Duplex scanning of normal or minimally diseased carotid arteries: Correlation with arteriography and clinical outcome

Transcription

1 Duplex scanning of normal or minimally diseased carotid arteries: Correlation with arteriography and clinical outcome R. Eugene Zierler, MD, Ted R. Kohler, MD, and D. Eugene Strandness, Jr., MD, Seattle, Wash. This study evaluated the role of duplex scanning in the management of patients with normal or minimally diseased cartoid arteries. Carotid duplex scans were interpreted according to previously established criteria and considered normal when pulsed Doppler spectral waveforms showed laminar flow or only minor flow disturbances. Normal flow patterns were noted by duplex scanning in 100 carotid bifurcations of 72 patients who also underwent carotid arteriography. Neurologic symptoms (amaurosis fugax, transient ischemic attack, or stroke) were present in relation to 23 arteries and absent in relation to 77 arteries. On the 23 symptomatic sides arteriography was interpreted as normal in eight, 1% to 15% stenosis in 14, and 16% to 40% stenosis in one. For the 77 asymptomatic sides, arteriography showed normal vessels in 15, 1% to 15% stenosis in 43, and 16% to 40% stenosis in 19. One symptomatic patient was treated by carotid endarterectomy for an irregular 1% to 15% stenosis. None of the asymptomatic lesions were in the range of 80% to 99% stenosis, which would justify endarterectomy for asymptomatic disease. Clinical follow-up for a mean interval of 28 months on 20 of the 22 symptomatic patients not undergoing surgery revealed no strokes and transient recurrent symptoms in two patients. Assuming that the single operation in this study was indicated, duplex scanning correctly identified lesions not requiring carotid endarterectomy in 96% (22/23) of the symptomatic patients. A normal duplex scan also predicted a benign clinical outcome without operation. Duplex scanning can reliably exclude surgically treatable carotid bifurcation lesions in asymptomatic patients, and endarterectomy is rarely indicated in symptomatic patients with normal duplex scan results. This study supports a nonoperative therapeutic approach for most patients with neurologic symptoms and a normal carotid duplex scan on the appropriate side. (J VAse SURG 1990;12: ) The principal role for duplex scanning of the cai rotid bifurcation has been in the screening of patients with suspected extracranial carotid artery disease. Carotid endarterectomy may be indicated in symptomatic patients with appropriate carotid bifurcation lesions and selected patients with asymptomatic highgrade carotid stenoses. 14 Duplex scanning has been especially valuable for identifying asymptomatic patients with severe carotid stenoses who are at particularly high risk for neurologic events? Patients who are candidates for carotid endarterectomy generally From the Division of Vascular Surgery, Department of Surgery, University of Washington School of Medicine, and the Section of Vascular Surgery, Seattle Veterans Administration Medical Center, Presented at the Fifth Annual Meeting of the Western Vascular Society, Coronado, Calif., Jan 25-28, Reprint requests: R. Eugene Zierler, MD, Department of Surgery (112), Seattle Veterans Administration Medical Center, 1660 South Columbian Way, Seattle, WA /6/22603 undergo preoperative arteriography to confirm the duplex scan findings and visualize the aortic arch and intracranial vessels that are not directly examined by the duplex method. Some surgeons have performed endarterectomy in selected patients based on the duplex scan results alone. 69 Interest in this approach has been stimulated by continued improvements in the technique of duplex scanning and an ongoing effort to minimize the costs and risks of carotid surgery. Although the diagnosis and treatment of carotid disease tend to emphasize the importance of severe carotid stenoses, patients are still encountered who have neurologic symptoms but no significant lesion by duplex scan in the corresponding carotid bifurcation. This study was carried out to determine the value of duplex scanning in the management of patients with normal or minimally diseased extracranial carotid arteries. It was designed to answer two din- 447

2 448 Zierler, Kohlerj and Strandness Journal of VASCULAR SURGF ~2( Table I. Criteria for classification of internal carotid artery disease by spectral analysis of pulsed Doppler signals Arteriographic lesion Spectral criteria A. 0% diameter reduction Peak systolic frequency less than 4 KHz; no spectral broadening B. 1% to 15% diameter re- Peak systolic freqnency less duction than 4 KHz; spectral broadening in deceleration phase of systole only C,!6%to 49% diameter ". Peak systolic frequency less reduction than 4 KHz; spectral broadening throughout systole D. 50% to 79% diameter Peak systolic frequency reduction greater than or equal to 4 KHz; end-diastolic frequency less than 4.5 KHz D +. 