Carotid My

Carotid endarterectomy: Preoperative evaluation; surgical technique; and complications

INTRODUCTION — Treatment aimed at carotid atherosclerotic lesions may be beneficial in both symptomatic and asymptomatic patients. This topic will review the preoperative evaluation, surgical technique, and complications of carotid endarterectomy (CEA). Nonsurgical approaches to carotid revascularization, such as angioplasty and stenting, are discussed separately. (See "Carotid angioplasty and stenting"). INDICATIONS — Carotid endarterectomy (CEA) is most commonly performed for symptomatic or asymptomatic high-grade (>60 or 70 percent) internal carotid artery stenosis, where the effectiveness of CEA has been established by large randomized clinical trials. The indications for CEA in patients with significant carotid atherosclerosis are discussed separately. (See "Carotid endarterectomy in symptomatic patients" and see "Carotid endarterectomy in asymptomatic patients"). Some patients have varying degrees of carotid disease bilaterally. No randomized clinical trials have evaluated the effectiveness of bilateral CEA for such patients. However, bilateral carotid occlusive disease appears to increase the risk for complications during and after unilateral CEA [1-3]. In one study of 700 patients undergoing CEA, for example, 15.4 percent had contralateral disease [2]. The combined death and stroke rates in these patients were almost twice as high as that in matched patients with unilateral disease (5.6 versus 2.4 percent). The impact of severe contralateral carotid artery stenosis or occlusion on the benefit and risk of unilateral CEA in patients with symptomatic and asymptomatic disease is discussed separately in the appropriate topic reviews. (See "Carotid endarterectomy in symptomatic patients", section on Contralateral carotid stenosis or occlusion, and see "Carotid endarterectomy in asymptomatic patients", section on Contralateral carotid stenosis or occlusion). Concomitant coronary artery bypass grafting — Patients with significant coronary artery disease are at high risk for a cardiac event during CEA. As a result, coronary artery bypass grafting (CABG) is often considered in conjunction with CEA. However, there have been no randomized studies examining the use of CEA in patients in patients having CABG. In addition, it is not clear if CABG should be combined with the CEA or should be staged (ie, performed before or after CEA). This issue is discussed separately. (See "Coronary artery bypass grafting in patients with cerebrovascular disease", section on Prophylactic carotid intervention). Endarterectomy prior to general surgery — Neurologic complications are second only to heart failure as a cause of morbidity and mortality following cardiac surgery. (See "Neurologic complications of cardiac surgery"). The incidence of stroke appears to be lower following general (nonvascular) surgical procedures than following cardiac surgery, with a reported incidence in patients undergoing general anesthesia of less than 0.5 percent [4-6]. The risk may be slightly higher (1 percent) in asymptomatic patients with a carotid bruit who undergo general surgery [7]. There have been no randomized studies examining the prophylactic use of carotid endarterectomy (CEA) in patients with carotid stenosis prior to general surgery. A retrospective review suggests that prophylactic CEA is probably not warranted in most patients with asymptomatic carotid disease [8]. This study was a chart review of 284 patients who had undergone nonvascular surgery requiring general anesthesia and had preoperative carotid ultrasound. While a previous history of stroke or TIA, a carotid bruit, or both were present in 250 patients, all were considered to have asymptomatic carotid stenosis [9]. Ten of 284 patients (3.5 percent) had perioperative ischemic strokes within 30 days of the index procedure, and 8 of 224 (3.6 percent) with >50 percent carotid stenosis had an

