Carotid endarterectomy (CEA) has been established as the gold standard for the treatment of both symptomatic and asymptomatic high-grade (≥ 70%) internal carotid artery (ICA) disease, based on the results of several prospective, randomized controlled trials [1–5]. However, though elderly people are typically seen in everyday clinical practice, an increasingly important group of patients, i.e. those aged ≥ 80 years, was rarely recruited [1, 2] or excluded from the major surgical trials [3, 6]. Due to concern about the excessive risk of complications deriving from concomitant disease or limited post-procedural life expectancy, these patients were labelled as "high risk" and diverted to medical treatment or a seemingly less invasive procedure, such as carotid angioplasty and stenting (CAS). Wherever it was evaluated, best medical care proved less effective than CEA [1–3, 7], however, and the results of CAS in elderly patients were hardly encouraging [8–10]. In fact, though many single-centers studies reported that CAS can be performed with acceptable periprocedural complication rates [11], advanced age was considered a significant predictor of periprocedural neurological events after unprotected and protected CAS procedures in several recently-published large single- and multicenter trials [8–10]. This finding is odd, but the authors did not attempt to explain why the outcome of CAS should be significantly worse in older than in younger patients. Indeed, if older age really does mean a higher risk, one would expect the less invasive procedure to be better tolerated by elderly patients and produce better results than CEA. Although reanalyzing outcome in the multicenter surgical trials on the basis of age revealed that patients aged ≥ 75 years benefited more from CEA than younger patients [12] (probably because elderly patients have a greater risk of ischemic stroke on medical treatment and a lower surgical risk), from the trial data it is impossible to say whether this also applies to over 80-year-olds.

The aim of this study was to analyze the perioperative (30-day) stroke risk and death rates of CEA in patients aged ≥ 80 years, as compared with the outcome after CAS reported in the recently-published literature.

A retrospective review was conducted on a prospectively compiled computerized database of all primary CEAs performed by a single surgeon at our institution from 1990 to 2003 in symptomatic and asymptomatic patients with high-grade ICA lesions according to the recommendations of the North American Symptomatic Carotid Endarterectomy Trial (NASCET) [1] and the Asymptomatic Carotid Atherosclerosis Study (ACAS) [3]. Patients scheduled for CEA with concomitant coronary artery bypass grafting, or with associated supra-aortic trunk lesions requiring concurrent surgery and patients requiring carotid surgery for recurrent disease were ruled out.

The patients' demographic data were collected, including any history of diabetes mellitus, cigarette smoking, hypertension and coronary artery disease (CAD), other clinical variables, indications for surgery, details of the operation and hospital stay.

The ICA lesion was diagnosed on preoperative traditional digital subtraction angiography (DSA) during the earlier part of this experience (with declining frequency over the years), while duplex ultrasonography scan (DUS) was the only preoperative ICA imaging study performed in most patients from mid-1998 onwards, combined in selected patients with either magnetic resonance (MR) angiography, computed tomography (CT) angiography, or traditional arteriography. The radiologist's estimate of any carotid bulb or ICA stenosis in the final DSA report was recorded using the NASCET method [1]. If no DSA was performed, stenosis was estimated from the findings at preoperative DUS, performed in our vascular laboratory. The velocity criteria used to classify the degree of stenosis revealed a satisfactory correlation with the DSA findings when the degree of stenosis was calculated as the percentage of diameter reduction in compliance with the NASCET method [1]. These criteria have been published elsewhere [13, 14], and are validated yearly. Clinical presentation was always classified by the consultant neurologist as transient ischemic attack (TIA; i.e., temporary hemispheric symptoms lasting no more than 24 hours, with complete recovery), amaurosis fugax (transient monocular visual loss), or stroke (neurological deficit persisting more than 24 hours, regardless of the mechanism and related to either cerebral hemisphere). Patients who had non-hemispheric symptoms, such as dizzy spells or vertigo, were included in the asymptomatic group. Preoperative cerebral CT was performed in all symptomatic patients. Preoperative cardiac work-up was tailored to each patient, on the basis of history, electrocardiographic (ECG) findings, and symptoms. Patients with evidence of clinically important CAD underwent echocardiography or dipyridamole-thallium stress tests followed by coronary arteriography as indicated. Pre- and postoperative cranial nerve assessment was done in all patients by a neurologist and an otolaryngologist. Vocal cord movements were assessed by direct fiberoptic laryngoscopy in patients presenting symptoms and/or signs of vagus nerve injury [15]. Preoperative patient preparation was standardized. To reduce the incidence of neck hematoma, antiplatelet therapy (aspirin or dipyridamole, and ticlopidine or clopidogrel in the final period) was suspended at least 1 week before the operation, and was not resumed until the patient was discharged from the hospital.

All CEA procedures involved either traditional CEA with patching (n = 302) or eversion CEA (n = 958). The technical details of both procedures have been described elsewhere [16, 17]. All CEAs were performed with patients under deep general anaesthesia and cerebral protection involving continuous perioperative electroencephalographic monitoring (EEG) for selective shunting. All perioperative EEGs were visually analyzed by a neurologist with an extensive experience in the interpretation of studies during sleep, either natural or induced by hypnotic or anaesthetic agents. Shunting criteria were based exclusively on EEG changes consistent with cerebral ischemia. Completion imaging studies were never performed.

Patients were usually monitored in the recovery room for 2 hours until their blood pressure and neurological status were considered acceptable before being transferred to a regular nursing unit specializing in vascular care and monitored for the next 12 to 24 hours after surgery. All patients with severe headache were observed for hyperperfusion syndrome, and hypertension was treated aggressively. Most patients were discharged 48–72 hours after CEA.

The endpoints of this study were perioperative stroke and death, and cardiac complications, which were prospectively recorded according to the guidelines of the Ad Hoc Committee on Reporting Standards for Cerebrovascular Disease, Society for Vascular Surgery/North American Chapter of the International Society of Cardiovascular Surgery [18]. Minor stroke was defined as a focal neurological deficit of acute onset, lasting more than 24 hours and not leading to disability. Major stroke was defined as a focal neurological deficit leading to disability and permanent handicap.

All patients were assessed postoperatively by a consultant neurologist at the awakening from the anaesthesia and before the discharge. After discharge, visiting nurses monitored the patients' blood pressure and neurological status. Clinical evaluation and DUS were performed systematically by a consultant neurologist and two experienced technologists in all surviving patients at 1, 6 and 12 months, and once every postoperative year thereafter, assessing any residual ICA stenosis, angulation, recurrent ICA disease, or occlusion with an Acuson Sequoia 512 ultrasound system (Mountain View, Calif). Cerebral CT or MR scans were performed in all patients presenting a new neurological deficit.

Cardiac complications included: 1) myocardial infarction (MI), with a diagnosis made on the basis of creatine kinase enzyme levels and ECG findings; 2) pulmonary edema confirmed by the official reading of the chest radiograph; 3) documented ventricular fibrillation or primary cardiac arrest; and 4) new complete heart block requiring a pacemaker. A postoperative ECG was routinely obtained in all patients who had a history of CAD, CHF, or arrhythmia (rhythm other than sinus). Cardiac isoenzymes were obtained in all patients who had new findings on the postoperative ECG.

The results of this study show that CEA for symptomatic and asymptomatic high-grade carotid disease can be performed safely in patients aged ≥ 80 years, with perioperative stroke risk and death rates comparable with those of younger patients. This finding is consistent and extends the results of our two previous studies on early and long-term outcomes in patients aged 75 or more [20] and in over 80 year-olds with contralateral carotid occlusion [21].

Older patients were significantly less likely than younger patients to have a history of diabetes mellitus and this could represent a selection bias, with only healthier older patients being referred for CEA. Though we cannot infer this from the current data, we do not believe this is the case. Even if this was, the significantly higher incidence of arterial hypertension in the older group should counterbalance any influence of this bias. The two groups had similar indications for surgery, but older patients had a statistically higher rate of intraoperative EEG changes suggesting cerebral ischemia and needed shunting statistically more often than younger patients. This may be owing to the significantly greater presence of contralateral ICA occlusion in elderly patients, making the cerebral collateral flow insufficient to withstand the stress of clamping [22]. Shunting did not appear to have a negative effect on perioperative outcome, however.

The suggestion that patients aged ≥ 80 years are good candidates for CAS procedures and that CAS is consequently preferable to CEA in this patient population appears to be based on two premises that remain to be seen, i.e. that CEA is a high risk procedure and that CAS is safer than CEA. The first premise is not supported by the perioperative outcome emerging from our own and many other institutional series [23–29]. The second is belied by the analysis of the periprocedural results of the three series considered [8–10], which revealed an alarming, unacceptably high non-fatal stroke and death rate among the very elderly, ranging from 11.3% of Hobson's series [10] to 16 % of Roubin's series [9], with an overall perioperative stroke risk and death incidence of 13.9% when the outcomes of the three series are pooled [8–10]. The conclusion reached by all three studies was that octogenarians should be considered high-risk patients for CAS procedures and the common practice of recommending CAS for older patients was questionable and should be abandoned until the results of controlled clinical trials become available.

The conviction that older age means higher risk needs to be revised. Patients aged ≥ 80 years can undergo CEA with no greater perioperative risks than younger patients. Those advocating CAS in "high risk" patients should bear this in mind, as well as the fact that periprocedural risk of stroke and death after CAS increases with age. In the light of our findings, and in good agreement with the low perioperative complication rates of many institutional series, CEA remains the best therapeutic option in elderly patients with high-grade symptomatic and asymptomatic carotid stenosis.