Distribution and Correlates of Sonographically Detected Carotid Artery Disease in the Cardiovascular Health Study

Transcription

1 1752 Distribution and Correlates of Sonographically Detected Carotid Artery Disease in the Cardiovascular Health Study Daniel H. O'Leary, MD; Joseph F. Polak, MD; Richard A. Kronmal, PhD; Steven J. Kittner, MD, MPH; M. Gene Bond, PhD; Sidney K. Wolfson Jr., MD; William Bommer, MD; Thomas R. Price, MD; Julius M. Gardin, MD; and Peter J. Savage, MD, on behalf of the CHS Collaborative Research Group Background and Purpose: This article describes the prevalence of extracranial carotid atherosclerosis assessed by ultrasonography, its association with risk factors, and its relation to symptomatic coronary disease and stroke in men and women aged >65 years. Methods: Maximum percent stenosis, maximum common carotid artery wall thickness, and maximum internal carotid artery wall thickness were assessed using duplex ultrasound in 5,201 men and women aged >65 years in the Cardiovascular Health Study, a study of the risk factors and natural history of cardiovascular disease in the elderly. Existing coronary disease and stroke were assessed by physical examination and participant history. Results: Detectable carotid stenosis was present in 75% of men and 62% of women, although the prevalence of >50% stenosis was low, 7% in men and 5% in women. Maximum stenosis and maximum wall thickness measurements increased with age and were uniformly greater at all ages in men than in women (p< ). Established risk factors for atherosclerosis (hypertension, smoking, diabetes) and indications of vascular disease (left ventricular hypertrophy, major electrocardiographic abnormality, bruits, and history of heart disease or stroke) related to all three carotid artery measures in the the elderly. Of the three ultrasound measures, the best correlate for a history of coronary disease was maximum internal carotid artery wall thickness. For stroke the best correlate was common carotid artery wall thickness. Multiple logistic regression models of prevalent coronary heart disease and stroke that included the ultrasound findings indicated, after adjustment for age and sex, that maximum internal wall thickness and maximum common carotid wall thickness were significant correlates of both. Maximum stenosis did not add significantly to the correlation. Conclusions: In the elderly the incidence of carotid atherosclerosis was high, although the frequency of severe disease was low. The prevalence and severity of carotid atherosclerosis continued to increase with age even in the late decades of life, and more disease was found in men than in women at all ages. Known risk factors for atherosclerosis continued to relate to carotid abnormalities in the later decades of life, both in symptomatic and asymptomatic subjects. (Stroke 1992;23: ) KEY WORDS aged carotid artery diseases epidemiology ultrasonics Whether atherosclerosis manifests itself as myocardial infarction or as stroke, this systemic disease is the major cause of morbidity and mortality among people aged >65 years in the United States. At the same time, however, the prevalence of asymptomatic atherosclerotic disease of the extracranial carotid arteries is largely unknown. 1 Previous data on the distribution of carotid atherosclerosis have come mostly from autopsy studies and from either angiograms or surgical specimens obtained in patients showing clinical evidence of cardiovascular disease. 2-8 Most studies using ultrasonography to estimate the prevalence of carotid atherosclerosis have been in hospitalbased populations or in patients with a high likelihood for having symptomatic atherosclerotic disease This biased sampling is further limited by the fact that individuals in very late decades of life tend to be excluded from such studies. Whether risk factors asso- From the Division of Radiology (D.H.O.), Geisinger Medical Center, Danville, Pa.; the Department of Radiology (J.F.P.), Brigham and Women's Hospital, Boston; the Department of Biostatistics (R.A.K.), University of Washington, Seattle, Wash.; the Department of Neurology (S.J.K, T.R.P.), University of Maryland, Baltimore, Md.; the Department of Anatomy (M.G.B.), Bowman Gray School of Medicine, Winston-Salem, N.C.; the Department of Neurosurgery (S.K.W.), University of Pittsburgh, Pittsburgh, Pa.; the Department of Medicine, University of California, Sacramento (W.B.), and University of California, Irvine (J.M.G.); and the Division of Epidemiology and Clinical Applications (P.J.S.), National Heart, Lung, and Blood Institute, Bethesda, Md. Supported in part by contracts NO 1-HC85079 through NO 1-HC85086 from the National Heart, Lung, and Blood Institute. Address for reprints: CHS Coordinating Center, University of Washington, JD 30, 1107 NE 45th Street, Room 530, Seattle, WA Received June 24, 1992; final revision received September 11, 1992; accepted September 14, 1992.