ipsilateral perioperative stroke (bilateral lesions were present in three patients). While this stroke risk exceeds that of the general population and of patients with carotid bruits, the increased risk appears to be insufficient to mandate prophylactic CEA for asymptomatic carotid stenosis in the general surgical population. PREOPERATIVE EVALUATION Carotid duplex ultrasound — Patients suspected of having carotid atherosclerosis typically undergo carotid duplex ultrasound (CDUS) screening. Other useful noninvasive methods for assessing stenosis of the internal carotid artery include magnetic resonance angiography (MRA), contrast enhanced magnetic resonance angiography (CEMRA), and computed tomography angiography (CTA). The utility of these noninvasive methods and cerebral angiography is discussed in detail separately. (See "Evaluation of carotid artery stenosis"). Role of other imaging studies — In patients with a hemodynamically significant atherosclerotic lesion on CDUS, it remains controversial if further imaging with conventional cerebral arteriography, MRA, or CTA is warranted for evaluation of arterial anatomy and verification of the degree of stenosis. The major concern with arteriography is that trials such as ACAS have found a 1.6 percent incidence of stroke associated with routine arteriography, although the risk was less in other reports [10]. Some physicians feel that this added risk outweighs the benefit of obtaining more anatomic detail. In addition, tolerance of carotid clamping cannot be reliably assessed via arteriography [11]. On the other hand, many surgeons feel that the sensitivity of noninvasive studies is insufficient to conclude that CEA is indicated; they typically quote lower stroke rates associated with arteriography or consider MRA and CTA as alternative imaging procedures. In support of this point of view is the fact that there are no uniform criteria for judging how to measure the degree of stenosis. As a result, comparison of results obtained by different ultrasound criteria may be difficult. Furthermore, in a study of 569 patients undergoing both angiography and noninvasive vascular imaging studies at an academic medical center and community hospital, the degree of carotid stenosis was misclassified in 28 percent by Doppler ultrasound alone, 18 percent by MRA alone, and 8 percent with both imaging studies combined [12]. A second community based study also raised doubts about the practice of performing surgery on the basis of carotid ultrasound alone [13]. In this study, 130 patients had duplex ultrasound performed in a community based center and were referred to an academic center for CEA. Angiography was performed at the referral hospital prior to considering surgery. Almost one-half of patients referred on the basis of 50 to 69 percent stenosis and 22 percent of patients with 70 to 99 percent stenosis by Doppler ultrasound did not have significant stenosis on angiography. The positive predictive value of the noninvasive study for identifying appropriate asymptomatic patients for carotid intervention (angiographic stenosis ≥60 percent) was 59 percent, with a false-positive rate of 41 percent. A study of 81 patients incorporated the costs of testing, surgery, and stroke [14]. The authors concluded that the combination of CDUS and MRA, followed by carotid arteriography if the results were disparate, was associated with the lowest morbidity and mortality and was cost-effective. Neurologists and vascular societies are currently addressing this issue; it is anticipated that formulae utilizing parameters such as peak systolic and end diastolic velocities will soon become standardized. Another possible role of arteriography is to assess for intracranial atherosclerotic disease, which is present in 20 to 50 percent of patients with stenosis of the extracranial internal carotid artery. In an analysis of a subset of patients from NASCET, the relative risk of stroke associated with intracranial atherosclerotic disease in medically treated patients varied from 1.3 to 1.8 with extracranial stenosis less than 50 and 85 to 99 percent, respectively [15]. CEA reduced this risk, suggesting that detection of intracranial atherosclerotic disease, particularly in those with moderate extracranial carotid stenosis, may help stratify patients into a group that is likely to achieve benefit from CEA.

Evidence of atherosclerosis elsewhere — A thorough vascular history and physical examination is an essential component of the evaluation of a patient being considered for CEA. A search should be made for other evidence of atherosclerosis, including abdominal aortic aneurysm and peripheral vascular disease. (See "Screening for abdominal aortic aneurysm" and see "Noninvasive diagnosis of peripheral arterial disease"). A thorough cardiac evaluation is also important since patients undergoing CEA are most likely to die from coronary heart disease. This evaluation may be performed with exercise stress testing, dobutamine echocardiography, or dipyridamole imaging, or, when warranted, coronary catheterization [16]. (See "Estimation of cardiac risk prior to noncardiac surgery", section on Noninvasive testing). The preoperative evaluation should also include computed tomography (CT) or magnetic resonance imaging (MRI) of the brain in the symptomatic patient. These imaging studies can assess the degree of cerebral infarction, if any, and exclude other disorders that might be responsible for symptoms such as subdural hematoma or a tumor. Perioperative risk factors — Identifying risk factors for morbidity and mortality associated with CEA is important in order to avoid surgery in patients who may face unacceptably high surgical risk. However, the data are conflicting regarding risk factors for CEA. One or more of the following characteristics have been associated with an increased risk of poor outcome (stroke, myocardial infarction, or death) at 30 days after CEA in some [17-24] but not all [25-29] studies: •

Age 80 or older

•

Severe heart disease

•

Severe pulmonary dysfunction

•

Renal insufficiency or failure

•

Stroke as the indication for endarterectomy

•

Anatomical issues including limited surgical access, prior cervical irradiation, prior ipsilateral CEA, and contralateral carotid occlusion

