- Email: [email protected]
Risk factors for site specific extracranial carotid artery plaque distribution as measured by B-mode ultrasound
J Clin Qidmtiol Vol. 42, No. 6, pp. 551-559,1989 Printedin Great Britain. All rights rcscrvcd
0895-4356/89 $3.00 + 0.00 Copyright 0 1989 Pergamon Press plc
RISK FACTORS FOR SITE SPECIFIC EXTRACRANIAL CAROTID ARTERY PLAQUE DISTRIBUTION AS MEASURED BY B-MODE ULTRASOUND GRETHES. TELL,‘,z* GEORGEHOWARD’*~ and WILLIAMM. MCKINNEY~ Center for Prevention Research and Biometry, *Stroke Center and 3Neurosonology, Department of Neurology, Bowman Gray School of Medicine, 300 South Hawthorne Road, Winston-Salem, NC 27103. U.S.A. (Received in revised form 17 October 1988)
Abstract-The effect of age, sex, diabetes, hypertension, pulse rate and cigarette smoking on extracranial carotid artery plaque thickness evaluated by B-mode ultrasonography was investigated in a cross-sectional study of 698 white men, 730 white women, 77 black men and 76 black women as part of a clinical ultrasound registry. Subjects were between 24 and 98 years of age, with a mean age of 63 years. Arterial locations evaluated were: proximal, mid and distal common carotid; the bifurcation; and the proximal and mid internal and external carotids. In a general linear multivariate analysis with plaque thickness at each site as the outcome variable, cigarette smoking and age were the two most consistent risk factors, and a!Iected plaque thickness at all the investigated sites. Hypertension affected more sites than diabetes, men had more plaques than women, and except for the common carotid, whites had more plaques than blacks. Thus, risk factors were not uniformly associated with atherosclerosis at all sites of the extracranial carotid arteries. Atherosclerosis
B-mode ultrasound
Carotid artery
INTRODUCMON
Autopsy studies have shown that atherosclerosis does not uniformly develop in all arteries or arterial segments in the human body. Even within the carotid system, there are certain anatomic sites of predilection for severe atherosclerotic lesions, including the bifurcation of the common carotid arteries, the sinus portion and the curved terminal part of the internal carotid artery: the cavernous part [1,2]. During the last years, noninvasive procedures employing high resolution B-mode scanning have become available to study even mild vascular lesions and wall irregularities of extracranial neck arteries [3,4]. This development has enabled investigators to study asymptomatic healthy subjects as well as symptomatic patients *Author for correspondence. rel. (919)-748-2981. Supported in part by Public Health Service Grant NINCDS-NSO6655 and Colonel CC Smith Research Fund.
Cerebrovascular disease
and thus cover a larger range of atherosclerotic lesions than what was previously possible. It also provides a method for the investigation of different arterial segments of the extracranial carotid arteries, allowing for in vivo and repeated comparisons of the effects of various risk factors on atherosclerosis formation. This article examines the distribution of atherosclerotic lesions in different regions of the extracranial carotid distribution in relation to age, sex, pulse rate, diabetes, hypertension and smoking history in a sample of 1428 white and 153 black men and women in the Southeast United States. MATERIALS
AND MRTI-IODS
Subjects and methodr In 1984 the Clinical Neurosonology Laboratory of the North Carolina Baptist Hospital/ Bowman Gray School of Medicine established a clinical database ultrasound registry of patients 551
552
G~THE S.
referred for diagnostic carotid ultrasonographic evaluation. This paper is based on 1428 white and 153 black men and women who between 1984 and 1986 were included in the registry. Patients were referred for a variety of reasons: stroke (lo%), transient ischemic attacks (TIA’s) (19%), carotid bruits (29%), a mixture of cerebrovascular symptoms including stroke, TIA’s and bruits (19%), and 23% who had no cerebrovascular symptoms (this group consisted mostly of patients with coronary artery disease). Patients were examined using a high-resolution Biosound Compact scanner (Biosound Inc., Indianapolis, Ind.). This instrument utilizes an 8 mHz transducer, mechanical sector type, with a spatial resolving capability of +0.3 mm. Measurements were made by the sonographer at the time of scanning using the instrument’s electronic calipers, with measurement readout in tenths of a millimeter. The minimal measurement with this system is 1.0 mm. Reproducibility of measurements were checked against an ultrasound phantom, RMI model 409A. Repeated measurements of target object separation, phantom vessel lumen, and simulated wall thickness varied by +O. 1 mm over an 8 month evaluation period. To facilitate scanning in multiple planes and to achieve best access to the neck vessels, patients were seated in a dental chair with their head resting against the headrest during the examinations. Longitudinal and transverse images were obtained and videotaped from anteroposterior, lateral and posterolateral
jl
TELL et al.
Table 1. Sex and age distribution of examined patients Whites Age (yr) 2&24 25-29 3cL34 3539 4w 4549 5&54 55-59 6@-64 6570 7&74 75-79 8684 85-89 90-94 95+ Total
Blacks
Male N
Female N
Male N
Female N
I
2 5 11 8 24 38 72 90 100 114 120 84 41 15 6 0 730
0 0 0 0 5 2 10 14 14 14 7 6 2 1 1 1 17
1 0 0 0 5 6 I I 10 8 13 10 I 2 0 0 16
2 146 18 40 64 100 134 116 107 59 28 6 3 0 698
projections. No discernible lumen and no appreciable pulsation were defined as obstruction. An evaluation was made at each of eight locations in the carotid arteries: (1) proximal common carotid artery (CCA), (2) mid CCA, (3) distal CCA, (4) bifurcation, (5) proximal internal carotid artery (ICA), (6) mid ICA, (7) proximal external carotid artery (ECA), and (8) mid ECA (Fig. 1). Because very few plaques were recorded at the mid ECA, (97.6% was classified as having no plaques), this site was not considered in the following evaluation. At each of the remaining seven sites, one of three physicians graded the artery wall as being normal, as having either wall thickening, minimal, moderate or severe plaque, or as being occluded (Table 2). The maximum plaque thickness at each site was recorded, whether it was at the near or far wall. For purposes of statistical analyses, we assigned a single number to each of these categories, generally the midpoint of the interval. In addition to the ultrasound scan, a brief medical history was obtained for each patient. A patient was considered hypertensive if he or she Table 2. Plaque thickness category and scores used for statistical analysis
Category
*numbers
indicate distance in cm from the bifurcation
Fig. 1. Extracranial carotid artery sites examined by B-mode ultrasound.
