Normal intima-media thickness on carotid ultrasound reliably excludes an ischemic cause of cardiomyopathy

Harmony R. Reynolds, MD, a David A. Steckman, MD, a Paul A. Tunick, MD, a Itzhak Kronzon, MD, a Iryna Lobach, PhD, b and Barry P. Rosenzweig, MD a New York, NY

Background Coronary artery disease (CAD) is the most common cause of left ventricular systolic dysfunction (LVSD). Patients with ischemia as the cause of LVSD may warrant revascularization. Angiography is the most accurate method of CAD diagnosis but is invasive, expensive, and associated with some risk. Noninvasive imaging for CAD often involves expensive equipment, radiation exposure, medication, and/or contrast administration. Carotid ultrasound with measurement of intimamedia thickness (IMT) is safe and inexpensive. Carotid IMT is well correlated with the presence of CAD. We assessed the accuracy of carotid ultrasound for identification of CAD as a potential etiology of LVSD. Methods Patients with LVSD (ejection fraction 40%) of uncertain etiology referred for angiography underwent carotid ultrasound. Patients with history of myocardial infarction were excluded. Two experienced cardiologists blinded to CAD status determined common carotid artery (CCA) IMT and plaque. Significant CAD was defined as 50% stenosis of any major artery. Ischemic LVSD was defined as (1) left main and/or proximal left anterior descending coronary artery 75% or (2) 2 major arteries with 75% stenosis. Results
Mean ejection fraction was 27% 10% in 150 patients. Significant CAD was found in 64 (42.7%) and ischemic LVSD in 40 (26.7%). Carotid plaque was seen in 95 (63.3%). Mean CCA IMT was 0.9 mm in 69 (46.0%). The combination of mean CCA IMT b0.9 mm and no plaque had negative predictive value for ischemic LVSD of 98%.

Conclusions

Carotid ultrasound with IMT measurement is a valuable screening tool for excluding an ischemic etiology of LVSD when CAD is suspected. (Am Heart J 2010;159:1059-66.)

Severe coronary artery disease (CAD) is the most common cause of heart failure with systolic dysfunction.1 Coronary artery disease is present in approximately 65% of patients with left ventricular systolic dysfunction (LVSD)2 and is associated with a poorer prognosis.3,4 However, CAD may not be detected in patients with LVSD without specific testing because not all patients with CAD have chest pain, history of myocardial infarction (MI), or abnormal functional testing results.5,6 The diagnosis of atherosclerosis among patients with LVSD is important for guiding medical therapy; for example, antiplatelet agents may not be indicated in patients with LVSD who do not have atherosclerosis. The demonstration of severe CAD with chronic ischemia as a potential cause of LVSD, often termed ischemic cardiomyopathy, makes revascularization a major therapeutic option.7

From the aDepartment of Medicine, New York University School of Medicine, New York, NY, and bDepartment of Biostatistics, New York University School of Medicine, New York, NY. Clinical trial registration information: ClinicalTrials.gov no. NCT00810550. Submitted March 9, 2010; accepted March 16, 2010. Reprint requests: Harmony R. Reynolds, MD, 530 First Avenue, SKI-9R New York, NY 10016. E-mail: harmony.reynolds@nyumc.org 0002-8703/$ - see front matter 2010, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.03.026

Coronary angiography remains the most accurate method for the diagnosis of CAD and is recommended in the American College of Cardiology/American Heart Association guidelines for the management of heart failure.8 However, this invasive and expensive procedure carries a risk of complications that is increased in patients with heart failure.9 A noninvasive method of reliable diagnosis would be preferable. Carotid ultrasound for the measurement of intimamedia thickness (IMT) is fast, relatively inexpensive, and highly reproducible10 and does not involve administration of contrast material or radiation exposure. It is performed using equipment already available at most centers and is easily interpreted. Carotid IMT measurements and identification of carotid plaque correlate with CAD extent11 and are predictive of cardiovascular events.12-15 Patients with ischemic LVSD have a large burden of atherosclerotic disease, and carotid IMT is correlated with CAD burden.16 We sought to assess the accuracy of carotid ultrasound for the identification of ischemia as a potential etiology of LVSD.

