The perceived clinical value of detection of a

11 Correlation of Carotid Bruits and Carotid Atherosclerosis Detected by B-Mode Real-Time Ultrasonography Virginia J. Howard, MSPH, George Howard, DrPh, Gary J. Harpold, MD, Catherine L. Nunn, BSN, Lawrence G. Myers, BS, William M. McKinney, MD, Kari E. Murros, MD, and James F. Toole, MD The extent of carotid atherosclerosis evaluated by B-mode real-time ultrasound and the presence of bruits established by carotid phonoangiography were determined in 1,107 patients. Unilateral bruit was associated with increased atherosclerosis compared with no bruit {p <0.0001). However, there was no association between laterality of the bruit and the degree of atherosclerosis (/7=0.66). There was marginal evidence that patients with bilateral bruits had more severe atherosclerosis than patients with unilateral bruit (p=0.06). The relation between bruit and atherosclerosis categorized by B-mode ultrasound was not sufficient to reliably predict the presence or absence of disease in an individual patient, though the presence of a bruit should be viewed as a risk factor for, or an indicator of, increased risk of systemic atherosclerosis. {Stroke 199;0:11-1) Downloaded from http://ahajournals.org by on January, 019 The perceived clinical value of detection of a carotid bruit as an indicator for an underlying atherosclerotic process has undergone a precipitous decline since 190. Before 196, Gilroy and Meyer 1 and Rennie et al reported arteriographic abnormalities corresponding to the site of bruits in >7% of patients with cerebrovascular disease who had bruits. In 1971, Ziegler et al found the positive predictive value of bruits to be 90% but also found a disturbingly low negative predictive value of 7%. Ueda et al reported that 7% of patients with transient ischemic attacks (TIAs) and bruits also had >60% carotid stenosis. Such reports led Barnes in 190 to conclude that bruits are the most common sign of extracranial carotid artery occlusive disease and to note that bruits rarely From the Department of Neurology (V.J.H., W.M.M., J.F.T.), Stroke Research Center (G.H., J.F.T.), Center for Prevention Research and Biometry (G.H., C.L.N.), and Center for Clinical Ultrasound (L.G.M., W.M.M.), Bowman Gray School of Medicine of Wake Forest University, Winston-Salem, North Carolina; the Department of Neurology, Lewis-Gale Clinic, Salem, Virginia (G.J.H.); and the Department of Neurology, Central Hospital of Central Finland, Jyvaskyla, Finland (K.E.M.). Supported by United States Public Health Service Grant NINDS-066 and the Colonel C.C. Smith Stroke Research Fund. Address for reprints: Virginia Howard, MSPH, Department of Neurology, Bowman Gray School of Medicine of Wake Forest University, 00 South Hawthorne Road, Winston-Salem, NC 710. Received January 0, 19; accepted May, 199. occur in the absence of carotid artery disease. In the same report, Barnes defined a bruit integrally related to the atherosclerotic process at that site as "the result of arterial wall vibrations induced by disturbed or turbulent blood flow immediately distal to a stenosis (narrowing), which reduces the vascular lumen by 0-%." Since 190, numerous investigators have reported that a bruit does not necessarily imply a plaque in the same arterial distribution. In population-based studies, Wolf et al 6 (for the Framingham population) and Heyman et al 7 (for the Evans County population) both reported a doubling of the stroke rate in patients with bruits, but these authors also reported that approximately one half the stroke and TIA events were not directly related to the site of the bruit. In hospital-based studies, authors such as Hennerici et al questioned the predictive value of bruits, noting site-specific bruit correlation in only % of patients with extracranial arterial disease established by direct continuous-wave Doppler ultrasonography. While Pessin et al 9 found bruits to predict underlying atherosclerosis reliably, most clinicians agreed with Murie et al 10 that "although a carotid bruit is often a marker of internal carotid intraluminal turbulence and stenosis, this is by no means always so." Similarly, by digital subtraction angiography Dugdale et al 11 showed the sensitivity of bruits as predictors of stenosis (in either carotid artery) to be 7%, with a specificity of %, but