Normal Wall thickening Minimal plaque Moderate plaque Severe plaque Occluded
Plaque thickness (mm) 0.0 0.1-1.0 1.1-2.0 2.1-4.0 24.1
Score for analysis 0.0 0.5 1.5 3.0 5.0 6.0
Risk Factors for Carotid Atherosclerosis reported ever being told by a physician that they were hypertensive and/or if their supine blood pressure taken before the ultrasound exam was above 140 mmHg systolic and/or 90 mmHg diastolic. Patients were considered diabetic if they reported that a physician had informed them they were so. Patients were categorized as being either non-smokers (never smoked or quit more than 10 years ago), past smokers (quit between 10 years and one month ago) or current smokers. The patient’s seated pulse rate was recorded for 30 seconds. The interpreting physician of the ultrasound scan was recorded. This report considers the above factors as possibly influencing the average (mean) plaque thickness in different regions of the extracranial carotid arteries. Data from patients with clearly visualized carotid scans, and with complete data for all of the above mentioned risk factors, were employed in the analysis presented here. Patients who had undergone endarterectomy were excluded. Table 1 shows the sex and age distribution of the patients included in the analysis presented in this paper. Statistical analysis An initial analysis using general linear multivariate model (GLMM) analysis (a generalixation of analysis of variance) was employed separately for each race. The outcome variable was the plaque score at all 14 sites (seven sites on each side of the neck), modeled as a function of the independent variables (demographic and risk factor variables). Testing for statistical significance was performed by standard linear contrasts. This large model was employed to test not only for the effects of the risk factors, but also for regional (site) and hemispheric differences (as well as their interactions). Finding no significant hemispheric effects or interactions, we then reduced the analysis by considering the region score as the mean of the region scores on the two sides. In addition, first order interactions between risk factors were considered, and when found insignificant removed from further modelling. Statistical testing using GLMM was then employed to test for regional, risk factor, and region by risk factor interactions. RESULTS For whites, sex, age, diabetes, hypertension, smoking history, and interpreting physician were all found to jointly influence the mean
553
plaque scores over all the sites combined (p < O.OOOl),while pulse rate had no such effect (p > 0.05). Since no first order interactions between any of the risk factors were highly significant, a main effects model was judged to describe the relationship between the risk factors and mean plaque score at each arterial site. For blacks, only age (p < O.OOOl), cigarette smoking (p = 0.022) and interpreting physician (p = 0.003) significantly influenced the mean plaque score in all the sites combined. As for whites, a main effects model was judged to describe the relationship between the risk factors and mean plaque score at each site. In both races there were no significant hemispheric differences in the plaque scores nor any interaction between region and hemisphere for any of the variables considered. Thus the mean values from the right and left sides are used in the following analysis. There were regional differences in mean plaque scores (p < 0.0001) in all the following analyses for both races. Whites Sex differences in whites (after adjustment for other model factors) can be seen in Fig. 2. There was a significant interaction between sex and arterial region (p < 0.0001). Compared to women, men had more atherosclerosis at all the sites, significantly so at the bifurcation, ICA proximal, and mid CCA. The adjusted effect of age on the mean plaque scores in whites can be seen in Fig. 3. As for sex, there was also a significant (p < 0.0001) interaction between region and age. For clarity of presentation, we separated the ages into three groups; less than 60 years, 60-70 years and 70 years and over. The less than 60 year-group had significantly less mean plaque scores at all the arterial sites compared with the two older age groups. Except for the mid ICA, there were also statistically significant differences between the two older age groups (0.002
554
GRETHE S. TELL et al.
2.5
w Men (N = 698) 0 Women (N = 730)
0.0001 2.0
Mean Plaque score bnmj
0.0001
1.5
1.0 0.035 0.5
0.018
0.088
0.178’
0
Proximal
I
Mid
Distal
I
Bifurcation
Common Carotid Artery
I
PrOYlmal
Mid
Proximal External Carotid Artery
1
Internal Carotid Artery
’ p-values for differences in adjusted means between men and women
Fig. 2. Mean plaque score by sex; whites. Adjusted for age, diabetes, hypertension and cigarette smoking.
??Less
than 80 (N = 495) p1 60-70 (N = 493) 0 Over 70 (N = 448)
r
2.0 Mean Plaque Score
, 5
’
(mm) 1.0
0
Proximal
I
Mid
Distal
Blfurcatlon
Proximal
Mid
Common Carotid Artery
Proxlmat External Carotld Artery
1
I
Internal Carotid Artery
p 4 0.0001 for all differences In adjusted means between 460 years and !70 years, and between 460 years and 60-70 years
Fig. 3. Mean plaque score by age group; whites. Adjusted for sex, diabetes, hypertension and cigarette
2.5
2.0
Mean Plaque Score (mm)
1
I
Non-dlabetlcs (N = 123! (N 193)
? D?labetlcs
??
r1
0.00’ I
l9
1.0 0.006
0.253 r
0.228
0.5 0.633
0
Proximal
Mid
Distal
I
Common Carotld Artery
Blfurcat Ion
_Proximal I
Mtd
_
J Internal Carotld Artery
Proximal External Carotid Artery
* p-values for differences in adjusted means between dlabetlcs and non-dlabctlcs
Fig. 4. Mean plaque score by diabetes status; whites. Adjusted for age, sex, hypertension and cigarette smoking.
Risk Factors for Carotid Atherosclerosis
r
??Normotemlve
O.OW
Mean Plaque score (mm)
555
0
(N = 464) Hypertensive (N 944)
T
??
r
0.000
1.5
1.0
prOXll"8l
Proximal
Mid
_“._..._.
Common Carotid Artery
Internal Carotid Artery
Carotid Artery
* p-values for dlffarcncer In adjusted means between normotenslves and hypartenslves
Fig. 5. Mean plaque score by hypertension status; whites. Adjusted for age, sex, diabetes and cigarette smoking.
(p < 0.0001). Unlike diabetes, hypertension ap pears to be associated with the extent of plaque not only in the immediate bifurcation region, but in the other investigated areas as well (except for the mid ICA). The adjusted effects of cigarette smoking on plaque scores among whites are provided in Fig. 6. There was an overall effect of smoking (p < O.OOOl), and there was also an interaction between smoking and region (p = 0.0015). The largest differences were observed between nonsmokers and current smokers (p < 0.0001 at all sites). Except for the distal CCA, current smokers had significantly higher plaque scores in each region compared with past smokers (0.0001 < p d 0.047), and, except for the mid ICA, past 2.5
2.0
Mefin Plrque Score (mm)
0
Non-rmokrrr
smokers had signticantly higher plaque scores than nonsmokers (0.0001
Except for sex, the investigated risk factors influenced plaque scores in the same direction as was seen for whites. However, because of smaller sample sixes and larger variance in the mean scores, the only significant differences between the various risk groups were found for age and cigarette smoking (graphic results for blacks are not presented). In contrast to whites, black women had more atherosclerosis than black men at all sites except for the mid ICA, although the differences were not statistically significant.
(N - 554) * 422) (N * 451)
?? Paat Smokers (N ??Current Smokers
1.5 1.0 0.046
_Proxlmal
Mid
Dlatu
I
. I
Common Carotid Artery
#mrcrtlon _Pfoximd I
’
Internal Cnotld Artery
p i 0.0001 or all dlfferencel In adjusted memr between non- and current smoker% p-values f( dlfferencer ln adJusted meam between non-smokers and past smoker% ‘* p-value8 f( dlfferencer k ?? dJu8ted meana between part and current smokers.
Proxknal External Carotid Artery
??
Fig. 6. Mean plaque score by cigarette smoking status; whites. Adjusted for age, sex, diabetes and hypertension.
556
GRETHE S. TELL et al. 2.5
0
r
1
2.0
Mean Plaque
Caucasians (N = 1428 (N = 153)
??Blacks
0.0001
0.0001
l5
SCOW (mm)
1.0
0.5
0
.Prowlmal
Mid
Blfurcatlon
Distal
I
I
Common Carotid Artery
. Proximal
Mid
L
’ Internal Carotid Artery
Proximal External Carotid Artery
* p-values for differences in adjusted means between blacks and Caucasians
Fig. 7. Mean plaque score by race. Adjusted for age, sex, diabetes, hypertension and cigarette smoking.