Methods
This study was approved by the New York University School of Medicine Institutional Review Board. Subjects

1060 Reynolds et al

American Heart Journal June 2010

provided written informed consent. The study was funded by the New York Cardiac Center Foundation. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Coronary angiography
Angiography of all coronary arteries was performed in multiple projections as per usual clinical practice. The angiographer was blinded to the results of carotid ultrasound in cases where carotid ultrasound was performed before coronary angiography. Significant CAD was defined as 50% stenosis in any major epicardial vessel. Ischemia as a potential etiology of LVSD (ischemic LVSD) was defined as at least one of the following: (1) left main and/or proximal left anterior descending coronary artery stenosis of 75% or (2) 2 major epicardial coronary arteries with 75% diameter stenosis each, based on a classification scheme that was derived from the Duke Databank for Cardiovascular Disease.18 As noted above, we excluded patients with clinical history of MI or revascularization.

Patient population
Subjects were identified after referral for coronary angiography for the evaluation of CAD as a potential cause of LVSD defined as an ejection fraction (EF) 40% as determined by any cardiac imaging technique, including echocardiography, gated single photon emission computed tomography (SPECT), left ventricular angiography, magnetic resonance imaging, or multigated acquisition scan. Both inpatients and outpatients were included. Patients with a history of MI and those with a prior diagnosis of obstructive coronary or carotid atherosclerosis were excluded. Patients were also excluded if they were admitted to hospital with a diagnosis of ST elevation or Q wave MI, or if peak troponin during a hospitalization prompting angiography was N10 ng/mL. Patients with regional wall motion abnormalities were not excluded because the presence of regional as opposed to global systolic dysfunction does not discriminate between ischemic and nonischemic cardiomyopathy.8 These criteria were intended to exclude patients with MI as a known cause of LVSD while allowing for the inclusion of patients with mild troponin elevations not considered etiologic for LVSD. Clinical history of risk factors for atherosclerotic vascular disease was collected; hyperlipidemia was defined by patient history and/or use of at least one lipid-lowering agent in this population with no history of coronary or carotid artery disease. Fasting lipid parameters were recorded when available and performed within 3 months of enrollment.

Statistical analysis
A receiver operating characteristic curve was constructed by calculating the sensitivity and specificity with respect to patient disease status for ischemic LVSD at each possible threshold of mean common carotid IMT. The threshold value of CCA IMT for dichotomous comparison with the ischemic LVSD definition was selected based on best specificity when combined with the presence or absence of carotid plaque. Because the aim of this study was to establish the use of carotid ultrasound as a preliminary screening tool for the detection of severe CAD as per the ischemic LVSD definition, a threshold with high sensitivity was considered desirable (90% or 95% sensitivity). That is, for a given sensitivity, a carotid ultrasound parameter with highest specificity was considered the best measure to set the threshold for the disease status prediction. The threshold with best specificity in all patients was then applied to prespecified subgroups based on presence or absence of diabetes and age. Various characteristics of those patients found to have a potential ischemic etiology of LVSD and those who did not were compared using t test after testing for normality and logarithmic transformation where appropriate (Excel 2003 SP3; Microsoft, Redmond, WA). Multivariate logistic regression analysis was performed (R 2.7.1, www.r-project.org) to explore the relationship between ischemic LVSD disease status, carotid ultrasound findings (IMT and plaque), and coronary heart disease (CHD) risk factors. First, we analyzed an association between the disease status and the number of risk factors and carotid ultrasound findings. The interaction effect of the number of CHD risk factors and the carotid ultrasound findings was not significant. Furthermore, we eliminated an interaction variable from the model; and P values of the main effects of IMT measurement and the number of CHD risk factors were .003 and .006, respectively, thus providing sufficient evidence to indicate an association between these factors and the disease status. To further elucidate the relationship between the CHD risk factors and the ischemic LVSD disease status, we performed an analysis based on a logistic regression that included all risk factors, IMT measurements, age, and sex.

