Outcome in Patients with Symptomatic Occlusion of the Internal Carotid Artery

Transcription

1 Eur J Vasc Endovasc Surg 19, (2000) doi: /ejvs , available online at on Outcome in Patients with Symptomatic Occlusion of the Internal Carotid Artery C. J. M. Klijn 1,5, P. A. van Buren 1,5, L. J. Kappelle 1,5, C. A. F. Tulleken 2,5, B. C. Eikelboom 3,5, A. Algra 1,4,5 and J. van Gijn 1,5 Departments of 1 Neurology, 2 Neurosurgery and 3 Vascular Surgery and the 4 Julius Center for Patient Oriented Research, University Medical Center Utrecht and the 5 Rudolf Magnus Institute for Neurosciences, The Netherlands Objectives: to assess whether the risk of recurrent ischaemic stroke in patients with symptomatic internal carotid artery (ICA) occlusion has changed over the past decades, to determine risk factors for the occurrence of ischaemic stroke and to assess the risk of endarterectomy (CEA) of a severe contralateral ICA stenosis. Design: retrospective cohort study. Patients and methods: patients with symptomatic ICA occlusion were identified from duplex registry files between 1991 and Information was obtained on vascular risk factors, performance of CEA for a contralateral ICA stenosis and on recurrence of ischaemic stroke. The rate of complications occurring within 30 days after CEA of the contralateral ICA in patients with symptomatic ICA occlusion was compared with the risk of CEA in patients with asymptomatic ICA occlusion and severe contralateral ICA stenosis (symptomatic or asymptomatic). Results: ninety-seven patients were identified. Mean follow-up time was 26 months. The annual risk of (non-)fatal stroke was 5.3% for all strokes (95% CI 2.9% 9.6%) and 3.8% for ipsilateral stroke (95% CI 1.9% 7.7%). Hyperlipidaemia and severe stenosis of the contralateral ICA were independent risk factors. Twenty-two of 32 patients with a severe stenosis of the contralateral ICA underwent CEA, of which one patient died and three suffered a minor ischaemic stroke. The perioperative risk of CEA in the control group of 20 patients with asymptomatic contralateral ICA occlusion was 0% (0 of 20). Conclusions: outcome in patients with symptomatic ICA occlusion has not substantially improved over the years. CEA for severe stenosis of the contralateral ICA carried a relatively high risk in our series, but deserves to be studied in a controlled design. Key Words: Carotid artery occlusion; Outcome; Carotid endarterectomy. Introduction that there might still be a role for EC/IC bypass surgery in a highly selected group of patients, based on Patients with symptoms of cerebral or retinal isch- the assumption that in some patients a compromised aemia associated with ipsilateral occlusion of the internal cerebral blood flow plays a role in causing transient carotid artery (ICA) have an annual risk of 5 8% ischaemic attacks (TIAs) and ischaemic stroke. 1,4 6 In of recurrent ischaemic stroke. 1,2 In 1985 the results this light, outcome in patients with symptomatic ICA of the extracranial intracranial (EC/IC) bypass study occlusion regains interest. Therefore, we studied the showed that superficial temporal artery to middle rate of recurrent ischaemic stroke and other vascular cerebral artery (MCA) bypass surgery does not prevent events in a consecutive series of patients to determine recurrent ischaemic stroke in these patients. 3 Consequently, if outcome has improved over the last decade, ir- the only treatment available for these respective of the haemodynamic state of the brain. patients is that applied in all patients with symptomatic This is important, because the possible risk of EC/IC atherosclerotic disease, e.g. antithrombotic bypass surgery has to be weighed against the risk of drugs and modification of risk factors for atherosclerosis. recurrent stroke with the treatment currently conment Over recent years, evidence has accumulated sidered optimal, e.g. antithrombotic medication, treat- of risk factors and probably CEA in case of severe stenosis of the contralateral ICA or the ip- Please address all correspondence to: C. J. M. Klijn, Department of Neurology, University Medical Center Utrecht, P.O. Box 85500, silateral external carotid artery (ECA). In addition, risk 3508 GA Utrecht, The Netherlands. factors for ischaemic stroke and vascular events were /00/ $35.00/ Harcourt Publishers Ltd.

