Surgical Management of Extracranial Carotid Artery Disease

Surgical Magnification

We perform the operation with 3.5× loupe–magnified technique. Microscopic repair of the internal carotid artery (ICA), which we have also tried, enables a primary repair that is unquestionably finer than that resulting from a loupe-magnified technique ^, but which, in our experience, does not alter the overall patient outcome or incidence of restenosis or acute occlusion. In the ongoing effort to reduce morbidity, we have instead adopted universal patch grafting with collagen-impregnated Dacron (Hemashield graft), which has essentially eliminated the problem of acute postoperative thrombosis or rapid restenosis. In our opinion, the graft procedure is more easily and expeditiously accomplished with 3.5× magnification rather than the microscope. There is no doubt that the suture lines are not as fine with this method; in our routine practice, however, the added lumen diameter with patch angioplasty renders the microscopic technique unnecessary.

Anesthetic Technique

General anesthesia and local anesthesia are both in common use for CEA. We routinely use general anesthesia with both full-channel electroencephalographic and concurrent somatosensory-evoked potential (SSEP) monitoring. Proponents of local anesthesia cite the advantages of patient response to questioning as a superior method of assessing the need for intraoperative shunting while minimizing anesthetic risks, reducing postoperative morbidity, and shortening length of stay. The patient has local anesthesia with light sedation, which allows the patient to perform a simple task with the contralateral hand during cross-clamping. The disadvantages include risk of contamination and patient movement during the procedure, along with the increased psychological stress of remaining awake. A recent review comparing our technique with that of an institutional vascular surgeon using local anesthesia showed a decreased incidence of electroencephalographic changes and intraoperative shunting with local anesthesia. However, there was no difference in stroke rate, complications, length of stay, or overall outcome. ^,

We prefer general anesthesia for a number of reasons, not the least of which is the controlled environment. Additionally, all commonly used inhalational anesthetic agents and intravenous barbiturates significantly reduce the cerebral metabolic rate of oxygen consumption, providing a theoretical protective effect in the setting of cerebral ischemia. We keep our patients normocapnic. Although there has been much interest in arterial levels of carbon dioxide, nonphysiologic hypercapnia and hypocapnia provide no cerebral protection. Gross and colleagues found that there was a 40% decrease in electroencephalographic (EEG) changes with cross-clamping in those patients receiving either one or two units of 6% hetastarch (500 to 1000 mL). They had acceptable outcomes with a postoperative stroke and mortality rate of 1.3%. We have not adopted this technique because of our policy of shunting quickly without hesitation for any hint of an EEG or SSEP change. Finally, blood pressure is maintained at normotensive levels with a tolerance of as high as a 20% increase in systolic pressure. Although some surgeons prefer to induce hypertension at cross-clamping if there are EEG changes and then to shunt if no improvement is seen in the EEG recordings, we are not trying to avoid shunt use. Moreover, as mentioned, we have a policy to shunt immediately if any monitoring changes become evident.

Monitoring Techniques

Monitoring techniques can be divided into two categories: (1) tests of vascular integrity—such as stump pressure measurements, xenon regional cerebral blood flow studies, transcranial Doppler and, to a lesser extent, intraoperative oculoplethysmography, Doppler/duplex scanning, and angiography—and (2) tests of cerebral function, such as EEG, EEG derivatives, and/or SSEP monitoring. The newly described near-infrared spectroscopy technique bridges both categories. We use full-channel EEG and concurrent SSEP interpreted by a neurologist. After completion of the arteriotomy closure, an intraoperative Doppler examination is performed of the common carotid artery (CCA), the ICA, and the external carotid artery (ECA). ^,

Intraoperative Shunting

Generally speaking, there are three schools of thought about intraoperative shunting. Carotid surgeons shunt in every case, shunt when indicated by some form of intraoperative monitoring, or never place a shunt. In our institution, we perform monitor-dependent shunting based on EEG criteria. We use a custom commercial shunt of our own design (Loftus shunt, Integra Neurocare, Plainfield, NJ). In our experience, we shunt approximately 15% of CEAs. This increases to approximately 25% if the contralateral carotid is occluded. After the shunt is placed, the monitoring should return to baseline. If this does not occur, the shunt must be inspected for possible kinking, thrombosis, or misplacement. We always auscultate the shunt with a Doppler probe that confirms patency and shunt flow.