2 O'Leary et al Sonographically Detected Carotid Artery Disease in CHS 1753 ciated with carotid stenotic lesions in a younger population continue to have the same importance as age increases beyond 65 years is poorly understood. Since clinical symptoms do not appear until the disease process is quite advanced and involves multiple portions of the arterial tree, it has been difficult to determine the developmental stages of carotid disease and to establish whether significant differences exist between symptomatic and asymptomatic adults. The goals of this study were to determine the prevalence of asymptomatic carotid atherosclerosis in the elderly using ultrasound to differentiate between those with and without disease, to compare the ultrasound findings with previously recognized risk factors for atherosclerosis, and to relate the ultrasound findings to existing cardiovascular disease. Data for the analyses presented here were collected as part of an ongoing multicenter study entitled the Cardiovascular Health Study (CHS). Subjects and Methods Subject Population At the time of this study, CHS had enrolled 5,201 men and women, aged ^65 years. The CHS cohort was recruited from a random sampling of the Health Care Financing Administration (HCFA) Medicare eligibility lists in four communities: Forsyth County, N.C.; Sacramento County, Calif.; Washington County, Md.; and Pittsburgh, Pa. Any person aged ^65 years who was living in the sampled person's household and was able to give informed consent was eligible to participate in the study. Potential participants were excluded if they were institutionalized, wheelchair-bound in the home, or receiving treatment for cancer. Details of the study design have been published. 12 Medical History Initial data collection included parameters such as age and sex as well as history of diabetes, hypertension, and smoking. Physical examination included identification and location of carotid bruits. Hypertension was determined from the average of two measures and was defined as average systolic blood pressure 2: 160 mm Hg or average diastolic pressure ^95 mm Hg or by subjectreported history of hypertension and current use of antihypertensive medication. Borderline hypertension was defined as average systolic blood pressure mm Hg or average diastolic blood pressure mm Hg. Previously known diabetes was defined on the basis of participant-reported prior physician diagnosis of diabetes, or use of insulin or oral hypoglycemic agents. Glucose tolerance tests were interpreted using World Health Organization criteria 13 : newly recognized diabetes was defined as fasting glucose ^140 mg/dl or 2-hour glucose ^200 mg/dl; impaired glucose tolerance was defined as fasting glucose <140 mg/dl and 2-hour glucose 2140 mg/dl and <200 mg/dl. Laboratory data included measurements of total low density lipoprotein (LDL) and high density lipoprotein (HDL) cholesterol and total trigh/cerides, creatinine, blood urea nitrogen, blood glucose, and uric acid. Coronary disease was defined as self-reported and confirmed history of angina, myocardial infarction, or coronary revascularization or old myocardial infarction documented on electrocardiography (ECG). Stroke was denned as selfreported history of stroke confirmed by physical examination or medical records whenever possible. Ultrasonography Carotid ultrasound was performed during the baseline clinic visit with all four centers using Toshiba SSA-270A imaging units (Toshiba America Medical Systems, Tustin, Calif.). The four machines were identically equipped with a phased-array imaging probe having a characteristic -3-dB cutoff point of 6.7 MHz. The pulsed Doppler frequency was 4.0 MHz. The imaging protocol involved obtaining a single longitudinal lateral view of the distal 10 mm of the right and left common carotid arteries and three longitudinal views in different imaging planes of each internal carotid artery. The internal carotid artery was defined as including both the carotid bulb, identified by the loss of the parallel wall present in the common carotid artery, and the 10-mm segment of the internal carotid artery distal to the tip of the flow divider that separates the external and internal carotid arteries. The examination concentrated on this site because studies have shown that plaques of the carotid artery occur primarily along the outer wall of the internal carotid artery in the region of the internal carotid artery opposite the bifurcation flow divider Studies were recorded and sent weekly to the Ultrasound Reading Center for standardized readings. The high-resolution images of the common and internal carotid arteries were analyzed to calculate combined intimal-medial thickness of the near and far wall and the lumen diameter at each arterial site. All measurements of lumen and wall thickness were calculated by use of a specially designed computer program. The reader estimated percent diameter stenosis for each internal carotid artery, using both image and Doppler data. Doppler peak systolicflowvelocities < 1.5 m/sec were assumed to indicate the absence of significant (50%) lumen stenosis. Gray-scale imaging data alone were then used to estimate percent diameter stenosis as absent (0%), mild (1-24%), or moderate (25-49%). Doppler peak flow velocities of m/sec were taken to represent stenosis of 50-74%, and Doppler flow velocities of 22.5 m/sec indicated 2:75% stenosis. The latter category included totally occluded internal carotid arteries. In all the analyses presented here, studies coded by the reader as having normal Doppler flow velocities but uninterpretable gray-scale images were included in the 0% stenosis category. Details of the scanning and reading protocols, as well as initial reproducibility results, have been published. 16 To quantify the degree of thickening of the carotid artery walls, the many measures of wall thickness were summarized into two variables, one for the common carotid and one for the internal carotid artery. The maximum wall thickness of the common carotid artery was defined as the mean of the maximum (m) wall thicknesses for near (N) and far (F) wall (W) on both the left (L) and right (R) sides: mlnw+mlfw+ mrnw+mrfw/4. The maximum wall thickness variable of the internal carotid artery was defined in the same way, and the results from the three scans were averaged.