2 580 C. J. M. Klijn et al. studied. Furthermore, we compared the risk of CEA Occurrence of outcome events within 30 days of for severe stenosis of the contralateral ICA in these CEA of the contralateral ICA was considered a com- patients with that of CEA in patients with an asympto- plication of surgery. Complications associated with matic ICA occlusion and a contralateral ICA stenosis CEA in the study cohort were compared with com- (symptomatic or asymptomatic). plications associated with CEA in patients with asymptomatic ICA occlusion and a (symptomatic or asymptomatic) severe stenosis of the contralateral ICA. Patients in this control group were identified from the Patients, Control Subjects and Methods same duplex registry in the same time period. Reports of surgical procedures were reviewed for use of an Patients with an occlusion of one or both ICAs were intraoperative shunt. selected from the registry of all duplex investigations CT or MRI scans of the brain performed at baseline of the carotid arteries performed in the vascular and in case of an outcome event were independently laboratory of the University Medical Center Utrecht, reviewed by two of the investigators (CJMK, LJK), The Netherlands between 1991 and Records of blinded for clinical data and side and severity of ICA 222 patients with an occluded ICA were reviewed for stenosis. Subsequently, clinical details were examined symptoms of cerebral or retinal ischaemia within 6 to assess if the scan abnormalities were related to the months prior to the demonstration of the ICA occlusion patients symptoms or signs. Cerebral infarcts were by two investigators (CJMK, PvB). Patients without categorised by the territories of the major cerebral symptoms attributable to ischaemia in the vascular arteries involved (anterior cerebral artery (ACA), MCA territory of the occluded ICA, patients with a diagnosis or posterior cerebral artery (PCA), (territorial infarcts) of dissection of the ICA, patients with a history of or as being in the border-zone area between the ACA radiotherapy in the region of the symptomatic ICA and and MCA, the MCA and PCA, between ACA and patients with severe disability after stroke (modified MCA and PCA or between the deep and superficial Rankin grade 4 or 5) were excluded. territory of the MCA (border-zone infarcts). Territorial Records were reviewed for gender, age, presenting infarcts were further divided into end-zone infarcts, symptoms, vascular risk factors, antithrombotic med- large subcortical infarcts and small deep (lacunar) ication at the time of presenting symptoms, con- infarcts. firmation of the occlusion of the ICA by angiography, Univariate analyses of risk factors for the primary stenosis of the contralateral ICA and performance of and secondary outcome events were performed by CEA of the contralateral ICA. Patients who underwent means of the Cox proportional hazards model resulting EC/IC bypass surgery were included until the time in hazard ratios (HR) with 95% confidence interval of surgery, while those patients who underwent CEA (CI). Variables with an at least borderline statistically because of severe stenosis of the contralateral ICA significant (p<0.10) relationship with an outcome event were followed until the end of the study period. were included in a multivariate model to determine The primary outcome event was fatal or non-fatal potential independent contributors to prognosis. ischaemic stroke. The secondary outcome event was the composite event of (non-)fatal stroke (ischaemic or haemorrhagic), (non-)fatal myocardial infarction, vascular death from other causes (sudden death, con- Results gestive heart failure, systemic bleeding, pulmonary embolism) or retinal infarction, whichever happened Ninety-seven patients (mean age 64 years, range 40 81) first. The diagnosis of ischaemic stroke was made were identified. Twenty-two presented with cerebral when sudden and focal neurological deficits caused TIAs, 49 with minor ischaemic stroke, 19 patients an increase in handicap of at least one grade on the with transient monocular blindness (TMB) and seven modified Rankin scale, which increase lasted for more patients with symptoms of chronic ocular ischaemia than 24 hours, and when haemorrhage was excluded (COI) alone. In 63 patients angiography was performed by computed tomography (CT) or magnetic resonance to confirm the diagnosis of ICA occlusion established imaging (MRI). Non-fatal ischaemic stroke was classified by duplex investigation. Baseline characteristics are as minor when the estimated modified Rankin shown in Table 1. At the time of their presenting score was Ζ3 and as major disabling when 4 or 5. symptoms, 27 patients were treated with low-dose Information about outcome events on follow up was aspirin, nine patients were on oral anticoagulation obtained by telephone interviews with patients, their (AC) therapy and one patient was treated with both relatives or their general practitioners. aspirin and AC.

