Atheromatous disease at the origin of the internal carotid artery is a significant cause of strokes. Its recognition as a cause of neurological symptoms is important, for it may be amenable to surgical correction by carotid endarterectomy. The usual indication for carotid endarterectomy is a transient ischaemic attack, although it may be performed in patients with an evolving stroke, a completed stroke, or an asymptomatic tight carotid artery stenosis. Transient stroke-like episodes, lasting minutes to hours, have been recognized for more than a century, but it was Chiari in 1905 who first described the association between thrombosis at the carotid bifurcation and distal embolism in the internal carotid artery, this association being further elaborated by Gunning and colleagues in Oxford.

The first carotid endarterectomy was performed by de Bakey in 1953, although several other carotid bifurcation reconstructions were performed during that decade. Over the next 30 years carotid endarterectomy became widely practised for both symptomatic and asymptomatic stenoses of the internal carotid artery. However, there were marked discrepancies in the operative rates in different parts of the world, there being, for example, 20 times more operations per million of population performed in North America compared to Great Britain in the 1980s. This difference reflected the controversy concerning the value of the operation in stroke prevention, remembering that the operation itself is associated with a significant risk of death and stroke. However, much of this controversy has now been resolved, at least in the management of the patient with transient ischaemic attacks. The interim results of two large multicentre trials in Europe and North America showed a marked beneficial effect of operation in preventing stroke in patients with transient ischaemic attacks and a tight (>70 per cent) stenosis of the internal carotid artery.

Atheromatous plaques form at the carotid bifurcation, and especially in the carotid sinus. The lesion in the internal carotid artery is restricted in most instances to the origin of the artery and the carotid sinus, the artery distal to the lesion and the common carotid artery proximal to the lesion being relatively normal. The predilection for this area is undoubtedly due to the turbulence in flow created by the bifurcation and the dilated carotid sinus. The plaque may ulcerate, giving rise to thrombus formation in the bed of the ulcer or, if the plaque eventually produces a tight stenosis of the artery origin thrombus may form at the stenosed area of the plaque (Figs. 1 and 2) 287,288. If thrombus is dislodged it will pass downstream, either to the retinal arteries via the ophthalmic artery or to the cerebrum via the middle cerebral artery, giving rise to a transient ischaemic attack, a complete stroke, amaurosis fugax, or a retinal infarct. An embolus from the same site in the plaque will tend to end up distally in the same place, for the bloodstream in the internal carotid artery is not random. Such patients suffer repeated attacks of amaurosis fugax, for example. Haemorrhage into an atheromatous plaque is also frequent and this can lead to sudden narrowing of the lumen as well as to ulceration. The latter may be associated with discharge of cholesterol crystals into the lumen which embolize distally, as well as thrombus formation in the bed of the ulcer. As the stenosis becomes more severe there is an increased risk of thrombosis and occlusion of the internal carotid artery, with a subsequent stroke, although occlusion does not always result in a stroke. Alternatively, flow past the stenosis may become so poor as to produce transient ischaemic attacks on a purely haemodynamic basis. However, the bulk of carotid artery transient ischaemic attacks are embolic rather than haemodynamic. Such attacks in the carotid distribution may be due to emboli from proximal lesions located in sites other than the internal carotid artery, such as the heart. Three types of emboli can be detected in the retina (Fig. 3) 289: white emboli of cardiac origin, causing segmental blockage and retinal infarcation, small white fibrin-platelet plugs which pass through the retinal circulation within minutes, and small refractile plaques (atheromatous debris).

Symptoms of transient ischaemia in the carotid artery territory reflect its distribution to the eye and the anterior two-thirds of the brain. The most common symptoms are weakness, numbness, and clumsiness of the limbs, especially the arm, contralateral to the side of the lesion, or loss of vision on the side of the lesion (amaurosis fugax). The patient characteristically describes the loss of vision as being as if a blind had been pulled down (occasionally across) and as vision returns, usually in a few minutes, the blind retreats in the opposite direction. Dysphasia associated with a transient ischaemic attack is also common, especially if the left hemisphere is involved in a right-handed person. Amaurosis fugax usually lasts for only minutes while transient neurological defects last from minutes to hours.