3 1754 Stroke Vol 23, 12 December 1992 The maximum stenosis was defined as the largest stenosis observed on the right or left side. When maximum stenosis was used as a continuous variable, it was coded as the midpoint of the interval to 0%, 1-25%, 26-50%, 51-75%, or %. Statistical Analyses Descriptive statistics, regression analysis, and analysis of variance were performed using the SPSS/PC statistical package (SPSS/PC PLUS V.4.0, SPSS Inc., Chicago). Analysis of variance was used to test for linearity of the relation between age and wall thicknesses with age treated as a discrete quantity with replication at each age. This method provided the capacity of identifying any significant deviations from a straight line. Relations between maximum stenosis and the maximum wall thicknesses were assessed in unadjusted form and after adjustment by linear regression for age and sex. In computation of odds ratios, both maximum common and internal carotid wall thicknesses were categorized into approximate quartiles of the total sample. All odds ratios for these two variables used the lowest quartile as the comparison group. For the maximum percent stenosis the comparison category was the nostenosis group. Logistic regression analysis was used to study the independent contributions of wall thickness measures and maximum stenosis, treated as continuous variables, to the prediction of a history of coronary disease or stroke after the adjustment for age and sex. In the logistic regression analysis, only values of/><0.01 were considered statistically significant. The tests and confidence intervals for the odds ratios were computed using EGRET statistical software (Statistics and Epidemiology Research Corp., Seattle, Wash.). Results Of the 9,024 persons who were eligible and invited, 5,201 (57.6%) agreed to participate. There were 2,946 (56.6%) women and 2,255 (43.4%) men. The CHS cohort was largely white (94.5%), with blacks accounting for 4.7% and other races for 0.8%. An ultrasound scan was done on 5,176 (99.5%) of the participants. From these, maximum percent stenosis estimates of either the right or left internal carotid artery were available for 5,114 (98%) of participants, maximum common carotid artery wall thickness measurements were available for 5,164 (99.8%), and maximum internal carotid artery wall thickness measurements were derived for 5,117 (98.9%). Tables 1 and 2 show the distribution of maximum percent stenosis in women and men by age. Overall, maximum stenosis increased significantly with age (/>< ). The prevalence of severe stenosis (75-100%) was quite low, 2.3% in men and 1.1% in women. Moderate stenosis (50-74%) also occurred infrequently, being present in only 5.3% of men and 4.0% of women. The group having mild stenosis (1-49%) was quite large, almost 70% in men and 60% in women. Only 25% of men and 38% of women had vessels without plaque deposition. Men had a greater frequency and severity of stenosis at all ages than women (/xo.qoool). Maximum internal carotid artery and common carotid artery wall thicknesses were greater in men than in women at every age, and both increased linearly with TABLE 1. Distribution of Maximum Stenosis* by Age for Women Age (years) Stenosis rmal 1-24% 25-49% 50-74% 75-99% 100% Total % % % 37 33% % 1 0.1% 1, % % % % 9 1.0% % % % % 5 0.9% 1 0.2% % % % % 4 1.7% % % % 7 8.0% 3 3.4% 87 Total 1, ,906 'Maximum percent diameter stenosis of either internal carotid artery. p< byx 2. age (Figure 1; test for deviation from linearity, NS; test for linear term, p< ). Wall thickness measurements of the internal carotid artery were uniformly greater than those of the common carotid artery. The unadjusted relations between risk factors and history of coronary disease and stroke to the three measures of carotid atherosclerosis are shown in Table 3. The R 2 values shown in the table provide a relative measure of the association of the variable and the ultrasound measurements, with larger values implying a greater association. Hypertension was strongly related to all three measures (/xo.ooool), with increased wall thickness associated with being borderline hypertensive and hypertensive. There was a stronger relation between smoking and the maximum internal carotid artery wall thickness than for the maximum common carotid artery wall thickness or the percent stenosis. Current smokers had the thickest arteries and most stenosis. The TABLE 2. Distribution of Maximom Stenosis* by Age for Men Age (years) Stenosis rmal i85 14 Total % 24.8% 19.4% 21.6% 143% 1-24% 25-49% 50-74% 75-99% % % % 5 0.7% % % % % % % % 4 0.9% % % % 6 2.4% % % 8 8.2% 2 2.0% % 2 03% 7 1.0% 4 0.9% 6 2,4% 4 4.1% 23 Total ,210 Maximum percent diameter stenosis of either internal carotid artery. ;>< by*