3 Outcome in Symptomatic ICA Occlusion 581 Table 1. Baseline characteristics of the study patients and results of univariate analyses of risk factors for occurrence of stroke and for the combined endpoint of vascular death, stroke of myocardial infarction. Ipsi- and contralateral ischaemic Vascular death, stroke or MI fatal or non-fatal stroke Characteristic No. of patients HR 95% CI HR 95% CI Demographics Male sex Age [65 years Type of symptoms Presenting symptoms TIA ref ref Minor stroke ref ref TMB 16 COI 7 Haemodynamic aspect presenting symptoms Symptoms after occlusion had been documented Classical risk factors History of stroke Ischaemic heart disease Intermittent claudication Diabetes mellitus Hypertension Hyperlipidaemia Current smoking Family history of vascular disease History of CEA History of CABG, PT(C)A, or vascular surgery of the legs Stenosis of contralateral ICA 0 30% % % % 8 Infarcts Relevant infarct, type Territorial 32 End-zone 21 Small deep 9 Large subcortical 2 Border-zone Irrelevant infarct only 3 No infarct 20 Imaging not performed 27 Scan performed, intractable 2 MI, myocardial infarction; HR, hazard ratio; CI, confidence interval; TIA, transient ischaemic attack; TMB, transient monocular blindness; COI, chronic ocular ischaemia; ICA, internal carotid artery; CEA, carotid endarterectomy; CABG, coronary arterial bypass graft; PT(C)A, percutaneous transluminal (coronary) angioplasty; ref1, cerebral TIA or minor stroke versus TMB or COI; ref2, minor stroke versus cerebral TIA, TMB or COI;, defined as patients with either a history of hyperlipidaemia, patients on drugs because of hyperlipidaemia, or patients with levels of cholesterol, triglycerides or high-density lipoprotein cholesterol outside the normal ranges., unknown in one of 97 patients.

4 582 C. J. M. Klijn et al. The history contained one or more haemodynamic sequently underwent a CEA without any complications. features in the presenting symptoms of 19 of 97 One of the two remaining patients had patients. Nine patients reported that their symptoms bilateral ICA occlusion, and only one of the 11 patients were related to rising from a lying or sitting position, who suffered a recurrent stroke did not have a stenosis in two patients symptoms were related to physical or occlusion of the contralateral ICA. exercise, 6 patients had symptoms consistent with limb Of the 11 patients with recurrent strokes the CT or shaking 7 9 and 5 with retinal claudication. 10,11 In two MRI of the brain was available for revision in 9. Four patients symptoms were related to low blood pressure patients showed border-zone infarction, 2 between the and one patient had symptoms subsequent to com- vascular territories of the MCA and PCA, 1 between pression of the contralateral carotid artery during the vascular territories of all three major cerebral ar- transcranial Doppler investigation. In six patients two teries and 1 between the vascular territories of the of these features occurred together. ACA and MCA. One of these four patients had an Mean follow-up time was 25.7 months (median 24.0; infarct in the territory of the ACA in addition to that range ). Follow-up was incomplete in only one in the border-zone area. Three patients had infarcts in patient, as he could not be traced after having moved the territory of the MCA; two were partial end-zone abroad. All patients were treated according to standard infarcts, one a large subcortical infarct. Two patients clinical practice, e.g. with antithrombotic agents (low- did not show a recent infarct on the CT of the brain, dose aspirin in the majority of patients) and modification in both cases performed within 48 hours of onset of of risk factors. symptoms. Thirty-two patients had a severe ([70%, based on Seven patients died of vascular causes other than Doppler criteria) stenosis of the contralateral ICA, that ischaemic stroke (six of myocardial infarction, one of was symptomatic in five. Twenty-two of the 32 patients stroke of undetermined type) and six patients suffered underwent CEA. Ten patients were not operated on non-fatal myocardial infarction. Retinal infarction was because the physician involved did not recommend not observed. Five more patients died, four of cancer the procedure (six patients, one of whom had COI and one of urosepsis. The annual risk of ischaemic alone), because the patient refused the