Carotid ultrasound
A linear array transducer (Philips model 11-7L, Bohtel, WA) was used by experienced sonographers to obtain highresolution images of the far wall of the right and left common carotid arteries, internal carotid arteries, and carotid bulbs, according to recommendations of the American Society of Echocardiography (ASE) Carotid Intima-Media Thickness Task Force.17 Sonographers and interpreting physicians were blinded to angiographic findings. At least 3 measurements were taken over a 1-cm length of each common carotid artery (CCA) segment, and these measurements on both sides were averaged to obtain the mean IMT in that segment. As per ASE recommendations, if plaque was present in the segment used for measurement of IMT, the plaque thickness was averaged into the IMT measurement.17 The presence and location of any plaque in the CCA, bulb, or internal carotid artery on the near or far wall were recorded.17 Plaque was defined as 50% increase over background IMT in any arterial segment or as a focal region with IMT 1.5 mm protruding into the vessel lumen, according to ASE recommendations.17 Carotid IMT assessed by ultrasound was verified by 2 separate trained cardiologists. In cases of disagreement for plaque or variation in mean CCA IMT of 10%, a third cardiologist reviewed the scan; and the average of values for mean CCA IMT from all readers was used as the final value.

Results
Patients were enrolled between August 2005 and October 2008. The demographics of the study

American Heart Journal Volume 159, Number 6

Reynolds et al 1061

Table I. Demographics, laboratory results, and results of imaging

All (N = 150) Demographics Age, y (mean SD) Age 50 y Age 65 y Male sex Race/ethnicity White (non-Hispanic) African American (non-Hispanic) Asian (non-Hispanic) Hispanic Other EF, % (mean SD) CAD risk factors Diabetes, % Hypertension, % Hyperlipidemia, % Current or former smoker, % No. of CAD risk factors 0 1 2 3 or 4 Laboratory tests Lipid profile Total cholesterol, mg/dL (median, IQR) HDL, mg/dL (median, IQR) LDL, mg/dL (median, IQR) Triglycerides, mg/dL (median, IQR) eGFR, mL/min (median, IQR) (n = 142) Carotid ultrasound CCA IMT, mean of right and left, mm (mean SD) Carotid plaque (n, %) Ischemic LVSD (n = 40) No ischemic LVSD (n = 110) P value

60.1 11.8 51 (22%) 33 (34%) 103 (69%) 26 (17.3%) 51 (34.0%) 11 (7.3%) 56 (37.3%) 6 (4.0%) 26.7 9.7 52 (34.9%) 115 (76.7%) 75 (50.0%) 72 (43.4%) 11 (7.3%) 29 (24.7%) 53 (33.3%) 53 (34.7%) (n = 115) 156 (130, 190) 43 (34.0, 54.0) 86 (64, 116) 93 (70, 120) 74 (58, 99) 0.91 0.31 95 (63.3%)

64.8 11.7 23 (57.6%) 5 (12.5%) 33 (83%) 11 (27.5%) 8 (20.0%) 5 (12.5%) 12 (30.0%) 4 (10.0%) 27.8 9.7 21 34 32 24 (52.5%) (85.0%) (80.0%) (61.5%)

58.3 11.4 28 (25.5%) 28 (27.3%) 70 (64%) 15 (13.6%) 43 (11.4%) 6 (5.5%) 44 (40.0%) 2 (1.8%) 26.4 9.7 31 81 43 48 (28.4%) (53.3%) (65.0%) (26.4%)

.003 b.001 .09 .03 .01

.006 b.001 .01 b.001 .002

3 (7.5%) 1 (2.5%) 9 (22.5%) 26 (65.0%) (n = 30) 153 (130, 191) 38 (33, 53) 83 (69, 117) 112 (86, 143) 66 (52, 87) 0.85 0.26 36 (90.0%)

8 (7.2%) 28 (25.5%) 44 (40.0%) 27 (24.5%) (n = 84) 158 (131, 187) 44 (36-56) 86.0 (63, 113) 92 (66,113) 79 (60, 108) 0.68 0.39 59 (53.6%)

.987 .159 .898 .013 .022 b.001 b.001

HDL, High-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular filtration rate. Ischemic versus nonischemic. Two-sided, 2-sample t test of means. m2 test of proportions. Two-sided, 2-sample Wilcoxon test.

population as well as the findings of carotid and coronary angiography are shown in Table I. Mean age was 60.1 11.8 years, 69% were male, and 35% had diabetes mellitus. The distribution of race/ethnicity represents the diverse patient population at our centers. The average EF was 26.7% 9.7%. Troponin was measured in 77 patients, of whom 62 (81%) had an abnormal value. Remaining patients underwent elective angiography, and troponin was not measured. The median peak troponin among those patients in whom it was abnormal was 0.144 ng/mL (interquartile range [IQR] 0.044-1.04, upper limit of normal 0.04 ng/mL). Left bundlebranch block was present on the electrocardiogram in 25 patients (17%).