One of the largest studies was a retrospective series of 3061 patients from a single center [18]. Patients defined as high risk by the presence of severe coronary artery disease, chronic obstructive pulmonary disease, or renal failure prior to CEA had a significantly elevated risk of stroke, myocardial infarction (MI), or death compared with the low-risk group (7.4 versus 2.9 percent respectively). This study has been criticized because the majority of patients defined as high risk had CEA combined with cardiac surgery, a setting that may be associated with higher death and complication rates. (See "Coronary artery bypass grafting in patients with cerebrovascular disease"). Nevertheless, another single center study of 1370 consecutive CEAs found that patients with two or more risk factors had significantly higher mortality compared with those who had no risk factors [19]. As already noted, a number of reports have NOT confirmed the association of these proposed risk factors with poor outcome after CEA [25,26,29,30]. As an example, a single center series of 2236 consecutive isolated CEA operations reported a 30-day stroke and death rate of 1.4 percent [30]. No single clinical variable was significantly associated with perioperative complications. Another study that reviewed 788 consecutive CEAs reported that stroke and death rates for patients identified as high risk and low risk were low and similar for both groups (1.3 and 1.1 percent, respectively) [25]. Advances in perioperative management have led at least some surgeons to conclude that the proportion of patients with unacceptable risk is extremely small and continuing to shrink [25,31]. Others have challenged the concept of a high-risk group of patients for CEA [29]. Modifications in surgical practice, especially the declining use of routine preoperative contrast angiography, may be driving a reduction in perioperative complication rates for CEA [30]. In the series of 2236 CEAs discussed above, the investigators noted that morbidity and mortality in the last five years of the study compared with the previous five years had declined by 36 percent [30]. Patients deemed unfit for general anesthesia can undergo CEA with regional anesthesia, a technique that has shown equally good perioperative outcomes after CEA in patients with and

without risk factors such as advanced age, diabetes, coronary artery disease, and contralateral internal carotid occlusion [32]. Intracranial aneurysm — Ipsilateral intracranial aneurysms that are distal to a cervical internal carotid artery stenosis may be susceptible to sudden hemodynamic changes with CEA. These hemodynamic changes could lead in turn to aneurysmal rupture. On the other hand, surgical clipping of an aneurysm distal to a severe internal carotid stenosis may increase the risk of ischemic stroke. Unfortunately, data for this situation are too sparse to allow firm conclusions as to which problem should be treated first. However, caution is advised if CEA is performed in this setting, especially if the unruptured ipsilateral aneurysm is 7 mm or larger in diameter or if there is a history of SAH from another aneurysm. (See "Unruptured intracranial aneurysms"). CONCURRENT MEDICAL THERAPY Aspirin — Antiplatelet therapy with aspirin reduces the risk of stroke of any cause in patients undergoing carotid endarterectomy (CEA) [33,34]. In addition, lower-dose aspirin (81 to 325 mg daily) is more effective than higher-dose aspirin (650 to 1300 mg daily). •

In a randomized, controlled trial involving 232 patients, aspirin 75 mg daily or placebo treatment was started preoperatively and continued for six months [35]. Patients assigned to aspirin had significantly fewer strokes at one month and six months than those assigned to placebo. However, this study was likely underpowered [36].

•

The ACE trial randomly assigned 2849 patients scheduled for endarterectomy to aspirin at doses of 81, 325, 650, or 1300 mg daily [37]. Aspirin was started before surgery and continued for three months. At three-month follow-up, the primary end point (stroke, MI, vascular death) was significantly reduced in the lower-dose aspirin group compared with the higher-dose group (6.2 versus 8.4 percent).

Consensus guidelines from the American Academy of Neurology (AAN) and the American College of Chest Physicians (ACCP) recommend aspirin for for symptomatic and asymptomatic patients undergoing CEA [36,38]. We recommend starting aspirin (81 to 325 mg daily) prior to CEA and continuing indefinitely in the absence of contraindications. (See "Antiplatelet therapy for secondary prevention of stroke"). Statins — Statin use on hospital admission for CEA in symptomatic patients may be associated with improved outcomes. In a retrospective observational study of 3360 CEAs, statin use by patients with symptomatic carotid stenosis was associated with reduced in-hospital mortality and combined inhospital ischemic stroke or death (adjusted odds ratio 0.25 [95% CI 0.07-0.90] and 0.55 [ 95% CI 0.32-0.95], respectively), but in-hospital cardiac outcomes were nonsignificantly reduced [39]. In contrast, statin use by patients with asymptomatic carotid stenosis was not associated with significantly different outcomes. Similar results were reported in another retrospective study involving 1566 patients with symptomatic and asymptomatic disease who received statins for at least one week before CEA [40]. These findings require confirmation in randomized clinical trials. Evidence is also emerging that statins may be of benefit in the perioperative period, and that this benefit might be lost if statins are discontinued. This issue is discussed elsewhere. (See "Perioperative medication management", section on Hypolipidemic agents). SURGICAL TECHNIQUE — Carotid endarterectomy (CEA) is performed through a neck incision either bordering the sternocleidomastoid muscle, or more esthetically, with a transverse incision in a skin crease at the level of carotid bulb. For the latter incision, preoperative imaging and palpation of the neck will guide the surgeon to the optimal location. The underlying platysma muscle and subcutaneous tissues are dissected and the carotid artery is carefully identified, from the common carotid to well beyond the bifurcation of its internal and external branches. After proximal and distal control of the artery is obtained, the patient is given systemic heparin. The internal, common, and external arteries are then clamped sequentially. A longitudinal arteriotomy is perform at the level of the bifurcation and extended proximally and distally. Some surgeons use a cerebral shunt which is inserted at this time. Alternatively, EEG monitoring, stump pressures, or physical examination in the awake patient can be used as a guide for selective shunt placement (see "Intraoperative monitoring" below). The carotid plaque, which is consistently found at the carotid bifurcation and the origin of the internal carotid artery, is then freed and removed through a dissection plane developed between the media and intima. Great care is taken to create a smoothly tapered transition between the