operation (two stroke (fatal or non-fatal) was 5.3% for all strokes (95% patients) or because EC/IC bypass surgery was ad- CI %) and 3.8% for ipsilateral ischaemic stroke vised (one patient). The tenth patient was scheduled (95% CI %). The annual risk of the combined for CEA, but suffered a minor stroke before the oper- vascular outcome event of death from any vascular ation took place. Shortly after the event this patient cause, non-fatal stroke or non-fatal MI was 10.7% (95% was operated on. Two further patients underwent CEA CI %). of a severe stenosis of the ipsilateral ECA. Risk factors Outcome The results of the univariate analyses of risk factors During the time of follow-up 11 patients suffered for the primary and secondary outcome event are a recurrent ischaemic stroke, eight of which were shown in Table 1. A severe contralateral stenosis of ipsilateral to the symptomatic occluded ICA (Table 2). the ICA, hyperlipidaemia and smoking were identified All 11 patients had been seen by a neurologist at the as risk factors for recurrent ischaemic stroke. Multitime of the recurrent stroke and all patients were variate analysis showed that a history of hyperusing antithrombotic medication: low-dose aspirin in lipidaemia (hazard ratio (HR) 5.0; 95% CI %) 9 patients, aspirin and dipyridamole in one patient and severe contralateral stenosis of the ICA (HR 12.6; (number 10) and one other patient (number 9) was 95% CI %) independently increased the risk of treated with heparin. Four strokes (three minor, one recurrent ischaemic stroke. A history of heart disease fatal) occurred within 30 days of CEA of the contra- (HR 4.5; 95% CI %) and smoking (HR 1.4; 95% lateral ICA, in fact within the first day of surgery. CI %) were independent determinants for the Patients 9 and 10 suffered a recurrent ipsilateral stroke occurrence of any vascular event. 6.5 and 3.5 months after their presenting symptoms, despite the fact that they underwent CEA of their severe stenosis of the contralateral ICA. Three strokes Risk of CEA of the contralateral ICA (patients 3, 4 and 6) occurred in the group of 10 patients who had a severe contralateral ICA stenosis Of the 22 patients who underwent CEA of a severe but were not operated on; patients 4 and 6 sub- stenosis contralateral to the symptomatic ICA oc-

5 Outcome in Symptomatic ICA Occlusion 583 Table 2. Characteristics of 11 patients who suffered a recurrent ischaemic stroke. Presenting Severe CEA Side recurrent stroke Artery Severity recurrent Time interval CEA Time interval symptoms contralateral (ipsi- or contralateral involved stroke recurrent stroke presenting symptoms stenosis to ICA occlusion) recurrent stroke Patient 1 Minor stroke Yes Yes Ipsilateral MCA Minor stroke 1 day 2 months Patient 2 Minor stroke No Ipsilateral Border-zone Minor stroke 43 months MCA-PCA Patient 3 Minor stroke Yes No Ipsilateral Border-zone Major stroke 6 months ACA-MCA and ACA Patient 4 Minor stroke Yes No Contralateral MCA Minor stroke 37 months Patient 5 TIA No Contralateral Border-zone Minor stroke 6 months (occlusion) MCA-PCA Patient 6 Minor stroke Yes No Ipsilateral MCA Minor stroke 1 month Patient 7 Minor stroke Yes Yes Contralateral Border-zone Minor stroke Several hours 2 months ACA-MCA-PCA Patient 8 TIA Yes Yes Ipsi- and ACA and Fatal stroke Several hours 21 months contralateral MCA? Patient 9 TIA Yes Yes Ipsilateral MCA Major stroke 6.5 months Patient 10 TMB only Yes Yes Ipsilateral MCA Minor stroke 3.5 months Patient 11 TIA Yes Yes Ipsilateral MCA Minor stroke 12 hours 6 days CEA, carotid endarterectomy; ICA, internal carotid artery; MCA, middle cerebral artery; PCA, posterior cerebral artery; ACA, anterior cerebral artery; TIA, transient ischaemic attack; TMB, transient monocular blindness;, severe stenosis of the contralateral ICA symptomatic;, CEA performed after recurrent stroke. In both cases without complications., not possible to determine side of stroke because the patient developed severe coma without any lateralising signs two hours after surgery and the CT of the brain showed no signs of recently developed ischaemia. Duplex investigation showed occlusion of the common carotid artery on the side of the CEA in addition to the earlier diagnosed occlusion of the ICA on the other side. Most likely ischaemia occurred in both hemispheres;, one day after coronary arterial bypass graft surgery.

