Regional Performance of Carotid Endarterectomy
Appropriateness, Outcomes, and Risk Factors for Complications
Background and Purpose Guided by the findings of randomized controlled trials evaluating carotid endarterectomy (CEA), we examined the appropriateness of CEAs performed in our city and determined the incidences and risk factors for postoperative stroke, death, and cardiac complications.
Methods Using health records, we retrospectively reviewed 291 consecutive CEAs performed in our region over 18 months. Based on randomized controlled trial results and standardized remeasurements of angiographic carotid stenoses, indications for CEA were considered appropriate for symptomatic carotid stenoses ≥70%, uncertain for <70% symptomatic or ≥60% asymptomatic stenoses, or inappropriate for <60% asymptomatic stenoses and for patients with preoperative neurological or medical instability.
Results We found that 41% of patients (118/291) were asymptomatic. Surgical indications were appropriate in 33% of cases (92/281), uncertain in 49% (138/281), and inappropriate in 18% (51/281). Stroke or death occurred within 30 days postoperatively in 5.2% (9/174) of symptomatic patients and 5.1% (6/117) of asymptomatic patients. At least one cardiac complication (angina, congestive heart failure, dysrhythmia, or myocardial infarction) developed in 8.9% (26/291). Independent preoperative risk factors for stroke or death were histories of angina or congestive heart failure and lack of antiplatelet medication; for cardiac complications, risk factors were age >75 years and a history of congestive heart failure.
Conclusions Almost 1 in 5 patients underwent CEA inappropriately, which was most commonly due to apparent overestimation of stenosis severity, and half had uncertain indications. Our high complication rate possibly negated any overall surgical benefit in the large group of asymptomatic patients.
Only a decade ago, the practice of CEA was seriously questioned.1 2 This was mainly due to its then unproven efficacy in stroke prevention, but contributing to these concerns were descriptions of its use for what appeared to be inappropriate indications,3 as well as reports of unacceptably high complication rates in several surgical audits.4 5 6 7 8 In response to this uncertainty, multicenter randomized controlled trials were launched that subsequently validated the use of CEA under certain circumstances.9 10 11 12 13 In these studies, carotid stenoses were measured in a standardized fashion, and treatment complications were carefully monitored and reported. Because the results of these trials have been disseminated throughout the medical community and have contributed to a large resurgence in the use of CEA,14 we wished to reexamine the issues of appropriateness and complications of CEA as currently practiced in our region.
Using the results of the recently reported randomized trials examining CEA as guidelines, the goal of this study was to determine the appropriateness and complication rates of all CEAs performed in our city over a recent 18-month period. Part of this analysis required an assessment of the accuracy of preoperative ultrasonography and angiography practiced in our region. We also wished to determine any risk factors in our patient population that were significantly predictive of postoperative stroke or death and cardiac complications.
Subjects and Methods
We performed a retrospective cohort study of all CEAs performed in Edmonton, Alberta, Canada, from April 1, 1994, through September 30, 1995. The cohort consisted of 291 consecutive cases of CEA performed on 265 patients by nine surgeons from the neurosurgical, general surgery, and vascular surgery services, from four teaching hospitals (two of which were tertiary-care centers). With regards to the 26 patients who underwent bilateral (staged) CEAs, each procedure was considered independent of the contralateral operated artery. Given that Canada’s universal healthcare system guarantees accessibility and the fact that our city’s hospitals serve as referral centers for about 1.6 million people from an expansive geographic area, this study may be considered a population-based analysis. Clinical and radiological information was collected primarily by examination of in-hospital medical records and, if necessary, supplemented through review of available office records of individual surgeons. For every patient, any in-hospital morbidity or mortality was determined by review of hospital charts. Additional follow-up for complications occurring within 30 days of CEA was achieved for 98% of patients (285/291) by telephone interview of patients or their families and/or survey of ambulatory-care records from the participating surgeons’ offices. Identification of any deaths occurring after hospital discharge for the entire patient cohort was independently determined through database searches of the provincial government mortality registry using linkage analyses with name and birth date.
