from other subclinical measures (22). Indeed, in the Framingham study, the correlation coefficients between coronary, aortic, and carotid subclinical atherosclerosis measures were all less than 0.3 (19).

Limitations There were several limitations of the present study. First, a relatively small number of cardiovascular events limited the statistical power to explore associations in important subgroups. Future studies with longer surveillance periods will be necessary to evaluate age, sex, and ethnic-specific differences in the prognostic value of these measurements. Second, the primary end point of our study included composite cardiovascular events,

and although analyses according to subtype of event (cardiac vs extracardiac) were prespecified, we did not account for multiple testing. Third, this study involved the use of only one MR imaging pulse sequence to measure aortic atherosclerosis, thereby limiting the assessment of plaque composition and possibly leading to underestimation of aortic plaque burden. The use of more advanced threeand four-dimensional sequences may allow superior in vivo analysis of aortic plaque morphology and plaque burden measurement (35). Future studies are planned to evaluate the prognostic value of various plaque constituents in an effort to characterize vulnerable aortic plaque. In conclusion, MR imaging measures of aortic atherosclerosis are predictive of future adverse cardiovascular events. MAWT is independently associated with increased risk for composite cardiovascular events, and both MAWT and

radiology.rsna.org  n Radiology: Volume 269: Number 1—October 2013

CARDIAC IMAGING: Abdominal Aortic Atherosclerosis at MR Imaging

aortic plaque burden are strongly associated with nonfatal extracardiac vascular events. Disclosures of Conflicts of Interest: C.D.M. No relevant conflicts of interest to disclose. E.R. No relevant conflicts of interest to disclose. C.A. No relevant conflicts of interest to disclose. R.M.P. Financial activities related to the present article: institution received a grant from the Donald W. Reynolds Foundation. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. A.K. No relevant conflicts of interest to disclose.

References 1. Li AE, Kamel I, Rando F, et al. Using MRI to assess aortic wall thickness in the multiethnic study of atherosclerosis: distribution by race, sex, and age. AJR Am J Roentgenol 2004;182(3):593–597. 2. Natural history of aortic and coronary atherosclerotic lesions in youth: findings from the PDAY Study. Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group. Arterioscler Thromb 1993;13(9):1291–1298. 3. Jaffer FA, O’Donnell CJ, Larson MG, et al. Age and sex distribution of subclinical aortic atherosclerosis: a magnetic resonance imaging examination of the Framingham Heart Study. Arterioscler Thromb Vasc Biol 2002;22(5):849–854. 4. Maroules CD, McColl R, Khera A, Peshock RM. Assessment and reproducibility of aortic atherosclerosis magnetic resonance imaging: impact of 3-Tesla field strength and parallel imaging. Invest Radiol 2008;43(9):656–662. 5. Rosero EB, Peshock RM, Khera A, Clagett GP, Lo H, Timaran C. Agreement between methods of measurement of mean aortic wall thickness by MRI. J Magn Reson Imaging 2009;29(3):576–582. 6. Witteman JC, Kok FJ, van Saase JL, Valkenburg HA. Aortic calcification as a predictor of cardiovascular mortality. Lancet 1986;2(8516):1120–1122. 7. Reaven PD, Sacks J; Investigators for the VADT. Coronary artery and abdominal aortic calcification are associated with cardiovascular disease in type 2 diabetes. Diabetologia 2005;48(2):379–385. 8. Wong ND, Lopez VA, Allison M, et al. Abdominal aortic calcium and multi-site atherosclerosis: the Multiethnic Study of Atherosclerosis. Atherosclerosis 2011;214(2):436–441. 9. Wong ND, Gransar H, Shaw L, et al. Thoracic aortic calcium versus coronary artery calcium for the prediction of coronary heart disease and cardiovascular disease events. JACC Cardiovasc Imaging 2009;2(3):319– 326. 10. Victor RG, Haley RW, Willett DL, et al. The Dallas Heart Study: a population-based probability sample for the multidisciplinary

study of ethnic differences in cardiovascular health. Am J Cardiol 2004;93(12):1473– 1480. 11. Deo R, Khera A, McGuire DK, et al. Association among plasma levels of monocyte chemoattractant protein-1, traditional cardiovascular risk factors, and subclinical atherosclerosis. J Am Coll Cardiol 2004;44(9):1812– 1818. 12. Khera A, de Lemos JA, Peshock RM, et al. Relationship between C-reactive protein and subclinical atherosclerosis: the Dallas Heart Study. Circulation 2006;113(1):38–43. 13. Jain T, Peshock R, McGuire DK, et al. African Americans and caucasians have a similar prevalence of coronary calcium in the Dallas Heart Study. J Am Coll Cardiol 2004;44(5):1011– 1017. 14. de Lemos JA, Drazner MH, Omland T, et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304(22):2503–2512. 15. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics—2012 update: a report from the American Heart Association. Circulation 2012;125(1):e2– e220.