6 584 C. J. M. Klijn et al. clusion, four (18%; 95% CI 5 40%) had a perioperative high, e.g. 8 of 11 (73%). In the literature, information stroke. In one of these (patient 7) the hemisphere on the severity of the recurrent strokes is rarely given. ipsilateral to the operated side had also given rise to Information on the prognostic value of risk factors for ischaemic symptoms and the infarct occurred on that atherosclerosis for recurrent vascular events in patients side. In 16 of the 22 operations (73%; 95% CI 50 89%) with symptomatic ICA occlusion is scarce. In two recent a shunt was used because of changes in the EEG after studies in patients with symptomatic ICA occlusion, clamping of the carotid artery. The group of 22 CEAs age, but no other cardiovascular risk factor indoes not include the two CEAs performed (without dependently predicted recurrent stroke. 6,15 Another complications) in patients 4 and 6 after a recurrent study found that a high systolic blood pressure inischaemic stroke (Table 2). creased the chance of recurrent stroke and that anti- A control group of 20 patients underwent CEA of hypertensive treatment diminished this chance. 14 A a severe stenosis of the ICA in the presence of an high haemoglobin concentration, a prognostic factor for asymptomatic contralateral ICA occlusion. Ten recurrent ischaemic stroke in the general population in patients had symptoms of the hemisphere or retina the Framingham study, 16 had no predictive value in the ipsilateral to the side of the stenosis and 10 patients medically treated patients randomised in the EC/IC were completely asymptomatic. In 10 of these 20 bypassstudy. 17 Althoughwerealisethattheresults from patients (50%; 95% CI 27 73%) a shunt was needed. the multivariate analysis in our study are based on rel- No complications (0%; 95% CI 0 17%) occurred within atively few endpoints as reflected in the rather wide 30 days of surgery. 95% CIs, it is of interest that we found that hyper- The absolute difference in complication risk after lipidaemia and a severe stenosis of the contralateral ICA CEA in the presence of a contralateral ICA occlusion were independent risk factors for recurrent stroke in between patients with and without symptoms atpatients with symptomatic ICA occlusion, whereas a tributable to the occluded ICA was 18% (95% CI history of heart disease and smoking were in- 2 34%). A shunt tended to be applied more often in dependently associated with any subsequent vascular patients with symptomatic ICA occlusion (RR 1.5; 95% event. We did not study haemoglobin concentration. CI %). Hypertension could not be shown to be a risk factor for recurrent ischaemic stroke, probably because it was so common in the entire group of patients. The presence Discussion of an infarct in a border-zone area as a predictor of poor outcome has been described, 18 but could not be confirmed in our study. The number of border-zone This study shows that the annual risk of recurrent ischaemic stroke in patients with symptomatic ICA infarcts as presenting stroke was 13 of 97 (13%) in our occlusion has not improved over the years. We found series, compared with 8 of 154 (5%) in the series pub- lished bybogousslavsky and Regli. 