Downloaded from http://ahajournals.org by on January, 019 1 Stroke Vol 0, No 10, October 199 these authors still concluded that bruits in the elderly are likely to be related to stenosis. Using posteroanterior films of the cervical spine, Imataka et al 1 reported evidence of carotid calcification in only % of patients with bruits, which, while > times the rate in patients without bruits, was a relatively low proportion with evidence of atherosclerosis. Imataka et al 1 noted a higher stroke rate among patients with bruits, but again the side of the stroke did not correlate with that of the bruit. Lusiani et al 1 found that <0% of patients with bruits had >0% stenosis related to the location of the bruit by duplex scanning, while 71% of patients with bruits had either normal carotid arteries or minimal (<0%) stenosis. Most recently, Ingall et al 1 have started the movement back toward the position of bruits being valuable predictors of carotid atherosclerosis. These authors reported moderate-to-severe carotid atherosclerosis (>0% stenosis) evaluated by angiography in >0% of 196 patients with bruits and similar stenosis in only % of 1,11 patients without bruit. The outcome of patients with asymptomatic bruits has been examined in follow-up studies by Roederer et al 1 and by Chambers and Norris. Both reports documented increased risk with a greater degree of asymptomatic stenosis and with progression of plaque. However, each report noted that the presence and location of bruits did not relate directly to underlying plaques. These reports contributed to the recent article by Feussner and Matchar 17 and to the accompanying position paper 1 suggesting that carotid bruits alone are sufficiently benign not to warrant investigation. We examine the relation of bruits and lumen narrowing determined by real-time B-mode ultrasound, with particular regard to laterality of the bruits. Subjects and Methods We evaluated,0 patients in the Neurosonology Laboratory at Bowman Gray School of Medicine of Wake Forest University Medical Center from January 19 through August 196. For each patient, basic demographic data, neurologic symptoms, and medical history (including history of stroke, TIA, cardiac disease, peripheral arterial disease, hypertension, diabetes, hyperlipidemia, smoking, and major surgical procedures) were collected by the nurse/sonographer. Carotid phonoangiography was performed on a Narco-Air Shields (Hatboro, Pennsylvania) carotid phonoangiograph, and a visual record of the bruit (if present) was made by photographing the oscilloscope display. The characteristics of the bruit were interpreted by one of three physicians. Bruits most pronounced at the clavicle and fading distally were interpreted to be referred from the heart or great vessels, and data from patients with such bruits were eliminated from the analysis. Data from patients with no bruit or bruits interpreted to be originating in the bifurcation area at the level of the superior thyroid cartilage and beyond were analyzed. The extracranial carotid system was ultrasonically imaged on a Biosound Compact B-mode instrument (Indianapolis, Indiana). High-resolution B-mode images were obtained using a.0-mhz mechanical sector-type transducer. Analog images were stored on videotape for later interpretation by a physician, and digitized freeze-frame images were used for plaque measurements. Scanning techniques included multiplanar imaging of the carotid system including anterior oblique, lateral, posterior oblique, and transverse images. For each patient, the maximum plaque extent in the common/ internal carotid artery pathway on the right and left sides was arbitrarily scored as 0, normal artery walls; 1, <1.0 mm;, 1.1-.0 mm;,.1-.0 mm;, >.1 mm; or, no blood flow detected. Data from patients whose B-mode plaque score could not be determined because of poorly visualized arteries or indistinct lumens are eliminated from the analysis. For purposes of analysis, the patients were grouped by the presence and location of their bruits; Group 1 patients had no bruits, Group patients had a bruit on the right only, Group patients had a bruit on the left only, and Group patients had bilateral bruits. The right and left B-mode plaque scores of patients within each bruit status group were subsequently analyzed using general linear multivariate modeling, which is closely related to profile analysis. 19 Calculations were performed using the REG procedure of SAS. 0 Results Of the,0 patients, we excluded 97 because they had a bruit interpreted to be referred from the heart or great vessels. Of the remaining 1,11 patients, we excluded 1 because of poorly visualized arteries or an indistinct lumen without Doppler evidence of occlusion. This analysis is therefore based on data from 1,107 patients, 917 without a bruit (Group 1), 6 with a bruit on the right only (Group ), with a bruit on the left only (Group ), and 71 with bilateral bruits (Group ). Table 1 provides a description of the study population. Patients with bruits (combined Groups,, and ) were less likely to be male, were slightly older, and were more likely to have a history of TIA and cardiac disease but less likely to have a history of stroke. For each bruit status group, Table provides the percentage of patients with a given B-mode plaque score for each carotid artery. Two patterns are shown: regardless of bruit status group there was little difference in right and left B-mode plaque scores, and Group and patients were more likely than Group 1 patients but less likely than Group patients to have a high B-mode plaque score. These two patterns were tested statistically.