Transient neurological deficits may be global rather than focal in nature and it is important to distinguish the two. Global symptoms include faintness, giddiness, vertigo, binocular visual loss, and syncope; these are due usually to a transient fall in blood pressure and hence cerebral perfusion. This seldom results in a focal deficit except where there is a tight carotid stenosis, as discussed earlier. The symptoms of vertebrobasilar ischaemia must also be recognized, but these are usually easily distinguished from transient ischaemic attacks in the carotid artery distribution.

Focal neurological deficits due to causes other than thromboembolism can present in a manner resembling a transient ischaemic attacks, but remembering that they exist will usually allow their exclusion (Table 1) 183.

Examination of the patient usually confirms that there is no residual neurological deficit. A bruit may be heard over the appropriate carotid bifurcation, but this is a relatively poor sign of an internal carotid artery stenosis, for an external carotid artery lesion may be the cause of a bruit. The absence of a bruit does not preclude a diagnosis of internal carotid artery disease as the cause of a transient ischaemic attack, as this is not uncommon in the presence of a very tight stenosis. Examination of the heart for evidence of valve disease as a possible source of emboli is also important. In patients who have had an attack of amaurosis fugax not long before being seen, examination of the fundi may be rewarding in that emboli may still be identified even though vision has recovered. If vision remains defective an embolus, usually refractile, can often be identified.

There have been major changes in the investigation of carotid distribution transient ischaemic attacks in recent years. The gold standard has always been carotid angiography with the addition of CT scanning or magnetic resonance imaging (MRI) of the brain in recent years. However, the risk of a stroke being precipitated by angiography is between 1 and 4 per cent when this is performed for extracerebrovascular disease, and this has provided enormous impetus in recent years for the development of non-invasive methods of determining the presence of significant carotid artery disease. Oculoplethysmography and oculopneumoplethysmography have served this role for several years, at least in determining which patients should have angiography (see Section 7.4) 37. However, the advent of the duplex scan, which combines real time B-mode scanning with Doppler sound spectral analysis, allows not only the carotid bifurcation to be imaged but also the degree of stenosis to be calculated. The addition of colour has further refined the technique: many vascular units perform a carotid endarterectomy on the basis of duplex scanning plus CT or MRI. In Oxford, for example, carotid endarterectomy is performed without angiography in 50 to 60 per cent of cases. Digital subtraction angiography has largely replaced conventional angiography but, in general, intra-arterial digital subtraction angiography is necessary to obtain satisfactory pictures, intravenous angiography having proved to be disappointing in that respect.

Transcranial Doppler analysis of blood velocity in the intracerebral arteries is a new approach to the evaluation of collateral cerebral blood flow, as is the development of probes for the duplex scanner which allow imaging of the intracerebral vessels. However, it is not yet possible to define their eventual role in diagnosis.

If there is any question of a cardiac focus for emboli an echocardiogram should be performed. Careful cardiac evaluation is always essential, however, for one-third of patients with carotid transient ischaemic attacks die of myocardial infarction within 5 years of undergoing carotid endarterectomy. Other investigations include all those that would be performed in any patient with atheromatous arterial disease, such as full blood count and blood lipid analysis.