18 an annual risk of recurrent ischaemic stroke of 5.3% CT-proven border- (95% CI %) and of recurrent ipsilateral ischaemic zone infarcts accounted for 72% (18 of 25) of recurrent stroke of 3.8% (95% CI %), which correspond strokes (all ipsilateral) in their study, whereas in our with the 5.5% (95% CI %) and 2.1% (95% CI series 44% (four of nine) of CT-proven recurrent strokes %) which were found in a review of 20 studies were of the border-zone type (two ipsi- and two contrapublished between 1961 and The annual stroke lateral). Haemodynamic features of presenting symp- risk found in other recent studies on the outcome of toms were not associated with increased risk of a patients with symptomatic ICA occlusion may be less recurrent vascular event. This is in accordance with the reliable, because most of these included measurements observations in a prospective study of 34 medically of cerebral blood flow or cerebrovascular reserve capwith treated patients with bilateral ICA occlusion. 19 Patients acity. 6,12 15 This may have caused selection bias in that ICA occlusion have lower cerebral blood flow/ patients perceived to be at high risk of recurrent stroke cerebral blood volume ratios when studied with pos- may have been more readily referred for ancillary itron emission tomography (PET) 20 and worse MCA investigations. Our study has the limitations inherent blood flow reactivity measured with PET after induced to the retrospective design, but the inclusion criteria hypercapnia, 21 but to date no prospective follow-up were stricter than in most previous studies in that study is available in which haemodynamic aspects of we included only patients with signs and symptoms symptoms in patients with unilateral ICA occlusion are attributable to the brain or eye on the side of the shown to have predictive value for recurrent stroke or ICA occlusion. In the current study the proportion of other vascular events. Because of the retrospective derecurrent strokes that were minor disabling was rather sign of our study, the presence of haemodynamic symp-

7 Outcome in Symptomatic ICA Occlusion 585 toms may have been underestimated. In concordance case decades ago, and that patients now are routinely with the results from a recent prospective study, 6 the treated with antithrombotic drugs, outcome in patients recurrence of presenting symptoms after the ICA ocimproved over recent decades. This information is with symptomatic ICA occlusion has not substantially clusion was documented had no predictive value for recurrent stroke in our series. A compromised cerebral important in the design of new studies to evaluate blood flow identified by PET, 6,13 by transcranial Doppler therapeutic strategies in these patients. CEA of severe with carbon dioxide challenge 22 or by stable xenon CT stenosis of the contralateral ICA carried a relatively with acetazolamide 23 has been shown to be an important high risk, but may still be a useful therapy and needs prognostic factor for recurrent ischaemic stroke in to be studied in a controlled design. patients with symptomatic ICA occlusion, but was not studied in our retrospective series. The presence of a severe stenosis of the contralateral ICA constitutes a special problem in patients with Acknowledgements symptomatic ICA occlusion. These patients carry a rel- Dr J. D. Blankensteijn gave helpful comments on an earlier version atively high risk of recurrent ischaemic stroke (9/32 of the manuscript. The help of Mrs F. A. C van Vliet with collection versus 2/65). In patients with occlusion of the ICA, of the data is gratefully acknowledged. C. J. M Klijn is supported by the Netherlands Heart Foundation (grant ). blood flow via the anterior communicating artery is an important collateral pathway in 45 70% of patients Although no randomised trials are available, CEA of a severe stenosis of the contralateral ICA seems a rational References option CEA in the presence of contralateral ICA occlusion was associated with a relatively high (up to 1Klijn CJM, Kappelle LJ, Tulleken CAF, van Gijn J. Symptomatic carotid artery occlusion. A reappraisal of hemodynamic 14%) risk in some series, but not in others. 