Author + information
- Received June 11, 2001
- Revision received January 3, 2002
- Accepted January 16, 2002
- Published online April 3, 2002.
- Laurent Belhassen, MD, PhD*,‡,
- Claudine Carville, MD*,‡,
- Gabriel Pelle, PhD*,
- Jean Luc Monin, MD†,
- Emmanuel Teiger, MD, PhD*,
- Anne-Marie Duval-Moulin, MD†,
- Patrick Dupouy, MD†,‡,
- Jean Luc Dubois Rande, MD, PhD† and
- Pascal Gueret, MD, FACC†,* ()
- ↵*Reprint requests and correspondence:
Dr. Pascal Gueret, Fédération de Cardiologie, Hôpital Henri-Mondor, 51 Avenue du Maréchal De Lattre de Tassigny, 94010 Créteil, France.
Objectives We assessed the value of carotid intima-media thickness (CIMT) and thoracic aorta intima-media thickness (AoIMT) in ruling out significant coronary artery disease (CAD) in patients scheduled for heart valve surgery.
Background Evaluation of CAD is needed in most patients undergoing heart valve surgery because of the high surgical morbidity in patients with significant CAD, raising the need for sensitive tests to exclude CAD. Coronary angiography is the reference standard, but this invasive procedure is not cost-effective, because more than two-thirds of these patients do not have significant CAD.
Methods In a pilot study, CIMT and AoIMT cutoff values separating low- from high-risk groups were determined in 96 patients by using receiver-operating characteristic curves. Then, a prospective study was conducted in 152 patients to determine the statistical power of these cutoff values used alone or in combination. In both studies, carotid artery ultrasonography and transesophageal echocardiography were performed before coronary angiography and valve surgery.
Results In the pilot study, CIMT <0.55 mm and AoIMT <3 mm were excellent predictors of the absence of CAD. In the prospective study, CIMT and AoIMT criteria were independent predictors of significant CAD in these patients, as assessed by logistic regression analysis. Carotid IMT criterion had 100% sensitivity and 100% negative predictive value. For the AoIMT criterion, sensitivity was 98%, and negative predictive value 99%. Combining the two criteria did not change sensitivity and negative predictive value but increased specificity to 78%.
Conclusions Measurements of CIMT and AoIMT may be useful in selecting patients who do not require coronary angiography before heart valve surgery.
Recent improvements in ultrasound technology provide reliable and accurate assessments of heart valve lesions and left ventricular function, challenging the need for evaluation of gradients by cardiac catheterization before heart valve surgery. However, coexistence of valvular heart disease and coronary artery disease (CAD) in adult patients raises diagnostic and therapeutic issues. The prevalence of CAD in patients undergoing heart valve surgery is ∼30% (1,2). Classic coronary risk factors and symptoms are weak predictors of CAD (3). Consequently, coronary angiography is still performed before valve surgery in the majority of patients >35 years old (4). This invasive investigation is associated with non-negligible morbidity, especially in patients with unstable hemodynamic variables. Therefore, there is a need for non-invasive screening methods capable of identifying those patients who may or may not need coronary angiography. Ultrasonographic evaluations of non-coronary atherosclerosis, such as mitral annulus calcifications, are currently being investigated for this purpose (5,6). The relationship between peripheral atherosclerosis and coronary atherosclerosis is now well documented. B-mode ultrasonography of the carotid artery has also been found useful and reproducible (7,8). In recent studies, the extent of carotid artery atherosclerosis, as evaluated by B-mode ultrasonography, was strongly and independently correlated with CAD (9–11). Moreover, carotid intima-media thickness (CIMT) was identified as a risk factor for myocardial infarction and stroke (12). Qualitative evaluation of thoracic aortic atherosclerosis by transesophageal echocardiography (TEE) has been related to CAD, with a good negative predictive value (NPV) for the disease (13).
If evaluations of carotid and aortic atherosclerosis are to be used to screen patients for CAD, quantitative variables are required to determine the positivity or negativity of tests in routine clinical practice, to help manage patients. Therefore, we assessed the efficacy of CIMT and aortic intima-media thickness (AoIMT) measurements for predicting the absence of CAD in patients scheduled for heart valve surgery, as well as their potential usefulness for avoiding unnecessary preoperative coronary angiography.
Pilot and prospective studies were performed. In the pilot study, 96 consecutive patients were evaluated to determine the cutoff values for CIMT and AoIMT. In the prospective study, the clinical significance of these cutoff values was evaluated in 152 consecutive patients. All patients in both groups underwent coronary angiography as a preliminary to heart valve replacement. Patients with a history of CAD, carotid artery surgery or stroke were excluded. In each patient, a medical chart review was used to determine the age, gender and risk factors for CAD: history of systemic hypertension with verified blood pressure >140/90 mm Hg; history of smoking, history of hypercholesterolemia with verified serum total cholesterol >2.4 g/liter; and history of diabetes with verified blood glucose >1.26 g/liter.