Maroules et al

Ethnic Study of Atherosclerosis [MESA]). Am J Cardiol 2008;102(4):491–496. 24. Botnar RM, Stuber M, Lamerichs R, et al. Initial experiences with in vivo right coronary artery human MR vessel wall imaging at 3 tesla. J Cardiovasc Magn Reson 2003;5(4):589–594. 25. Polak JF, Pencina MJ, Pencina KM, O’Donnell CJ, Wolf PA, D’Agostino RB Sr. Carotid-wall intima-media thickness and cardiovascular events. N Engl J Med 2011;365(3):213–221. 26. Karim R, Hodis HN, Detrano R, Liu CR, Liu CH, Mack WJ. Relation of Framingham risk score to subclinical atherosclerosis evaluated across three arterial sites. Am J Cardiol 2008;102(7):825–830. 27. Chan SK, Jaffer FA, Botnar RM, et al. Scan reproducibility of magnetic resonance imaging assessment of aortic atherosclerosis burden. J Cardiovasc Magn Reson 2001;3(4):331–338. 28. Oyama N, Gona P, Salton CJ, et al. Differential impact of age, sex, and hypertension on aortic atherosclerosis: the Framingham Heart Study. Arterioscler Thromb Vasc Biol 2008;28(1):155–159.

16. Dallas–Fort Worth Hospital Council. http:// www.dfwhcfoundation.org/. Accessed April 29, 2013.

29. Witteman JC, Grobbee DE, Valkenburg HA, van Hemert AM, Stijnen T, Hofman A. Cigarette smoking and the development and progression of aortic atherosclerosis: a 9-year population-based follow-up study in women. Circulation 1993;88(5 Pt 1):2156–2162.

17. Budoff MJ, Nasir K, Katz R, et al. Thoracic aortic calcification and coronary heart disease events: the multi-ethnic study of atherosclerosis (MESA). Atherosclerosis 2011;215(1):196– 202.

30. Kovacic JC, Lee P, Baber U, et al. Inverse relationship between body mass index and coronary artery calcification in patients with clinically significant coronary lesions. Atherosclerosis 2012;221(1):176–182.

18. Pencina MJ, D’Agostino RB Sr, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med 2011;30(1):11– 21.

31. McGill HC Jr, McMahan CA, Malcom GT, Oalmann MC, Strong JP. Relation of glycohemoglobin and adiposity to atherosclerosis in youth. Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group. Arterioscler Thromb Vasc Biol 1995;15(4):431–440.

19. Kathiresan S, Larson MG, Keyes MJ, et al. Assessment by cardiovascular magnetic resonance, electron beam computed tomography, and carotid ultrasonography of the distribution of subclinical atherosclerosis across Framingham risk strata. Am J Cardiol 2007;99(3):310–314. 20. Tison GH, Blaha MJ, Nasir K. Atherosclerosis imaging in multiple vascular beds— enough heterogeneity to improve risk prediction? Atherosclerosis 2011;214(2):261–263. 21. Grodin JL, Powell-Wiley TM, Ayers CR, et al. Circulating levels of matrix metalloproteinase-9 and abdominal aortic pathology: from the Dallas Heart Study. Vasc Med 2011;16(5):339–345. 22. Gupta S, Berry JD, Ayers CR, et al. Left ventricular hypertrophy, aortic wall thickness, and lifetime predicted risk of cardiovascular disease: the Dallas Heart Study. JACC Cardiovasc Imaging 2010;3(6):605–613. 23. Malayeri AA, Natori S, Bahrami H, et al. Relation of aortic wall thickness and distensibility to cardiovascular risk factors (from the Multi-

Radiology: Volume 269: Number 1—October 2013  n  radiology.rsna.org

32. Okuno S, Ishimura E, Kitatani K, et al. Presence of abdominal aortic calcification is significantly associated with all-cause and cardiovascular mortality in maintenance hemodialysis patients. Am J Kidney Dis 2007;49(3):417– 425. 33. Adame IM, van der Geest RJ, Bluemke DA, Lima JA, Reiber JH, Lelieveldt BP. Automatic vessel wall contour detection and quantification of wall thickness in in-vivo MR images of the human aorta. J Magn Reson Imaging 2006;24(3):595–602. 34. Rosero EB, Peshock RM, Khera A, Clagett P, Lo H, Timaran CH. Sex, race, and age distributions of mean aortic wall thickness in a multiethnic population-based sample. J Vasc Surg 2011;53(4):950–957. 35. Koktzoglou I, Kirpalani A, Carroll TJ, Li D, Carr JC. Dark-blood MRI of the thoracic aorta with 3D diffusion-prepared steady-state free precession: initial clinical evaluation. AJR Am J Roentgenol 2007;189(4):966–972.

91