There was a significant interaction between age and region (p < 0.0001) in blacks, and the difference in plaque score between the youngest and the oldest age group was statistically significant at all the sites except at the proximal CCA and mid ICA. The only significant difference between the age group ~60 and 60-70 years was seen at the bifurcation (p = 0.007). Differences between 60-70 years and > 70 were found at the mid CCA (p = 0.027) and at the bifurcation (p = 0.0007). Although the overall effect of cigarette smoking was significant in blacks, the only significant difference at any site was seen at the proximal ICA where current smokers had a statistically significant larger plaque score than nonsmokers (p = 0.044). No interactions between region and smoking status were seen. Black-white d@erences
Figure 7 provides the difference in mean plaque scores between black and white patients, after control for sex, age, diabetes, hypertension and cigarette smoking. There was clear evidence (p < 0.0001) of racial plaque score differences between the regions. There was no hemispheric effect (p = 0.39) or a region by hemisphere interaction (p = 0.18). Compared to blacks, whites had a significantly higher mean plaque score at the bifurcation, the proximal ICA, the mid ICA and the proximal ECA. For both the bifurcation and the proximal ICA this difference was estimated to be approximately 0.4mm. DISCUSSION
The patchy distribution of atherosclerosis in human arteries implies that its development
cannot be explained solely by global circulating factors but suggests that each site might have its own predisposing factors. The disease preferentially affects regions of vascular branching and curvature, suggesting that local systemic factors as well as the local blood-flow pattern influence its development. This study investigated whether recognized atherogenic risk factors would influence the distribution of plaque formation in a relatively small area of the vascular tree; the common, internal and external extracranial carotid arteries as well as the bifurcation. Solberg and Eggen [2] investigated lesions in 961 autopsied cases aged 25-69 years. The common carotid and internal carotid arteries were each divided into five equal segments. Lesions were most prevalent near the bifurcation of the common carotid, in the siphon region as well as in the curved terminal part of the internal carotid artery, the cavernous part. The trend with age and the distribution from segment to segment indicated that the atherosclerotic process begins at the bifurcation and in the siphon region of the internal carotid artery and progresses proximally from there along the common or the internal carotid artery. Similarly, in an autopsy study of 100 cases over 50 years of age, Martin et al. [5] found that the maximal degree of atherosclerosis was in the carotid sinuses and proximal 2 cm of the internal carotid arteries. These results are consistent with our findings of most severe atherosclerosis at the bifurcation, and gradually less atherosclerosis from the proximal to the mid to the distal internal carotid artery. Of all known factors, age has shown the strongest and most consistent association with
Risk Factors for Carotid Atherosclerosis
atherosclerotic lesions of all arteries including the carotid, and the average involvement with raised atherosclerotic lesions increases with age in every population sampled for autopsy [6]. However, because lesions vary greatly among individuals of a given population at each age, the association of age with atherosclerosis is probably due to continued exposure to etiological agents and not to “aging” as an intrinsic process in the arteries. In our study sample, the mean plaque score in the bifurcation and the proximal ICA was approximately 0.05 mm greater with each year of advancing age, while in the CCA proximal the difference was 0.01 mm per year. Coronary atherosclerosis is generally considered to be more extensive in men than in women, at least in white populations [6]. Carotid atherosclerosis has also been shown to be more prevalent in men than in women in autopsy studies, although the differences are smaller and not as consistent as in the coronary arteries [2]. In our study, after controlling for the other measured risk factors, white men had more plaque than white women at all of the investigated regions, significant at the bifurcation, proximal ICA and mid CCA. Although no statistically significant sex differences were seen in our black patients, black women had interestingly consistently more plaque than black men. The International Atherosclerosis Project [A established that diabetes accelerates atherosclerosis in the coronary arteries and abdominal aorta independent of sex, age, race or the presence of hypertension. Our study confirms these findings for the carotid arteries in whites. In contrast, the history of diabetes was not significantly associated with extracranial carotid atherosclerosis score in the Italian multicenter study [8]. Similarly, in a study by Javid et al. [9], incidence of diabetes did not appear to influence the rate of angiographically determined stenotic changes in the carotid bifurcation. However, with reference to the latter study, it is important to note that due to arterial dilatation, atherosclerosis may occur without narrowing of the lumen [lo], in which case a possible relationship would be missed. Although diabetes has been found to increase calcification of existing coronary lesions [l 11, the mechanisms of accelerated atherogenesis in diabetic patients remain to be elucidated, and the role of glucose in atheroma formation, if any, is poorly understood.
557
Antecedent hypertension has been shown to increase the severity of atherosclerosis found at autopsy [12,13], both in the carotid arteries and in other vascular beds. The association between hypertension and atherosclerosis is thought to influence atherosclerosis formation in the intracranial arteries more than in extracranial neck arteries [14, 151. Although statistically significant only among whites, in our study hypertension affected plaque thickness at most of the regions studied, only current smoking and age influenced equally many, or more, regions. Although hypertension is a major risk factor for cardiovascular complications, the mechanism by which hypertension predisposes to accelerated atherogenesis is not clear. There is, however, some evidence for altered medial metabolism and for increased intimal permeability under conditions of increased mural stretching [16]. Also, the medial and intimal thickening characteristic of hypertension may disturb the rate of passage or “clearance” of transintimal perfusates through the wall, favoring accumulation in the intima of particular substances and continued intimal proliferation [ 171. A preponderance of evidence indicates that cigarette smoking has a significant positive association with atherosclerosis in the coronary, aortic, abdominal and peripheral arteries [ 17-2 I]. However, the relationship between cigarette smoking and atherosclerosis in the cerebral vasculature has not been extensively evaluated. In a study of elderly adults with isolated systolic hypertension, cigarette smoking was positively related to the prevalence of carotid atherosclerosis, although not significantly so [22]. Stemby [23] reported that cigarette smokers had significantly more extensive raised lesions in the basilar artery than had nonsmokers, while Holme et al. [24] found no significant correlation between raised lesions in the cerebral vessels and the number of cigarettes smoked. Candelise et al. [25] reported that smoking was significantly associated with severity of extracranial disease in 11 arterial segments, independent of the effect of age. These investigators also reported a doseresponse effect of cigarette smoking; the larger the number of cigarettes smoked per day, the greater the atherosclerosis score. We report here that the extent of atherosclerosis in cigarette smokers is significantly greater at all the investigated areas compared with nonsmokers, at least in whites. The differences in mean plaque scores
558
GIWHE S. TELLetal.
between past smokers and nonsmokers were smaller than the differences between current smokers and nonsmokers, a finding that we have discussed in more detail previously [26]. Severai studies have shown that in general, whites have more atherosclerotic lesions in extracranial vessels, while blacks have a predominance of intracranial vascular lesions [ 14,27,28]. Our study found racial differences in atherosclerosis distribution within the extracranial carotid circulation. Whites had more disease at the bifurcation and at the origin of both ECA and ICA compared to blacks. These differences were not explained by racial differences in the prevalence of hypertension, diabetes, smoking or age, and are in contrast to findings from an autopsy study in New Orleans, where no consistent differences in mean involvement of total carotid raised lesions between white and black men within each lo-year age group was found [29]. Several methodological questions arise in ultrasonographic studies of carotid artery disease. The inherent limitations of B-mode ultrasound in assessing atherosclerotic lesions need to be considered. Variability exists in their interpretation, with between-reader agreement reported to range from 34% in the ECA to 65% in the ICA and within-reader agreement ranging from 64% in the ECA to 73% in the ICA [30]. In addition, as compared with pathology data of endarterectomy specimens, B-mode measurements of lesion width underestimate the amount of disease, and the two measurement methods has been found to differ with more than 1 mm in as much as 64% of investigated vessels [31]. Although the major focus of this report is the varying degree with which risk factors for atherosclerosis are associated with different arterial sites, it is important to note that the patients in this study constituted a select group. Thus, our findings may not be generalizable to the general population. However, possible selection biases in the current study could in fact have resulted in an underestimation of the actual relation of carotid atherosclerotic disease to risk factor levels, due to the under-representation of people without carotid disease making our study sample less heterogeneous. The fact that the vast majority of patients included in this registry are sick may also attenuate the observed relationships between risk factors and atherosclerosis compared with studies including both sick and healthy people. In spite of inherent shortcomings in using hospital-based registry
data, our study is nevertheless the first to present data on carotid atherosclerosis distribution in vivo on a large number of persons. These data are currently being used to generate hypotheses to be tested in population-based studies of the natural course of carotid atherosclerosis, as well as in intervention studies of atherosclerosis progression. Acknowledgements-We are indebted to the ultrasonographers at the Center for Medical Ultrasound, Bowman Gray School of Medicine, without whose skillful work this study could not have been done: Catherine Nunn, R.N., R.V.T.; Lawrence Myers, R.D.M.S., R.V.T.; Dana Meads, R.T.-R., R.V.T.