80% to 99% diameter End-diastolic frequency reduction greater than or equal to 4.5 KHz E. Occlusion (100% diame- No internal carotid flow sigter reduction) nal; flow to zero in common carotid artery Criteria are based on a pulsed Doppler with a 5 MHz transmitting frequency, a sample volume that is small relative to the internal carotid artery, and a 60 degree beam-to-vessel angle ofinsonation. ically relevant questions: (1) How reliable is a normal duplex scan for excluding a surgically treatable lesion at the carotid bifurcation? and (2) What is the clinical outcome in patients with neurologic symptoms and a normal carotid duplex scan on the appropriate side? MATERIAL AND METHODS Patients. The records of the Noninvasive Vascular Laboratory at the Seattle Veterans Administration Medical Center were reviewed to identify all patients who had both a normal carotid duplex scan result on one or both sides and a carotid arteriogram within 3 months of the duplex examination. For these selected patients, the hospital charts and radiology reports were also reviewed. Patients with a history of carotid endarterectomy on the side of interest were excluded. Data collected included the duplex scan results, arteriographic findings, presence or absence of neurologic symptoms at the time of the initial evaluation, treatment, and cfinicfil outcome. Neurologic symptoms consisted of transient ischemic attack (TIA), amaurosis fugax, and stroke. For the purpose of this study TIA was defined as an episode of aphasia or a unilateral motor or sensory deficit lasting less than 24 hours. Amaurosis fugax was similarly defined as monocular blindness lasting less than 24 hours. When any of the preceding hemi- spheric symptoms persisted for more than 24 hours, the event was classified as a stroke. Patients with nonhemispheric symptoms were considered to be asymptomatic. Duplex scanning. Carotid duplex scans were performed with an ATL 600 ultrasound system (Advanced Technology Laboratories, Bothell, Wash.). The details of the technique and interpretation of carotid duplex scanning have been described previously. 1 -~2 In summary, the duplex scanner combines B-mode imaging and pulsed Doppler flow detection in a single instrument. The B-mode image is used to identify the vessel of interest and allows precise placement of the pulsed Doppler sample volume. Spectral waveform analysis of the Doppler signal is then used to characterize the flow pattern at selected sites. Center stream spectral waveforms are routinely obtaig'~d from the proximal and distal common carotid, external carotid, and proximal, middle, and distal internal carotid arteries. Prior comparisons between various spectral features and carotid arteriography allow classification of stenosis severity according to the criteria given in Table I? aa For distinguishing between normal and diseased carotid arteries, these, criteria have provided a sensitivity of 99% and a specificity of 84%; the overall accuracy for detecting 50% to 99% diameter stenosis or occlusion is 93%. 12 In this study a carotid bifurcation with spectral waveforms classified in categories A or B of Table I was considered normal. Category A indicates a normal laminar flow pattern without increased velocity or turbulence. Some bifurcations in category A also exhibit flow separation in the proximal internal carotid artery, a finding that correlates with normal bulb geometry} 4 Category B represents a minor G* gree of flow disturbance characterized by spectralf, broaderfing late in the cardiac cycle and normal peak systolic velocities. These spectral features signify either normal or minimally diseased vessels and generally correlate with lesions of no more than 15% diameter reduction by arteriography. Since B-mode imaging alone has not been reliable for the detection of ulcerated carotid plaques, no attempt was made to determine plaqu~ composition or surface features by duplex scanning./s,16 Arteriography. All arteriograms were intraarterial procedures that included aortic arch views, selective injections of contrast in the common carotid arteries, and visualization of the carotid siphon and intracranial vessels. The maximum degree of diameter reduction at the carotid bifurcation was expressed as a percentage of the estimated normal carotid bulb diameter? 7 This is different from the more common approach of calculating percent diameter rednctic_'~

3 Volume 12 Number 4 O~ober 1990 Duplex scanning of normal carotid arteries 449 Table II. Duplex scan results on the contralateral sides in 44 patients with one normal carotid bifurcation by duplex scan [see Table I for duplex scan classification) Contralateral duplex scan (% diameter reduction) Neurologic 0% to 16% to 50% to 80% to status 15% 49% 79% 99% 100% Symptom- 1 1 atic (2) Asymptom- 1 ~ atic (42) Neurologic status refers to carotid territory on side of normal bifurcation. ~This side showed a category B lesion after a carotid endarterectomy. relative to the diameter of the normal distal internal carotid artery. The former method was used to characterize early or minimal lesions of the carotid bulb more precisely. Although the geometry of the bulb at the origin of the internal carotid artery is quite variable, the normal bulb diameter generally exceeds that of the internal carotid distal to the bulb. Therefore, considerable accumulation of atherosclerotic plaque can occur in the bulb before the smaller distal lumen is compromised The method used in this study tends to give a higher estimate of percent diameter reduction for a particular lesion compared to the value obtained by use of the normal distal internal carotid artery diameter as the site of reference. For example, the carotid bulb lesion shown in Fig. 1 was inter- ;eted as a 20% diameter stenosis, even though the lumen at the stenotic site is not narrowed relative to the distal internal carotid artery. RESULTS During