endarterectomized portion of the artery and its normal distal extent. This maneuver avoids intimal flaps which might lead to arterial dissection after flow is reestablished. After meticulous inspection of the endarterectomized surface to remove any plaque or debris, attention is directed at repair. Some surgeons choose to repair primarily, while others patch the artery with saphenous vein or prosthetic material such as Dacron or polytetrafluoroethylene (PTFE). (See "Patch angioplasty versus primary closure" below). Just prior to completion of the arterial closure, the internal carotid artery is unclamped distally and flushed free of debris in a retrograde fashion. This vessel is then reclamped and the common and external arteries are opened. In this fashion, debris travels out the external carotid artery prior to reinstituting antegrade flow. Thrombin soaked gel foam may be used over the suture line to slow any oozing of blood. Once hemostasis is achieved, a Jackson Pratt drain is placed to minimize neck hematomas and the platysma and skin are closed. Eversion endarterectomy is a variant of the above procedure. It involves a circumferential transverse incision made in the carotid artery. The artery is then everted or turned inside out to create the exposure otherwise afforded by a traditional vertical arteriotomy. This technique may lower the incidence of restenosis while also being faster to perform [41]. Upon awakening from anesthesia, a neurologic assessment is performed and repeated every hour during recovery. Manipulation of the carotid bulb during carotid endarterectomy not infrequently results in hemodynamic instability intraoperatively and in the early postoperative period. Adequate cerebral perfusion pressure should be maintained during periods of hemodynamic instability to avoid low cerebral blood flow and cerebral ischemia. Pressors may be added to avoid low flow states. Patients with intracerebral small vessel disease are at especially increased risk for an ischemic event should low flow occur. Local versus general anesthesia — The choice of local (LA) versus general anesthesia (GA) for CEA is generally decided by individual surgeon preference and patient characteristics. A systematic review of evidence found seven randomized trials and 41 nonrandomized trials comparing CEA under LA compared with GA [42]. Meta-analysis of the randomized studies showed that, within 30 days of operation: •

LA was associated with a nonsignificant trend towards decreased mortality.

•

LA was associated with a significant reduction in local postoperative hemorrhage.

•

There was no difference between LA and GA in the risk of postoperative stroke.

Meta-analysis of the nonrandomized studies showed that LA was associated with significant reductions in the odds of death (35 studies), stroke (31 studies), stroke or death (26 studies), myocardial infarction (22 studies), and pulmonary complications (7 studies) within 30 days of operation [42]. The methodologic quality of many of the nonrandomized trials was questionable. The reviewers concluded that there was insufficient evidence from randomized trials to draw reliable conclusions regarding the use of LA or GA for CEA [42]. Results from a large ongoing randomized trial (GALA) are awaited to shed more light on this question. Patch angioplasty versus primary closure — As noted above, some surgeons choose to repair primarily, while others patch the artery with saphenous vein or prosthetic material such as Dacron or polytetrafluoroethylene (PTFE). The trials that have been performed suggest two benefits from use of a patch: a marked reduction in the frequency of ≥50 percent restenosis and a lower rate of ipsilateral stroke [1,43-45]. In a randomized trial of 399 procedures, for example, the respective values of patch versus nonpatch were 2 to 9 versus 34 percent for restenosis and 1 versus 5 percent for ipsilateral stroke [45]. These conclusions were confirmed by a systematic review of patch angioplasty versus primary closure during CEA that was published in 2004 [46]. The review identified seven eligible randomized controlled trials involving 1281 operations that compared primary closure with routine patch closure, and eight trials involving 1480 operations that compared different patch materials; two trials compared both. Many of the trials were limited by significant methodological flaws; most were small, and none could be analyzed on a true intention-to-treat basis because of losses to follow-up. The following observations were made: •

Patch angioplasty was associated with a significant risk reduction for stroke of any type, ipsilateral stroke, and both perioperative and long-term stroke or death.