REFERENCES 1. Samuel KC. Atherosclerosis and occlusion of the internal carotid artery. J Path Bnct 1956; LXXI: 391-401. 2. Solberg LA, Eggen DA. Localization and sequence of development of atherosclerotic lesions in the carotid and vertebral arteries. Circulation 1971; 43: 71 l-724. 3. von Reutem GM. Functional and morphological evaluation of the cerebral cirulation by ultras&d. In: Poeck K. Freund HJ. Ganshirt H. Eds. Neumlorzv-ProceedIngs ‘of tbe XIIIth World Congress of Ne&&gy. Berlin: Springer Verlag; 1986, 441452. 4. Hennerici M, Reifschneider G, Trockel U, Aulich A. Detection of early atherosclerotic lesions by Duplex scanning of the carotid artery. J CBn UItrasound 1984; 12: 455464. 5. Martin MJ, Whisnant JP, Sayre GP. Occlusive vascular disease in the extracranial cerebral circulation. Areh Neurol 1960; 5: 64-72. 6. Strong JP, Eggen DA, Oalmann MC. The natural history, geographic pathology and epidemiology of atherosclerosis. In: Wissler RW, Gear JC, Eds. The Pathogenesis of Atherosclerosis. Baltimore: Williams & Wilkins; 1972: 164-199. 7. Robertson WB, Strong JP. Atherosclerosis in persons with hypertension and diabetes mellitus. Lab Invest 1968; 18: 538-551. 8. Passer0 S, Rossi G, Nardini M, Bonelli G, D’Ettore MD, Martini A, Battistini N, Albanese V, Bono G, Brambilla GL, Candelise L, De Zanche L, Inzitari D; Rasure M, Fieschi C. Italian multicenter studv of reversible cerebral ischemic attacks. Part 5. Risk-factors and cerebral atherosclerosis. Atbermcleragis 1987; 63: 21 l-224. 9. Javid H, Ostermiller WE, Hengesh JW, Dye WS, Hunter JA, Najafi H, Julian OC. Natural history of carotid bifurcation atheroma. Surgery 1970; 67: 8&86. 10. Bond MG, Sawyer JK, Bullock BC, Barnes RW, Ball MR. Animal studies of atherosclerosis progression and regression. In: Bond MG, Instill W, Glagov S, Chandler AB, Comhill JF, Eds. CLinIcal Magnu& of Atberosclerosiq Quantitative Methods of Evaluation. New York: Springer Verlag; 1982: 435451. 11. Waller BF, Palumbo PJ, Lie JT et al. The heart in diabetes mellitus as viewed from a morphologic perspective. In: Scott RC, Ed. ClInIcal Cardiology and Diabetes Part 1. Fundamental Considerations in Cardiology ~IUIDiabetes. Mount Kisco, New York: Futura; 1983: 83-125. 12. McGill HC, Ed. The Geographic Pathology of Atberosclemis. Baltimore: Williams & Wilkins; 1968.
Risk Factors for Carotid Atherosclerosis 13. Solberg LA, McGarry PA. Cerebral atherosclerosis in persons with selected diseases. Lab Invest 1968; 18: 153-159. 14. Solberg LA, McGarry PA. Cerebral atherosclerosis in 1972; 16: Negroes and Caucasians. Adwodeda 141-154. 15. Caplan LR, Gorelick PB, Hier DB. Race, sex and occlusive cerebrovascular disease: a review. Stroke 1986; 17: 648-655. 16. Duncan LE Jr, Buck K, Lynch A. Effect of pressure and stretching on the passage of labelled albumin into canine aortic wall. J Atheroselv Res 1965; 5: 69-79. 17. Glacov S. Hemodynamic risk factors: Mechanical stress, mural architecture, medial nutrition, and the vulnerability of arteries to atherosclerosis. In: Wissler RW, Geer JC, Eds. The Pat&gene& of Atheroscleresis. Baltimore: Williams & Wilkins: 1972: 164-199. 18. Auerback 0, Hammond EC, Garfinkel L. Smoking in relation to atherosclerosis of the coronary arteries. N Eaal J Med 1965: 273: 775-779. 19. S&rig JP, Richards ML, McGill HC, Eggen DA, McCurry M. Gn the association of cigarette smoking with coronary and aortic atherosclerosis. J Athemacler Res 1969; IO: 303-317. 20. Sackett DL, Gibson RW, Brass IDJ, Pickren JW. Relation between aortic atherosclerosis and the use of cigarettes and alcohol. N EngI J Med 1968; 279: 1413-1420. 21. Strong JP, Richards ML. Cigarette smoking and atherosclerosis in autopsied men. Atherosclerosis 1976; 23: 451-476. 22. Sutton KC, Wolfson SK, Kuller LH, Carotid and lower extremity arterial disease in elderly adults with isolated systolic hypertension. Stroke 1987; 18: 817-822. 23. Stemby NH. Atherosclerosis, smoking and other risk factors. In: Gotto AM Jr, Smith LC, Allen B, Eds. A&re&em&V.ProcPUthMSymponAtheros&roais, Houston, Texas, 6-9 November 1979. New York: Springer-Verlag; 1980: 67-70.
24.
25.
26.
27.
28.
29.
30.
31.
559
Holme I, Enger SC, Helgeland A, Hjermann I, Leren P, Lund-Larsen PG, Solberg LA, Strong JP. Risk factors and raised atherosclerotic lesions in coronary and cerebral arteries. Statistical analysis from the Oslo study. &t&&em& 1981: 1: 250-256. Candelise L, Bianchi F, Gahigoni F, Albanese V, Bonelli 0. Boxzao L. Inxitari D. Mariani F. Rasura M, Rognbni F, Sangiovanni G, Fieschi d. Italian Multicenter Study on Reversible Cerebral Ischemic Attacks: III-Infhtence of age and risk factors on cerebrovascular atherosclerosis. Stroke 1984; 15: 379-382. Tell GS, Howard G, &Kinney WM, Toole JF. Cigarette smoking cessation and extracranial carotid atherosclerosis. JAMA 1989; 261: 11781180. Gorelick PB, Caplan LR, Hier DB, Parker SL, Pate1 D. Racial differences in the distribution of anterior circulation occlusive disease. NeuroIogy 1984; 34: 54-59. Heyden S, Heyman A, Goree JA. Nonembolic occlusion of the middle cerebral and carotid arteries-A comparison of predisposing factors. Stroke 1970; 1: 363-369. McGarry P, Solberg LA, Guxman MA, Strong JP. Cerebral atherosclerosis in New Orleans. Comparisons of lesions by age, sex, and race. Lab Invest 1985; 52: 533-539. O’Leary D, Bryan FA, Goodison MW, Ritkin MD, Gram& R, Bail M, Bond MG, Dunn RA, Goldberg BB. Toole JF. Wheeler HG. Gus&on NF. Ekholm S. Raines JK. ’Measurement variability of carotid atherosclerosis: Real-time (B-mode) uhrasonography - _ _ and angiography. Stroke 1987; 18: -101I-1017. Schenk EA. Bond G. Aretx TH. Anaelo JN. Choi HY. Rynalski T, Gustaf&m NF, Betso; AS, Ricotta JJI Goodison MW, Bryan FA, Goldberg BB, Toole JF, O’Leary DH. Multicenter validation study of real-time ultrasonography, arteriography and pathology: Pathologic evaluation of carotid endarterectomy specimens. Stroke 1988; 19: 289-296.