the 4-year-period from 1984 through 1987, 72 patients met the criteria for inclusion in this study. In 44 patients duplex scans showed just one normal carotid bifurcation, and 28 patients had normal carotid bifurcations bilaterally. Thus, a total of 100 normal carotid bifurcations were identified by duplex scanning and evaluated by arteriography. Twenty-three of these were associated with neurologic symptoms, and 77 were asymptomatic. Two of the patients with unilateral normal duplex scan findings had neurologic symptoms appropriate to the normal artery, whereas the remaining 42 were asymptomatic with regard to that carotid territory. Table II gives the duplex scan classification of the ~0ntralateral carotid bifurcation for the 44 patients Fig. 1. Asymptomatic patient with normal (category B) duplex scan and carotid arteriogram interpreted as showing a 20% stenosis. who had only one normal side. Hemodynamicaily significant lesions (categories D, D +, and E in Table I) were found in 77% (34/44) of these contralateral sides. As indicated in Table III, neurologic symptoms appropriate to the contralateral side were noted in 68% (30/44) of these patients. It is apparent that the presence of severe contralateral carotid stenosis or neurologic symptoms related to the contralateral side prompted the clinical evaluation in many patients. Twenty-one of the 28 patients with normal duplex scans bilaterally had neurologic symptoms associated with one carotid bifurcation only, and seven were asymptomatic. The arteriogram results for all 100 carotid bifurcations are shown in Table IV. Twenty bifurcations were considered to have 16% to 40% diameter stenosis by arteriography, and all but one of these lesions were in the asymptomatic group. These consisted of seven 20% stenoses, eleven 30% stenoses, and two 40% stenoses. One asymptomatic normal internal carotid artery was associated with a 90% stenosis of the ipsilateral carotid siphon. No other

4 450 Zierler, Kohler, and Strandness Journal of VASCULAR SURGEP~Y Table III. Prevalence of neurologic symptoms related to the contralateral carotid territory in 44 patients with one normal carotid bifurcation by duplex scan Neurologic symptoms Neurologic status None TIA AF Stroke Symptom- 1 1 atic (2) Asymptom atic (42) Neurologic status refers to carotid territory on side of normal bifurcation. TIA, Transient ischemic attack; AF, amaurosis fugax. Fig. 2. Carotid bifurcation in an asymptomatic patient classified as normal (category B) by duplex scanning and 1% to 15% diameter stenosis by arteriography. significant aortic arch vessel or intracranial lesions were reported. The carotid bifurcation in Fig. 2 was classified in category B by duplex scanning and interpreted as 1% to 15% diameter stenosis by arteriography. In this example there is evidence of smooth plaque along the outer wall of the carotid bulb without significant narrowing of the vessel lumen. Fig. 1 shows a carotid bifurcation that also had a category B lesion noted on duplex scan, but the arteriogram was interpreted as a 20% stenosis of the proximal internal carotid artery. Although plaque is dearly present at that location, there is minimal reduction in lumen size compared to the normal distal segment of the internal carotid artery. Among the 23 patients with neurologic symptoms related to a carotid bifurcation that was normal by duplex scan, TIA occurred in eight, amaurosis fugax in seven, and stroke in eight. Treatment at the time of initial presentation included carotid endarterectomy in one patient, aspirin in eight, warfarin (Coumadin) in three, steroids for presumed temporal arteritis in one, and no specific treatment in 10. The one carotid endarterectomy was done in a patient with a history of stroke appropriate to a lesion c(7~sifted by arteriography as 1% to 15% diameter reducing and irregular (Fig. 3). At surgery an ulcerated atherosclerotic plaque with overlying thrombus was found. Clinical follow-up data were available on 20 of the 22 symptomatic patients not undergoing surgery for a mean interval of 28 months (range 6 to 53 months). No neurologic events were reported during the follow-up period in 18 patients. One patient who originally presented with amaurosis fugax had several recurrent episodes over the first 2 years but no other neurologic symptoms for the remainder of a 40- month follow-up period. Another patient with a history of stroke and atrial fibrillation reported recurrent TIAs during a 34-month follow-up period in spite of treatment with Coumadin. Six symptomatic patients had follow-up duplex scans at a mean interval of 31 months (range 20 to 38 months), and none ~ the symptomatic carotid bifurcations showed evidence of disease progression. DISCUSSION The main purpose of arteriography for extracranial carotid disease is to confirm the presence of a surgical lesion at the carotid bifurcation and reveal any additional abnormalities that might make the patient unsuitable for endarterectomy. Carotid duplex scanning has been used primarly as a noninvasive screening test to avoid unnecessary arteriography. If a surgically treatable carotid lesion can be excluded by noninvasive testing, then arteriography is generally not indicated. Numerous studies have been done on patients with normal carotid arteriograms and neurologic symptoms; 1822 however, the clinical significance of a normal carotid duplex scan result has not been established.