Howard et al Carotid Bruit and Atherosclerosis 1 TABLE 1. Characteristics of Study Population Characteristic Demographics Caucasian Male Median age (yr) Preexisting risk factors Stroke Transient ischemic attack(s) Cardiac disease Myocardial infarction Coronary artery disease Valvular disease Murmur Irregular heart beat Angina Peripheral arterial disease Hypertension Diabetes High concentration of blood lipids Smoking Without (n =917) No. 799 19 17 1 71 6 1 1 9 1 19 7 6 % 7 1 19 1 7 10 7 1 1 60 Bruit status With (n = 190) No. 17 66 7 1 1 6 9 9 0 11 % 91 1 1 1 1 0 9 9 Downloaded from http://ahajournals.org by on January, 019 For each bruit status group, the mean B-mode plaque scores are provided for the right and left carotid arteries in Table. General linear multivariate modeling indicated significant differences between bruit status groups (p<0.0001). However, there was no significant difference between sides within any group (p=0.7). There was also no significant (/>>0.0) evidence that the mean B- mode plaque scores in Groups and were not symmetrical, that is, the right score for Group was similar to the left score for Group, and the left score for Group was similar to the right score for Group. In an a posteriori analysis to determine which bruit status group(s) differed, we found the mean B-mode plaque score for Groups and to be higher than that for Group 1 (/?<0.0001) and lower than that for Group (p=0.06); however, there was no difference between the scores for Groups and (p>0.0). TABLE. Percentage of Patients With B-Mode Plaque Score by Bruit Status Group Bruit status group Group 1 Group Group Group p-nvxip plamip (n=917) (#i=6) («=) score 0 1 1 7 6 11 7 0 7 7 0 7 7 1 17 1 1 t 6 1 0 1 1 0 Groups: 1, no bruit;, bruit on right only;, bruit on left only;, bilateral[ bruits. Score: 0, normal artery walls; 1, sl.o mm;, 1.1-.0 mm;,..1-.0 mm;, >.1 mm;, no blood flow. TABLE. B-Mode Plaque Scores on Each Side by Bruit Status Group Bruit status group Group 1 Group Group Group Side (»=917).±0.0.±0.0 (#i=6).7±0.1.±0.1 (i=).7±0.1.7±0.1.1±0.1.0±0.1 Data are mean±sem. Groups: 1, no bruit;, bruit on right only;, bruit on left only;, bilateral bruits.