Measurement of CIMT
B-mode ultrasound measurements were performed with a 7-MHz linear-array transducer connected to a 128 XP/10 Acuson echograph (Mountain View, California). Patients were examined in a supine position, with the head in the axis of the body. The operator directed the sound beam perpendicularly to the arterial surface of the far wall of the vessel to obtain two parallel echogenic lines corresponding to the blood-intima and media-adventitia interfaces. Carotid IMT was measured as the distance between these two lines, as previously described by Pignoli et al. (8). The scanning method used was similar to that reported by Salonen and Salonen (14). The image was focused on the posterior far wall; a segment was chosen on the linear portion of the artery to avoid disturbance of CIMT related to curves; and four images of the left and right common carotid arteries were recorded at least 15 mm proximal to the bifurcation. Carotid IMT was calculated as the mean of eight measurements. Carotid IMT >1.2 mm was considered indicative of plaque and consequently was not included in the calculation of CIMT. Log compression and magnification settings were kept constant throughout the study. However, gain settings could be changed from one patient to the next. Because the artery wall is subjected to blood pressure variations, all ultrasound measurements were electrocardiographically gated and recorded in end diastole. Two investigators participated in the measurements. Investigators had no knowledge of the results of TEE or angiography. Intraobserver and interobserver reproducibilities were good (1%).
Measurement of AoIMT
All patients underwent TEE before cardiac catheterization. Transesophageal echocardiography was performed using an Acuson 128 XP or Sequoia C256 ultrasonograph (Acuson) equipped with a 3.5- to 7.0-MHz multiplane probe. To ensure imaging of the entire thoracic aorta, the probe was rotated posteriorly and advanced to the distal esophagus and withdrawn slowly to scan the descending aorta and aortic arch. The probe was then rotated and advanced again to image the ascending aorta. All scans were recorded on S-VHS videotapes for later off-line analysis. The maximal value of AoIMT was collected in each patient. Transesophageal echocardiography was performed by experienced investigators who had no knowledge of the results of other studies. The S-VHS videotapes were read by two experienced investigators. Intraobserver and interobserver reproducibilities were good (4.5% and 4.9%, respectively).
Selective coronary angiograms were obtained using the Judkin’s technique, following standard procedures in our laboratory. Identified lesions were visualized in at least two orthogonal views by a CGR (General Electric, Issy Les Moulineaux, France) X-ray unit connected to a DPS PLUS system (ADAC Laboratories, Milpitas, California). This system allowed fully automatic edge-detection and quantitation of coronary segments, as previously described (15). Automatic measurements were validated by two experienced operators who had no knowledge of the rest of the study. Interobserver and intraobserver reproducibilities were good (4.3% and 4.7%, respectively). Significant CAD was defined as stenosis ≥70% of at least one major branch (left anterior descending, left circumflex or right coronary artery) or stenosis ≥50% of the main left coronary artery. Coronary angiographic findings were used as the reference standard against which CIMT and AoIMT were evaluated.
All descriptive data are expressed as the mean value ± SD. Intraobserver and interobserver variabilities were estimated using the paired ttest. Coronary risk factors were examined; and hypercholesterolemia, hypertension, smoking and diabetes were dichotomized by the presence or absence of the risk factor in the medical chart review, because the risk factors had been corrected before inclusion in the study. Continuous variables—CIMT, AoIMT and age—were compared using the Student ttest. Nominal variables were compared using the chi-square test. Correlations between nominal variables were analyzed using least-squares regression techniques. Logistic regression was used to compare nominal and continuous variables. Univariate and multivariate analyses were performed. The presence of CAD constituted the dependent variable (x) in our model. Age, gender and the presence of CIMT ≥0.55 mm, AoIMT ≥3 mm, dyslipidemia, diabetes, hypertension and smoking represented the independent variables (y). Sensitivity, specificity, positive predictive value and NPV were calculated for CIMT, AoIMT and the combined criteria. Receiver operating characteristic (ROC) curves were constructed to assess the relationship between sensitivity and specificity of the criteria. The ROC curves were used to determine the true-positive rate (sensitivity) versus the false-positive rate (1 − specificity) at various levels of the index and to identify the cutoff values yielding the largest number of correctly classified patients (16,17). P values <0.05 were considered significant.
Table 1summarizes the characteristics of the patients in the pilot study. Approximately 60% of the patients were men. The prevalence of coronary artery stenosis was 22%. Aortic stenosis was the main valvular lesion observed (70%). Mean CIMT was significantly higher in patients with significant CAD than in patients without significant CAD (0.70 ± 0.1 vs. 0.58 ± 0.1 mm, p < 0.001). Analysis of the ROC curve for CIMT (Fig. 1A) showed that a cutoff value of 0.55 mm had 96% sensitivity, 49% specificity and 98% NPV (Table 2). Carotid IMT was ≥0.55 in 25 of the 26 patients with CAD and in 38 of the 70 patients without CAD.