Figure 1

Carotid ultrasonic and coronary angiographic findings One hundred forty-seven of 150 patients had interpretable common carotid IMT studies on both sides. There was 100% agreement regarding the presence or absence of plaque among interpreting physicians. Mean

Prevalence of carotid ultrasound abnormalities among all patients and those with and without ischemic LVSD.

1062 Reynolds et al

American Heart Journal June 2010

Figure 2

Diagnostic accuracy of carotid ultrasound for CAD in the entire cohort and in selected subgroups based on age and presence of absence of diabetes. These subgroups were prespecified. The age threshold of 55 years was selected to separate the youngest tertile from the older 2 tertiles of patients. *P values are based on m2 test. eGFR, Estimated glomerular filtration rate; NPV, negative predictive value; PPV, positive predictive value; SN, sensitivity; SP, specificity.

CCA IMT measurements differed by 10% in 5 (3.3%) of 150 cases; these cases were interpreted by a third physician, and all measurements were averaged to yield a final value. Mean CCA IMT was 0.91 0.31 mm. The average number of plaques per patient with plaque was 1.5. Plaque was found in 95 (63.3%), and 69 (46.0%) had mean CCA IMT 0.9 mm. The mean thickness of plaque was 2.5 1.1 mm. The percentage of patients with either plaque or CCA IMT 0.9 mm was 65.3% (n = 98). Severe CAD meeting

the ischemic LVSD definition was found in 40 patients (26.7%). Left main stenosis 75% was present in 1 patient (20.7%), proximal left anterior descending coronary artery stenosis 75% in 18 (12%), and multivessel stenosis 75% in 32 (21.3%). Some patients met N1 criterion for ischemic LVSD by angiography. All 4 patients with left main stenosis 50% to 74% met another criterion for ischemic LVSD. See Figure 1 for the proportion of patients with carotid plaque and mean CCA IMT 0.9 mm, overall and with and without criteria for ischemic LVSD by angiography. Significant

American Heart Journal Volume 159, Number 6

Reynolds et al 1063

Figure 3

True negative carotid ultrasound. Representative images of carotid ultrasound and coronary angiography in the same patient. Top panels: normal carotid ultrasound. Bottom panels: normal coronary arteries in same patient.

Figure 4

True positive carotid ultrasound for ischemic LVSD. Representative images of carotid ultrasound and coronary angiography in the same patient. Top panels: abnormal common carotid IMT with plaque in right carotid bulb. Bottom panels: severe, 3-vessel CAD in the same patient.

1064 Reynolds et al

American Heart Journal June 2010

Figure 5

Table II. Multivariate analysis of correlates of ischemic LVSD by angiography

Standard error of the regression coefficient 1.09

Variable Carotid ultrasound (mean CCA 0.9 mm and/or plaque) Hyperlipidemia Male sex eGFR* (per 10 mL/min)

Regression coefficient 2.30

P value .03

2.16 1.24 0.99

0.57 0.63 0.47

.0001 .05 .04

Variables analyzed included: age, sex, smoking, hyperlipidemia, diabetes, hypertension, eGFR, carotid ultrasound results. *Log(1 + eGFR) transformation was used to ensure that the distribution of values is Gaussian.

Receiver operating characteristic curve for mean CCA IMT versus ischemic heart failure by coronary angiography. Area under curve = 0.68, indicating suboptimal discrimination on the basis of IMT alone; as can be seen in Figure 2, the addition of plaque status improves discrimination substantially.

coronary disease (50% stenosis in any major epicardial vessel) was identified in 64 patients (42.7%) overall and in 24 of 110 patients (21.8%) who did not meet the criterion for ischemic LVSD.

Multivariate logistic regression analysis Abnormal carotid ultrasound results (based on the thresholds identified as above) remained a significant predictor of ischemic LVSD on multivariate analysis, with an odds ratio of 10 (corresponding regression coefficient 2.3, standard error = 1.09) for ischemic LVSD. Additional multivariate correlates of ischemic LVSD are presented in Table II.