29,33 36 In a factors. Stroke 1997; 28: recent review perioperative complications associated 2Hankey GJ, Warlow CP. Prognosis of symptomatic carotid with CEA were more common in the presence of a artery occlusion. An overview. Cerebrovasc Dis 1991; 1: contralateral ICA occlusion (odds ratio 1.91 (95% CI 3The EC/IC Bypass Study Group. Failure of extracranial intracranial arterial bypass to reduce the risk of ischemic stroke %)). 37 In many previous studies patients with Results of an international randomized trial. N Engl J Med 1985; symptoms attributable to the hemisphere ipsilateral to 313: the ICA occlusion were not described separately, and 4Widder B, Kornhuber HH. Extra intracranial bypass surgery in carotid artery occlusion; Who benefits? Neurol Psychiat Brain most studies included patients with asymptomatic oc- Res 1994; 2: clusion as well. Although the number of patients who 5Vorstrup S, Paulson OB. Extracranial intracranial bypass re- visited. Cerebrovasc Dis 1992; 2: underwent CEA was relatively small, our data indicate 6Grubb RL Jr, Derdeyn CP, Fritsch SM et al. Importance of that the risk of CEA of a severe ICA stenosis in the hemodynamic factors in the prognosis of symptomatic carotid presence of an ICA occlusion is higher in patients who artery occlusion. JAMA 1998; 280: Fisher CM. Concerning recurrent transient cerebral ischemic have suffered ischaemic symptoms of the eye or brain attacks. Can Med J 1962; 86: ipsilateral to the ICA occlusion than in those who have 8Baquis GD, Pessin MS, Scott RM. Limb shaking A carotid always been asymptomatic. If haemodynamic com- TIA. Stroke 1985; 16: Tatemichi TK, Young WL, Prohovnik I et al. Perfusion inpromise indeed plays an important role in causing sufficiency in limb-shaking transient ischemic attacks. Stroke cerebral ischaemia in patients with symptomatic ICA 1990; 21: occlusion, 1 then susceptibility to perioperative changes 10 Furlan AJ, Whisnant JP, Kearns TP. Unilateral visual loss in bright light. An unusual symptom of carotid artery occlusive in blood pressure could possibly be one explanation for disease. Arch Neurol 1979; 36: the high risk of surgery in these patients. Whether the 11 Ross Russell RW, Page GR. Critical perfusion of brain and risk of CEA of a contralateral severe ICA stenosis is retina. Brain 1983; 106: Derlon JM, Bouvard G, Viader F et al. Impaired cerebral outweighed by the benefit of the procedure in later years hemodynamics in internal carotid occlusion. Cerebrovasc Dis 1992; can not be concluded from our uncontrolled data. Also, 2: Yamauchi H, Fukuyama H, Nagahama Y et al. Evidence for EC/IC bypass surgery is not without risk, and further misery perfusion and risk for recurrent stroke in major cerebral study is needed to identify patients at high risk for fur- arterial occlusive diseases from PET. J Neurol Neurosurg Psychiatry ther cerebral ischaemic events. A low-flow state of the 1996; 61: Yokota C, Hasegawa Y, Minematsu K, Yamaguchi T. Effect of brain is likely to be one, but probably not the only, imacetazolamide reactivity and long-term outcome in patients with portant determinant. major cerebral artery occlusive diseases. Stroke 1998; 29: We conclude that, despite the fact that much more 15 Vernieri F, Pasqualetti P, Passarelli F, Rossini PM, Silvestrini M. Outcome of carotid artery occlusion is predicted by attention is being paid to treatment of risk factors such cerebrovascular reactivity. Stroke 1999; 30: as hypertension and hyperlipidaemia than was the 16 Kannel WB, Gordon T, Wolf PA, McNamara P. Hemoglobin