4 O'Leary et al Sonographlcally Detected Carotid Artery Disease in CHS 1755 FIGURE 1. Line graph of maximum internal carotid artery (ICA) and common carotid artery (CCA) wall thicknesses by age for women (F) and men (M) Ag. former smokers had intermediate values, and those who had never smoked had the smallest values. Diabetes and impaired glucose tolerance were associated with thicker walls and greater stenosis. Obesity, as defined by weight 130% above ideal, showed no consistent association with carotid abnormalities. Left ventricular hypertrophy on ECG had similar associations with all three ultrasound measures. For the presence of a major ECG abnormality, both wall thickness measures had larger adjusted R 2 values than maximum stenosis. Bruits, on the other hand, were associated with maximum stenosis more strongly than they were with the two wall thicknesses. Clinically recognized coronary heart disease and stroke had highly significant associations with all three ultrasound measures, but the relative strength of the associations varied considerably. The strongest correlate of the three for a history of coronary heart disease was the maximum internal carotid artery wall thickness. For a history of stroke the strongest correlate was common carotid artery wall thickness. Relations of the continuous risk factors with wall thickness and percent stenosis were computed using the partial correlation coefficient after adjustment for age and sex (Table 4). The largest correlations were seen for systolic blood pressure, whereas diastolic blood pressure was not associated with maximum common wall thickness (r=-0.01,/?=ns) and only weakly correlated with the other measures. Total and LDL cholesterol and trigryceride levels were all positively associated with carotid abnormalities, whereas HDL cholesterol was negatively associated with them. Glucose and insulin levels were positively associated with the ultrasound measures. Left heart ventricular mass measured by echocardiography was positively associated with all three measures. Relations of dichotomous variables were assessed by logistic regression analysis. Coronary disease was reported and confirmed in 25.6% of subjects. The prevalence of coronary disease was examined by both stenosis grade and quartile of wall thickness of the internal and the common carotid arteries (Table 5). Maximum stenosis was highly associated with coronary disease. The prevalence of coronary disease increased from 17.8% in those with 0% stenosis to 45.8% in those with 75% stenosis. The odds ratio for coronary disease comparing those with stenosis to those with 0% stenosis increased from 1.50 for those with 1-24% stenosis to 3.90 for those with 75% stenosis (x 1 test for trend=116, p<0.001). Similar results were obtained for maximum wall thickness, although the differences between categories were less striking for the common wall thickness (x 2 test for trend=45, /xo.ooool). The differences between categories were impressive for the internal wall thickness (x 2 test for trend=153, /xo.00001), even though nearly all subjects represented in the three highest stenosis categories are grouped together in the highest wall thickness quartile. Multiple logistic regression was used to assess independent relations of carotid measures to prevalent coronary disease. A model of predicting the risk of coronary disease that included maximum stenosis, the two measures of wall thickness, age, and sex showed that after adjustment for age and sex, both the maximum internal wall thickness and the maximum common carotid wall thickness were significant correlates of coronary disease (p< and/?<0.02, respectively). The likelihood of coronary disease was estimated to increase by 36% for a 1-SD (0.69-mm) increase in the maximum internal wall thickness. The relation was somewhat weaker for the maximum common carotid wall thickness with a 1-SD (0.22-mm) increase estimating a 9% increase in risk. Maximum stenosis did not add significantly to the prediction of risk of prior coronary disease, however. Prior stroke was reported and confirmed in 4.7% of participants. All three carotid ultrasound measures were associated with a history of stroke (Table 6). The incidence of prior stroke was 3.3% in those with 0% stenosis and increased to 10.8% in those with 75% stenosis. The odds ratio for prior stroke comparing

5 1756 Stroke Vol 23, 12 December 1992 TABLE 3. Relation of Ultrasound Measurements to Risk Factors and Manifestations of Coronary Heart Disease and Stroke Maximum common IMT Maximum internal IMT Maximum percent stenosis Variable Mean SD n Mean SD n Mean SD n Hypertension rmal , ^ ,248 Borderline Hypertensive , , , * Smoking status Never , Former , , ,130 Current IP(P) Diabetes rmal Impaired Diabetic ^76 1,427 1, ^63 1,407 1, ^64 1,405 1, (< ) >130% of ideal weight , , , , , ,408 LVH on ECG Major ECG abnormality Bruits (NS) (<0.0005) , ,700 1,463 4, (<0.005) (NS) , ,668 1,448 4, (<0.01) (NS) , ,673 1,442 4, History of heart disease , , ,818 1, (< ) IMT, mean maximum intimal-medial thickness, near and far wall, right and left carotid arteries; LVH, left ventricular hypertrophy; ECG, electrocardiogram. ' for age and sex.