0895-4356/89 $3.00 + 0.00 Copyright 0 1989 Pergamon Press plc
RISK FACTORS FOR SITE SPECIFIC EXTRACRANIAL CAROTID ARTERY PLAQUE DISTRIBUTION AS MEASURED BY B-MODE ULTRASOUND GRETHES. TELL,‘,z* GEORGEHOWARD’*~ and WILLIAMM. MCKINNEY~ Center for Prevention Research and Biometry, *Stroke Center and 3Neurosonology, Department of Neurology, Bowman Gray School of Medicine, 300 South Hawthorne Road, Winston-Salem, NC 27103. U.S.A. (Received in revised form 17 October 1988)
Abstract-The effect of age, sex, diabetes, hypertension, pulse rate and cigarette smoking on extracranial carotid artery plaque thickness evaluated by B-mode ultrasonography was investigated in a cross-sectional study of 698 white men, 730 white women, 77 black men and 76 black women as part of a clinical ultrasound registry. Subjects were between 24 and 98 years of age, with a mean age of 63 years. Arterial locations evaluated were: proximal, mid and distal common carotid; the bifurcation; and the proximal and mid internal and external carotids. In a general linear multivariate analysis with plaque thickness at each site as the outcome variable, cigarette smoking and age were the two most consistent risk factors, and a!Iected plaque thickness at all the investigated sites. Hypertension affected more sites than diabetes, men had more plaques than women, and except for the common carotid, whites had more plaques than blacks. Thus, risk factors were not uniformly associated with atherosclerosis at all sites of the extracranial carotid arteries. Atherosclerosis
B-mode ultrasound
Carotid artery
INTRODUCMON
Autopsy studies have shown that atherosclerosis does not uniformly develop in all arteries or arterial segments in the human body. Even within the carotid system, there are certain anatomic sites of predilection for severe atherosclerotic lesions, including the bifurcation of the common carotid arteries, the sinus portion and the curved terminal part of the internal carotid artery: the cavernous part [1,2]. During the last years, noninvasive procedures employing high resolution B-mode scanning have become available to study even mild vascular lesions and wall irregularities of extracranial neck arteries [3,4]. This development has enabled investigators to study asymptomatic healthy subjects as well as symptomatic patients *Author for correspondence. rel. (919)-748-2981. Supported in part by Public Health Service Grant NINCDS-NSO6655 and Colonel CC Smith Research Fund.
Cerebrovascular disease
and thus cover a larger range of atherosclerotic lesions than what was previously possible. It also provides a method for the investigation of different arterial segments of the extracranial carotid arteries, allowing for in vivo and repeated comparisons of the effects of various risk factors on atherosclerosis formation. This article examines the distribution of atherosclerotic lesions in different regions of the extracranial carotid distribution in relation to age, sex, pulse rate, diabetes, hypertension and smoking history in a sample of 1428 white and 153 black men and women in the Southeast United States. MATERIALS
AND MRTI-IODS
Subjects and methodr In 1984 the Clinical Neurosonology Laboratory of the North Carolina Baptist Hospital/ Bowman Gray School of Medicine established a clinical database ultrasound registry of patients 551
552
G~THE S.
referred for diagnostic carotid ultrasonographic evaluation. This paper is based on 1428 white and 153 black men and women who between 1984 and 1986 were included in the registry. Patients were referred for a variety of reasons: stroke (lo%), transient ischemic attacks (TIA’s) (19%), carotid bruits (29%), a mixture of cerebrovascular symptoms including stroke, TIA’s and bruits (19%), and 23% who had no cerebrovascular symptoms (this group consisted mostly of patients with coronary artery disease). Patients were examined using a high-resolution Biosound Compact scanner (Biosound Inc., Indianapolis, Ind.). This instrument utilizes an 8 mHz transducer, mechanical sector type, with a spatial resolving capability of +0.3 mm. Measurements were made by the sonographer at the time of scanning using the instrument’s electronic calipers, with measurement readout in tenths of a millimeter. The minimal measurement with this system is 1.0 mm. Reproducibility of measurements were checked against an ultrasound phantom, RMI model 409A. Repeated measurements of target object separation, phantom vessel lumen, and simulated wall thickness varied by +O. 1 mm over an 8 month evaluation period. To facilitate scanning in multiple planes and to achieve best access to the neck vessels, patients were seated in a dental chair with their head resting against the headrest during the examinations. Longitudinal and transverse images were obtained and videotaped from anteroposterior, lateral and posterolateral
jl
TELL et al.
Table 1. Sex and age distribution of examined patients Whites Age (yr) 2&24 25-29 3cL34 3539 4w 4549 5&54 55-59 6@-64 6570 7&74 75-79 8684 85-89 90-94 95+ Total
Blacks
Male N
Female N
Male N
Female N
I
2 5 11 8 24 38 72 90 100 114 120 84 41 15 6 0 730
0 0 0 0 5 2 10 14 14 14 7 6 2 1 1 1 17
1 0 0 0 5 6 I I 10 8 13 10 I 2 0 0 16
2 146 18 40 64 100 134 116 107 59 28 6 3 0 698
projections. No discernible lumen and no appreciable pulsation were defined as obstruction. An evaluation was made at each of eight locations in the carotid arteries: (1) proximal common carotid artery (CCA), (2) mid CCA, (3) distal CCA, (4) bifurcation, (5) proximal internal carotid artery (ICA), (6) mid ICA, (7) proximal external carotid artery (ECA), and (8) mid ECA (Fig. 1). Because very few plaques were recorded at the mid ECA, (97.6% was classified as having no plaques), this site was not considered in the following evaluation. At each of the remaining seven sites, one of three physicians graded the artery wall as being normal, as having either wall thickening, minimal, moderate or severe plaque, or as being occluded (Table 2). The maximum plaque thickness at each site was recorded, whether it was at the near or far wall. For purposes of statistical analyses, we assigned a single number to each of these categories, generally the midpoint of the interval. In addition to the ultrasound scan, a brief medical history was obtained for each patient. A patient was considered hypertensive if he or she Table 2. Plaque thickness category and scores used for statistical analysis
Category
*numbers
indicate distance in cm from the bifurcation
Fig. 1. Extracranial carotid artery sites examined by B-mode ultrasound.