Downloaded from http://ahajournals.org by on January, 019 1 Stroke Vol 0, No 10, October 199 Discussion These data imply that although the presence of a bruit is related to an increased risk of carotid atherosclerosis as defined by B-mode real-time ultrasound, a bruit's location is not closely related to the laterality of atherosclerosis. That is, a unilateral bruit is associated with similar increases in B-mode plaque score for either carotid artery. This finding is consistent with the population studies of Wolf et al 6 and Heyman et al, 7 who found increased stroke and coronary mortality associated with bruits but a poor correlation between the location of the bruit and the hemisphere affected by stroke. In both population studies, the authors concluded that the carotid bruit should be considered a risk factor for systemic atherosclerosis. That a bruit is associated with a general increase in atherosclerotic disease, but not necessarily with the location of the disease, also may explain the frequently reported lack of correlation between bruits and site-specific carotid atherosclerosis. - 10-1 As other authors have noted, the bruit may arise from a wide range of conditions, some as extreme as a normal carotid artery with accelerated blood flow to compensate for a contralateral high-grade stenosis or occlusion..,,11,1 Because of the lack of laterality in the bruit-stenosis relation, bruits should be considered to predict stenosis somewhere in the carotid system. Under this approach, the reported low predictive power of bruits may be greatly improved. Consider, for example, the paper by Hennerici et al, who reported on 7 patients with carotid artery examination before major vascular surgery on the aorta or iliac arteries; 1 (%) had abnormal Doppler results. Of these 1 patients, only (%) had a bruit in the distribution in which the Doppler results were abnormal; hence, Hennerici et al claimed that "if bruits are taken as the only signs of associated EAD [extracranial arterial disease] in patients with systemic atherosclerosis, only 7.6% in a group of 1 patients would have been correctly diagnosed." However, among these 1 patients, an additional 9 (%) had a bruit contralateral to the Doppler abnormality, increasing to 6 the number of patients who had abnormal Doppler results and a bruit, for a total sensitivity of detecting atherosclerosis by Doppler ultrasonography of 1%, nearly double the reported rate. Therefore, the sensitivity is consistent with previous reports of >0%. - n Coupled with the reported specificity of 7%, the presence of a bruit may be seen to be a reasonable screening test for the presence of atherosclerosis as determined by Doppler ultrasonography. This is particularly interesting because Doppler ultrasonography tends to provide positive results only when stenosis exceeds 0%. Our results are also similar to those of Ingall et al, 1 who recently reported non-site-specific atherosclerosis of >0% stenosis (measured angiographically) in % of patients with no bruit but in approximately % of those with a unilateral bruit. However, Ingall et al 1 detected no further increase in patients with bilateral bruits. We document a similarly large difference between average B-mode plaque score in patients with no bruit and in those with a unilateral bruit, but a much smaller (though significant) difference between patients with a unilateral bruit and those with bilateral bruits. In both studies, the difference in risk of atherosclerosis between patients with no bruit and those with a unilateral bruit is much greater than that between patients with a unilateral bruit and those with bilateral bruits. Our report suggesting that bruits are related to an increased risk of atherosclerosis in the carotid system, coupled with the outcome studies of Roederer et al 1 and of Chambers and Norris, who have documented that both high-grade stenosis and advancing stenosis are highly significant predictors of subsequent events in asymptomatic patients, calls into question the recent recommendations against noninvasive testing in asymptomatic bruit patients. There are three specific points of concern. First, only by studying and monitoring patients suspected of having advanced or advancing stenosis can any informed treatment decisions be made. The identification of patients susceptible to the development of atherosclerosis allows intervention in a number of ways. Feussner and Matchar 17 assume that the only treatment for patients with documented atherogenic plaques is carotid endarterectomy, hence discounting the merit of aggressive risk factor modification to prevent the advance of carotid plaque. These interventions may not prove useful once advanced plaques are established; however, the interventions could play a role in preventing plaque progression. Similarly, patients with high-grade stenosis of the carotid system are at increased risk of coronary events, and evaluation and intervention by cardiologists are likely to be