Once a diagnosis of a transient ischaemic attack in the carotid distribution has been made or even suspected then the patient should be given aspirin, 300 mg/day, and referred for evaluation by either a neurologist or vascular surgeon specializing in this area. Duplex scanning of the carotid arteries will usually enable a diagnosis of a carotid lesion to be established or excluded. A CT scan or MRI of the brain is advisable not only to exclude any other intracerebral lesion but also to show any existing areas of infarction. If a tight stenosis (>70 per cent) is present carotid endarterectomy is indicated, provided the patient's general health, in particular their cardiac status, is sound (Table 2) 184. Age of itself is not a contraindication to carotid endarterectomy but the majority of patients will be between 65 and 75 years of age. If the stenosis is minimal, (0–29 per cent), surgery is not indicated: these patients have a worse outcome in terms of stroke than the non-surgical patients because of the perioperative mortality/stroke rate and the very low risk of a subsequent stroke associated with minimal lesions. These patients should be maintained on aspirin therapy. The appropriate treatment for moderate stenoses (30–70 per cent) is unclear at this time: both the North American and European trials are continuing to enter patients in this category. If not entered into trials carotid endarterectomy is probably best withheld at this time and the patients should be treated with aspirin. If transient ischaemic attacks continue to occur during treatment with aspirin, surgery or anticoagulation with warfarin or coumadin should be considered.

Whatever the management, any hypertension should be well controlled and hyperlipidaemia treated by diet or, if severe, by cholesterol lowering agents. All patients are encouraged to stop smoking.

A transient ischaemic attack is defined as an acute loss of focal cerebral or ocular function with symptoms lasting less than 24 h and which, after adequate investigation is presumed to be due to embolic or thrombotic vascular disease. This definition not only includes transient neurological defects but also transient episodes of blindness (amaurosis fugax). The 24-h time period is arbitrary, but has been accepted for many years and indeed the majority of attacks last from minutes to hours. Those due to an atheromatous plaque at the origin of the internal carotid artery (Fig. 1) 287, are by far the most common indication for carotid endarterectomy.

If an embolus lodges long enough in the retinal arteries, infarction of part or all of the retina may occur. This is regarded in the same light as amaurosis fugax with respect to indications for surgery.

A patient whose neurological deficit lasts longer than 24 h but makes a full recovery or who is left with a permanent but stable neurological deficit and who has a stenosis of the internal carotid artery appropriate to the neurological signs is also a candidate for carotid endarterectomy, especially if a good recovery has been made.

This is a controversial indication. Early experience of carotid endarterectomy in this group of patients was accompanied by a high perioperative mortality (40–60 per cent) and the procedure was abandoned in this situation. However, cautious evaluation of the value of carotid endarterectomy in carefully selected patients with an evolving stroke or a fluctuating neurological deficit, where urgent investigations have confirmed the presence of an appropriate lesion, is now taking place in several centres.

This is another controversial area. It has been an indication for carotid endarterectomy in a significant proportion of patients in the United States, but hardly at all in Great Britain. Retrospective and prospective studies of patients with a known asymptomatic stenosis of the internal carotid artery suggest that the risk of stroke in these patients is relatively low, and indeed less than the risk of myocardial infarction. Furthermore, when strokes do occur they are usually preceded by transient ischaemic attacks. The associated risk of death and stroke from the operation itself is, therefore, probably not outweighed by any benefit of subsequent stroke protection. If cardiac surgery is required, this is generally performed first, followed by the carotid surgery at a later date, rather than as a combined procedure, although again the indication for the carotid endarterectomy remains unresolved. There are a number of multicentre trials in progress comparing medical management plus carotid endarterectomy versus medical management alone in asymptomatic patients with a tight stenosis of the internal carotid artery, the first of which suggests, but not conclusively, that there may be a modest benefit of surgery. However, for the moment, until results of later trials are available, the controversy remains unresolved.

This is an extremely complicated clinical problem, in that there is no established method of determining whether a tight stenosis of the internal carotid artery, especially if bilateral, is responsible for transient neurological deficits by causing hypoperfusion rather than emboli. Certainly in some instances it must be, and the development of transcranial Doppler analysis and duplex scanning of the circle of Willis may in time allow patients with cerebral hypoperfusion, who would benefit from carotid endarterectomy, to be better defined. It is not likely to be a large number, based on the experience of the intracerebral–extracerebral reconstruction trial of several years ago.