5 Volume 12 Number 4 O :~'ober 1990 Duplex scanning of normal carotid arteries 451 Table IV. Arteriographic findings in 100 carotid bifurcations with normal duplex scans Arteriography (% diameter reduction) 0% 1% to 15% 16% to 40% Neurologic rtatus Stenosis Stenosis Stenosis Symptomatic (23) Asymptomatic (77) ~Neurologic status refers to carotid territory on side of normal bifurcation. The accuracy of duplex scanning in the diagnosis of carotid artery disease has been determined by comparing noninvasive test results for individual carotid b; ~rcations with independently interpreted arteriograms Thus the reported accuracy figures apply to single carotid bifurcations rather than to patients. Similarly, natural history studies relating carotid bifurcation disease as documented by duplex scanning to clinical outcome have considered neurologic risk in terms of the single carotid bifurcation and ipsi- ;lateral cerebral circulation. 3.s Consequently, both the accuracy of carotid duplex scanning and the risk of neurologic events have been characterized for single carotid bifurcations, without regard for the status of the contralateral carotid system. This approach of considering each carotid bifurcation independently was used in the present study. It is acknowledged that many of the patients in our study with unilateral normal findings on carotid duplex scanning had either severe contralateral stenoses or neurologic symptoms appropriate to the ~[oontralateral carotid bifurcation that justified the use of arteriography. However, inclusion of these cases is valid, since the results of arteriography and the clinical outcome related to the sides with normal duplex scans were determined for each individual side, irrespective of disease involving tlae contralateral carotid bifurcation. A basic problem in selecting the duplex scan criteria for normal carotid arteries is the subjectivity of interpreting spectral waveforms, particularly when the flow disturbances are minor. It has been shown that variability in the results of carotid duplex scanning is related mostly to the normal (A), minimal disease (B), and moderate disease (C) categories listed in Table I; more consistent agreement has been observed in the identification of greater than 50% diameter re&acing stenoses and occlusions. 23 Since there is significant variability in differentiating between normal and minimally diseased carotid arteries Fig. 3. Carotid bifurcation of a patient with a prior history of stroke. Duplex scan showed a category B lesion, and the arteriogram was interpreted as an irregular 1% to 15% diameter stenosis. by duplex scanning, spectral waveform categories A and B were combined in this study. When there is disagreement between the duplex scan classification and arteriography, it is rarely by more than one spectral waveform category. In the present study, 20 carotid bifurcations with category A or B spectral wajeeforms had stenoses of 16% to 40% diameter reduction by arteriography. All but two of these were in the 20% to 30% range. Although these differences may be strictly regarded as duplex scan errors, the normal diversity of carotid bulb geometry complicates the interpretation of arteriograms, and such discrepancies can be attributed largely to observer variability in determination of percent diameter reduction by carotid arteriography. 17 Given the variability in assessing less severe carotid lesions by both duplex scanning and arteriography, it may be more useful to emphasize the flow patterns associated with normal or minimally diseased arteries rather than adhere to any rigid criteria based on percent diameter reduction alone. The con-

6 452 Zierler, Kohler, and Strandness Journal of VASCULAR SURGEB~Y cept of predicting clinical outcome on the basis of arterial flow patterns has been reported previously. For example, the finding of flow separation in the carotid bulb by duplex scanning predicts that any neurologic symptoms are not related to carotid bifurcation disease.~4 On the other extreme, the severe flow disturbance of spectral waveform category D + in Table I correlates with a high probability of TIA, stroke, or internal carotid occlusion in patients who are initially asymptomatic. 