•

Patching was associated with reduced risk for perioperative arterial occlusion and decreased long-term recurrent stenosis.

•

Seven trials that compared different patch types showed no difference in the risk for stroke, death, or recurrent arterial stenosis during perioperative or one-year follow-up time points. The only high-quality trial that compared use of Dacron and PTFE in CEA found that Dacron was associated with increased risk for perioperative stroke and increased risk for both perioperative and late recurrent stenosis, compared with PTFE [47,48]. Nonetheless, the reviewers concluded that more data are needed to establish differences between various patch materials.

Intraoperative monitoring — The benefit derived from CEA is partially dependent on a low intraoperative morbidity. Although 80 to 85 percent of patients tolerate clamping of the carotid artery without symptoms, assessment of collateral circulation occurring contralaterally via the circle of Willis is needed in all patients. Establishing which patients will be intolerant may be performed in a number of ways: •

Patients who undergo regional anesthesia may be followed clinically throughout the procedure by monitoring speech and upper extremity function.

•

For surgeons preferring general anesthesia because of an intolerant awake patient, EEG monitoring may be performed. Neurologists can then ascertain cerebral ischemia intraoperatively (theta and delta waves or disorganized rhythms) and prompt the need for selective shunting (see "Selective shunting" below) [49].

•

Alternatively, "stump pressures" may be obtained after clamping of the proximal common and external carotid arteries. Pressures greater than 30 to 50 mmHg generated from retrograde arterial flow via the circle of Willis down the ipsilateral carotid artery implies adequate collateralization and is associated with strokes rates below 0.5 percent [50]. Lower stump pressures are an indication for shunt placement.

Critics of this technique caution that pressures are only obtained after initial clamping and therefore represent a "snapshot" in time. In addition, the above criteria should be used with caution in patients who have suffered prior ipsilateral strokes since there is a poor correlation between adequate perfusion pressures and outcomes in this setting. Selective shunting — Those patients demonstrating evidence of cerebral ischemia by any of these monitoring techniques are selectively shunted. A temporary prosthetic shunt is placed beyond the proximal and distal extent of the arteriotomy into the common and internal carotid artery. Blood flow is therefore bypassed through the shunt, providing continuous cerebral perfusion during the procedure. Neurologic reassessment is then performed. Many surgeons feel that the advantages of shunting warrants its routine use and avoids the need for intraoperative neurologic monitoring. There are, however, rare complications that can result from shunt placement: •

An intimal flap may arise during shunt insertion, resulting in arterial dissection

•

Plaque emboli may be dislodged during vessel manipulation or air embolism may occur

Studies targeted at defining the best approach in this regard have been equivocal and the selective versus routine use of carotid shunts is largely a matter of surgeon preference [51-53]. Bilateral CEA — When the extent of contralateral carotid disease is significant enough to warrant bilateral CEA, it should be approached as a staged procedure. The risk of respiratory compromise secondary to neck hematomas or laryngeal nerve injury plus the frequent difficulty with blood pressure control after bilateral manipulation of the carotid sinus makes the combined approach prohibitive. In addition, the effect of bilateral cerebral ischemia, even though temporary, is not clear despite advances in neurologic monitoring and shunt placement. Thus, most physicians advocate a delay of at least one to two weeks between subsequent procedures. (See "Indications" above). ACUTE COMPLICATIONS — While a number of controlled trials have highlighted the patient population most likely to benefit from CEA, this operation is not without risk [54]. The perioperative mortality associated with CEA ranges from <0.5 to 3 percent. As mentioned above, mortality rates are higher when this procedure is performed at non-tertiary care centers [55-57]. The majority of

deaths are due to cardiac events, placing emphasis on the appropriate cardiac workup and perioperative management in these patients. The American Heart Association consensus statement states that indications for surgery are proven in symptomatic patients in whom morbidity and mortality rates associated with CEA are less than 6 percent and in the asymptomatic patient when the rates are less than 3 percent [58,59]. As a result, surgeons are encouraged to keep accurate records of their individual stroke rates to ensure that these standards are upheld. Low patient volume (<3 CEAs performed every two years) and a greater number of years since licensure of the surgeon are associated with worse outcomes following CEA [60]. Postoperative stroke — Stroke is the second most common cause for mortality due to CEA. Stroke rates associated with CEA range from less than 0.25 to more than 3 percent, with the experience of the surgeon again being important. Multiple factors can contribute to postoperative stroke in patients who have undergone CEA. These include: •