Normal Wall thickening Minimal plaque Moderate plaque Severe plaque Occluded
Plaque thickness (mm) 0.0 0.1-1.0 1.1-2.0 2.1-4.0 24.1
Score for analysis 0.0 0.5 1.5 3.0 5.0 6.0
Risk Factors for Carotid Atherosclerosis reported ever being told by a physician that they were hypertensive and/or if their supine blood pressure taken before the ultrasound exam was above 140 mmHg systolic and/or 90 mmHg diastolic. Patients were considered diabetic if they reported that a physician had informed them they were so. Patients were categorized as being either non-smokers (never smoked or quit more than 10 years ago), past smokers (quit between 10 years and one month ago) or current smokers. The patient’s seated pulse rate was recorded for 30 seconds. The interpreting physician of the ultrasound scan was recorded. This report considers the above factors as possibly influencing the average (mean) plaque thickness in different regions of the extracranial carotid arteries. Data from patients with clearly visualized carotid scans, and with complete data for all of the above mentioned risk factors, were employed in the analysis presented here. Patients who had undergone endarterectomy were excluded. Table 1 shows the sex and age distribution of the patients included in the analysis presented in this paper. Statistical analysis An initial analysis using general linear multivariate model (GLMM) analysis (a generalixation of analysis of variance) was employed separately for each race. The outcome variable was the plaque score at all 14 sites (seven sites on each side of the neck), modeled as a function of the independent variables (demographic and risk factor variables). Testing for statistical significance was performed by standard linear contrasts. This large model was employed to test not only for the effects of the risk factors, but also for regional (site) and hemispheric differences (as well as their interactions). Finding no significant hemispheric effects or interactions, we then reduced the analysis by considering the region score as the mean of the region scores on the two sides. In addition, first order interactions between risk factors were considered, and when found insignificant removed from further modelling. Statistical testing using GLMM was then employed to test for regional, risk factor, and region by risk factor interactions. RESULTS For whites, sex, age, diabetes, hypertension, smoking history, and interpreting physician were all found to jointly influence the mean
553
plaque scores over all the sites combined (p < O.OOOl),while pulse rate had no such effect (p > 0.05). Since no first order interactions between any of the risk factors were highly significant, a main effects model was judged to describe the relationship between the risk factors and mean plaque score at each arterial site. For blacks, only age (p < O.OOOl), cigarette smoking (p = 0.022) and interpreting physician (p = 0.003) significantly influenced the mean plaque score in all the sites combined. As for whites, a main effects model was judged to describe the relationship between the risk factors and mean plaque score at each site. In both races there were no significant hemispheric differences in the plaque scores nor any interaction between region and hemisphere for any of the variables considered. Thus the mean values from the right and left sides are used in the following analysis. There were regional differences in mean plaque scores (p < 0.0001) in all the following analyses for both races. Whites Sex differences in whites (after adjustment for other model factors) can be seen in Fig. 2. There was a significant interaction between sex and arterial region (p < 0.0001). Compared to women, men had more atherosclerosis at all the sites, significantly so at the bifurcation, ICA proximal, and mid CCA. The adjusted effect of age on the mean plaque scores in whites can be seen in Fig. 3. As for sex, there was also a significant (p < 0.0001) interaction between region and age. For clarity of presentation, we separated the ages into three groups; less than 60 years, 60-70 years and 70 years and over. The less than 60 year-group had significantly less mean plaque scores at all the arterial sites compared with the two older age groups. Except for the mid ICA, there were also statistically significant differences between the two older age groups (0.002
554
GRETHE S. TELL et al.
2.5
w Men (N = 698) 0 Women (N = 730)
0.0001 2.0
Mean Plaque score bnmj
0.0001
1.5
1.0 0.035 0.5
0.018
0.088
0.178’
0
Proximal
I
Mid
Distal
I
Bifurcation
Common Carotid Artery
I
PrOYlmal
Mid
Proximal External Carotid Artery
1
Internal Carotid Artery
’ p-values for differences in adjusted means between men and women
Fig. 2. Mean plaque score by sex; whites. Adjusted for age, diabetes, hypertension and cigarette smoking.
??Less
than 80 (N = 495) p1 60-70 (N = 493) 0 Over 70 (N = 448)
r
2.0 Mean Plaque Score
, 5
’
(mm) 1.0
0
Proximal
I
Mid
Distal
Blfurcatlon
Proximal
Mid
Common Carotid Artery
Proxlmat External Carotld Artery
1
I
Internal Carotid Artery
p 4 0.0001 for all differences In adjusted means between 460 years and !70 years, and between 460 years and 60-70 years
Fig. 3. Mean plaque score by age group; whites. Adjusted for sex, diabetes, hypertension and cigarette
2.5
2.0
Mean Plaque Score (mm)
1
I
Non-dlabetlcs (N = 123! (N 193)
? D?labetlcs
??
r1
0.00’ I
l9
1.0 0.006
0.253 r
0.228
0.5 0.633
0
Proximal
Mid
Distal
I
Common Carotld Artery
Blfurcat Ion
_Proximal I
Mtd
_
J Internal Carotld Artery
Proximal External Carotid Artery
* p-values for differences in adjusted means between dlabetlcs and non-dlabctlcs
Fig. 4. Mean plaque score by diabetes status; whites. Adjusted for age, sex, hypertension and cigarette smoking.
Risk Factors for Carotid Atherosclerosis
r
??Normotemlve
O.OW
Mean Plaque score (mm)
555
0
(N = 464) Hypertensive (N 944)
T
??
r
0.000
1.5
1.0
prOXll"8l
Proximal
Mid
_“._..._.
Common Carotid Artery
Internal Carotid Artery
Carotid Artery
* p-values for dlffarcncer In adjusted means between normotenslves and hypartenslves
Fig. 5. Mean plaque score by hypertension status; whites. Adjusted for age, sex, diabetes and cigarette smoking.
(p < 0.0001). Unlike diabetes, hypertension ap pears to be associated with the extent of plaque not only in the immediate bifurcation region, but in the other investigated areas as well (except for the mid ICA). The adjusted effects of cigarette smoking on plaque scores among whites are provided in Fig. 6. There was an overall effect of smoking (p < O.OOOl), and there was also an interaction between smoking and region (p = 0.0015). The largest differences were observed between nonsmokers and current smokers (p < 0.0001 at all sites). Except for the distal CCA, current smokers had significantly higher plaque scores in each region compared with past smokers (0.0001 < p d 0.047), and, except for the mid ICA, past 2.5
2.0
Mefin Plrque Score (mm)
0
Non-rmokrrr
smokers had signticantly higher plaque scores than nonsmokers (0.0001
Except for sex, the investigated risk factors influenced plaque scores in the same direction as was seen for whites. However, because of smaller sample sixes and larger variance in the mean scores, the only significant differences between the various risk groups were found for age and cigarette smoking (graphic results for blacks are not presented). In contrast to whites, black women had more atherosclerosis than black men at all sites except for the mid ICA, although the differences were not statistically significant.
(N - 554) * 422) (N * 451)
?? Paat Smokers (N ??Current Smokers
1.5 1.0 0.046
_Proxlmal
Mid
Dlatu
I
. I
Common Carotid Artery
#mrcrtlon _Pfoximd I
’
Internal Cnotld Artery
p i 0.0001 or all dlfferencel In adjusted memr between non- and current smoker% p-values f( dlfferencer ln adJusted meam between non-smokers and past smoker% ‘* p-value8 f( dlfferencer k ?? dJu8ted meana between part and current smokers.
Proxknal External Carotid Artery
??
Fig. 6. Mean plaque score by cigarette smoking status; whites. Adjusted for age, sex, diabetes and hypertension.
556
GRETHE S. TELL et al. 2.5
0
r
1
2.0
Mean Plaque
Caucasians (N = 1428 (N = 153)
??Blacks
0.0001
0.0001
l5
SCOW (mm)
1.0
0.5
0
.Prowlmal
Mid
Blfurcatlon
Distal
I
I
Common Carotid Artery
. Proximal
Mid
L
’ Internal Carotid Artery
Proximal External Carotid Artery
* p-values for differences in adjusted means between blacks and Caucasians
Fig. 7. Mean plaque score by race. Adjusted for age, sex, diabetes, hypertension and cigarette smoking.
