Cilostazol for peripheral arterial disease

BACKGROUND

Peripheral arterial disease (PAD) affects 4% to 12% of people aged 55 to 70 years and 20% of people over 70 years. The most common complaint is intermittent claudication (IC) characterised by pain in the legs or buttocks that occurs with exercise and which subsides with rest. Compared with age-matched controls, people with IC have a three- to six-fold increase in cardiovascular mortality. Symptoms of IC, walking distance, and quality of life can be improved by risk factor modification, smoking cessation, and a structured exercise program. Antiplatelet treatment is beneficial in patients with IC for the reduction of vascular events but has not been shown to influence claudication distance.

OBJECTIVES

To determine the effect of cilostazol on improving walking distance and in reducing vascular mortality and cardiovascular events in patients with stable IC.

SEARCH STRATEGY

The Cochrane Peripheral Vascular Diseases Group searched their specialised register (last searched August 2007) and the Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 3, 2007). We searched MEDLINE (1966 to November 2005), EMBASE (1980 to November 2005), several more specialised databases, and reference lists of articles.

SELECTION CRITERIA

Double-blind, randomised controlled trials of cilostazol versus placebo, or versus other antiplatelet agents in patients with stable IC or patients undergoing vascular surgical intervention for PAD.

DATA COLLECTION AND ANALYSIS

Two authors independently assessed trials for selection and all three authors independently extracted data.

MAIN RESULTS

Seven randomised controlled trials comparing cilostazol with placebo were included. The weighted mean difference (WMD) for the initial claudication distance (ICD) was improved following treatment with cilostazol 100 mg twice daily (WMD 31.1 m; 95% confidence interval (CI): 21.3 to 40.9 m) and 50 mg twice daily (WMD 41.3 m; 95% CI: -7.1 to 89.7 m) compared with placebo. Participants receiving cilostazol 150 mg twice daily had an increased ICD (WMD 15.7 m; 95% CI: -9.6 to 41.0 m) compared with those receiving placebo. One study also included a comparison with pentoxifylline. In this study, participants receiving cilostazol had significant improvement in ICD compared with placebo. There was no increase in major adverse events including cardiovascular events or mortality in patients receiving cilostazol compared with placebo.

AUTHORS' CONCLUSIONS

Patients with IC should receive secondary prevention for cardiovascular disease. Cilostazol has been shown to be of benefit in improving walking distance in people with IC. There are no data on whether it results in a reduction of adverse cardiovascular events.

Cilostazol for peripheral arterial disease

BACKGROUND

Peripheral arterial disease (PAD) affects 4% to 12% of people aged 55 to 70 years and 20% of people over 70 years. The most common complaint is intermittent claudication (IC) characterised by pain in the legs or buttocks that occurs with exercise and which subsides with rest. Compared with age-matched controls, people with IC have a three- to six-fold increase in cardiovascular mortality. Symptoms of IC, walking distance, and quality of life can be improved by risk factor modification, smoking cessation, and a structured exercise program. Antiplatelet treatment is beneficial in patients with IC for the reduction of vascular events but has not been shown to influence claudication distance.

OBJECTIVES

To determine the effect of cilostazol on improving walking distance and in reducing vascular mortality and cardiovascular events in patients with stable IC.

SEARCH STRATEGY

The Cochrane Peripheral Vascular Diseases Group searched their specialised register (last searched August 2006) and the Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 3, 2006). We searched MEDLINE (1966 to November 2005), EMBASE (1980 to November 2005), several more specialised databases, and reference lists of articles.

SELECTION CRITERIA

Double-blind, randomised controlled trials of cilostazol versus placebo, or versus other antiplatelet agents in patients with stable IC or patients undergoing vascular surgical intervention for PAD.

DATA COLLECTION AND ANALYSIS

Two authors independently assessed trials for selection and all three authors independently extracted data.

MAIN RESULTS

Eight randomised controlled trials comparing cilostazol with placebo were included. The weighted mean difference (WMD) for the initial claudication distance (ICD) was improved following treatment with cilostazol 100 mg twice daily (WMD 31.1; 95% confidence interval (CI): 21.4 to 40.9) and 50 mg twice daily (WMD 41.3; 95% CI: -7.1 to 89.7) compared with placebo. Participants receiving cilostazol 150 mg twice daily had an increased ICD (WMD 15.7; 95% CI: -9.6 to 41.0) compared with those receiving placebo. One study also included a comparison with pentoxifylline. In this study, participants receiving cilostazol had significant improvement in ICD compared with placebo. There was no increase in major adverse events including cardiovascular events or mortality in patients receiving cilostazol compared with placebo.