405 Patients with TIAs in a carotid artery territory have a 10% to 40% risk of stroke over five years Although arteriography shows significant atherosclerotic occlusive disease in most of these patients, the proportion that have normal or minimally diseased carotid arteries is between 7% and 43% y,30-32 Clinical follow-up studies of patients with TIAs and normal carotid arteriograms suggest that the risk of subsequent TIA varies from 6% to 20%, and the risk of stroke is between 5% and 13%. ~8,2~,22 The two studies by Mendelowitz et al.20 and Evans et al.21 are of particular interest, since the original article indicated a relatively benign prognosis, whereas a later report with a larger number of patients and longer mean follow-up period documented the occurrence of TIA or stroke in 13% of patients. For all these studies, the cause of neurologic events in patients with normal carotid arteriograms is difficult to determine; however, cardiac disease and platelet abnormalities have been implicated in many patients. 18,22,33-35 The present study evaluated the arteriographic findings and clinical outcome in symptomatic and asymptomatic patients with normal carotid arteries by duplex scanning. Although 19 of the 77 asymptomatic carotid lesions were classified as 16% to 40% diameter reducing by arteriography, none was in the range of 80% to 99% stenosis, which would justify endarterectomy for asymptomatic disease. 4,5 Thus a normal duplex scan, as defined in this study, reliably excludes the presence of a surgical lesion in an asymptomafic patient. All but one of the symptomatic patients in this study were treated without operation. The one carotid endarterectomy was considered necessary to remove a potential source of emboli at the carotid bifurcation. Although this decision may have been correct, the ability of ultrasonography or arteriography to detect ulcerated carotid plaques is limited, ls'~6'36 and there was no obvious difference between the arteriographic appearance of the surgically treated artery and many of those treated without operation. Furthermore, although it is recognized that duplex scanning cannot reliably identify arterial ul- cerations, pathologic studies of symptomatic carotid disease indicate that ulcerated lesions, and presumably sources of emboli, are most commonly associated with moderate to severe stenoses, rather than normal or minimally diseased arteries. 37 Assuming that only the single operation in this study was indicated, a duplex scan showing normal carotid arteries correctly identified lesions not requiring endarterectomy in 96 % (22 / 2 3 ) of the symptomatic patients. A normal finding on carotid duplex scan also predicted a benign clinical outcome in patients with neurologic symptoms who did not un-, dergo endarterectomy. No strokes occurred, and only two symptomatic patients not undergoing surgery had recurrence of transient neurologic events during the follow-up period. Although duplex scanning does not directly evaluate the aortic arch and intracra~al" vessels, significant lesions were uncommon in thcsc segments. Furthermore, there is no clear relationship between the presence of intracranial vascular disease and the severity of stenosis at the carotid bifurcation or the clinical outcome of carotid endarterectomy.38'39 One limitation of this retrospective study was the,, inability to establish a definite cause for neurologic events in the symptomatic patients with normal resuits on carotid duplex scanning. Unfortunately, investigations for cardiac disease and hematologic abnormalities were not consistently obtained in our patients. However, as noted previously, these patients~ should be evaluated for conditions other than carotid artery disease that are associated with neurologic symptoms, particulary cardiac disorders. It is difficult to reconcile the relatively high incidence of TIA and stroke reported in symptom~ patients with normal findings on carotid arteriography and the low incidence of recurrent neurologic events for the symptomatic patients in the present study. Among the symptomatic patients with followup data who did not undergo operation there were no strokes, and the estimated ctunulative incidence of TIA at 24 months was 13%. Although the number of symptomatic patients in our study was smaller than some of the other series, our mean follow-up interval of 28 months was substantially longer than the 16 months in the original report by Mendelowitz et al.20 The principal difference between arteriography and duplex scanning is that the former is a strictly anatomic investigation, whereas the latter provides physiologic information on flow patterns. This study suggests that normal carotid flow patterns are associated with clinically benign lesions, irrespective of the extent of stenosis demonstrated by arteriography.