Plaque emboli

•

Platelet aggregates

•

Improper flushing

•

Poor cerebral protection

•

Relative hypotension

However, neurologic changes in the patient after CEA must be considered graft thrombosis until proven otherwise and technical error must be ruled out. Evaluation and treatment — Evaluation of the CEA patient with new neurologic deficits varies among surgeons. Some advocate recovery room or intraoperative ultrasound to assess potential thrombosis, while others immediately return to the operating room and open the anastomosis for direct visual inspection. The optimal time to heparinize the symptomatic postoperative patient is also controversial [61]. Some surgeons heparinize immediately upon suspicion of the diagnosis, while others first obtain a head CT to rule out hemorrhagic stroke. Head CT performed immediately after an embolic event is frequently normal; follow-up CT in a few days may reveal injury. Percutaneous transluminal carotid angioplasty with stenting is an alternative therapy for elective treatment of carotid artery disease. (See "Carotid angioplasty and stenting"). It may also be effective for perioperative stroke after CEA, particularly when the cause is a flow-limiting dissection. As an example, one study evaluated 13 patients with major or minor neurologic complications after CEA who underwent emergency carotid angiography and stent placement, the angiographic success was 100 percent and 11 patients had complete resolution of neurologic symptoms [62]. In contrast, only one of five patients undergoing surgical reexploration had neurologic recovery Intraarterial thrombolytic therapy may become a valid treatment option in patients with a postoperative stroke proven by arteriography to be due to thrombosis. The rationale for the administration of tissue-type plasminogen activator (alteplase) in this setting is based upon trials in acute stroke in which benefit has been demonstrated if therapy is initiated within 4.5 hours in highly selected patients. (See "Fibrinolytic (thrombolytic) therapy for acute ischemic stroke"). However, the incidence of intracranial hemorrhage in patients treated with thrombolytic therapy for acute stroke has been in the range of 6 percent [63]. Furthermore, it is not known if the results obtained from the intravenous systemic administration of alteplase can be extrapolated to localized intraarterial therapy. Intraarterial thrombolysis for patients with postoperative stroke has only been described in case reports and retrospective studies and is , currently experimental. There are as yet no controlled trials and its use is therefore not justified. Nevertheless, some neurologists advocate searching for distal thrombosis via arteriogram and, if found, proceed with intraarterial thrombolytic therapy. Hyperperfusion syndrome — The cerebral hyperperfusion syndrome is probably the cause of most postoperative intracerebral hemorrhages and seizures in the first two weeks after CEA. The clinical manifestations of hyperperfusion occur in only a small percentage of patients after carotid revascularization (from less than 1 to as high as 3 percent in various reports) [64-67]. The mechanism of hyperperfusion is related to changes that occur in the ischemic or low-flow carotid vascular bed. To maintain sufficient cerebral blood flow, small vessels compensate with chronic maximal dilatation.

After surgical correction of the carotid stenosis, blood flow is restored to a normal or elevated perfusion pressure within the previously hypoperfused hemisphere. The dilated vessels are unable to vasoconstrict sufficiently to protect the capillary bed because of disrupted autoregulation of cerebral blood flow. Breakthrough perfusion pressure then causes edema and hemorrhage, which in turn results in the clinical manifestations. The hyperperfusion syndrome is characterized by the following clinical features: •

Headache ipsilateral to the revascularized internal carotid, typically improved in upright posture, may herald the syndrome in the first week after endarterectomy.

•

Focal motor seizures are common, sometimes with postictal Todd's paralysis mimicking post endarterectomy stroke from carotid thrombosis.

•

Intracerebral hemorrhage is the most feared complication, occurring in about 0.6 percent of patients after CEA, usually within two weeks of surgery [68].

Neuroimaging studies including head CT and MRI with T2 or FLAIR sequences typically show cerebral edema, petechial hemorrhages, or frank intracerebral hemorrhage. Post recanalization ipsilateral cerebral blood flow (CBF) is markedly increased compared with preprocedure flow [69]. Ipsilateral CBF after recanalization may be two to three times that of homologous regions in the contralateral hemisphere [70]. However, hyperperfusion syndrome may develop in the presence of only moderate (20 to 44 percent) increases in ipsilateral cerebral blood flow, as measured by perfusion magnetic resonance imaging (PWI), and in the absence of increases in middle cerebral artery flow velocity, as measured by transcranial Doppler (TCD) [71]. The hyperperfusion syndrome appears to be more likely with revascularization of a high-grade (80 percent or greater stenosis) carotid lesion, and it may be more likely when CEA is performed after recent cerebral infarction [72]. Reduced CBF or cerebral vasoreactivity prior to CEA may also be a risk factor for postoperative hyperperfusion [73]. Transcranial Doppler techniques have been used to monitor flow velocities of the middle cerebral artery in order to predict the occurrence of hyperperfusion syndrome [74-76], but the utility of these methods for this indication is not clearly established. Treatment — The best treatment of cerebral hyperperfusion may be prevention. Strict control of postoperative hypertension is paramount. Systolic blood pressure must be maintained at or below 150 mmHg. Aggressive measures including intravenous labetalol, nitroprusside, and nitroglycerin may be necessary to achieve this goal. Therapy should begin at the time of restoration of internal carotid flow and be maintained vigilantly during the hospital stay and for the first 10 to 14 days postprocedure. Fortunately, most postoperative blood pressure lability resolves in the first 24 hours. (See "Labile blood pressure" below). Seizures related to hyperperfusion are usually successfully treated with standard antiepileptic drugs such as phenytoin [77]. Intracerebral hemorrhage from hyperperfusion is often devastating. Hypertension must be strictly controlled and anticoagulant and antithrombotic drugs should be discontinued. For patients on aspirin, platelet transfusions may be useful to reverse the antiplatelet effect. Nerve injury — A number of nerve injuries can complicate CEA. •