AUTHORS' CONCLUSIONS

Patients with IC should receive secondary prevention for cardiovascular disease. Cilostazol has been shown to be of benefit in improving walking distance in people with IC. There are no data on whether it results in a reduction of adverse cardiovascular events.

Dextran 40 reduces in vitro platelet aggregation in peripheral arterial disease

Dextran 40 has been used to prevent post-operative thrombosis. However, its antithrombotic and antiplatelet effects in peripheral arterial disease (PAD) are poorly understood. We studied the in vitro effects of Dextran 40 on platelet function in control subjects and PAD patients using whole blood methods. Platelet function was assessed in 20 control subjects and 20 PAD patients. Spontaneous platelet aggregation (SPA) and agonist-induced platelet aggregation in response to increasing concentrations of Dextran 40 in vitro were measured by whole blood aggregometry. Flow cytometric measurements of platelet P-selectin, GpIIb/IIIa, GpIb and PAC-1 binding were also performed. There was no difference in SPA or ADP-induced aggregation in control patients with Dextran 40 in vitro. However, Dextran 40 inhibited collagen-induced aggregation in control patients (P < 0.05, Friedman test). In PAD patients, SPA and ADP (1 microM)-induced aggregation were significantly reduced by Dextran 40 in vitro (P < 0.001, Friedman test). In PAD patients, collagen-induced platelet aggregation (1 and 5 microg/ml) was significantly reduced by Dextran 40 in vitro (P < 0.01, Friedman test). GpIIb/IIIa, PAC-1 and P-selectin expression were significantly reduced in whole blood samples from PAD patients following incubation with Dextran 40 (P < 0.05, Wilcoxon rank test) but not in samples from control patients. Dextran 40 reduces spontaneous and agonist-induced platelet aggregation as well as the surface expression of markers of platelet activation in PAD patients. This antiplatelet effect may be of benefit to patients undergoing vascular surgical procedures where thrombosis is a significant risk.

Comparison of POSSUM with P-POSSUM for prediction of mortality in infrarenal abdominal aortic aneurysm repair

The Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM) is a simple and valid scoring system in predicting mortality and morbidity rates. The Portsmouth predictor equation (P-POSSUM) has been shown to be a more accurate predictor of death than the POSSUM in vascular patients. The length of hospital stay (LOS) equation has been suggested to be of value in predicting total length of stay. The aim of this study was to test the validity of the POSSUM, P-POSSUM, and LOS in predicting outcome of patients undergoing abdominal aortic aneurysm (AAA) repair. POSSUM scores in 118 patients who underwent AAA repair by a single consultant were recorded retrospectively. Observed rates of mortality, morbidity, and length of hospital stay were correlated with the rates predicted by POSSUM, P-POSSUM, and LOS equations in three groups: all cases, 93 elective repairs, and emergency AAA repairs. The POSSUM and the P-POSSUM performed similarly in terms of accuracy of prediction, with all predicted values being not significantly different from those observed. The POSSUM tended to overpredict mortality compared to the P-POSSUM. The POSSUM predicted morbidity well. The LOS equation failed to predict significantly observed total hospital stay. POSSUM and P-POSSUM outcome risk equations are thus valid in predicting mortality for all cases and emergency AAA repairs. The POSSUM morbidity equation predicts complications quantitatively.

Systematic review of antiplatelet therapy for the prevention of myocardial infarction, stroke or vascular death in patients with peripheral vascular disease

BACKGROUND

Antiplatelet agents may prevent vascular events and death in patients with peripheral vascular disease (PVD).

METHODS

A systematic review of 39 randomized controlled trials of antiplatelet therapy in patients with PVD was performed.

RESULTS

For patients with PVD the number suffering a non-fatal myocardial infarction, non-fatal stroke or vascular death in the antiplatelet group was 6.5 per cent compared with 8.1 per cent in the placebo group (odds ratio 0.78 (95 per cent confidence interval (c.i.) 0.63--0.96); P = 0.02), favouring antiplatelet treatment. For infrainguinal bypass surgery (ten trials) and balloon angioplasty (two) the differences were still in favour of antiplatelet therapy, but they did not reach statistical significance. In five trials of aspirin against another antiplatelet agent, 8.4 per cent in the aspirin group suffered a vascular event compared with 6.6 per cent in the second antiplatelet group (odds ratio 0.76 (95 per cent c.i. 0.64--0.91); P < 0.01), favouring ticlopidine/clopidogrel/aspirin + dipyridamole against aspirin alone.

CONCLUSION

Antiplatelet therapy reduces serious vascular events and vascular death in patients with PVD. For infrainguinal arterial surgery or balloon angioplasty the benefit remains unproven, but the number of trials to date is small. There is also evidence to support the use of antiplatelet drugs other than aspirin for the prevention of vascular events in those with PVD.

Correlates of embolic events detected by means of transcranial Doppler in patients with carotid atheroma

PURPOSE

This study identified in patients with carotid plaques the associations of emboli detected by means of transcranial Doppler (TCD) with cerebrovascular symptoms, brain computed tomography (CT) infarction patterns, and the attributes of plaques (echodensity, degree of stenosis).

METHODS

Eighty carotid plaques (in 59 patients), producing 50% to 99% stenosis, were imaged on duplex scanning and analyzed echomorphologically in a computer with the gray scale median (GSM). The GSM facilitated the quantitative distinction of dark (low GSM) from bright (high GSM) plaques. Stenosis was assessed with duplex scanning. Emboli were counted on TCD in the ipsilateral middle cerebral artery for half an hour. The brain CT infarction patterns (pattern A: discrete subcortical and cortical; pattern B: hemodynamic, diffuse white matter lesions, basal ganglia infarctions, lacunes) and normal CT and cerebrovascular symptoms on the ipsilateral hemisphere were noted.

RESULTS

Emboli were more frequent in symptomatic (median count, 3) than asymptomatic (median count, 0) hemispheres (Mann-Whitney U test, P =.031) and in hemispheres with pattern A infarction (median count, 3.5) than in hemispheres with pattern B infarction or normal CT (median count, 0; Kruskal-Wallis test, P =.047). The increased embolic count was associated with decreased GSM (Spearman correlation, P =.045, r = -0.22), but not with high degrees of stenosis (Spearman correlation, P =.44, r = 0.086).

CONCLUSION

Emboli were more frequent in symptomatic than asymptomatic hemispheres and in CT pattern A harboring hemispheres than in CT pattern B or normal hemispheres. They were more frequent in the presence of low-plaque echodensity, but not in the presence of a high degree of stenosis. These data support the embolic nature of cerebrovascular symptomatology and CT pattern A infarctions.

Echomorphologic and histopathologic characteristics of unstable carotid plaques

BACKGROUND AND PURPOSE

Our hypothesis was that the carotid plaques associated with retinal and cerebrovascular symptomatology and asymptomatic presentation may be differ from each other. The aim of this study was to identify the sonographic and histopathologic characteristics of plaques that corresponded to these three clinical manifestations.

METHODS

The echo process involved duplex preoperative imaging of 71 plaques (67 patients, 21 plaques were associated with retinal, 25 with cerebrovascular symptoms, and 25 were asymptomatic), which was performed in a longitudinal fashion. Appropriate frames were captured and digitized via S-video signal in a computer and digitized sonograms were normalized by two echo-anatomic reference points: the gray scale median (GSM) of the blood and that of the adventitia. The GSM of the plaques was evaluated to distinguish dark (low-GSM) from bright (high-GSM) plaques. Subsequent to endarterectomy, the plaques were sectioned transversely, and a slice at the level of the largest plaque area was examined for the relative size of necrotic core and presence of calcification and hemorrhage.

RESULTS

Retinal symptomatology was associated with a hypoechoic plaque appearance (median GSM: 0), asymptomatic status with a hyperechoic plaque appearance (median GSM: 34), and cerebrovascular symptomatology with an intermediate plaque appearance (median GSM: 16) (P = .001). The histopathologic characteristics did not disclose differences between the three clinical groups. The hypoechoic plaque appearance was associated only with the presence of hemorrhage (median GSM for the hemorrhagic plaques, 6, and for the non-hemorrhagic ones, 20 [P = .04]). The relative necrotic core size and the presence of calcification did not show any echomorphologic predilection.

CONCLUSION

Our results showed that distinct echomorphologic characteristics of plaques were associated with retinal and cerebrovascular symptomatology and asymptomatic status. Histopathologically, only the presence of hemorrhage proved to have an echomorphologic predilection.

The significance of the cerebral collateral capacity in patients with carotid atheroma

OBJECTIVES

to identify the echodensity, stenosis of carotid plaques and cerebral collateral capacity that were associated with various ipsilateral presentations (retinal, cerebrovascular, asymptomatic).

DESIGN

cross-sectional study.

MATERIALS

forty-four patients, with 44 plaques associated with various presentations, were studied.

METHODS

the duplex images of the plaques were analysed echomorphologically in a computer by means of Grey Scale Median (GSM) [hypoechoic (low GSM), hyperechoic (high GSM)]. The percentage (%) reduction of the mean velocity in the middle cerebral artery (PRMCA) on transcranial Doppler, during clamping in carotid endarterectomy, was evaluated to distinguish the competent cerebral collateral supply (low PRMCA) from the non-competent one (high PRMCA).

RESULTS

the retinal symptoms were associated with plaques of low median GSM (0), severe median stenosis (90%) and low median PRMCA (0.31) as contrasted with the cerebrovascular symptoms (17, 84%, 0.47, respectively) and asymptomatic status (32, 83%, 0.4, respectively) [(p =0.038 (GSM), p =0.67 (stenosis), p=0.15 (PRMCA)]. The retinal and the cerebrovascular symptoms were distinct in terms of PRMCA (p=0.045).

CONCLUSIONS

the retinal symptoms were produced by hypoechoic and possibly embologenic plaques, whereas the cerebrovascular ones possibly by the combination of carotid embolism and a non-competent cerebral collateral circulation. Asymptomatic status was associated with the absence of any relevant mechanism.

Vascular surgical society of great britain and ireland: carotid angiography is used more selectively in the asymptomatic carotid surgery trial

BACKGROUND: The ongoing Asymptomatic Carotid Surgery Trial (ACST) has randomized more than 1900 patients to determine whether carotid endarterectomy prolongs stroke-free survival compared with best medical treatment alone. Previously the Asymptomatic Carotid Atherosclerosis Study demonstrated that preoperative angiography caused stroke or death in 1 per cent of patients, and many centres have now reduced or abandoned this practice. This study determined the changing practice of carotid angiography in the ACST. METHODS: Collaborating surgeons completed questionnaires annually on their use and method of angiography. Information on patients in the ACST who had angiography at randomization was also obtained. RESULTS: In 1993, 77 per cent of responding collaborators always performed preoperative angiography and 23 per cent used angiography selectively. This trend has reversed: by 1997, 26 per cent always used preoperative angiography, 70 per cent of respondents used preoperative angiography selectively and 4 per cent had abandoned angiography (P < 0.001, chi2 for trend). Information on carotid angiography at randomization has to date been obtained on 1141 patients in the ACST. Some 44 per cent (497 of 1141) had carotid angiography at randomization. Surgical patients had angiography more commonly than those in the medical group (49 versus 42 per cent; P < 0.03, chi2 test). Changes in carotid angiography were analysed by year of randomization. In 1993, 61 per cent of patients randomized had carotid angiography compared with 42 per cent in 1996 and 1997 (P < 0.001, chi2 for trend). The use of carotid angiography was not related to degree of stenosis estimated by Doppler ultrasonography. CONCLUSION: In the ongoing ACST, there is increasingly selective use of carotid angiography. Less than half the patients in this study had carotid angiography and the use of angiography is now decreasing. This has important implications for validation of carotid duplex in this trial and in future studies.

Vascular surgical society of great britain and ireland: platelet function during carotid endarterectomy and the antiplatelet effect of dextran 40

BACKGROUND: Dextran 40 has been shown to reduce cerebral embolization following carotid endarterectomy (CEA). This study aimed to determine changes in platelet function during CEA and the antiplatelet effect of dextran 40. METHODS: Platelet function was measured by whole-blood flow cytometry in the peripheral arterial blood of patients during CEA and 24 h later. The binding index of P-selectin and PAC-1 expression were measured as markers of activation and aggregation. Patients were kept on aspirin 75-150 mg until the day of surgery and received an intravenous bolus of 5000 units unfractionated heparin before carotid artery clamping. High-intensity transient signals (HITS) in the ipsilateral middle cerebral artery were measured with transcranial Doppler (TCD) ultrasonography before, during and after operation. Results are presented as median (interquartile range) and statistical significance was determined using the Mann-Whitney U test. RESULTS: Thirty-eight patients undergoing CEA were studied. The P-selectin binding index rose significantly from incision (0.9 (0.2-2.7)) after carotid clamping (1.5 (0.6-3.6); P < 0.01), clamp release (1.7 (0. 3-3.0); P < 0.01), 1 h after operation (1.5 (0.3-2.6); P < 0.05) and 24 h after operation (1.3 (0.6-2.5); P < 0.05). PAC-1 binding index increased from incision (0.4 (0.1-0.8)) after carotid clamping (2.0 (0.4-4.2); P < 0.01) and clamp release (1.8 (0.3-2.9); P < 0.05). TCD monitoring showed an increase in preoperative HITS per 30 min (2 (0-3)) during dissection (8 (1-15); P < 0.05), after clamp release (16 (2-27); P < 0.01) and during recovery (10 (2-29); P < 0.01). After operation, patients with more than 50 HITS per 30 min were started on an infusion of dextran 40. Six patients had a dextran 40 infusion. The P-selectin binding index decreased from 1.6 (0.7-1.9) to 0.6 (0.3-1.5) 1 h after dextran (P < 0.05) and 0.1 (0.1-0.2) 24 h after dextran (P < 0.05). PAC-1 expression decreased from 0.4 (0.3-0. 5) to 0.1 (0.1-0.1) 24 h after dextran (P < 0.05). CONCLUSION: Significant platelet activation and aggregation occurs during CEA despite the use of antiplatelet treatment. A simultaneous increase in HITS was demonstrated with TCD. Dextran 40 was shown to have an antiplatelet effect after CEA; this is further evidence that it may have a role in reducing thromboembolic complications.

Carotid duplex ultrasonography: importance of standardisation. Asymptomatic Carotid Surgery Trial Collaborators and Vascular Laboratories

BACKGROUND

Carotid duplex ultrasonography (CDUS) is an established non-invasive tool for assessing patients with suspected carotid bifurcation disease. Current trends show an increased dependence on CDUS in recommending patients for stroke prevention surgery. The aim of the study was to evaluate routine practice in vascular laboratories across 26 countries participating in The Asymptomatic Carotid Surgery Trial (ACST), and to determine the areas which are in need of future standardisation if CDUS is to be the primary tool in recommending patients for CEA.

METHODS

Retrospective. Information was gathered from questionnaires, and 22 on-site visits of vascular laboratories between August 1996 to September 1997.

SETTING

Clinical vascular laboratory practice.

PARTICIPANTS

Eligible participants were vascular laboratories of ACST collaborators.

MEASURES

Laboratories were compared in 7 categories: ultrasound equipment, operators, experience, protocols, stenosis evaluation, interpretation criteria, and reporting.

RESULTS

Information on 117 respondents showed that (i) experience: at least one operator in each laboratory had more than 3 years experience; (ii) equipment: 88% (103/117) had colour duplex capability; (iii) operators: 54% of laboratories had exams performed by technologists, 33% vascular surgeons, 28% radiologists, and 35% other. The most significant findings were in (iv) stenosis evaluation: only 29% (33/117) were using a standardised Doppler angle (this single factor may greatly alter exam results); and (v) interpretation criteria; with >41 different criteria reported. These specific laboratory variations can affect those patients considered appropriate for CEA.

CONCLUSIONS

This study highlights the most significant areas for future standardisation to be Doppler angle and interpretation criteria, if CDUS is to be a primary tool in recommending patients for CEA, when indicated by clinical trial results.

Prevalence of asymptomatic CT infarcts in the ongoing Asymptomatic Carotid Surgery Trial (ACST)

BACKGROUND

The ACST has randomised over 1670 patients to determine if carotid endarterectomy (CEA) prolongs stroke free survival when compared with best medical treatment alone. This study aims to assess the prevalence of asymptomatic cerebral infarcts in patients grouped by sex, age, mean blood pressure (MBP), total cholesterol, diabetes and coronary artery disease (CAD).

METHODS

The type of study was retrospective. Patients with preoperative CT scans were included. In those with previous stroke, only the contralateral cerebral hemisphere was studied for presence of infarcts. One thousand one hundred and forty-two patients were studied.

RESULTS

The proportion of patients with CT infarcts was 19% (211/1142). Men (155/777) had a higher incidence of CT infarction (20% vs 15%, p = 0.01). Mean age was 68 years and mean MBP was 107 mmHg. The incidence of CT infarcts was not associated with age or MBP. Diabetics or those with definite CAD had no difference in the incidence of cerebral infarction. Diabetics with definite CAD (51/796) had a greater incidence of CT infarcts (33% vs 21%, p = 0.03) compared with those without evidence of CAD and diabetes together (745/796).

CONCLUSIONS

In ACST, CT infarcts are more prevalent in men compared to women, and in diabetics with CAD. There was no difference in the incidence of CT infarction by age, mean BP, or by the presence of diabetes or CAD alone.

The prevalence of cerebral infarcts in the Asymptomatic Carotid Surgery Trial (ACST) in relation to prior contralateral symptoms

BACKGROUND

The ACST has randomised over 1670 patients to determine if carotid endarterectomy (CEA) prolongs stroke free survival versus best medical treatment alone. Some patients have had contralateral symptoms to the side under investigation, for which CEA may have been performed. This study aims to determine the prevalence of hemispheric cerebral infarcts in relation to prior contralateral symptoms.

METHODS

Patients with preoperative CT or MR scans were divided into those with prior contralateral stroke, cortical TIA, amaurosis fugax or no symptoms.

RESULTS

There were 1144 patients with preoperative CT and 170 patients with MR scans. Incidence of contralateral hemispheric CT and MR infarcts were 19% (222/1144) and 20% (34/170) respectively. Those with prior contralateral stroke (141) had the highest incidence of hemispheric CT infarcts (62%). Those with TIA (129) had a 30% incidence of CT infarct. Incidence of hemispheric CT infarcts were 17% with amaurosis fugax (46) and 10% with no prior symptoms (803). Mantel-Haenszel test for linear association was significant (p < 0.001) for increasing severity of symptoms. Considering MR scans, those with prior contralateral stroke (17) had the highest incidence of hemispheric MR infarcts (53%). Patients with TIA (23) or amaurosis fugax (6) had a 27% incidence of MR infarction. Those with no prior symptoms (121) had a 14% incidence of MR infarcts. Mantel-Haenszel test was significant (p < 0.001).

CONCLUSIONS

In the ACST, currently asymptomatic patients have an incidence of contralateral hemispheric CT and MR infarcts proportional to the severity of prior contralateral symptoms.

Are we detecting and operating on high risk patients in the asymptomatic carotid surgery trial? The Asymptomatic Carotid Surgery Trial Collaborators

OBJECTIVE

This study aims to determine whether asymptomatic carotid surgery trial (ACST) centres have entered and can identify high risk patients using duplex.

DESIGN

Retrospective study.

MATERIALS AND METHODS

Eighty-six vascular laboratories collaborating in ACST were studied, Equipment, operator experience, methodology and interpretation criteria were assessed. The ACST randomisation data were examined to determine whether patients believed to be at higher risk of stroke because of tight stenosis, contralateral occlusion or echolucent plaque were randomised.

RESULTS

Laboratories (92%) had colour duplex and 62% of all operators had > 3 years experience in carotid evaluation. The Doppler angle used to obtain peak velocity was 30-60 degrees in 65%, 60 degrees in 28% and 60-80 degrees in 6% of laboratories. Sixty-two per cent reported diameter reduction, 27% area reduction, and 11% used both methods. One-third of 1657 randomised patients were reported to have ipsilateral echolucent plaque. Median ipsilateral stenosis was 80%, 8% had contralateral occlusion and 8.5% had bilateral > 80% stenosis.

CONCLUSIONS

Centres in ACST use experienced operators, high quality equipment and conscientious data recording. Variations in methods of determining carotid stenosis exist, but can be smoothed by simple data collection. Patients at higher perceived risk of stroke are being entered and with continued recruitment it should be possible to determine whether surgery improves disabling stroke-free survival.