7 Volume 12 Number 4 OCtober 1990 Duplex scanning of normal carotid arteries 453 Returning to the questions raised at the beginning of this report, the data presented indicate that duplex scanning is extremely reliable for excluding surgically treatable carotid bifurcation lesions in asymptomatic patients. In addition, carotid endarterectomy is rarely indicated in symptomatic patients with normal findings on duplex scanning. With regard to clinical outcome in patients with neurologic symptoms, a normal duplex scan result correlated with a very low incidence of recurrent events. This study supports a nonoperative therapeutic approach for most patients With neurologic symptoms and a normal carotid duplex scan on the appropriate side. ~teriography can still be obtained to look for a source of emboli at the carotid bifurcation in patients with repetitive neurologic symptoms, particularly v ~en the symptoms occur while taking antiplatelet medication. However, such lesions are uncommon, and arteriography should not be necessary in most patients with normal carotid arteries demonstrated by duplex scanning. :~ REFERENCES 1. West H, Burton R, Roon AJ, Malone IM, Goldstone J, Moore WS. Comparative risk of operation and expectant management for carotid artery disease. Stroke 1979;10: ilubin JR, Goldstone J, Mclntyre KE, Malone JM, Bernhard VM. The value of carotid endarterectomy in reducing the morbidity and mortality of recurrent stroke. ] VAse SURG 1986;4: Healy DA, Clowes AW, Zierler RE, et al. Immediate and long-term results of carotid endarterectomy. Stroke 1989; 20: Moneta GL, Taylor DC, Nicholls SC, et al. Operative versus nonoperative management of asymptomatic high-grade internal carotid artery stcnosis: improved results with endarterectomy. Stroke 1987;18: Roederer GO, Langlois YE, Jager KA, et al. The natural history of carotid arterial disease in asymptomatic patients with cervical bruits. Stroke 1984;15: Flanigan DP, Schuler JJ, Vogel M, Borozan PG, Gray B, Sobinsky KR. The role of carotid duplex scanning in surgical decision making, l VASe SURG 1985;2: Walsh J, Markowitz I, Kerstein MD. Carotid endartereetomy for amaurosis fugax without arteriography. Am J Surg 1986;152: Geuder ~r, LampareUo PJ, Riles TS, Giangola G, Imparato AM. Is duplex scanning sufficient evaluation before carotid endarterectomy? ] VASe St~RG 1989;9: Moore WS, Ziomek S, Quinones-Baldrich WJ, Machleder HI, Busuttil RW, Baker ]D. Can clinical evaluation substitute for arteriography in the evaluation of carotid artery disease? Ann Surg 1988;208: Fell G, Phillips D], Chikos PM, Harley JD, Thiele BL, Strandness DE Jr. Ultrasonic duplex scanning for disease of the carotid artery. Circulation 1981;64: Langiois Y, Roederer GO, Chan A, e~ al. Evaluating carotid artery disease: the concordance between pulsed Doppler/spectral analysis and angiography. Ultrasound Med Biol 1983;9: Zierler RE. Carotid artery evaluation by duplex scanning. Semin Vase Surg 1988;1: Koederer GO, Langlois YE, Jager KA, et al. A simple spectral parameter for accurate classification of severe carotid disease. Bruit 1984;8: Nicholls SC, Phillips DI, Primozich IF, et al. Diagnostic significance of flow separation in the carotid bulb. Stroke 1989;20: O%eary DH, Holen J, Ricotra JJ, et al. Carotid bifurcation disease--prediction of ulceration with B-mode US. Radiology 1987;162: Comerata AJ, Katz ML, White J-V, Grosh JD. The preoperative diagnosis of the ulcerated carotid atheroma. J VAsc Suite 1990;11: Chikos PM, Fisher LD, Hirsch JH, Harley JD, Thiele BL, Strandness DE Jr. Observer variability in evaluating extracranial carotid artery stenosis. Stroke 1983;14: Marshall J, Wilkinson MS. The prognosis of carotid transient ischemic attacks in patients with normal angiograms. Brain 1971;94: Toole JF, Yuson CP. Transient ischemic attacks with normal arteriograms: serious or benign prognosis.~ Ann Neurol 1977;1: Mendelowitz DS, Kimmins S, Evans WE. Prognosis of patients with transient ischemic attacks and normal angiograms. Arch Surg 1981;116: Evans WE, Hayes ]P. Life history of patients with transient ischemic attacks and essentially normal angiograms. ) VAse SURG 1987;6: Shuaib A, Hachinski VC, Oczkowski WJ. Transient ischemic attacks and normal cerebral angiograms: a follow-up study. Stroke 1988;19: Kohler T, Langlois Y, Roederer GO, et al. Sources of variability in carotid duplex exanfination: a prospective study. Ultrasourld Med Biol 1985;1i: Ziegler DK, Hassanein KS. Prognosis in patients with transient ischemic attackl Stroke 1973;4: Baker RN, Karuseyer JC, Schwartz WS. Prognosis in patients with transient cerebral ischemic attacks. Neurology 1968; 18: Toole JF, Yuson CO, Janeway R, et al. Transient ischemic attacks: a prospective study of 225 patients. Neurology 1978;28: Heyrnan A, Leviton A, Millikan CH, ct al. Transient focal cerebral ischemia: epidemiological and clinical aspects. Stroke 1974;5: Millikan CH.Transient cerebralischemia: definitions andnatural history. Prog Cardiovasc Dis 1980;22: Whisnant JP, Matsumoto N, Elveback LR. Transient cerebral ischernic attacks in a community, Rochester, Minnesota, 1959 through Mayo Clin Proc 1973;48: Pessin MS, Duncan GW, Mohr JP, Poskanzer DC. Clinical and angiographic features of carotid transient ischemic attacks. N Engl J Med 1977;296: DeBono DP, Warlow CP. Potential sources of embofi in patients with presumed transient cerebral or retinal ischemia. Lancet 1981;1: Candelise L, Albanese V, Bono G, et al. Prognosis of treated recent reversible ischemic attacks (RIAs): analysis of recur-