Mean AoIMT was significantly higher in patients with significant CAD than in patients without significant CAD (4.8 ± 0.8 vs. 2.7 ± 1.1 mm, p < 0.001). The ROC curve for AoIMT (Fig. 1B) showed that a 3-mm cutoff value had 100% sensitivity, 53% specificity and 100% NPV (Table 2). The AoIMT was ≥3 mm in all 26 patients with significant CAD and in 36 of the 70 patients without significant CAD.
The characteristics of the 152 patients included in the prospective study are reported in Table 1. No significant differences were observed, as compared with the pilot study group. Fifty-seven percent of the patients were men. The prevalence of coronary artery stenosis was 31%. Aortic stenosis was the most common valvular lesion (found in 65% of patients), followed, in decreasing order, by mitral regurgitation (n = 23), aortic regurgitation (n = 14) and mitral stenosis (n = 11). Patients without CAD were significantly younger than patients with CAD (65 ± 10 and 72 ± 7 years, p < 0.05). Mean CIMT was 0.62 ± 0.13 mm overall, 0.67 ± 0.11 mm in the patients with significant CAD and 0.59 ± 0.13 mm in the patients without significant CAD (p < 0.001). Mean CIMT was ≥0.55 mm in all of the patients with CAD and in 55 of the 111 patients without CAD. Mean AoIMT was 3.0 ± 1.4 mm overall, 4.3 ± 0.7 mm in the patients with significant CAD and 2.5 ± 1.2 mm in the patients without significant CAD (p < 0.001). Aortic IMT was ≥3 mm in 40 of 41 patients with significant CAD and in 72 of 111 patients without significant CAD. Linear regression analysis showed a significant relationship between CIMT and the degree of stenosis (r = 0.45, p < 0.001) (Fig. 2) and between AoIMT and the degree of stenosis (r = 0.65, p < 0.001) (Fig. 3).
Univariate logistic regression analysis showed that CIMT, AoIMT, age and diabetes were correlated to CAD (Table 3). Among these variables, CIMT ≥0.55 mm, AoIMT ≥3 mm and age were the only independent predictors of CAD, as revealed by multivariate logistic regression analysis.
Of the risk factors studied, only age was correlated with the CIMT and AoIMT criteria (r = 0.37, p < 0.001 and r = 0.45, p < 0.001, respectively).
Carotid IMT showed 100% sensitivity and 50% specificity for the detection of significant CAD (Table 2). The NPV for exclusion of CAD was 100%. Aortic IMT was also very sensitive. Only one of the 41 patients with CAD had an AoIMT value under the 3-mm cutoff (2.7 mm). Specificity was 65% and NPV was 99%. Combining CIMT and AoIMT criteria significantly increased the specificity (78%), while preserving high sensitivity and NPV (Table 2).
In this study, we evaluated the usefulness of quantitative CIMT and AoIMT measurements in excluding significant CAD in patients scheduled for heart valve surgery. In a pilot study, we determined the cutoff values for CIMT and AoIMT by using ROC analysis. Then, in a prospective study, we showed that CIMT <0.55 mm or AoIMT <3 mm had an excellent NPV for excluding significant CAD. Combining both criteria improved the specificity without significantly affecting the sensitivity or NPV. These data suggest that CIMT and AoIMT may be helpful in identifying patients who do not require coronary angiography before heart valve surgery. They could be used in the stepwise strategy proposed in Figure 4.
Peripheral atherosclerosis and CAD
Special attention is being given to the correlation between CAD and peripheral atherosclerosis. Atherosclerosis is a diffuse disease initially characterized by impairment of functional properties of the vascular wall (18). Many studies have demonstrated a correlation between the presence of CAD and peripheral atherosclerosis. Coronary artery disease has been reported to be positively correlated with CIMT (10), the presence of atherosclerotic plaque on the aorta (19), the existence of mitral annulus calcification (5)and inferior limb atherosclerosis (20). However, there has been no attempt to define quantitative variables that would help to exclude CAD.