Discussion
Our study shows that carotid ultrasound with measurement of IMT and identification of plaque has excellent negative predictive value (98%) for the detection of CAD severe and extensive enough to cause LVSD. Identification of carotid plaque was achieved without variance among observers, and carotid IMT measurement had very low interobserver variance in this study. It is well known that atherosclerosis is a systemic disease; and therefore, it is not surprising that increasing carotid IMT and plaque are correlated with CAD burden.11,16 The IMT threshold used in our analysis corresponds to values 90th percentile for most middle-aged men and women in the Atherosclerosis Risk in Communities study.19 Coronary artery disease remains the most common cause of LVSD in the United States; and coronary angiography remains the standard for definitive exclusion of CAD according to current guidelines, recognizing that perfusion deficits and segmental wallmotion abnormalities suggestive of CAD are commonly present in patients with a non-ischemic cardiomyopathy on noninvasive imaging.8 In an effort to avoid the risks, inconvenience, and expense of coronary angiography, clinicians often screen patients for CAD using noninvasive modalities that do not image directly for atherosclerosis, such as stress echocardiography and nuclear imaging. Each of these techniques has limitations when compared with carotid ultrasound. Gated SPECT is expensive, is time

Diagnostic accuracy of carotid ultrasound for detection of CAD The absence of plaque and/or mean CCA IMT 0.9 mm had a negative predictive value of 98% for ischemic LVSD. See Figure 2 for diagnostic accuracy of different carotid ultrasound parameters for CAD. Figures 3 and 4 demonstrate representative cases of true positive and true negative carotid ultrasound with corresponding angiographic images. The receiver operating curve constructed for different values of mean CCA IMT versus ischemic LVSD is depicted in Figure 5. Subgroup analyses There was no difference in the diagnostic accuracy of carotid ultrasound for ischemic LVSD based on the presence or absence of diabetes, or based on age b55 or 55 years (Figure 2). There was no difference in diagnostic accuracy by sex. Midleft anterior descending coronary artery lesions Among those patients who were classified as not having ischemic LVSD, there were 3 patients with 75% or greater stenosis of the middle segment of the left anterior descending coronary artery. All 3 had plaque on carotid ultrasound.

American Heart Journal Volume 159, Number 6

Reynolds et al 1065

consuming, and requires the administration of radioactive material. Dobutamine stress echocardiography requires intravenous administration of a medication with potential for adverse effects including arrhythmia and hypotension, and is likewise time consuming. Both of these techniques are less accurate in the presence of left bundle-branch block, which was present in 17% of patients in the present study. Magnetic resonance imaging is expensive, not widely available, and timeconsuming and requires highly specialized readers. Computed tomographic (CT) coronary angiography has also been evaluated for the purpose of noninvasive diagnosis of an ischemic etiology of LVSD.20 Based on the requirement for iodinated contrast, patients with serum creatinine 1.6 mg/dL were excluded from this study of CT. Of note, 11% of our study cohort would have been excluded from the CT study based on serum creatinine measurement. In clinical practice, this excluded proportion might be even higher because enrollment into the present study was based on referral for angiography and patients with impaired renal function might have been referred for noninvasive testing in lieu of or before angiography. Gated SPECT, dobutamine stress echocardiography, and delayed gadolinium-enhanced cardiovascular magnetic resonance imaging have the potential to provide information about myocardial viability, which may be associated with improved ventricular function after revascularization.21 However, their sensitivity for a potential ischemic etiology of LVSD appears to be lower than that demonstrated in the current study for carotid ultrasound.22-28 Viability techniques might be applicable after the identification of a potential ischemic etiology of LVSD by coronary angiography. Clearly, less extensive CAD than that described by the ischemic LVSD definition may cause LVSD via MI. In patients with MI, there is no uncertainty about the diagnosis of CAD. We excluded by design those patients with a history of MI or admission for acute MI. Therefore, the prevalence of a potential ischemic etiology of LVSD in our population, 26.7%, is lower than that seen in other cohorts of patients with systolic dysfunction.2,29 The sensitivity of the carotid ultrasound parameters identified in this study for significant CAD is not as high as that for the more severe degree of CAD that is included in the ischemic LVSD definition. However, the importance of significant CAD that does not meet the ischemic LVSD definition in this population is arguable. For example, single-vessel disease in the right coronary artery would be an unlikely cause of severe LVSD in the absence of a clinical infarction. Based on our findings, one might consider carotid ultrasound with IMT measurement and plaque screening as an initial evaluation for atherosclerosis for patients found to have LVSD of unknown etiology. A positive test result would then be followed by referral