Once a firm diagnosis is established an elective operation can be planned as soon as possible. The risks of a stroke before surgery after the patient has started taking aspirin are relatively small and the risks of perioperative complications are reduced by taking the time for a careful evaluation of the patient and in particular his or her cardiac status.

Although some surgeons perform the operation under local anaesthetic, this is uncomfortable for the patient over the 1 to 2 h required (longer if a patch is used to close the arteriotomy) and most surgeons prefer a general anaesthetic. It is essential that blood pressure control is well established before the patient comes to surgery and it needs to be carefully monitored via an intra-arterial line throughout the procedure.

Where the operation is performed under local anaesthesia it is possible for the patient to report any neurological deficit, provided they have not required so much sedation as to lose their awareness of events. In patients under general anaesthetic, electroencephalography is valuable in units where the appropriate facilities exist (simple three-channel EEG monitoring is of no value). More recently monitoring of middle cerebral artery velocity on the side of the operation with transcranial Doppler is being widely adopted (Fig. 4) 290. Whether this will prove an adequate measure of the maintenance of collateral flow remains to be seen.

The carotid bifurcation is exposed through an oblique incision along the anterior border of the sternomastoid (Fig. 5) 291. Having exposed the carotid bifurcation, arterial pressures are measured in the common carotid artery proximal to the lesion and in the internal carotid artery distal to the lesion to establish the presence of a pressure gradient, and then the common carotid artery is clamped and the internal carotid pressure recorded. This is the stump pressure and is a measure of the collateral flow from the other side and the vertebrobasilar system via the circle of Willis. If the pressure is below 50 to 60 mmHg, surgeons who shunt selectively would consider that an indication for a shunt. Blood flow in the internal carotid artery may also be measured, using either an electromagnetic flow meter or more recently an ultrasound probe (OpDop). This provides a baseline for comparison with a similar measurement made after reconstruction.

The patient is then heparinized and clamps are applied to the common, internal, and external carotid arteries. An arteriotomy starting in the common carotid artery below the lesion is carried through the lesion into the internal carotid artery, distal to the disease (Fig. 6) 292. If a shunt is to be used, it is inserted at this time. The two most widely used shunts are the Javid shunt and the Pruitt shunt (Fig. 7) 293. The latter is more easily inserted and can be moved more easily out of the way during the endarterectomy (Fig. 8) 294.

The endarterectomy requires identification of a plane between the plaque and the media: generally this lies somewhere between the outer one-third and inner two-thirds of the media. If the correct plane is identified the lesion comes away without difficulty. It is essential to obtain a tightly adherent distal flap in the internal carotid artery and if this is not the case it should be tacked down with some fine 7–0 prolene sutures (Fig. 9) 295.

Having completed the endarterectomy the arteriotomy can be closed directly with a continuous prolene suture (Fig. 10) 296 or with a vein (internal jugular vein or saphenous vein), polytetrafluoroethylene (Gore-Tex or Impra), or Dacron patch (Fig. 11) 297. In recent years many vascular surgeons have routinely closed most arteriotomies with a patch, as the recognition of a relatively high incidence of recurrent stenoses, usually at the distal end of the endarterectomy has become apparent. Occasionally if the plaque is localized to the origin of the internal carotid artery and the external carotid is free of disease, the bifurcation can be reconstructed by suturing the posterior walls of these two arteries together and then suturing the anterior walls together; in effect this reconstructs a higher bifurcation, but widens the artery at the end of the endarterectomy (Fig. 12) 298.

Following surgery the patient is kept in the recovery room until awake and then returned to the ward. As the vast majority of perioperative strokes occur in the first 24 h after operation neurological observations must be made every 30 min over the first 12 h and then every hour until 24 h. It is also important to maintain a stable blood pressure avoiding any periods of hypo- or hypertension. This should not generally be a problem if the patient has come to operation with well controlled blood pressure.