There was a significant interaction between age and region (p < 0.0001) in blacks, and the difference in plaque score between the youngest and the oldest age group was statistically significant at all the sites except at the proximal CCA and mid ICA. The only significant difference between the age group ~60 and 60-70 years was seen at the bifurcation (p = 0.007). Differences between 60-70 years and > 70 were found at the mid CCA (p = 0.027) and at the bifurcation (p = 0.0007). Although the overall effect of cigarette smoking was significant in blacks, the only significant difference at any site was seen at the proximal ICA where current smokers had a statistically significant larger plaque score than nonsmokers (p = 0.044). No interactions between region and smoking status were seen. Black-white d@erences
Figure 7 provides the difference in mean plaque scores between black and white patients, after control for sex, age, diabetes, hypertension and cigarette smoking. There was clear evidence (p < 0.0001) of racial plaque score differences between the regions. There was no hemispheric effect (p = 0.39) or a region by hemisphere interaction (p = 0.18). Compared to blacks, whites had a significantly higher mean plaque score at the bifurcation, the proximal ICA, the mid ICA and the proximal ECA. For both the bifurcation and the proximal ICA this difference was estimated to be approximately 0.4mm. DISCUSSION
The patchy distribution of atherosclerosis in human arteries implies that its development
cannot be explained solely by global circulating factors but suggests that each site might have its own predisposing factors. The disease preferentially affects regions of vascular branching and curvature, suggesting that local systemic factors as well as the local blood-flow pattern influence its development. This study investigated whether recognized atherogenic risk factors would influence the distribution of plaque formation in a relatively small area of the vascular tree; the common, internal and external extracranial carotid arteries as well as the bifurcation. Solberg and Eggen [2] investigated lesions in 961 autopsied cases aged 25-69 years. The common carotid and internal carotid arteries were each divided into five equal segments. Lesions were most prevalent near the bifurcation of the common carotid, in the siphon region as well as in the curved terminal part of the internal carotid artery, the cavernous part. The trend with age and the distribution from segment to segment indicated that the atherosclerotic process begins at the bifurcation and in the siphon region of the internal carotid artery and progresses proximally from there along the common or the internal carotid artery. Similarly, in an autopsy study of 100 cases over 50 years of age, Martin et al. [5] found that the maximal degree of atherosclerosis was in the carotid sinuses and proximal 2 cm of the internal carotid arteries. These results are consistent with our findings of most severe atherosclerosis at the bifurcation, and gradually less atherosclerosis from the proximal to the mid to the distal internal carotid artery. Of all known factors, age has shown the strongest and most consistent association with
Risk Factors for Carotid Atherosclerosis
atherosclerotic lesions of all arteries including the carotid, and the average involvement with raised atherosclerotic lesions increases with age in every population sampled for autopsy [6]. However, because lesions vary greatly among individuals of a given population at each age, the association of age with atherosclerosis is probably due to continued exposure to etiological agents and not to “aging” as an intrinsic process in the arteries. In our study sample, the mean plaque score in the bifurcation and the proximal ICA was approximately 0.05 mm greater with each year of advancing age, while in the CCA proximal the difference was 0.01 mm per year. Coronary atherosclerosis is generally considered to be more extensive in men than in women, at least in white populations [6]. Carotid atherosclerosis has also been shown to be more prevalent in men than in women in autopsy studies, although the differences are smaller and not as consistent as in the coronary arteries [2]. In our study, after controlling for the other measured risk factors, white men had more plaque than white women at all of the investigated regions, significant at the bifurcation, proximal ICA and mid CCA. Although no statistically significant sex differences were seen in our black patients, black women had interestingly consistently more plaque than black men. The International Atherosclerosis Project [A established that diabetes accelerates atherosclerosis in the coronary arteries and abdominal aorta independent of sex, age, race or the presence of hypertension. Our study confirms these findings for the carotid arteries in whites. In contrast, the history of diabetes was not significantly associated with extracranial carotid atherosclerosis score in the Italian multicenter study [8]. Similarly, in a study by Javid et al. [9], incidence of diabetes did not appear to influence the rate of angiographically determined stenotic changes in the carotid bifurcation. However, with reference to the latter study, it is important to note that due to arterial dilatation, atherosclerosis may occur without narrowing of the lumen [lo], in which case a possible relationship would be missed. Although diabetes has been found to increase calcification of existing coronary lesions [l 11, the mechanisms of accelerated atherogenesis in diabetic patients remain to be elucidated, and the role of glucose in atheroma formation, if any, is poorly understood.
557
Antecedent hypertension has been shown to increase the severity of atherosclerosis found at autopsy [12,13], both in the carotid arteries and in other vascular beds. The association between hypertension and atherosclerosis is thought to influence atherosclerosis formation in the intracranial arteries more than in extracranial neck arteries [14, 151. Although statistically significant only among whites, in our study hypertension affected plaque thickness at most of the regions studied, only current smoking and age influenced equally many, or more, regions. Although hypertension is a major risk factor for cardiovascular complications, the mechanism by which hypertension predisposes to accelerated atherogenesis is not clear. There is, however, some evidence for altered medial metabolism and for increased intimal permeability under conditions of increased mural stretching [16]. Also, the medial and intimal thickening characteristic of hypertension may disturb the rate of passage or “clearance” of transintimal perfusates through the wall, favoring accumulation in the intima of particular substances and continued intimal proliferation [ 171. A preponderance of evidence indicates that cigarette smoking has a significant positive association with atherosclerosis in the coronary, aortic, abdominal and peripheral arteries [ 17-2 I]. However, the relationship between cigarette smoking and atherosclerosis in the cerebral vasculature has not been extensively evaluated. In a study of elderly adults with isolated systolic hypertension, cigarette smoking was positively related to the prevalence of carotid atherosclerosis, although not significantly so [22]. Stemby [23] reported that cigarette smokers had significantly more extensive raised lesions in the basilar artery than had nonsmokers, while Holme et al. [24] found no significant correlation between raised lesions in the cerebral vessels and the number of cigarettes smoked. Candelise et al. [25] reported that smoking was significantly associated with severity of extracranial disease in 11 arterial segments, independent of the effect of age. These investigators also reported a doseresponse effect of cigarette smoking; the larger the number of cigarettes smoked per day, the greater the atherosclerosis score. We report here that the extent of atherosclerosis in cigarette smokers is significantly greater at all the investigated areas compared with nonsmokers, at least in whites. The differences in mean plaque scores
558
GIWHE S. TELLetal.
between past smokers and nonsmokers were smaller than the differences between current smokers and nonsmokers, a finding that we have discussed in more detail previously [26]. Severai studies have shown that in general, whites have more atherosclerotic lesions in extracranial vessels, while blacks have a predominance of intracranial vascular lesions [ 14,27,28]. Our study found racial differences in atherosclerosis distribution within the extracranial carotid circulation. Whites had more disease at the bifurcation and at the origin of both ECA and ICA compared to blacks. These differences were not explained by racial differences in the prevalence of hypertension, diabetes, smoking or age, and are in contrast to findings from an autopsy study in New Orleans, where no consistent differences in mean involvement of total carotid raised lesions between white and black men within each lo-year age group was found [29]. Several methodological questions arise in ultrasonographic studies of carotid artery disease. The inherent limitations of B-mode ultrasound in assessing atherosclerotic lesions need to be considered. Variability exists in their interpretation, with between-reader agreement reported to range from 34% in the ECA to 65% in the ICA and within-reader agreement ranging from 64% in the ECA to 73% in the ICA [30]. In addition, as compared with pathology data of endarterectomy specimens, B-mode measurements of lesion width underestimate the amount of disease, and the two measurement methods has been found to differ with more than 1 mm in as much as 64% of investigated vessels [31]. Although the major focus of this report is the varying degree with which risk factors for atherosclerosis are associated with different arterial sites, it is important to note that the patients in this study constituted a select group. Thus, our findings may not be generalizable to the general population. However, possible selection biases in the current study could in fact have resulted in an underestimation of the actual relation of carotid atherosclerotic disease to risk factor levels, due to the under-representation of people without carotid disease making our study sample less heterogeneous. The fact that the vast majority of patients included in this registry are sick may also attenuate the observed relationships between risk factors and atherosclerosis compared with studies including both sick and healthy people. In spite of inherent shortcomings in using hospital-based registry
data, our study is nevertheless the first to present data on carotid atherosclerosis distribution in vivo on a large number of persons. These data are currently being used to generate hypotheses to be tested in population-based studies of the natural course of carotid atherosclerosis, as well as in intervention studies of atherosclerosis progression. Acknowledgements-We are indebted to the ultrasonographers at the Center for Medical Ultrasound, Bowman Gray School of Medicine, without whose skillful work this study could not have been done: Catherine Nunn, R.N., R.V.T.; Lawrence Myers, R.D.M.S., R.V.T.; Dana Meads, R.T.-R., R.V.T.