Carotid angiograms were remeasured in a standardized fashion by one investigator (J.H.W.) who was blinded to patient identity and the original reported degree of stenosis. Angiograms that were difficult to interpret were evaluated in consultation with the senior author (J.M.F.) in a similar blinded manner. The angiographic method of measurement used was that of NASCET, ie, comparing the greatest degree of linear diameter stenosis of the ICA at the carotid bifurcation to the diameter of the distal normal ICA.15 These results were compared with the angiographic percentage of stenosis appearing on the original radiologist’s report and the reported range of stenosis determined by ultrasonography. The presence or mention of an angiographic flow-limiting critical stenosis resulting in the appearance of a distally thin ICA, or a delay in contrast filling of the ICA compared with the adjacent external carotid artery, was classified as a 95% stenosis.16 Carotid bifurcation plaques were graded according to their angiographic appearance as being ulcerated, irregular, or smooth.17 The distal ipsilateral cerebral vasculature was assessed for arterial stenosis, defined as narrowing >50% of the local luminal diameter. Ultrasonography data were collected from the actual ultrasound report as obtained from the in-hospital record or referring diagnostic facility. No attempt was made in this study to establish standards of ultrasonographic equipment or technique. Angiography or ultrasonography reports using only subjective descriptors such as “severe” to quantify carotid stenosis were excluded from comparison analyses.
The appropriateness of operative indications was determined on the basis of a review of five recent randomized controlled trials examining CEA9 10 11 12 13 and clinical practice guidelines established by both the American Heart Association18 and the Canadian Neurosurgical Society (cerebrovascular writing group for the Canadian Neurosurgical Society).18A With use of these criteria in conjunction with our NASCET remeasurement values, the appropriateness of CEA for each patient examined in this study was classified into one of the following categories: (1) appropriate for symptomatic patients with isolated and surgically accessible carotid stenoses ≥70%, (2) uncertain for patients with either symptomatic stenoses of <70% or asymptomatic stenoses ≥60%, or (3) inappropriate for patients with asymptomatic stenoses of <60% and for those patients with significant neurological instability or high-risk medical conditions. Preoperative neurological instability, based on criteria as originally described by Sundt et al19 and more recently validated in a retrospective study,20 was defined as a progressing neurological deficit or a neurological deficit within 1 day before CEA. High-risk preoperative medical conditions were defined as unstable angina (defined as angina developing at rest or of new onset), myocardial infarction within 3 months before CEA, or uncontrolled CHF. Patients were considered symptomatic from their carotid disease if there was a documented history of any prior cerebral or retinal ischemia referable to the vascular territory of the stenotic carotid artery (ie, ipsilateral transient ischemic attack, stroke, amaurosis fugax, or retinal infarction). Patients with nonhemispheric or nonretinal symptoms, such as dizziness, syncope, or bilateral visual difficulties, were regarded as being asymptomatic from their carotid disease.
The primary outcome event in this analysis was stroke or death occurring within 30 days of surgery. Any hemispheric neurological deficit lasting more than 24 hours was classified as a stroke. A major stroke was defined as one producing a significant functional deficit. For example, a patient rendered incapable of independent activities of daily living or requiring further rehabilitation was classified as having a major deficit, whereas mild weakness in a patient otherwise capable of independent ambulation and returning home on discharge was classified as a minor stroke. The degree of clinical recovery after hospital discharge was not assessed. A secondary outcome event was the development of at least one postoperative in-hospital cardiac complication (namely, angina, CHF, dysrhythmia, or myocardial infarction).
Data were collected using a computer database (FileMaker Pro 2.1, Claris Corp) and analyzed with statistical software (SPSS 6.1, SPSS Inc). Results were expressed as mean±SEM. The pooled chance-corrected measure of agreement between the original interpretation and the NASCET remeasurement value of the angiograms was expressed as a κ summary statistic. Univariate testing used χ2 and Fisher’s exact tests for comparison of proportions. Measures of association between risk factors and outcome events were expressed as odds ratios with 95% CI. Level of significance was set at P≤.05, and all tests were two-tailed. The associations between risk factors and outcome events were also evaluated with multivariate analyses using logistic regression models. Dependent variables were the dichotomous outcomes of stroke or death and cardiac complications. Independent variables used in these models included those listed in Table 5⇓, as well as the following dichotomous variables: history of diabetes mellitus, history of claudication, history of valvular heart disease, preoperative neurological instability (present or absent),19 20 preoperative systolic blood pressure (>160 mm Hg or ≤160 mm Hg), carotid plaque irregularity (present or absent),17 carotid plaque ulceration (present or absent),17 intraoperative shunting (yes or no), surgical service (neurosurgical or nonneurosurgical), and postoperative bradycardia (<60 or ≥60 beats per minute). A stepwise procedure was used to include variables in the model using values of P≤.05 for variable inclusion and P≤.10 for exclusion.