The vagus nerve, which lies posterolaterally in the carotid sheath, is identified during dissection of the carotid from the internal jugular vein and is at risk for injury.

•

The recurrent laryngeal nerve branches of the vagus are distal to the area of carotid artery dissection; injury to this nerve may result in unilateral vocal cord paralysis. The occasionally nonrecurrent nerve places this branch at even higher risk.

•

The facial nerve exits the stylomastoid foramen and courses along the inferior portion of the ear. The most common branch affected during CEA is the marginal mandibular branch, which may be damaged during improper or prolonged retraction. The resulting paresis of the lateral aspect of the obicularis oris muscle may be exacerbated during bedside examination with a revealing asymmetric smile.

•

The glossopharyngeal nerve is more cephalad than the extent of the typical neck dissection during CEA. A branch of this nerve, the nerve of Hering, has great clinical significance since it innervates the carotid sinus and is responsible for the bradycardic and hypotensive responses that may be seen with manipulation of this structure. Some

surgeons anesthetize the carotid sinus with lidocaine to avoid this complication, which is typically manifested intraoperatively. •

Damage to the hypoglossal nerve, also identified routinely during a CEA, is the complication with which most are familiar. Injury to this nerve may result from inadvertent transection or retraction; it results in tongue deviation to the side of injury.

•

Branches of the trigeminal nerve may be transected during dissection, resulting in sensory loss in the area of distribution

•

The ansa hypoglossus nerve innervates the strap muscles of the neck and is typically seen coursing along the carotid sheath. Unlike the other nerves, this nerve may be sacrificed with impunity.

•

The superior laryngeal nerve is rarely injured during CEA. The internal branch supplies sensation to the larynx, while the external branch innervates the cricopharyngeal muscle. Changes in voice quality may result from nerve injury.

The vast majority of cranial nerve injuries associated with CEA resolve over the first few months after surgery, and the risk of permanent cranial nerve deficit is very low. Among the 1739 patients who had CEA in the European Carotid Surgery Trial (ECST), immediate motor cranial nerve injury occurred in 5.1 percent, all ipsilateral to the side of the operation [78]. By hospital discharge, the cranial nerve injury rate had declined to 3.7 percent, and the involved cranial nerves included hypoglossal (n=27), marginal mandibular (n=17), recurrent laryngeal (n=17), accessory (n=1), and Horner syndrome (n=3). The rate of persistent cranial nerve injury at four-month follow-up declined to 0.5 percent. Duration of surgery longer than two hours was the only independent risk factor for cranial nerve injury. Bleeding — Postoperative bleeding, resulting in neck hematoma, occasionally occurs after CEA. It is typically self-limited and can be treated conservatively. However, there must be a low threshold to reexplore the neck and search for a surgically correctable source of bleeding. The incidence of neck hematoma is higher in those patients who remain on anticoagulation postoperatively. Bleeding is also more likely in patients who have undergone a combined CEA and CABG, generally because of a coagulopathy. As noted above, when these procedures are combined, the CEA is performed first and the neck is packed and left open; CABG is then performed and, when completed, the neck wound is closed. In this way, adequate hemostasis is ensured in these coagulopathic patients. Infection — Wound infection rarely occurs following CEA. Some proponents of primary repair site reports of an increased incidence of infection when a prosthetic patch has been used. Parotitis — Parotitis is an unusual complication after CEA that results from manipulation of the parotid gland during the procedure. For this reason, most surgeons use this landmark as the cephalad extent of their dissection. Labile blood pressure — Adequate blood pressure control in the postoperative period is of paramount importance. Hypertension will increase the likelihood of neck hematoma or suture line disruption, while relative hypotension compared to the patient's baseline value will increase the likelihood of inadequate cerebral perfusion. Hypotension is more likely to result in cerebral ischemia and neurologic deficits in those patients who also have intracerebral small vessel disease. Because blood pressure lability is common in the first 12 to 24 hours postoperatively, it is standard care for CEA patients to be placed in a monitored setting with an arterial line. These patients often require dopamine or nitroglycerin drips to maintain adequate hemodynamics. RESTENOSIS — Restenosis of the carotid artery after CEA was reported in up to 20 percent of patients in earlier studies [79] although lower values (2.6 to 10 percent at five years) have been reported in subsequent studies [1,80]. Pathology — The pathology of the restenotic lesion is related to the time of presentation after initial surgery [81]. •