for conventional (or CT) coronary angiography. Such a strategy has several advantages over other noninvasive techniques for the identification of patients with a potential ischemic etiology of LVSD. One third of our patients had neither a mean CCA IMT 0.9 mm nor plaque. Therefore, using carotid ultrasound for screening would have spared these patients coronary angiography. Among those referred for angiography, approximately 39% would have an ischemic etiology of LVSD based on our results. Using a simplified cost analysis, taking into account only the 2010 Medicare reimbursement rates for these 2 tests, the frequency of abnormal carotid ultrasound, and the diagnostic accuracy determined in this study, screening with carotid ultrasound followed by conventional angiography in patients with a positive test result would cost $127,556 for 100 patients as compared with a cost of $165,061 for routine conventional angiography in 100 patients. This cost savings would likely be amplified in a rigorous economic analysis that takes into account complications such as contrast nephropathy, access site bleeding, allergic reactions, stroke, and peripheral embolization. Eliminating other noninvasive techniques in current clinical use for screening this population would reduce costs even further.

Limitations The specificity of carotid ultrasound for ischemic LVSD is limited. However, this is consistent with its use as a screening technique. Not all patients with severe and extensive CAD on angiography have myocardial hibernation as the cause of LVSD; and it is possible that some such patients had prior, unrecognized MI. The sample size of this study, although similar to or larger than that of previous studies of noninvasive imaging for detection of CAD in patients with LVSD, is not large. The diagnostic accuracy represents interpretation by experienced cardiologists at our center and may not be generalizable to other, less experienced centers.

Conclusions
The reliable exclusion of CAD in patients with LVSD and no history of MI using a low-cost, no-risk method is highly desirable. Carotid ultrasound is a rapid, relatively inexpensive technique without known risk, applicable to nearly all patients without regard to renal function, electrocardiogram findings, or the presence of implanted devices. Carotid ultrasound with IMT measurement appears to be a valuable screening tool for the exclusion of an ischemic etiology of LVSD in patients without known CAD or history of MI. Further research is needed to determine whether carotid ultrasound with IMT measurement may be used as an initial screening test to exclude CAD as the etiology of LVSD.

1066 Reynolds et al

American Heart Journal June 2010

Acknowledgements
Dr Reynolds is a recipient of a Doris Duke Charitable Foundation Clinical Scientist Development Award.

Disclosures
The authors have no conflicts of interest.

References
1. Gheorghiade M, Bonow RO. Chronic heart failure in the United States: a manifestation of coronary artery disease. Circulation 1998; 97:282-9. 2. Heart Failure Society of America. HFSA 2006 comprehensive heart failure practice guideline. J Card Fail 2006;12:e1-e2. 3. Likoff MJ, Chandler SL, Kay HR. Clinical determinants of mortality in chronic congestive heart failure secondary to idiopathic dilated or to ischemic cardiomyopathy. Am J Cardiol 1987;59:634-8. 4. Bart BA, Shaw LK, McCants Jr CB, et al. Clinical determinants of mortality in patients with angiographically diagnosed ischemic or nonischemic cardiomyopathy. J Am Coll Cardiol 1997;30:1002-8. 5. Hare JM, Walford GD, Hruban RH, et al. Ischemic cardiomyopathy: endomyocardial biopsy and ventriculographic evaluation of patients with congestive heart failure, dilated cardiomyopathy and coronary artery disease. J Am Coll Cardiol 1992;20:1318-25. 6. Wallis DE, O'Connell JB, Henkin RE, et al. Segmental wall motion abnormalities in dilated cardiomyopathy: a common finding and good prognostic sign. J Am Coll Cardiol 1984;4:674-9. 7. O'Connor CM, Velazquez EJ, Gardner LH, et al. Comparison of coronary artery bypass grafting versus medical therapy on long-term outcome in patients with ischemic cardiomyopathy (a 25-year experience from the Duke Cardiovascular Disease Databank). Am J Cardiol 2002;90:101-7. 8. Hunt SA. ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). J Am Coll Cardiol 2005;46:e1-e82. 9. Ammann P, Brunner-La Rocca HP, Angehrn W, et al. Procedural complications following diagnostic coronary angiography are related to the operator's experience and the catheter size. Catheter Cardiovasc Interv 2003;59:13-8. 10. Stensland-Bugge E, Bonaa KH, Joakimsen O. Reproducibility of ultrasonographically determined intima-media thickness is dependent on arterial wall thickness. The Tromso Study. Stroke 1997;28: 1972-80. 11. Rohani M, Jogestrand T, Ekberg M, et al. Interrelation between the extent of atherosclerosis in the thoracic aorta, carotid intima-media thickness and the extent of coronary artery disease. Atherosclerosis 2005;179:311-6. 12. van der Meer IM, Bots ML, Hofman A, et al. Predictive value of noninvasive measures of atherosclerosis for incident myocardial infarction. The Rotterdam Study. Circulation 2004;109:1089-94. 13. Chambless LE, Heiss G, Folsom AR, et al. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities (ARIC) Study, 1987-1993. Am J Epidemiol 1997;146:483-94. 14. Bots ML, Hoes AW, Koudstaal PJ, et al. Common carotid intimamedia thickness and risk of stroke and myocardial infarction: the Rotterdam Study. Circulation 1997;96:1432-7.