Carotid IMT is higher in patients with severe CAD than in normal subjects and is positively and significantly correlated with CAD (21). However, the use of CIMT as a screening test for CAD has generated controversy. Recently, Adams et al. (22)found a significant but weak correlation between CIMT and the extent of CAD in a group of patients with chest pain and a high prevalence of CAD. In their study, CIMT did not help to predict the severity of CAD. However, CIMT values were >0.60 mm in most patients with CAD, indicating that CIMT might have been helpful in predicting non-significant CAD. Adams et al. (22)studied patients with chest pain, and their results cannot be extrapolated to asymptomatic patients (23); however, CIMT is intended mainly as a mean of screening asymptomatic patients with coronary risk factors, with the goal of identifying those who do not need further coronary artery evaluation. In our prospective study group, chest pain was present in only 18% of patients, most of whom had aortic stenosis, a lesion that detracts from the value of clinical symptoms for the diagnosis of CAD. Most of these patients would not have undergone coronary angiography if they had not been scheduled for surgery. Moreover, patients with hemodynamically non-significant aortic stenosis have an increased risk of myocardial infarction (24), but do not undergo coronary angiography unless they have angina.
CIMT and AoIMT as independent predictors of CAD
Carotid IMT ≥0.55 mm and AoIMT ≥3 mm were independent predictors of coronary stenosis. Both criteria were extremely sensitive for detecting significant CAD. The CIMT criterion missed only one patient with significant CAD in the pilot study and none in the prospective study. Specificity was about 50%, so that coronary angiographic screening of patients, based on CIMT alone, would have resulted in coronary angiography in 56 patients without CAD (55 true negative results). The AoIMT criterion missed no patient in the pilot study and only one patient in the prospective study. The missed patient had a value close to the 3-mm cutoff. The specificity of AoIMT was higher (65%) than that of CIMT, and screening based on AoIMT alone would have resulted in coronary angiography in 39 patients without CAD. With the two criteria in combination, specificity increased significantly, to 78%, and only 25 coronary angiograms would have been performed in patients without CAD (86 true negative results). For each criterion, specificity was lower than sensitivity, probably because only significant atherosclerosis was considered, and high CIMT or AoIMT values can occur in mild atherosclerosis. The sensitivity and NPV of these tests are more important than specificity because it is essential not to miss any patients with significant coronary stenosis.
Strategy for excluding CAD in patients undergoing heart valve surgery
The results of the present study suggest that a stepwise strategy could be used to identify patients who do not require coronary angiography before undergoing heart valve surgery (Fig. 4). It should be emphasized that it may not apply to other categories of patients. In this strategy, coronary arteriography is considered unnecessary if CIMT is <0.55 mm. If CIMT is ≥0.55 mm, TEE is performed to determine AoIMT. If this variable is <3 mm, coronary angiography is unnecessary. Patients with CIMT ≥0.55 mm and AoIMT ≥3 mm should undergo coronary angiography before surgery. This strategy may be particularly helpful in minimizing exposure to the non-negligible risks of coronary angiography in patients with an unstable hemodynamic condition and in elderly patients. Furthermore, it may decrease the cost of the preoperative evaluation in patients scheduled for heart valve surgery.
- aortic intima-media thickness
- coronary artery disease
- carotid intima-media thickness
- intima-media thickness
- negative predictive value
- receiving operating characteristic
- transesophageal echocardiography
- Received June 11, 2001.
- Revision received January 3, 2002.
- Accepted January 16, 2002.
- American College of Cardiology Foundation
- Morrison G.W.,
- Thomas R.D.,
- Grimmer S.F.,
- Silverton P.N.,
- Smith D.R.
- Bonow R.O.,
- Carabello B.,
- de Leon A.C. Jr.,
- et al.
- Crouse J.R.,
- Harpold G.H.,
- Kahl F.R.,
- Toole J.F.,
- McKinney W.M.
- Pignoli P.,
- Tremoli E.,
- Poli A.,
- Oreste P.,
- Paoletti R.
- Handa N.,
- Matsumoto M.,
- Maeda H.,
- et al.
- Heiss G.,
- Sharrett A.R.,
- Barnes R.,
- Chambless L.E.,
- Szklo M.,
- Alzola C.
- Wofford J.L.,
- Kahl F.R.,
- Howard G.R.,
- McKinney W.M.,
- Toole J.F.,
- Crouse J.R. 3rd.
- the Cardiovascular Health Study Collaborative Research Group,
- O’Leary D.H.,
- Polak J.F.,
- Kronmal R.A.,
- Manolio T.A.,
- Burke G.L.,
- Wolfson S.K. Jr.
- Tribouilloy C.,
- Peltier M.,
- Colas L.,
- Rida Z.,
- Rey J.L.,
- Lesbre J.P.
- the FROST Study Group,
- Lafont A.,
- Dubois-Rande J.L.,
- Steg P.G.,
- et al.
- Altman D.G.,
- Bland J.M.
- Allan P.L.,
- Mowbray P.I.,
- Lee A.J.,
- Fowkes F.G.
- Craven T.E.,
- Ryu J.E.,
- Espeland M.A.,
- et al.
- Adams M.R.,
- Nakagomi A.,
- Keech A.,
- et al.