8 454 Zierler, Kohler, and Strandness Journal of VASCULAR SUKGEB~ rence of RLAs, stroke and death according to the pathogenesis, angiographic findings and risk factors. Path Biol 1982; 30: AI-Mefty O, Marano G, Rajarman S, Nugent GR, Rodman N. Transient ischemic attacks due to increased platelet aggregation and adhesiveness- ultrastructural and functional correlation. J Neurosurg 1979;50: Poole CJM, Russell RWR, Harrison P, Savidge GF. Amaurosis fugax under the age of 40 years. I Neurol Neurosurg Psych 1987;50: Shuaib A, Hachinski VC. Carotid transient ischemic attacks and normal investigations: a follow-up study. Stroke 1990; 21: Edwards ]H, Kricheff II, Riles T, Imparato A. Angiograph- ically undetected ulceration of the carotid bifurcation as a cause of embolic stroke. Radiology 1979;132: Bassiouny HS, Davis H, Masawa N, Gewertz BL, Glagov S, Zarins CK. Critical carotid stenosis: morphologic and chemical similarity between symptomatic and asymptomati9 plaques. ] VASC SUV, G 1989;9: , Mackey WC, O'Donnell TF, Callow AD. Carotid endarterectomy in patients with intracranial vascular disease: shortterm risk and long-term outcome. I VASC SURG 1989;10: Roederer GO, Langlois YE, Chan ARW, et al. Is siphon disease important in predicting outcome of carotid endarterectomy? Arch Surg 1983;118: DISCUSSION Dr. David V. Cossman (Los Angeles, Calif.). This article demonstrates that duplex scanning of normal or minimally diseased carotid arteries correlates well with arteriography in both symptomatic and asymptomatic patients, so that duplex scanning can reliably exclude surgically treatable carotid bifurcation lesions in asymptomatic patients, and endarterectomy is rarely indicated in symptomatic patients with normal duplex scans. The most interesting group of arteries in the study group of 100 carotid bifurcations are the 20 in patients with ipsilateral neurologic symptoms who were followed for a mean of 28 months. The authors conclude that symptomatic patients with normal or nearly normal scans do not need arteriograms and may be followed safely. They do caution, however, that the follow-up period was short, and that a similar study done by Evans following patients with TIAs and essentially normal arteriograms revealed a 13% incidence of TIA or stroke if the patients were fouowed for longer periods of time. I have several questions regarding the authors' advice to follow without arteriography the symptomatic patient with a nearly normal finding on duplex scanning. Transient ischemic attacks and strokes may be caused by valxa~lar heart disease, cardiac chamber disease, cardiac arrhythmias, lacunar thrombotic infarcts, and thromboemboli from the extracranial cerebral vessels out of the view of the duplex scanner. We do not know from the data presented regarding the symptomatic group how many of the TIAs came from or were strongly suspected of coming from the carotid bifurcation. The authors note that there is a large discrepancy between the lack of recurrent symptoms in their small group of symptomatic patients and the 5% to 20% incidence of recurrent TIAs and strokes reported by other authors in patients with TIAs and normal arteriograms. Perhaps the explanation for this is the difference in origins of the symptoms. What if the 23 symptomatic patients with normal duplex scan results had large ulcerations without stenosis? Despite certain reports in the literature, we have had no success in calling ulcerations on duplex scans and therefore continue to advise arteriography in patients with significant hemispheric symptoms and normal scans. We have also not put a great deal of reliance on carotid bulb flow separation to predict whether or not T/As are caused by carotid bifurcation disease. This is in contrast to symptomatic patients with tight appropriate lesions whom we are operating on increasingly without arteriography. However, an arteriogram should only be obtained in the symptomatic patient with a suspected ulcer if the surgeon believes that such a patient warrants an operation, and more importantly, if the neurologist and the patient would allow it. The chances of all three occurring are increasingly remote, and perhaps justifiably so. In summary, I have the following questions. (1) Of the 22 symptomatic patients with normal scan findings, how many do you suspect had noncarotid bifurcation origins for their symptoms? (2) Do you worry about missing large nonstenotic ulcers in symptomatic patients, ~,~ would you recommend carotid endarterectomy in SUCh patients? Dr. Zierler. As I mentioned in the presentation, we were unable to establish definite causes for the neurologic symptoms in our patients. I think many patients had noncarotid bifurcation causes, but that