If any neurological deficit appears relevant to the side of operation, a rapid return to the operating room is indicated, with exploration of the neck and exposure of the carotid vessels. The endarterectomized site should be re-explored with removal of any thrombus; a careful inspection of the distal flap is mandatory. If thrombosis and occlusion of the internal carotid artery has occurred restoration of the cerebral flow may increase the likelihood of recovery or limit the extent of the neurological deficit. Although a CT scan of the brain and duplex scan of the operated side would be desirable, it is not practical to do this because of the loss in time in restoring flow in an occluded vessel that would result. Thrombus is found in less than 50 per cent of patients re-explored, and whether removing this will limit the evolving deficit is unknown. As any one surgeon's experience in this area is small it probably never will be, but it would instinctively seem an appropriate course to follow.

The major complications of this procedure are death, either from stroke or a myocardial infarction, or a stroke. A stroke occurring during the operation or generally within the first 24 h of surgery may be transient or more severe, leading to a permanent disability. The combined perioperative mortality and stroke rate as defined in the European and North American trials (death or neurological deficit resulting in a permanent disability and occurring during and within 30 days of surgery) should be less than 5 per cent (Table 5) 187. A neurological deficit may be due to thrombosis at the site of endarterectomy, with either occlusion or embolization, or an intracerebral haemorrhage.

As most patients are receiving aspirin significant haematomas are relatively common, but rarely need draining. Local nerve pareses are also not uncommon but are usually transient, recovering within weeks to months. Retraction is the usual cause of these local nerve pareses, which is often unavoidable, especially in the case of the hypoglossal nerve and a high carotid bifurcation. Retraction of the vagus may result in paresis of the superior or recurrent laryngeal nerve. In general the patient can be reassured that recovery will take place within a month or two.

The availability of duplex scanning after carotid endarterectomy has shown that the incidence of recurrent stenosis, commonly at the distal end of the endarterectomy, is much greater than hitherto believed, perhaps of the order of 15 to 30 per cent at 5 years. However, the development of symptoms related to the recurrent stenosis is very uncommon. Restenosis may be due to myointimal hypertrophy or atheroma. In the absence of symptoms no treatment other than continued daily aspirin is necessary. The development of transient ischaemic attacks is an indication for reoperation. Whether the incidence of restenosis is less with patch closures is not known at this time.

As carotid endarterectomy for stenosis of the internal carotid artery is an operation carried out in the expectation that it will prevent a subsequent stroke, it is perhaps surprising that it has taken nearly 30 years to establish that at least in the case of a tight symptomatic stenosis (>70 per cent) it does indeed do so. The simultaneous interim analyses of the European Carotid Surgery Trial (ECST) and the North American Symptomatic Carotid Endarterectomy Trial (NASCET) produced very similar results.

In the ECST 2518 patients who had had a carotid-territory non-disabling ischaemic stroke, transient ischaemic attacks, or retinal infarct and a stenotic lesion in the relevant internal carotid artery were randomized either to best medical management or best medical management plus carotid endarterectomy. All patients received antiplatelet therapy. For the 778 patients with a severe stenosis (70–99 per cent) the total risk of surgical death, surgical stroke, ipsilateral ischaemic stroke, or any other stroke was 12.3 per cent, compared to 21.9 per cent in the non-surgical group (2p <0.01) (Fig. 13) 299. In contrast, in the minimal stenosis group (0–29 per cent) there was little risk of ipsilateral ischaemic stroke in the non-surgical group so that any possible benefits of surgery were outweighed by the perioperative risks of surgery. It was not clear in the patients with a moderate stenosis (30–69 per cent) whether surgery was beneficial or not, and entry into the trial of such patients is continuing.

In the NASCET 659 patients with a severe stenosis (70–99 per cent) who had had a hemispheric or retinal transient ischaemic attack or a non-disabling stroke were randomized to best medical management or best medical management plus carotid endarterectomy. All patients received antiplatelet therapy. The cumulative risk of stroke at 2 years was 26 per cent in the non-surgical group but only 9 per cent in the surgical group (p<0.001) (Fig. 14) 300. In the moderate stenosis group (30–69 per cent), just as in ECST, there was no clear answer and entry to the trial is continuing. Patients with minimal stenoses were not entered into this trial.