"Early" restenosis is that which occurs within two to three years after CEA. Patients with early restenosis frequently have highly cellular and minimally ulcerated intimal hyperplasia, similar to that which occurs after angioplasty or with stent placement. As a result, there is a low likelihood of symptomatic embolization.

•

"Late" restenosis occurs more than two to three years after CEA and generally results from progression of atherosclerotic disease. It is frequently associated with irregular plaques that may serve as an embolic source.

Risk factors — Patients at risk for restenosis include those below age 65, smokers, and women, probably due to the smaller size of their carotid arteries [81,82]. Elevated creatinine has been associated with the development of early restenosis, and elevated serum cholesterol with late restenosis [80]. The cellular features of the atheroma at the time of CEA may predict the occurrence of restenosis. In a prospective study of 500 patients that examined target lesion atherosclerotic plaque composition from specimens obtained at carotid endarterectomy, both low macrophage infiltration and a small or absent lipid core were associated with an increased risk of restenosis at one year [83]. In another study of 150 patients, an abundance of smooth muscle cells and a scarcity of macrophages were seen in the primary lesion of those who had neointima development six months after surgery; in those who did not develop neointima, the lesions were rich in lymphocytes and macrophages [84]. Patch angioplasty appears to be associated with a decreased risk of long-term recurrent stenosis compared with primary closure (see "Patch angioplasty versus primary closure" above) [46]. Lipid lowering drugs may be protective for both early and late restenosis [80], although this finding requires confirmation. Reoperation — Once the diagnosis of restenosis has been made, a decision has to be made about the need for reoperation. This decision is not one to be made lightly since reoperative CEA may be associated with a significant incidence of complications, although the evidence is retrospective and conflicting. The following studies illustrate the range of perioperative complications reported for redo CEA: •

An earlier series described 69 patients (48 percent men, 66 percent symptomatic) who had 82 reoperative CEA procedures [85]. Nine patients had two reoperative CEAs and two patients had three reoperative CEAs for either bilateral recurrence or a second recurrence on the same side. The average time to presentation with recurrent carotid stenosis was 6.5 years. The incidence of postoperative stroke (4.8 percent), transient ischemic attack (7.3 percent), and hematomas (7.3 percent) were nearly twice as high as reported for a first CEA [85].

•

In a series of 401 reoperative CEAs in 10 states in the United States, the 30-day combined risk of stroke or death from mid 1998 to mid 1999 was 5.7 percent [86].

•

A subsequent series described 145 patients (56 percent men, 36 percent symptomatic) who had 153 reoperative CEA procedures [87]. The incidence of perioperative stroke (1.9 percent) and death (O) was very low. While the average time from primary to reoperative CEA was 6.1 years in this series, 41 percent of the cohort were patients with early (<2 years) restenosis, which is typically due to intimal hyperplasia and carries a low risk of symptomatic disease compared with late restenosis (see "Pathology" above).

An additional major concern is that there are no controlled studies establishing the efficacy of reoperative CEA in patients with restenosis. The presumed benefits of surgery in this group of patients are merely an extrapolation of the results of trials performed on patients at initial presentation. PREDICTORS OF LONG TERM OUTCOME — The predictors of long-term outcome after surgical therapy for atherosclerotic occlusive disease of the carotids were evaluated in one study of 1982 patients who underwent surgery at a single center and were followed for ≥25 years. Predictors of mortality included [88]: •

Age, with a relative risk of 1.51 for each 10 year increase in age

•

Male sex, relative risk 1.58

•

Diabetes mellitus, relative risk 1.48

•

Systemic hypertension, relative risk 1.31

•

Cigarette smoking, relative risk 1.13

Predictors of recurrence of symptoms or progression of disease were established by analysis of a subset of 886 patients who underwent one or more postoperative angiogram; these included [88]: •

Total cholesterol, with a relative risk of 1.13 for each 50 mg/dL (1.3 mmol/L) increase

•

Systemic hypertension, relative risk 1.42

•

Cigarette smoking, relative risk 1.47