fruitful. Second, while we agree that the efficacy of carotid endarterectomy has not been established, denying the patient and his physician the knowledge that a plaque is rapidly progressing to occlusion is ethically questionable. Provided with the current information on the prognosis of advancing high-grade plaques and estimates for stroke following occlusion as high as %/yr, 1 we believe the patient and the physician must make an informed decision concerning the treatment of advancing plaques. The identification of such a high-risk subgroup may simultaneously delineate that segment of the population in which carotid endarterectomy is most efficacious. There are several large-scale randomized endarterectomy trials in both asymptomatic and symptomatic patients that are directed to resolving this ongoing scientific debate. Finally, a number of the assumptions employed by Feussner and Matchar 17 may be in error. For

Downloaded from http://ahajournals.org by on January, 019 example, the relatively low estimate of 0% for the positive predictive value of a bruit would be increased if the bruit were used to predict systemic rather than site-specific disease. Conversely, criticism can also be directed at our study. Our patients were not a sample of the general population, and selection bias must be considered in our interpretation of the results. However, our study does have the advantage of a relatively large sample size and the identification of relatively many patients with a bruit. Even though there is a statistical relation between bruit and atherosclerosis, for an individual patient, the relation does not predict the site of disease. Table shows that the chance of having a plaque >.0 mm in width increases from approximately % for no bruit to 6% for a unilateral bruit and to 79% for bilateral bruits. While these differences are statistically significant, they are not sufficient to conclude that a bruit implies increased carotid atherosclerosis; rather, they suggest that the presence of a bruit places the patient at increased risk of having more severe atherosclerosis. References 1. Gilroy J, Meyer JS: Auscultation of the neck in occlusive cerebrovascular disease. Circulation 196;:00-10. Rennie L, Ejrup B, McDowell F: Arterial bruits in cerebrovascular disease. Neurology 196;1:71-76. Ziegler DK, Zileli T, Dick A, Sebaugh JL: Correlation of bruits over the carotid artery with angiographically demonstrated lesions. Neurology 1971;1:60-6. Ueda K, Toole JF, McHenry LC: Carotid and vertebrobasilar transient ischemic attacks: Clinical and angiographic correlation. Neurology 1979;9:109-1101. Barnes R: Noninvasive evaluation of the carotid bruit. Annu /?eviwerfl90;1:01-1 6. Wolf PA, Kannel WB, Sorlie P, McNamara P: Asymptomatic carotid bruit and risk of stroke. JAMA 191; :1-1 Howard et al Carotid Bruit and Atherosclerosis 1 7. Heyman A, Wilkinson WE, Heyden S, Helms MJ, Bartel AG, Karp HR, Tyroler HR, Hames CG: Risk of stroke in asymptomatic persons with cervical artery bruits. N Engl J Med 190;0:-1. Hennerici M, Aulich A, Sandmann W, Freund HJ: Incidence of asymptomatic extracranial arterial disease. Stroke 191; 1:70-7 9. Pessin MS, Panis W, Prager RJ, Millan VG, Scott RM: Auscultation of cervical and ocular bruits in extracranial carotid occlusive disease: A clinical and angiographic study. Stroke 19;1:6-9 10. Murie JA, Sheldon CD, Quin RO: Carotid artery bruit: Association with internal carotid stenosis and intraluminal turbulence. BrJSurg 19;71:0-11. Dugdale LM, Gray AV, Mitchell A: The carotid bruit: Is it a useful sign of vessel disease? A clinical and digital angiographic study. Australas Radiol 19;9:60-6 1. Imataka K, Seki A, Takahashi N, Fugii J: Carotid bruits and their clinical significance.// Heart J 19;:7-71 1. Lusiani L, Visona A, Castellani V, Ronsisvalle G, Bononome A, Pagnan A: Prevalence of atherosclerotic lesions at the carotid bifurcation in patients with asymptomatic bruits. An echo-doppler (duplex) study. Angiology 19;-9 1. Ingall TJ, Homer D, Whisnant JP, Baker HL, O'Fallon WM: Predictive value of carotid bruit for carotid atherosclerosis. Arch Neurol 199;6:1-1. Roederer GO, Langlois YE, Jager KA, Primozich JF, Beach KW, Phillips DJ, Strandness DE: The natural history of carotid arterial disease in asymptomatic patients with cervical bruits. Stroke 19;1:60-61. Chambers BR, Norris JW: Outcome in patients with asymptomatic neck bruits. N Engl J Med 196;1:60-6 17. Feussner JR, Matchar DB: When and how to study the carotid arteries. Ann Intern Med 19;109:0-1 1. Health and Public Policy Committee, American College of Physicians: Diagnostic evaluation of the carotid arteries. Ann Intern Med 19;109:-7 19. Morrison DF: Multivariate Statistical Models. New York, McGraw-Hill Book Co, 1967 0. SAS Institute Inc: SAS User's Guide: Statistics, 19 Edition. Cary, NC, SAS Institute Inc, 19 1. Cote R, Caron J: Management of carotid artery occlusion. Stroke 199;0:1- KEY WORDS bruit carotid artery diseases ultrasonics