6 O'Leary et al Sonographically Detected Carotid Artery Disease in CHS 1757 TABLE 3. Continued Maximum common IMT Mean SD n Maximum percent stenosis Mean SD n Variable History of stroke Maximum internal IMT Mean SD n , , , (<0.0005) (<0.005) (NS) those with stenosis to those with 0% stenosis increased from 1.33 for those with 1-24% stenosis to 3.60 for those with 275% stenosis (^2 test for trend=11, p<0.001). Subjects in the highest quartile of common carotid wall thickness had a threefold increased incidence of prior stroke compared with those in the lowest quartile (x 2 test for trend=17, /><0.0001). Subjects in the highest quartile of internal carotid artery maximum wall thickness were 2.6 times as likely to have had a history of stroke as subjects in the lowest quartile (x 2 test for trend=13,p<0.0001). When ultrasound measures were included in a multiple logistic regression model for the prediction of prior stroke, adjusting for age and sex, both measures of wall thickness were significantly associated with a history of stroke and had similar risk ratios associated with a 1-SD change in wall thickness. Each of the wall thickness measures showed about a 20% increase in risk for a 1-SD change (p<0.005). As in the instance of prior coronary disease, maximum stenosis did not add significantly to the prediction of risk of prior stoke once adjustment was made for age, sex, maximum common carotid artery wall thickness, and maximum internal carotid artery wall thickness. Discussion Our data show that carotid atherosclerosis in a freeliving, healthy elderly population was commonplace, although the incidence of severe disease was low. The prevalence and severity of carotid atherosclerosis continued to increase with age among the elderly, and more disease was found in men than in women at all ages. Recognized risk factors for atherosclerosis remain related to carotid atherosclerosis in the later decades of life, both in our symptomatic and asymptomatic individuals. Thickness of the common carotid wall was related to internal carotid stenotic lesions and was a marker for symptomatic coronary disease and stroke in our elderly subjects. Several ongoing cohort studies using ultrasound have reported similar patterns of carotid atherosclerosis in the asymptomatic elderly, although none are of the magnitude presented here. Salonen et al 17 showed that the prevalence of detectable atherosclerotic plaques in the carotid arteries of 412 Eastern Finnish men increased from 1% in those aged 42 years to 28% in their oldest group, those aged 60 years. Only 5% of this latter group were found to have more than 20% stenosis. 17 Examination of 1,189 members of the Framingham cohort, aged years, revealed no disease in 30%, <50% stenosis in 62%, and ^50% stenosis in 8%, 18 results nearly identical to those found for CHS participants. The Rotterdam Elderly Study recently reported that of the initial 954 participants aged 55 years who underwent successful duplex examination of the right carotid artery, only 3.0% had 16-49% stenosis and 1.4% had 250% stenosis. 19 However, since this study used Doppler criteria rather than gray-scale findings to categorize all levels of disease, and since only the right carotid artery was examined, it is likely that their results underestimated the prevalence of carotid atherosclerosis in their population. They found, as we did, that the prevalence of carotid stenosis continues to increase in the later decades of life. Our focus on deriving reproducible wall thickness measurements from digitized high-resolution gray-scale images using a specially designed computer program and analyzing these both separately and in combination with Doppler estimates of stenosis represents an evolution in the use of ultrasound to categorize carotid abnormalities. We found that wall thickness increased in a linear fashion with increasing age. On average, intimal-medial thickness of the common carotid artery increased by 0.01 mm for both men and women for each year beyond age 65, and intimal-medial thickness of the internal carotid artery increased by 0.02 mm. Whereas ours was a cross-sectional study, these results give an estimate of expected wall thickening over time. Because natural selection would presumably eliminate those with higher levels of disease over time, our findings are likely conservative estimates of disease progression. The relations we found between carotid atherosclerosis and risk factors such as hypertension, smoking status, diabetes, major ECG abnormality, and the presence of carotid bruits were similar to those previously described " 27 The partial correlations for our elderly population may be smaller than in other studies because of premature death among subjects with more severe disease. The previously described inverse relation between carotid atherosclerosis and HDL cholesterol was seen in our elderly cohort , We found that systolic but not diastolic blood pressure was positively associated with ultrasound measurements. A similar finding has been previously published by Salonen and Salonen. 29 In fact, in our study diastolic blood pressure had a weak inverse association with intimalmedial thickness in the internal carotid artery and maximal stenosis. This finding, and the association between all three ultrasound measures and left ventricular mass, will be discussed in a future report. All measures of carotid artery involvement by the atherosclerotic process were related to clinically manifest coronary heart disease. The association between atherosclerosis in the carotid and coronary systems is