8 586 C. J. M. Klijn et al. and the risk of cerebral infarction: The Framingham Study. Stroke 28 Sterpetti AV, Feldhaus RJ, Schultz RD, Farina C. Operative 1972; 3: strategies in patients with symptomatic internal carotid artery 17 Wade JPH, Taylor DW, Barnett HJM, Hachinski VC, for the occlusion. Surgery 1989; 105: EC/IC bypass study group. Hemoglobin concentration and 29 Meyer FB, Fode NC, Marsh WR, Piepgras DG. Carotid enprognosis in symptomatic obstructive cerebrovascular disease. darterectomy in patients with contralateral carotid occlusion. Stroke 1987; 18: Mayo Clin Proc 1993; 68: Bogousslavsky J, Regli F. Borderzone infarctions distal to in- 30 Friedman SG, Riles TS, Lamparello PJ, Imperato AM, Sakwa ternal carotid artery occlusion: prognostic implications. Ann MP. Surgical therapy for the patient with internal carotid artery Neurol 1986; 20: occlusion and contralateral stenosis. J Vasc Surg 1987; 5: Wade JPH, Wong W, Barnett HJM, Vandervoort P. Bilateral 31 Gasecki AP, Eliasziw M, Ferguson GG, Hachinski V, Barnett occlusion of the internal carotid arteries. Brain 1987; 110: HJM, North American Symptomatic Carotid Endarter ectomy Trial Collaborators. Long-term prognosis and effect 20 Gibbs JM, Wise RJS, Leenders KL, Jones T. Evaluation of cerebral of endarterectomy in patients with symptomatic severe carotid perfusion reserve in patients with carotid occlusion. Lancet 1984; stenosis and contralateral carotid stenosis or occlusion: results 1: from NASCET. J Neurosurg 1995; 83: Levine RL, Lagreze HL, Dobkin JA et al. Cerebral vasoarterectomy contralateral to an occluded carotid artery: peri- 32 Mackey WC, O Donnell TF, Callow AD. Carotid endcapacitance and TIA s. Neurology 1989; 39: operative risk and late results. J Vasc Surg 1990; 11: Kleiser B, Widder B. Course of carotid artery occlusion with 33 Mattos MA, Barkmeier LD, Hodgson KJ, Ramsey DE, Sumner impaired cerebrovascular reactivity. Stroke 1992; 23: DS. Internal carotid artery occlusion: Operative risks and long- 23 Webster MW, Makaroun MS, Steed DL et al. Compromised term stroke rates after contralateral carotid endarterectomy. Surcerebral blood flow reactivity is a predictor of stroke in patients gery 1992; 112: with symptomatic carotid artery occlusive disease. J Vasc Surg 34 Deriu GP, Franceschi L, Milite D et al. Carotid artery end- 1995; 21: arterectomy in patients with contralateral carotid artery oc- 24 Muller M, Hermes M, Bruchmann H, Schimrigk K. Trans- clusion: perioperative hazards and late results. Ann Vasc Surg cranial Doppler ultrasound in the evaluation of collateral blood 1994; 8: flow in patients with internal carotid artery occlusion: Correlation 35 Cao P, Giordano G, De Rango P et al. Carotid endarterectomy with cerebral angiography. AJNR 1995; 16: contralateral to an occluded carotid artery: A retrospective case- 25 Furst G, Steinmetz H, Fischer H et al. Selective MR angio- control study. Eur J Vasc Endovasc Surg 1995; 10: graphy and intracranial collateral blood flow. J Cumput Assist 36 Hammacher ER, Eikelboom BC, Bast TJ, De Geest R, Ver- Tomogr 1993; 17: meulen FEE. Surgical treatment of patients with a carotid artery 26 Baumgartner RW, Baumgartner I, Mattle HP, Schroth G. occlusion and a contralateral stenosis. J Cardiovasc Surg 1984; 25: Transcranial color-coded duplex sonography in the evaluation of collateral flow through the circle of Willis. Am J Neuroradiol 37 Rothwell PM, Slattery J, Warlow CP. Clinical and an- 1997; 18: giographic predictors of stroke and death from carotid endarterectomy: 27 Harrison MJG, Marshall J. The variable clinical and CT systematic review. BMJ 1998; 315: findings after carotid occlusion: the role of collateral blood supply. J Neurol Neurosurg Psychiatry 1988; 51: Accepted 14 February 2000