Thus, these two trials convincingly demonstrate the beneficial effect of carotid endarterectomy in a patient with a tight symptomatic internal carotid artery. Furthermore the ECST shows that surgery is not indicated in patients with a symptomatic minimal stenosis. Studies of patients with a symptomatic moderate stenosis are continuing. It is not known whether the operation has a place in patients with a tight asymptomatic stenosis in reducing the risk of a subsequent stroke, but trials in progress at the present time should provide an answer to that question in due course.

Barnett HJM, Haines SJ. Carotid endarterectomy for asymptomatic carotid stenosis. N Engl J Med 1993; 328: 276–8.

Chambers BR, Norris JW. Outcome in patients with asymptomatic neck bruits. N Engl J Med 1986; 315: 860–5.

Chiari H. Uber des Verhalten des Teilungswinkels der Carotis communis bei der endarteritis chronica deformans. Verl Dtsch Ges Pathol 1905; 9: 326–30.

DeBakey ME. Successful carotid endarterectomy for cerebrovascular insufficiency. JAMA 1975; 233: 1083–5.

Eastcott HHG, Pickering GW, Robb CG. Reconstruction of internal carotid artery in a patient with intermittent attacks of hemiplegia. Lancet 1954; ii: 994–6.

European Carotid Surgery Trialist's Collaborative Group (1991). MRC European Carotid Surgery Trial: interim results for symptomatic patients with severe (70–99 per cent) or with mild (0–29 per cent) carotid stenosis. Lancet 1991; 337: 1235–43.

Grotta JC. Current medical and surgical therapy for cerebrosvascular disease. N Engl J Med 1987; 317: 1505–16.

Gunning AJ, Pickering GW, Robb-Smith AH, Ross Russell R. Mural thrombosis of the internal carotid artery and subsequent embolism. Q J Med 1964; 33: 155–95.

Hankey GJ, Warlow CP, Sellar RJ. Cerebral angiographic risk in mild cerebrovascular disease. Stroke 1990; 21: 209–22.

Hertzer NR, Bevin E, O'Hara PJ, Krajewski LP. A prospective study of vein patch angioplasty during carotid endarterectomy. Ann Surg 1987; 206: 628–35.

Heyman A, et al. Risk of stroke in asymptomatic persons with cervical arterial bruits. N Engl J Med 1980; 302: 838–42.

Hobson RW, et al. Efficacy of carotid endarterectomy for asymptomatic carotid stenosis. N Engl J Med 1993; 328: 221–7.

Meissmer I, Wiebers DO, Whisnant JP, O'Fallon WM. The natural history of asymptomatic carotid artery occlusive lesions. JAMA 1987; 258: 2704–7.

Moore WS, Ziomek S, Quinones-Baldrich WJ, Machleder HI, Busuttil RW, Baker JD. Can clinical evaluation and non-invasive testing substitute for arteriography in the evaluation of carotid artery disease? Ann Surg 1988; 208: 91–4.

Murie JA, Morris PJ. Carotid endarterectomy in Great Britain and Ireland. Br J Surg 1986; 76: 867–70.

North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J Med 325: 445–53.

Sterpetti AV, et al. Surg Gynecol Obstet 1989; 168: 217–23.

Thompson JE, Talkinton CM. Carotid endarterectomy. Adv Surg 1992; 26: 99–131.

Warlow C, Morris PJ. Transient Ischaemic Attacks. New York: Marcel Dekker, 1982.

Winslow CM, Solomon DH, Chassin MR, Kosecoff J, Merrick NJ, Brook RH. The appropriateness of carotid endarterectomy. N Engl J Med 1988; 318: 721–7.