7 1758 Stroke Vol 23, 12 December 1992 TABLE 4. Age- and Sex- Partial Correlations Between Selected Risk Factor* and the Three Ultrasound Measures Maximum Maximum CCA wall ICA wall Maximum Correlations thickness thickness stenosis Systolic blood pressure (mmhg) 0.18* 0.13* 0.11* Diastolic blood pressure (mmhg) t -0.07* Cholesterol level 0.09' 0.10' 0.08' Triglyceride level (log) 0.08* 0.10* 0.09* HDL level -0.12' -0.12* -0.11* LDL level 0.12* 0.12' 0.10* Glucose level (log) 0.12' 0.08* 0.07* Insulin level (log) 0.08' 0.07* 0.05* Left ventricular mass (from echocardiogram) 0.14* 0.10* 0.07* CCA, common carotid artery, ICA, internal carotid artery, HDL, high density lipoprotein; LDL, low density lipoprotcin. V< ; tp<0.01; *p< well documented in the pathology and clinical literature Recent studies, using ultrasound, have shown that carotid atherosclerosis, expressed as average maximum wall thickness measurements from the internal and common carotid arteries, can correlate with coronary disease. 910 These studies involve participants who were preselected by virtue of having either coronary or carotid artery disease. Ours is a study of relatively healthy elderly men and women recruited from a random sample of HCFA Medicare eligibility lists lacking this bias. Nevertheless, this association remains, with internal carotid artery stenosis shown to be a stronger correlate of coronary disease than wall thickness. Our study presently permits only cross-sectional analysis of ultrasound findings and prevalent coronary heart dis- TABLE 5. Odds Ratios for Coronary Heart Disease for Levels of Maximum Stenosis and Quartilef of Common Carotid Artery and Internal Carotid Artery Wall Thickness Coronary heart disease Odds Confidence Category ratio interval Stenosis (%) 0 1^ , CCA wall thickness (mm) Quartile 1 (<0.870) 1, Quartile 2 (<0.972) Quartile 3 (<1.102) Quartile ICA wall thickness (mm) Quartile 1 (<0.915) 1, Quartile 2 (< 1.382) 1, Quartile 3 (< 1.960) Quartile CCA, common carotid artery, ICA, internal carotid artery. TABLE 6. Odds Ratios for Stroke for Levels of Maximum Stenosis and Quartiles of Common Carotid Artery and Internal Carotid Artery Wall Thickness Stroke 95% Odds Confidence Category ratio interval Stenosis (%) 0 1, , , CCA wall thickness (mm) Quartile 1 (<0.870) Quartile 2 (<0.972) 1, Quartile 3 (<1.102) 1, Quartile 4 1,198 % ICA wall thickness (mm) Quartile 1 (<0.915) 1, Quartile 2 (< 1382) 1, Quartile 3 (<1.960) 1, Quartile 4 1, Z CCA, common carotid artery, ICA, internal carotid artery. ease. Obviously of great interest is the relation between the ultrasound measures and incident coronary disease. The Kuopio Ischemic Heart Disease Risk Factor Study reported that the risk of acute coronary events in 1,288 men followed for a period varying between 1 month and 2.5 years was 3.29 times greater for those having any structural changes in the carotid artery compared with those with normal vessels. 31 As we report in relation to prevalent disease, they found that carotid stenosis carried a higher risk for future coronary events than did wall thickening. Because CHS is a prospective, observational, epidemiological study of risk factors for coronary heart disease and stroke, we will be able to explore this topic in a future report. Although all carotid artery measures were related to a history of stroke, the association was not as strong as expected. The likelihood of stroke was 2.7 times greater for those with 50-74% stenosis compared with those having 0% stenosis, but it only increased to 2.9 for those having % stenosis. It is an underappreciated fact that only 9% of ischemic strokes are thought to be caused by extracranial carotid disease. 32 The majority of patients with incident stroke do not have significant extracranial carotid disease by noninvasive study- 33 Despite these considerations, the association of carotid disease with ischemic stroke is almost certain to be substantially greater than is suggested by the data from our prevalence series. First, approximately 15% of all strokes are either subarachnoid or intracerebral hemorrhage and did not involve an atherosclerotic mechanism. Second, and more importantly, patients with ischemic stroke and significant carotid atherosclerosis have a much worse survival likelihood than the average ischemic stroke patient, leading to a survivor bias in any prevalence series. In the Stroke Data Bank, the inhospital mortality was only 1% for lacunar infarction compared with 18% for atherosclerotic infarction. 32 Furthermore, there is evidence that the mortality rate