is impossible to prove in retrospect. When we reviewed the arteriograms we did look for aortic arch and intracranial lesions, and only one. significant siphon lesion was noted. Regarding the issue of ulceration, this is a difficult problem from the diagnostic point of view. We believe that duplex scanning, whether one looks at the B-mode image or the Doppler information, is really not an adequate way of identifying ulcerated plaques. Arteriography also may not be a reliable method. The standard for diagnosis of ulcerated lesions is really the surgical specimen, when and if it is available. Our experience would suggest that significant ulcerations are relatively uncommon in nonstenotic plaques. The concept of exploring these lesions surgically strikes me as somewhat dangerous, particularly at a time

9 Volume 12 Number 4 Oc-9ber 1990 Duplex scanning of normal carotid arteries 455 when carotid surgery is being subjected to such close scrutiny. The way this particular experience has influenced our thinking and practice is that when we encounter a patient,~vith normal or minimally diseased carotid arteries by duplex scanning who also has convincing neurologic symptoms, we have a higher threshold for getting an arteriogram. What we usually do is put the patient on aspirin and get appropriate consultations or investigations to rule out other causes for the symptoms. If results of all these studies are negative, and the patient continues to have symptoms on aspirin, we can then make a strong case for proceeding with arteriography. However, we see relatively few patients who fall into this category. Dr. Ronald Stoney (San Francisco, Calif.). My concern with this conclusion is that some physicians, particularly the neurologists, would decide on the basis of a aormal flow pattern on duplex scanning that the arteriogl~rn can be eliminated, since there is no carotid disease. I think we must be careful not to eliminate the arteriogram and miss ulcerative disease in proximal arteries like the hmominate artery. Two distal carotid lesions occur not so uncommonly, particularly in young women. These are dissection and fibromuscular hyperplasia of the internal carotid artery. When you get normal duplex information,from the bifurcation, you could be doing these people a disservice if you're not going to obtain arteriographic studies of the carotid arterial system and areas adjacent to the bifurcation. This is essential in symptomatic patients even when the duplex scan of the carotid bifurcation is normal. Dr. Zierler. I think that is a very good point to emphasize. I am certainly not opposed to doing arteriograms 'in these patients. My message is simply to think in terms of noncarotid bifurcation causes for symptoms when the duplex scan result is was normal. You mentioned fibromuscular dysplasia of the carotid arteries, and we can usually see some indication of this on the duplex scan by ~amining the flow patterns in the mid and distal internal carotid arteries. Dr. Kenneth E. McIntyre, Jr. (Tucson, Ariz.). I recently had a patient just like Dr. Stoney described who had an innominate lesion that was discovered after an evaluation for one episode of amaurosis permanans. Both the duplex image and spectral analysis were normal, and we proceeded with an arteriogram and found this innominate lesion. We know that the natural history of symptomatic patients has a favorable outcome when they are treated with aspirin. Although you mentioned it briefly, I would like to know if the favorable outcome in those patients with symptoms can be attributed to the fact that they were on aspirin therapy. We know that aspirin therapy may reduce the instance of stroke. It does not alleviate it. So I am interested in your comments in that regard. Dr. Zierler. In our study only eight patients were treated with aspirin, so based on these data it is difficult to make any definite conclusions. However, I think that it makes sense if you cannot establish a definite cause in these patients to put them on aspirin as a therapeutic trial. Dr. Albert Hall (Greenbrae, Calif.). I am concerned about the patients in your series who had TIAs. Did some of these have visual defects with field cuts and ophthalmoscopic evidence of retinal embolization? Were you willing to follow those patients solely because they had "normal" duplex scans? In my opinion, patients who have ophthalmoscopic evidence of retinal branch artery embolization are at risk for losing vision because they are susceptible to multiple episodes that can lead to blindness unless the source of embolization is identified and treated. Are you willing to follow patients in that category without studying them futvjacr? Dr. Zierler (closing). Seven patients had amaurosis fugax in our study, based on the clinical presentation of a transient monocular visual deficit. One of those seven patients did have several recurrent episodes during followup. We do not have ophthalmoscopic evaluations to document embolic material in the retinal vessels in every case. However, a formal ophthalmology consultation is valuable to exclude primary eye disease as a cause for symptoms.