15. Hodis HN, Mack WJ, LaBree L, et al. The role of carotid arterial intima-media thickness in predicting clinical coronary events. Ann Intern Med 1998;128:262-9. 16. Cerne A, Kranjec I. Atherosclerotic burden in coronary and peripheral arteries in patients with first clinical manifestation of coronary artery disease. Heart Vessels 2002;16:217-26. 17. Stein JH, Korcarz CE, Hurst RT, et al. 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 189-90]. 18. Felker GM, Shaw LK, O'Connor CM. A standardized definition of ischemic cardiomyopathy for use in clinical research. J Am Coll Cardiol 2002;39:210-8. 19. Howard G, Burke GL, Evans GW, et al. Relations of intimal-medial thickness among sites within the carotid artery as evaluated by B-mode ultrasound. ARIC Investigators. Atherosclerosis Risk in Communities. Stroke 1994;25:1581-7. 20. Ghostine S, Caussin C, Habis M, et al. Non-invasive diagnosis of ischaemic heart failure using 64-slice computed tomography. Eur Heart J 2008. 21. Lorenz MW, von Kegler S, Steinmetz H, et al. Carotid intima-media thickening indicates a higher vascular risk across a wide age range: prospective data from the Carotid Atherosclerosis Progression Study (CAPS). Stroke 2006;37:87-92. 22. Danias PG, Ahlberg AW, Clark 3rd BA, et al. Combined assessment of myocardial perfusion and left ventricular function with exercise technetium-99 m sestamibi gated single-photon emission computed tomography can differentiate between ischemic and nonischemic dilated cardiomyopathy. Am J Cardiol 1998;82:1253-8. 23. Yao SS, Qureshi E, Nichols K, et al. Prospective validation of a quantitative method for differentiating ischemic versus nonischemic cardiomyopathy by technetium-99 m sestamibi myocardial perfusion single-photon emission computed tomography. Clin Cardiol 2004; 27:615-20. 24. Danias PG, Papaioannou GI, Ahlberg AW, et al. Usefulness of electrocardiographic-gated stress technetium-99 m sestamibi single-photon emission computed tomography to differentiate ischemic from nonischemic cardiomyopathy. Am J Cardiol 2004; 94:14-9. 25. Vigna C, Russo A, De Rito V, et al. Regional wall motion analysis by dobutamine stress echocardiography to distinguish between ischemic and nonischemic dilated cardiomyopathy. Am Heart J 1996;131: 537-43. 26. Sharp SM, Sawada SG, Segar DS, et al. Dobutamine stress echocardiography: detection of coronary artery disease in patients with dilated cardiomyopathy. J Am Coll Cardiol 1994; 24:934-9. 27. Cohen A, Chauvel C, Benhalima B, et al. Is dobutamine stress echocardiography useful for noninvasive differentiation of ischemic from idiopathic dilated cardiomyopathy? Angiology 1997;48: 783-93. 28. Duncan A, Francis D, Gibson D, et al. Usefulness of electrocardiographic-gated stress technetium-99 m sestamibi single-photon emission computed tomography to differentiate ischemic from nonischemic cardiomyopathy. Am J Cardiol 2004;94. 29. Sweitzer NK, Lopatin M, Yancy CW, et al. Comparison of clinical features and outcomes of patients hospitalized with heart failure and normal ejection fraction (Nor = 55%) versus those with mildly reduced (40% to 55%) and moderately to severely reduced (b40%) fractions. Am J Cardiol 2008;101:1151-6.