8 O'Leary et al Sonographically Detected Carotid Artery Disease in CHS 1759 after discharge is higher for stroke patients with coronary artery disease as well as atherogenic risk factors such as diabetes and hypertension. 34 Wall thickness measurements showed stronger associations with a history of stroke than the more standard measure of percent diameter stenosis. This finding, which has not been previously reported, was true for both internal and common carotid artery wall measurements. The value of common carotid wall thickness measurements for evaluating the severity of carotid atherosclerosis was expected. Poli et al 35 have reported that the carotid far walls of 36 hypercholesterolemic patients were thicker than those of 31 normolipidemic subjects (/j<0.001) of similar ages. Although common carotid artery wall measurements are easier to make and more reproducible than internal carotid artery wall measurements, 16^6 it was not known whether thickening of the common carotid artery wall would prove an independent marker of atherosclerosis in the elderly. Our results suggest that this measurement is as strong a correlate of coronary disease and stroke as stenosis or internal carotid artery wall measurements. This is not to suggest that either the internal or the common carotid artery wall measurements should be used exclusively. Although the odds ratios for the presence of coronary disease or stroke are greater for the higher quartile of wall thickness measurements of both common carotid and internal carotid arteries, these measurements have slightly different predictive value. For example, logistic regression for coronary disease risk shows that the internal carotid artery wall thickness has a higher correlation value than the common carotid artery wall measurement. A similar model for stroke risk shows that both measures contribute. At the same time, a smaller change in common carotid wall thickness is necessary to significantly increase the risk of prior stroke. For example, a 0.22-mm change in common carotid wall thickness contributes approximately the same as a 0.69-mm change in internal carotid wall thickness. Nevertheless, it is striking that the strength of the relation of the common and internal carotid arteries is so similar in terms of 1-SD change. Each of the wall thickness measures showed about a 20% increase in risk for a 1-SD change. This additive effect will be magnified in many people because thick walls in the common carotid are associated with thick walls in the internal carotid. For example, someone with both the common and the internal wall thicknesses 1-SD larger than a comparable person would be estimated to have a 47% greater risk of stroke. There are significant advantages in using wall thickness measurements in population studies rather than stenosis categories traditionally used in clinical studies. Studies of the variability associated with measurement of stenosis by ultrasound or angiography have shown that within- and between-observer agreement is better when arteries are normal or highly diseased and worse when they are mildly or moderately diseased Our results demonstrate that the majority of elderly subjects have mild disease. Also, wall thickness measurements are continuous variables as opposed to stenosis categories. Most importantly, only small numbers of participants will fall into the higher levels of disease categories when defined as percent stenosis. Of the 5,176 subjects analyzed in our study, only 316 had % stenosis. Yet, when the population was divided into quartiles based on wall thickness, the more than 2,500 participants in the higher two quartiles showed increased probability of prevalent coronary heart disease and stroke when compared with those in the lowest quartile of similar magnitude to that of the participants with % stenosis compared with those with 0% stenosis. Thus, wall thickness identified over 2,000 participants with significant disease who were missed when assessed by percent stenosis criteria. In this cross-sectional analysis of an elderly population, both established risk factors for atherosclerosis and coronary disease or stroke correlated significantly with measures of carotid artery disease by ultrasound. The presence of increases in wall thickness beyond those predicted by age and sex may identify individuals at higher risk for coronary disease and stroke. Future follow-up of the CHS cohort will help determine if these measures of subclinical disease are better independent predictors of clinical disease than the traditional cardiovascular disease risk factors. References 1. Chimowitz MI, Mancini GBJ: Asymptomatic coronary artery disease in patients with stroke: Prevalence, prognosis, diagnosis, and treatment. Stroke : Young W, Gofman JW, Tandy R, Malamud N, Waters ESG: The quantification of atherosclerosis: III. The extent of correlation of degrees of atherosclerosis within and between the coronary and cerebral vascular beds. Am J Cardiol 1960;6: Mathur KS, Kasbyap SK, Kumar V: Correlation of the extent and severity of atherosclerosis in the coronary and cerebral arteries. Circulation 1963^7: Solberg LA, Eggen DA: Localization and sequence of development of atherosclerotic lesions in the carotid and vertebral arteries. Circulation 1971;43: Hertzer NR, Young JR, Beven EG, Graor RA, O'Hara PJ, Ruschhaupt WF, dewolfe VB, Maljovec LC: Coronary angiography in 506 patients with extracranial cerebrovascular disease. Arch Intern Mtd 1985;145: Svindland A, Torvik A: Atherosclerotic carotid disease in asymptomatic individuals: An histologic study of 53 cases. Ada Neural Scand 1988;78:5O Bassiouny HS, Davis H, Massawa N, Gewertz BL, Glagov S, Zarins CK: Critical carotid stenoses: Morphological and chemical similarity between symptomatic and asymptomatic plaques. J Vase Surg 1989^9: Chimowitz MI, Lafranchise EF, Furlan AJ, Dorosti K, Paranandi L, Beck GJ: Evaluation of coexistent