REFERENCES 1. Samuel KC. Atherosclerosis and occlusion of the internal carotid artery. J Path Bnct 1956; LXXI: 391-401. 2. Solberg LA, Eggen DA. Localization and sequence of development of atherosclerotic lesions in the carotid and vertebral arteries. Circulation 1971; 43: 71 l-724. 3. von Reutem GM. Functional and morphological evaluation of the cerebral cirulation by ultras&d. In: Poeck K. Freund HJ. Ganshirt H. Eds. Neumlorzv-ProceedIngs ‘of tbe XIIIth World Congress of Ne&&gy. Berlin: Springer Verlag; 1986, 441452. 4. Hennerici M, Reifschneider G, Trockel U, Aulich A. Detection of early atherosclerotic lesions by Duplex scanning of the carotid artery. J CBn UItrasound 1984; 12: 455464. 5. Martin MJ, Whisnant JP, Sayre GP. Occlusive vascular disease in the extracranial cerebral circulation. Areh Neurol 1960; 5: 64-72. 6. Strong JP, Eggen DA, Oalmann MC. The natural history, geographic pathology and epidemiology of atherosclerosis. In: Wissler RW, Gear JC, Eds. The Pathogenesis of Atherosclerosis. Baltimore: Williams & Wilkins; 1972: 164-199. 7. Robertson WB, Strong JP. Atherosclerosis in persons with hypertension and diabetes mellitus. Lab Invest 1968; 18: 538-551. 8. Passer0 S, Rossi G, Nardini M, Bonelli G, D’Ettore MD, Martini A, Battistini N, Albanese V, Bono G, Brambilla GL, Candelise L, De Zanche L, Inzitari D; Rasure M, Fieschi C. Italian multicenter studv of reversible cerebral ischemic attacks. Part 5. Risk-factors and cerebral atherosclerosis. Atbermcleragis 1987; 63: 21 l-224. 9. Javid H, Ostermiller WE, Hengesh JW, Dye WS, Hunter JA, Najafi H, Julian OC. Natural history of carotid bifurcation atheroma. Surgery 1970; 67: 8&86. 10. Bond MG, Sawyer JK, Bullock BC, Barnes RW, Ball MR. Animal studies of atherosclerosis progression and regression. In: Bond MG, Instill W, Glagov S, Chandler AB, Comhill JF, Eds. CLinIcal Magnu& of Atberosclerosiq Quantitative Methods of Evaluation. New York: Springer Verlag; 1982: 435451. 11. Waller BF, Palumbo PJ, Lie JT et al. The heart in diabetes mellitus as viewed from a morphologic perspective. In: Scott RC, Ed. ClInIcal Cardiology and Diabetes Part 1. Fundamental Considerations in Cardiology ~IUIDiabetes. Mount Kisco, New York: Futura; 1983: 83-125. 12. McGill HC, Ed. The Geographic Pathology of Atberosclemis. Baltimore: Williams & Wilkins; 1968.
Risk Factors for Carotid Atherosclerosis 13. Solberg LA, McGarry PA. Cerebral atherosclerosis in persons with selected diseases. Lab Invest 1968; 18: 153-159. 14. Solberg LA, McGarry PA. Cerebral atherosclerosis in 1972; 16: Negroes and Caucasians. Adwodeda 141-154. 15. Caplan LR, Gorelick PB, Hier DB. Race, sex and occlusive cerebrovascular disease: a review. Stroke 1986; 17: 648-655. 16. Duncan LE Jr, Buck K, Lynch A. Effect of pressure and stretching on the passage of labelled albumin into canine aortic wall. J Atheroselv Res 1965; 5: 69-79. 17. Glacov S. Hemodynamic risk factors: Mechanical stress, mural architecture, medial nutrition, and the vulnerability of arteries to atherosclerosis. In: Wissler RW, Geer JC, Eds. The Pat&gene& of Atheroscleresis. Baltimore: Williams & Wilkins: 1972: 164-199. 18. Auerback 0, Hammond EC, Garfinkel L. Smoking in relation to atherosclerosis of the coronary arteries. N Eaal J Med 1965: 273: 775-779. 19. S&rig JP, Richards ML, McGill HC, Eggen DA, McCurry M. Gn the association of cigarette smoking with coronary and aortic atherosclerosis. J Athemacler Res 1969; IO: 303-317. 20. Sackett DL, Gibson RW, Brass IDJ, Pickren JW. Relation between aortic atherosclerosis and the use of cigarettes and alcohol. N EngI J Med 1968; 279: 1413-1420. 21. Strong JP, Richards ML. Cigarette smoking and atherosclerosis in autopsied men. Atherosclerosis 1976; 23: 451-476. 22. Sutton KC, Wolfson SK, Kuller LH, Carotid and lower extremity arterial disease in elderly adults with isolated systolic hypertension. Stroke 1987; 18: 817-822. 23. Stemby NH. Atherosclerosis, smoking and other risk factors. In: Gotto AM Jr, Smith LC, Allen B, Eds. A&re&em&V.ProcPUthMSymponAtheros&roais, Houston, Texas, 6-9 November 1979. New York: Springer-Verlag; 1980: 67-70.
24.
25.
26.
27.
28.
29.
30.
31.
559
Holme I, Enger SC, Helgeland A, Hjermann I, Leren P, Lund-Larsen PG, Solberg LA, Strong JP. Risk factors and raised atherosclerotic lesions in coronary and cerebral arteries. Statistical analysis from the Oslo study. &t&&em& 1981: 1: 250-256. Candelise L, Bianchi F, Gahigoni F, Albanese V, Bonelli 0. Boxzao L. Inxitari D. Mariani F. Rasura M, Rognbni F, Sangiovanni G, Fieschi d. Italian Multicenter Study on Reversible Cerebral Ischemic Attacks: III-Infhtence of age and risk factors on cerebrovascular atherosclerosis. Stroke 1984; 15: 379-382. Tell GS, Howard G, &Kinney WM, Toole JF. Cigarette smoking cessation and extracranial carotid atherosclerosis. JAMA 1989; 261: 11781180. Gorelick PB, Caplan LR, Hier DB, Parker SL, Pate1 D. Racial differences in the distribution of anterior circulation occlusive disease. NeuroIogy 1984; 34: 54-59. Heyden S, Heyman A, Goree JA. Nonembolic occlusion of the middle cerebral and carotid arteries-A comparison of predisposing factors. Stroke 1970; 1: 363-369. McGarry P, Solberg LA, Guxman MA, Strong JP. Cerebral atherosclerosis in New Orleans. Comparisons of lesions by age, sex, and race. Lab Invest 1985; 52: 533-539. O’Leary D, Bryan FA, Goodison MW, Ritkin MD, Gram& R, Bail M, Bond MG, Dunn RA, Goldberg BB. Toole JF. Wheeler HG. Gus&on NF. Ekholm S. Raines JK. ’Measurement variability of carotid atherosclerosis: Real-time (B-mode) uhrasonography - _ _ and angiography. Stroke 1987; 18: -101I-1017. Schenk EA. Bond G. Aretx TH. Anaelo JN. Choi HY. Rynalski T, Gustaf&m NF, Betso; AS, Ricotta JJI Goodison MW, Bryan FA, Goldberg BB, Toole JF, O’Leary DH. Multicenter validation study of real-time ultrasonography, arteriography and pathology: Pathologic evaluation of carotid endarterectomy specimens. Stroke 1988; 19: 289-296.