Proponents of universal shunting tout the benefits of the maximal degree of cerebral protection in every case while eliminating dependence on specialized intraoperative monitoring techniques. They assert that shunt placement is benign and allows extra time to ensure meticulous intimal dissection and arteriotomy repair.

Proponents of nonshunting believe that shunt placement is not benign. In one series, there was a higher stroke rate with shunting compared with nonshunting, indicating that embolization from shunt placement, especially by surgeons inexperienced in the procedure, is a real risk. Another documented concern is distal intimal damage leading to embolization or carotid artery dissection.

Many surgeons, in part because of the concerns previously discussed, choose not to shunt. There have been multiple series that have shown good surgical results with no shunts being used. These authors do not deny the existence of postoperative stroke, but they strongly believe that neurologic deficits from carotid artery surgery are invariably embolic rather than hemodynamic in nature and that intraoperative monitoring and/or shunt placement will not further reduce the already low morbidity in their series.

As discussed previously, we prefer to shunt when there are changes in the EEG and/or SSEP with cross-clamping. This policy has been well supported by several reports of large series of patients. ^, Of note, there are also several authors who normally practice selective shunting but advocate shunting all patients who have had recent strokes or reversible neurologic events (due to their belief that intraoperative monitoring is unreliable in the face of recent ischemic events). Although we understand their concerns, this has not been our practice.

Patch Graft Angioplasty

Almost all carotid surgeons perform patch angioplasty for recurrent carotid stenosis. Many will also use selective patching in cases where the internal carotid is small, where the arteriotomy may have extended far up the ICA, or in any similar case in which compromise of the lumen and a high risk of thrombosis are anticipated. We have taken this policy one step further and, for a number of years, have used a Hemashield patch graft primarily in all our patients. We have not encountered any restenosis or acute occlusions since using the patch universally. Several other synthetic grafts are available depending on surgeon preference, and of course autologous saphenous vein grafts can be used, but we prefer Hemashield, mindful of the real concern of central patch rupture with autologous vein, especially in women and patients with diabetes. Rupture risk can be reduced by harvesting the graft from a high femoral site rather than at the ankle, but at either location donor site discomfort remains an issue.

Heparinization

A single dose of intravenous heparin is given to the patient at some point before cross-clamping. This dose is between 2500 and 10,000 U of heparin, depending on the surgeon’s preference. There are no published reports to support one dose versus another; however, Poisik and colleagues recently reported their results using weight-based dosing of heparin at 85 U/kg. Although they did not see any statistically significant differences between fixed-dose heparinization (5000 U) and weight-based dosing, there were trends of decreased complications of hematoma formation and neuropsychometric testing differences. Some individuals reverse the heparinization with protamine after the operation. ^, We have not found any benefit in this practice. For those patients who come to the operating room on a continuous heparin drip, we continue the infusion until the arteriotomy closure is finished. With meticulous technique, bleeding in these cases has not been a problem.

Tacking Sutures

Tandem sutures to secure the distal intima in the ICA after plaque removal are considered a great advance by some and are deemed unnecessary by others. ^{, ,} The concern with tacking sutures is that they may narrow the lumen, but to us this risk seems small compared with the concern of intimal dissection from an unsecured intimal flap. Several authors ^{, ,} state that if the arteriotomy is carried far enough to see normal intima distal to the plaque, the tacking sutures are unnecessary. We strongly agree with an arteriotomy that extends past the plaque, but we are not always satisfied with how the intimal plaque tapers at the distal endpoint. In recent years, because of negative experiences with plaque that does not feather cleanly when pulled down from the distal ICA, we have adopted the use of fine scissors to “trim” the plaque cleanly in the ICA as it is removed. When this is done, tacking sutures are rarely necessary. We estimate that we now selectively place tacking sutures in the distal ICA in approximately 20% of cases.