carotid and coronary disease in combined angiography. J Stroke Certbrovasc Dis 1991;l: Crouse JR, Toole JF, McKinney WM, Dignan MB, Howard G, Kahl FR, McMahan MR, Harpold GH: Risk factors for extracranial carotid atherosclerosis. Stroke 1987;18: Grotta JC, Yatsu FM, Pettigrew LC, Rhoades H, Bratina P, Vital D, Alam R, Earls R, Picone C: Prediction of carotid stenosis progression by lipid and hematologic measurements. Neurology : Craven TE, Ryu JE, Espeland MA, Kahl FR, McKinney WM, Toole JF, McMahan MR, Thompson CJ, Heiss G, Crouse JR: Evaluation of the associations between carotid artery atherosclerosis and coronary artery stenosis: A case-control study. Circulation 199O;82:123O Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH, Manolio TA, Mittelmark MB, Newman A, O'Leary DH, Psaty B, Rautaharju P, Tracy RP, Weiler PG, for the CHS Collaborative Research Group: The Cardiovascular Health Study: Design and rationale. Ann Epidemiol 1991;l: WHO: Diabetes mellitus: Report of a WHO study group. WHO Technical Report Series, Geneva, WHO, Tell GS, Howard G, McKinney WM: Risk factors for site specific extracranial carotid artery plaque distribution as measured by B-mode ultrasound. / Clin Epidemiol 1989;

9 1760 Stroke Vol 23, 12 December Zarins CK, Glagov S, Giddens DP: What do we find in human atherosclerosis that provides insight into the hemodynamic factors in atherogenesis, in Glagov S, Newman WP, Schaffer SA (eds): Pathobiology of the Human Atherosclerotic Plaque. New York, Springer-Verlag New York, Inc, 1990, pp O'Leary DH, Polak JF, Wolfson SK, Bond MG, Bommer W, Sheth S, Psaty BM, Sharrett AR, Manolio TA: Use of sonography to evaluate carotid atherosclerosis in the elderly: The Cardiovascular Health Study. Stroke 1991;22: Salonen R, Seppanen K, Rauramaa R, Salonen JT: Prevalence of carotid atherosclerosis and serum cholesterol levels in Eastern Finland. Arteriosclerosis 1988;8:l O'Leary DH, Anderson KM, Wolf PA, Evans JC, Poehlman HW: Cholesterol and carotid atherosclerosis in older persons: The Framingham Study. Ann Epidemiol 1992;2: Bots ML, Breslau PJ, Briet E, de Bruyn AM, van Vliet HHDM, van den Ouweland FA, de Jong PTVM, Hofman A, Grobbee DE: Cardiovascular determinants of carotid artery disease: The Rotterdam Elderly Study. Hypertension 1992;19: Tell G, Howard G, McKinney WM: Risk factors for site-specific extracranial carotid artery plaque as measured by B-mode ultrasound. J Clin Epidemiol 1989;42: Handa N, Matsumoto M, Maeda H, Hougaku H, Ogawa S, Fukunaga R, Yoneda S, Kimura K, Kamada T: Ultrasonic evaluation of early atherosclerosis. Stroke 1990;21: Salonen R, Salonen JT: Progression of carotid atherosclerosis and its determinants: A population-based ultrasonography study. Atherosclerosis 1990;81: Salonen R, Salonen JT: Determinants of carotid intima-media thickness: A population-based ultrasonography study in Eastern Finnish men. J Intern Med 1991;229: Pujia A, Rubba P, Spencer MP: Prevalence of extracranial carotid artery disease detectable by echo-doppler in an elderly population. Stroke 1992;23: Dempsey RJ, Moore RW: Amount of smoking independently predicts carotid artery atherosclerosis severity. Stroke 1992;23: Ricci S, Flamini FO, Celani MG, Marini M, Antonini D, Bartolini S, Ballatori E: Prevalence of internal carotid artery stenosis in subjects older than 49 years: A population study. Cerebrovasc Dis 27. Gostomzyk JG, Heller WD, Gerhardt P, Lee PN, Keil U: B-scan ultrasound examination of the carotid arteries within a representative population (MONICA Project Ausburg). Klin Wochenschr 1988;66(suppl XI): Ryu JE, Howard G, Craven TE, Bond MG, Hagaman AP, Crouse JR: Postprandial triglyceridemia and carotid atherosclerosis in middle-aged subjects. Stroke 1992;23: Salonen R, Salonen JT: Carotid atherosclerosis in relation to systolic and diastolic blood pressure: Kuopio Ischemic Heart Disease Risk Factor Study. Ann Med 1991;23: Di Pasquale GD, Andreoli A, Pinelli G, Grazi P, Manini G, Tognetti F, Testa C: Cerebral ischemia and asymptomatic coronary artery disease: A prospective study of 83 patients. Stroke 1986;17: Salonen JT, Salonen R: Ultrasonographically assessed carotid morphology and the risk of coronary heart disease. Arterioscler Thromb 1991;11: Foulkes MA, Wolf PA, Price TR, Mohr JP, Hier DB: The Stroke Data Bank design, methods and baseline characteristics. Stroke 1988;19: Zhu CZ, rris JW: The role of carotid stenosis in ischemic stroke. Stroke 1990;21: Sacco RL, Hauser WA, Mohr JP, Foulkes MA: One-year outcome after cerebral infarction in whites, blacks, and Hispanics. Stroke 1991;22: Poli A, Tremoli E, Colombo A, Sirtori M, Pignoli P, Paoletti R: Ultrasonographic measurement of the common carotid arterial wall thickness in hypercholesterolemic patients. Atherosclerosis 1988;70: Crouse JR, Harpld GH, Kahl FR, Toole JF, McKinney WM: Evaluation of a scoring system for extracranial carotid atherosclerosis extent with B-mode ultrasound. Stroke 1986;17: O'Leary DH, Bryant FA, Goodison MW, Rifkin M, Gramiak R, Ball M, Bond MG, Dunn RA, Goldberg BB, Toole JF, Wheeler HG, Gustafson N, Ekholm S, Raines J: Measurement variability of carotid atherosclerosis: Real time (B mode) ultrasound and angiography. Stroke 1987;18: Chikos PM, Fisher LD, Hirsch JH, Harley JD, Thiele BL, Strandness DE: Observer variability in evaluating carotid artery stenosis. Stroke 1983; 14: