Risk factors for intra-operative neurological deficit during carotid endarterectomy

Relationship between absolute mean cerebral transit time and absolute mean flow velocity on transcranial Doppler ultrasound after ischemic stroke

BACKGROUND AND PURPOSE

Previous studies of transcranial Doppler (TCD) sonography in acute stroke have used the relative difference between the symptomatic and asymptomatic arteries to assess arterial occlusion. However, a simple measure of absolute mean flow velocity might provide a direct assessment of "perfusion reserve" in acute ischemic stroke.

METHODS

In a prospective study, 62 patients with ischemic stroke had TCD and a mean cerebral transit time examination within 48 hours of stroke. Absolute intracranial arterial mean flow velocities were correlated with the corresponding absolute mean transit times.

RESULTS

The authors found a significant correlation between middle cerebral artery (MCA) mean flow velocity and transit time in the symptomatic (Spearman rank correlation coefficient [rho] = -0.65, P < .01) but not in the asymptomatic (rho = -0.04, P = ns) MCA territory. Equations relating absolute mean flow velocity to absolute transit time were derived.

CONCLUSION

The findings suggest that in the normal hemisphere (with intact autoregulation on the horizontal portion of the autoregulation curve), flow velocity and transit time are not closely related to each other, but in the symptomatic hemisphere (on the downward slope of the autoregulation curve), flow velocity is directly proportional to the transit time and, therefore, to its inverse, perfusion reserve. The use of absolute mean flow velocity values on TCD should be further explored as a simple way of assessing "perfusion" in acute ischemic stroke.

Reducing the risk of carotid surgery: a 7-year audit of the role of monitoring and quality control assessment

BACKGROUND AND PURPOSE

The current risk of stroke after carotid endarterectomy may be worse than reported in the international trials. Because studies have suggested that most operative strokes follow surgeon error, the aim of the current study was to audit the impact of introducing a strategy of perioperative monitoring and quality control assessment on outcome.

METHODS

A total of 500 patients underwent carotid endarterectomy with intraoperative transcranial Doppler scan monitoring, completion angioscopy, and 3 hours of postoperative transcranial Doppler scan monitoring. The last of these guided selective dextran therapy in patients with high rates of postoperative embolization, which in previous series has been shown to be highly predictive of progression to thromboembolic stroke.

RESULTS

Intimal flaps were repaired in 3% of patients and luminal thrombus removed in 4% of patients. The rate of intraoperative stroke was 0.2%. A total of 313 patients had more than one embolus detected postoperatively (96% within 2 hours of flow restoration), but only 22 patients had sustained embolization requiring dextran. Embolization ceased in all but one patient receiving dextran, although the dose had to be increased in seven patients (36%). One patient was unable to receive adequate dextran therapy because of severe cardiac failure. Overall, the 30-day death/stroke rate was 2.2%, no patient had a perioperative stroke because of carotid thrombosis, and the rate of ipsilateral embolic stroke was 0.8%. Most complications resulted from cardiac pathology or intracranial hemorrhage.

CONCLUSIONS

A program of monitoring and quality control assessment has been associated with a 60% decrease in the operative risk in comparison with that observed before implementation of the protocol.

A policy of quality control assessment helps to reduce the risk of intraoperative stroke during carotid endarterectomy

OBJECTIVES

A pilot study in our unit suggested that a combination of transcranial Doppler (TCD) plus completion angioscopy reduced incidence of intra-operative stroke (i.e. patients recovering from anaesthesia with a new deficit) during carotid endarterectomy (CEA). The aim of the current study was to see whether routine implementation of this policy was both feasible and associated with a continued reduction in the rate of intraoperative stroke (IOS).

MATERIALS AND METHODS

Prospective study in 252 consecutive patients undergoing carotid endarterectomy between March 1995 and December 1996.

RESULTS

Continuous TCD monitoring was possible in 229 patients (91%), while 238 patients (94%) underwent angioscopic examination. Overall, angioscopy identified an intimal flap requiring correction in six patients (2.5%), whilst intraluminal thrombus was removed in a further six patients (2.5%). No patient in this series recovered from anaesthesia with an IOS, but the rate of postoperative stroke was 2.8%.

CONCLUSIONS

Our policy of TCD plus angioscopy has continued to contribute towards a sustained reduction in the risk of IOS following CEA, but requires access to reliable equipment and technical support. However, a policy of intraoperative quality control assessment may not necessarily alter the rate of postoperative stroke.

Incidence, timing, and causes of cerebral ischemia during carotid endarterectomy with regional anesthesia

PURPOSE

Controversy exists regarding the best technique to identify cerebral ischemia during carotid endarterectomy (CEA). Regional anesthesia allows continuous evaluation of neurologic function and therefore can help determine the incidence, timing, and causes of cerebral ischemia.

METHODS

The timing and clinical manifestations of any neurologic event during CEA and as long as 30 days afterward was determined by review of operative reports, hospital charts, and outpatient records of consecutive patients who underwent CEA under regional anesthesia over a 68-month period.

RESULTS

Two hundred patients underwent CEA; indications were asymptomatic stenosis > 60% in 25%, transient ischemic attack with stenosis > 50% in 52%, and prior stroke with stenosis > 50% in 23%. Eight patients (4%) were converted to general anesthesia for non-ischemic reasons. Of the remaining 192 patients, 183 (95.5%) underwent the procedure with regional anesthesia and no shunt, 2% had cerebral ischemia and underwent shunt placement, and 2.5% had cerebral ischemia, were converted to general anesthesia, and underwent shunt placement. Cerebral ischemia developed in nine patients after carotid cross-clamping, manifested by loss of consciousness in four, confusion in two, dysarthria and confusion in one, and decreased contralateral motor strength in two. Immediate cerebral ischemia developed in four of the nine patients within 1 minute of cross-damping; all four underwent shunt placement. In five of the nine patients, cerebral ischemia occurred between 20 and 30 minutes after cross-clamping; all occurred during relative intraoperative hypotension (average reduction of 35 mm Hg in the systolic pressure). All awake patients in whom ischemic symptoms developed immediately regained and maintained normal neurologic function with shunt placement. Five of 26 patients (19%) with contralateral occlusion required a shunt; none had postoperative ischemia. The mean carotid cross-clamp time was 27 minutes. Postoperative (30 day) complications included a 0.5% stroke rate, a 0.5% rate of postoperative transient ischemic attack, a 0.5% rate of worsening of preexisting acute stroke, and a 0.5% rate of myocardial infarction (no deaths). Of the nine patients who had intraoperative ischemic changes, none had a postoperative neurologic deficit; the three patients who had postoperative neurologic changes had no intraoperative ischemic symptoms.

CONCLUSIONS

CEA with regional anesthesia allows continuous neurologic monitoring and can be performed safely even when contralateral occlusion coexists; intraoperative shunting for ischemia is necessary in 4.5% of all cases and in 19% of patients with contralateral occlusion. Intraoperative ischemia was flow-related in our patients; it occurred early from ipsilateral carotid clamping and late from reduced collateral flow as a result of hypotension. Monitoring should be continued throughout cross-clamping to identify late cerebral ischemia. Postoperative cerebral ischemia is not associated with intraoperative ischemia, if corrected.

Prevention of postoperative thrombotic stroke after carotid endarterectomy: the role of transcranial Doppler ultrasound

PURPOSE

To determine the incidence of particulate embolization after carotid endarterectomy (CEA), the effect of dextran-40 infusion in patients with sustained postoperative embolization, and the impact of transcranial Doppler (TCD) monitoring plus adjuvant dextran therapy on the rate of postoperative carotid thrombosis.

METHODS

Prospective study in 100 patients who underwent CEA with 6-hour postoperative monitoring using a TCD that was modified to allow automatic, intermittent recording from the ipsilateral middle cerebral artery waveform (10 minute sample every 30 minutes). An incremental dextran-40 infusion was commenced if 25 or more emboli were detected in any 10-minute period.

RESULTS

Overall, 48% of patients had one or more emboli detected in the postoperative period, particularly in the first 2 hours. However, sustained embolization that required Dextran therapy developed in only five patients. In each case, the rate of embolization rapidly diminished.

CONCLUSIONS

A small proportion of patients have sustained embolization after CEA, which in previous studies has been shown to be highly predictive of thrombotic stroke. Intervention with dextran reduced and subsequently stopped all the emboli in those in whom it was used and contributed to a 0% perioperative morbidity and mortality rate in this series.

Carotid Endarterectomy and Intraoperative Emboli Detection: Correlation of Clinical, Transcranial Doppler, and Magnetic Resonance Findings

Carotid endarterectomy (CEA) performed with continuous transcranial Doppler (TCD) monitoring provides a unique opportunity to determine the number of cerebral microemboli and to relate their occurrence to the surgical technique. The purpose of this study was to assess in CEA the impact of cerebral microembolism on clinical outcome and brain morphology. We also evaluated the influence of the audible TCD signal on the surgeon and his/her technique. In a prospective series of 301 endarterectomies, brain function was monitored with electroencephalography and TCD ultrasonography. Preoperative and intraoperative risk factors were entered in a logistic regression analysis program to assess their correlation with cerebral outcome. To evaluate the impact of cerebral microembolism on brain morphology, we compared preoperative and postoperative magnetic resonance (MR) images of the brain in a subgroup of 40 patients. Microemboli (>10) observed during the dissection phase of the operation were related to intraoperative (P < 0.002) and postoperative (P < 0.02) cerebral complications. Microemboli that occurred during shunting were also related to intraoperative complications (P < 0.007). Finally, the phenomenon of >10 microemboli during dissection was significantly (P < 0.005) related to new hyperintensive lesions on the postoperative T(2)-weighted MR images. The presence of microembolism (>10) during the dissection phase of carotid endarterectomy shows a statistically significant relationship with perioperative cerebral complications and with new ischemic lesions on MR images of the brain. Microembolism during shunting is also related to intraoperative complications. Surgeons can be guided by the audio Doppler and emboli signals by changing their technique, which may result in a decline of the intraoperative stroke rate. (ECHOCARDIOGRAPHY, Volume 13, September 1996)

Computerised tomography findings as a risk factor in carotid endarterectomy: early and late results

OBJECTIVES

To evaluate whether preoperative CT evidence of brain infarction is associated with an increased risk of early and late stroke and death in patients undergoing CEA.

DESIGN

Retrospective clinical study.

MATERIALS AND METHODS

We evaluated 844 CT scanning records from 893 patients undergoing CEA from 1986-1994: 43% (367) CT positive for cerebral infarction and 57% (477) negative. Univariate and multivariate analysis was performed for risk factors and preoperative symptoms in patients with positive and negative CT scans, and Kaplan Meier survival curves for late events.

RESULTS

A positive CT was significantly more frequent in males vs. females (p < 0.0001; O.R. 2.52; C.I. 1.73-3.73), diabetics vs. non-diabetics (p = 0.03; O.R. 1.52; C.I. 1.03-2.26), symptomatics vs. asymptomatics (p < 0.001; O.R. 2; C.I. 1.93-3.53) and contralateral occlusion vs. patency (p < 0.001; O.R. 2; C.I. 1.30-3.10). The perioperative disabling stroke/ death rate was higher in patients with a positive CT (p = 0.002; O.R. 6.27; C.I. 1.73-34.20); in asymptomatic patients this difference was striking (5 patients vs. O, p = 0.0002). Multiple logistic regression analysis for risk factors, CT findings, symptoms preceding surgery, and congruity of brain infarction confirmed a significantly higher incidence of perioperative stroke/death rate (p = 0.003; O.R. 6.37; C.I. 5.12-7.63) and early and late stroke (p = 0.02; O.R. 1.95; C.I. 1.38-2.53) and death (p = 0.0005; O.R. 2.38; C.I. 1.89-2.88) in patients with brain lesions. After 7 years, the survival rate (p = 0.0009) and stroke-free interval (p = 0.003) were lower in patients with a positive CT. After 5 years, in asymptomatic patients the survival rate (p = 0.003) and stroke-free interval (p = 0.01) were lower in the positive CT group.

CONCLUSIONS

A positive CT finding, regardless of congruity of the lesion, should be regarded as an indicator of an increased risk of stroke and death in patients scheduled for carotid surgery, especially in those with asymptomatic stenosis.

Influence of patient-related variables on the outcome of carotid endarterectomy

PURPOSE

Variability in outcome after carotid endarterectomy is well recognized. This study examines the importance of patient-related factors in determining outcome.

METHODS

Four hundred and sixty consecutive patients undergoing carotid endarterectomy for symptomatic severe (60% to 99%) internal carotid stenosis performed by one vascular surgeon have been studied prospectively. Patients were followed-up at 3, 6, 9, and 12 months and then yearly. Pre-, intra-, and perioperative details and follow-up information were entered on a database.

RESULTS

Multiple logistic regression identified a number of factors significantly associated with death and stroke. A history of crescendo transient ischemic attacks (TIAs) (p = 0.003, p = 0.0002) and being female (p = 0.03, p = 0.0001) were associated with both perioperative death and stroke within 30 days of operation, respectively. Deaths between 1 and 36 months were associated with ischemic heart disease (p = 0.03) and diabetes (p = 0.04), whereas stroke was associated with small internal carotid diameter (p = 0.02). The importance of symptoms at presentation on outcome was further emphasized by life-table analysis. In 98% of patients with amaurosis fugax, only 67% of those with crescendo TIAs were alive at 18 months (p < 0.01). The survival of patients with amaurosis was significantly better than those with TIAs (p < 0.01), transient stroke (p < 0.01), and progressive, stroke (p < 0.05). Similarly, postoperative stroke was significantly more common for patients with crescendo TIAs than those with amaurosis (p < 0.01), established stroke (p < 0.05), and TIA (p < 0.05). Transient stroke was associated with a poor outcome, with only 66% of patients being alive in 36 months and 14% having suffered a stroke (p < 0.05 compared with established stroke).

CONCLUSION

Presenting symptoms significantly predict outcome after carotid endarterectomy. This should be considered both in patient selection and comparison of patient series.

The significance of microemboli detection by means of transcranial Doppler ultrasonography monitoring in carotid endarterectomy

PURPOSE

Carotid endarterectomy (CEA) performed with continuous transcranial Doppler monitoring provides a unique opportunity to determine the number of cerebral microemboli and to relate their occurrence to the surgical technique. The purpose of this study was to assess in CEA the impact of cerebral microembolism on clinical outcome and brain architecture. We also evaluated the influence of the audible transcranial Doppler signal on the surgeon and his or her technique.

METHODS

In a prospective series of 301 patients, CEA was monitored with electroencephalography and transcranial Doppler ultrasonography of the ipsilateral middle cerebral artery. Preoperative and intraoperative risk factors were entered in a logistic regression analysis program to assess their correlation with cerebral outcome. To evaluate the impact of cerebral microembolism on brain architecture, we compared preoperative and postoperative computed tomography scans or magnetic resonance images of the brain in two subgroups of 58 and 40 patients, respectively.

RESULTS

Seven (2.3%) patients had intraoperative transient ischemic symptoms, three (1%) had intraoperative strokes, 1 (0.3%) had transient ischemic symptoms after operation, and 10 (3.3%) had postoperative strokes. Four (1.3%) patients died. Microemboli (> 10) noticed during dissection were related to both intraoperative (p < 0.002) and postoperative (p < 0.02) cerebral complications. Microemboli that occurred during shunting were also related to intraoperative complications (p < 0.007). Microembolism never resulted in new morphologic changes on postoperative computed tomography scans. On the contrary, the phenomenon of more than 10 microemboli during dissection was significantly (p < 0.005) related to new hyperintense lesions on postoperative T2-weighted magnetic resonance images.

CONCLUSIONS

During CEA the presence of microembolism (> 10 microemboli) during dissection shows a statistically significant relationship with perioperative cerebral complications and with new ischemic lesions on magnetic resonance images of the brain. Moreover, microembolism during shunting is also related to intraoperative complications. Surgeons can be guided by the audio Doppler and emboli signals by changing their technique. This change may result in a decrease of microembolism and consequently in a decline of the intraoperative stroke rate.

Prevalence of impaired cerebrovascular reserve in patients with symptomatic carotid artery disease

Cerebrovascular reserve (CVR) was studied in 104 consecutive patients with symptomatic carotid territory disease and ipsilateral internal carotid artery stenosis. Overall, 30 of 104 patients (29 per cent) had impaired CVR. The frequency of CVR impairment increased with the severity of internal carotid artery stenosis: impairment was present in none of 11 patients with stenosis of less than 50 per cent, four of 24 with stenosis of 50-69 per cent, 14 of 41 with stenosis of 70-89 per cent and 12 of 28 with stenosis of 90-99 per cent. Patients presenting with a stroke were significantly more likely to have impaired CVR than those with transient ischaemic attacks and/or amaurosis fugax (odds ratio 3.7 (95 per cent confidence interval (c.i.) 1.5-9.0)), as were those with a residual neurological deficit (odds ratio 4.3 (95 per cent c.i. 1.6-11.5)) and evidence of infarction from computed tomography (odds ratio 3.8 (95 per cent c.i. 1.6-9.4)).

Serial imaging of the carotid bifurcation and cerebrovascular reserve after carotid endarterectomy

A radioisotopic method of quantifying mean cerebral transit time was used to assess the immediate effects of carotid endarterectomy on cerebrovascular reserve (CVR) in 69 patients. In addition, serial postoperative data were acquired on CVR, clinical status and non-invasive imaging of the internal carotid arteries in 56 patients over a period of 6-48 (median 24) months. Twenty-one patients (30 per cent) had preoperative evidence of impaired CVR in the symptomatic hemisphere. Within 4 days of surgery, however, reserve had returned to normal in 17 of the 21 patients. During follow-up, four of the 56 patients developed recurrent stenosis (> 50 per cent) or occlusion of the artery operated on but only two of these had impairment of CVR and none was symptomatic. Three patients suffered recurrent transient ischaemic attacks (TIAs) but none had recurrent internal carotid artery disease or impaired CVR. One patient suffered a TIA in the territory of the non-operated artery during follow-up in association with disease progression and CVR impairment. However, the TIA preceded recognition of either of these changes. Twelve other patients had (or developed) stenosis (> 50 per cent) in the non-operated artery during follow-up but none was symptomatic or developed impairment of CVR. Although assessment of CVR provided useful information on the frequency of haemodynamic compromise before carotid endarterectomy and on the natural history of disease progression, neither serial assessment of reserve nor non-invasive imaging of the carotid bifurcation influenced clinical practice during follow-up.

Carotid endarterectomy with transcranial Doppler and electroencephalographic monitoring. A prospective study in 130 operations

BACKGROUND AND PURPOSE

We report the results of combined recording of hemodynamic and thromboembolic phenomena during carotid endarterectomy by means of computerized electroencephalography as well as transcranial Doppler ultrasonography. The study focuses on the additional value of transcranial Doppler to detect ischemia during surgery.

METHODS

Combined monitoring was performed in 130 consecutive operations, using standard anesthesiological, surgical, and neurophysiological procedures.

RESULTS

A reduction of > or = 70% of blood flow velocities in the middle cerebral artery during cross-clamping was measured in 16 patients. In seven of these cases there were no severe electroencephalographic changes and a shunt was not used, but one of the patients developed a subcortical infarct with slight disability. In 55 patients, 75 episodes of embolization were detected by transcranial Doppler. In one of these, with massive embolization after release of the clamp, an intraoperative stroke occurred without changes on cranial computerized tomography or neurological disability on follow-up. In the other 54 patients, intraoperative embolization did not cause clinical or neuroradiological symptoms. Electroencephalographic changes occurred in only two of the 75 episodes. In addition to the two nondisabling strokes during surgery (1.5%), six strokes occurred within 5 days of operation, including one hemorrhage. There was no significant relation between contralateral carotid occlusion and stroke (p = 0.6).

CONCLUSIONS

During carotid endarterectomy, transcranial Doppler immediately provides information about thromboembolism and hemodynamic changes that are not detected by electroencephalography alone. Acoustic feedback from the transcranial Doppler monitoring unit has a direct influence on the surgical technique. Transcranial Doppler ultrasound may be a useful tool in the study and prevention of intraoperative stroke.

Immediate effects of carotid clamp release on middle cerebral artery blood flow velocity during carotid endarterectomy

Transcranial Doppler ultrasound was employed in 44 patients undergoing carotid endarterectomy in order to identify factors associated with the greatest increases in middle cerebral artery blood flow velocity (MCAV) immediately after carotid clamp release and restoration of flow. Previous reports have suggested that such increases might reflect post-ischemic hyperaemia. Overall, the median increase in MCAV on restoration of flow [58 cm/s (95% Cl 44-68)] was greater than the median decrease in MCAV after initial carotid clamping [36 cm/s (95% Cl 26-43), p < 0.0001]. However, the immediate increase in MCAV tended to be transient and, in one patient, may have represented a hyperaemic response to a haemodynamically induced intraoperative neurological deficit. There was no association between the magnitude of MCAV increase and clinical presentation, degree of carotid stenosis, computed tomography scan findings, type of anaesthesia, nor the presence or absence of impaired cerebral vascular reserve, nor the occurrence of intraoperative air embolisation. The greatest increases in MCAV on clamp release were observed in patients with the greatest decreases in MCAV at clamping and those with the lowest internal carotid artery stump pressures. The most likely explanation for the findings is that they represent a transient hyperaemic response to carotid occlusion. It remains unclear whether this phenomenon is mediated by some degree of ischaemic injury or simply by the effect of a sudden surge of blood through low resistance arterioles that have dilated in order to maintain the collateral circulation during carotid clamping.

Monitoring and cerebral protection during carotid endarterectomy

Two recently published multicentre trials have confirmed the overall benefit of carotid endarterectomy in symptomatic patients with severe carotid artery disease. The key to improving further the long-term advantages of carotid endarterectomy, however, remains the continued reduction of the initial operative risk. While the principal responsibility for this continues to be borne by the surgeon, specifically in reducing technical error, the time is perhaps approaching when he or she might also be able to apply some of the recent advances in cerebrovascular research to reduce operative morbidity still further in the future. This article summarizes the aetiology and pathophysiology of operation-related neurological deficits and reviews current approaches towards intraoperative monitoring, cerebral protection and assessment of quality control.

Transcranial Doppler monitoring during carotid endarterectomy

Transcranial Doppler monitoring of the middle cerebral artery blood flow velocity was used as an adjunct to routine methods of cerebral monitoring in a prospective study of 30 consecutive patients undergoing carotid endarterectomy to investigate whether transcranial Doppler monitoring provided information influencing operative technique. Application of carotid clamps caused a significant fall in middle cerebral artery velocity and there was a linear relationship between middle cerebral artery velocity and internal carotid artery stump pressure. Assuming a stump pressure of less than 50 mmHg to be an indication for shunting, this would correspond to a systolic middle cerebral artery velocity of less than 42 cm/s and a mean velocity of less than 30 cm/s. Transcranial Doppler monitoring immediately identified problems with shunt function and demonstrated a higher frequency of intraoperative embolization than had been anticipated, particularly after shunt insertion and final restoration of flow. With revision of operative technique this phenomenon is rarely encountered now. Two of the 30 patients exhibited a minor neurological deficit on recovery of consciousness, and transcranial Doppler monitoring was able to identify the probable underlying cause in both cases. Unnecessary and potentially hazardous re-exploration was avoided. In the absence of transcranial Doppler monitoring the neurological deficit in one of the patients might have been much worse.

The role of imaging in the management of cerebral and ocular ischaemia

The last decade has seen several major technological advances in vascular neuroradiology, the most clinically significant of which have been the facility to image the brain and the extracranial carotid bifurcation noninvasively with accuracy and safety. Another major advance has been unequivocal evidence from formal statistical overviews that antiplatelet therapy, particularly aspirin, reduces the risk of serious vascular events by about 25%. These advances have changed clinical practice such that most patients presenting with symptoms suggestive of cerebral ischaemia should now have cranial CT to exclude intracerebral hemorrhage, not only because the causes and prognosis of cerebral ischaemia differ from those of intracerebral hemorrhage, but because many patients with cerebral ischaemia should be considered for antiplatelet therapy. Besides the use of long term antiplatelet therapy and control of vascular risk factors, other acute treatment options are limited with the possible exception of anticoagulation, thrombolysis, cytoprotective agents and carotid endarterectomy. If, as seems likely, the current clinical trials show that carotid endarterectomy plus medical therapy improve upon the stroke-free survival of patients treated medically, at least in symptomatic patients with severe stenosis, the number of carotid endarterectomies performed will increase considerably because carotid bifurcation disease is the most common cause of cerebral and ocular ischemic events. It will then be even more important to be able to obtain accurate anatomical and physiological information about the extracranial and intracranial circulations with utmost safety. Duplex ultrasound is currently the noninvasive screening method of choice for carotid bifurcation disease because it is available, relatively cheap, and reasonably accurate. It not only images the vessel lumen and degree of stenosis, but also the morphology of the vessel wall and associated plaque, the relevance of which is still uncertain in the pathogenesis of cerebral and ocular ischaemia. A major limitation of duplex sonography is that it cannot reliably distinguish tight stenosis from occlusion and it does not image the proximal or distal carotid circulation. The aim of newer techniques will be to distinguish tight extracranial carotid stenosis from occlusion and to provide anatomical, physiological and pathological information about the intracranial circulation and ischemic lesions (in view of potential for thrombolytic therapy of major intracranial vessel occlusion) with safety and reproducible accuracy.