The mean patient age was 67.8±0.5 years, and 58% of patients were men (170/291). Patient characteristics are summarized in Table 1⇓. Forty-one percent of operations (118/291) were for asymptomatic stenosis, and the remainder were for symptomatic disease. One of two neurosurgeons operated on 122 patients, and 169 patients were operated on by one of seven general or vascular surgeons.
All patients had undergone carotid or cerebral angiography before surgery, and 97% (281/291) of cases had angiograms available for remeasurement. For this analysis, comparisons were made with respect to each artery (both operated and nonoperated) rather than per patient. Thirty arteries were excluded because the original radiologist did not numerically quantify the degree of stenosis narrowing. When the angiographic stenoses on both operated and nonoperated sides were remeasured with the NASCET method, there was an overall moderate agreement with the original radiologist’s interpretation (symptomatic stenosis: κ=0.72 and 95% CI, 0.64 to 0.80; asymptomatic stenosis: κ=0.67 and 95% CI, 0.57 to 0.77) (see Table 2⇓). The method of angiographic measurement used by the original radiologist could not be reliably determined in this study.
Stenoses on ultrasonography in our region are reported in percentage ranges rather than specific values. Of the 248 patients who underwent carotid ultrasonography as the initial diagnostic test, 219 had ultrasonography reports available for review (88%), and 214 had both ultrasonographic and angiographic studies examined (86%). Ultrasonographic accuracy was investigated per artery (both operated and nonoperated) rather than per patient. After reports that did not quantify carotid disease were excluded, 392 arteries were left for analysis. As shown in Table 3⇓, preoperative ultrasonography generally had a high sensitivity (89% and 97%) but only moderate specificity (58% and 67%) for carotid disease of ≥50% and ≥80%, respectively.
With our criteria of appropriateness in conjunction with the NASCET remeasurement values, the appropriateness of CEA was determined for our patient population; the results are summarized in Table 4⇓. Appropriate indications for CEA were found in 33% of patients (92/281), whereas 49% had indications considered uncertain (138/281), and inappropriate surgical indications were found in 18% of patients (51/281). Of the 51 cases with inappropriate indications, 23 (45%) were asymptomatic patients who were originally incorrectly measured as having ≥60% stenosis, 14 (27%) were asymptomatic patients who underwent surgery despite stenoses correctly measured at <60%, 8 (16%) were patients with preoperative neurological instability, and 6 (12%) were patients at high preoperative medical risk. Recognizing the marked benefit of CEA for severe symptomatic disease that was found in NASCET9 and therefore the possible tendency to operate on patients with stenoses that approached this degree of narrowing, we reanalyzed the data using modified criteria by which symptomatic stenoses ≥65% were considered appropriate for CEA. However, this only minimally increased the proportion of appropriate indications to 37% (104/281) and decreased uncertain indications to 45% (126/281).
Inappropriate use of CEA was found to vary widely among individual surgeons, ranging from 0% to 33% (P=.07, df=8, χ2 test). Of CEAs performed by the neurosurgical service, 14% were classified as inappropriate versus 21% inappropriate CEAs for the nonneurosurgical services, a difference that was not statistically significant (P=.15, χ2 test). Uncertain indications for CEA were found in 33% to 67% of individual surgeons’ series (P=.26, df=8, χ2 test). When stratified according to surgical service, neurosurgeons used CEA for uncertain indications significantly less often than their surgical colleagues (40% versus 55%, P=.006, χ2 test).
Overall, in this series there were six major strokes and five minor postoperative strokes occurring within 30 days of surgery. (Two additional minor strokes that developed as a result of angiography in previously asymptomatic patients were not included in the outcome analysis.) All strokes that occurred in this study were ischemic, and all were in the territory of the operated carotid artery. Of the four deaths that occurred in this series, two were due to myocardial infarction, one was due to ischemic bowel secondary to cardiovascular causes, and one was due to interstitial pneumonitis. The combined stroke or death rate in this series was 5.2±1.7% for symptomatic patients (9/174) and 5.1±2.0% for asymptomatic patients (6/117). Although a broad range of stroke or death rates for individual surgeons was observed (0% to 33%), the relatively small number of operations performed by the large number of surgeons in this analysis may have contributed to an inability to statistically associate stroke or death rates to individual surgeons (P=.22, df=8, χ2 test) or to surgical specialty. The combined stroke or death rate was 4.1% for neurosurgery and 5.9% for all other surgical specialties combined (P=.49, χ2 test).
Significant risk factors for stroke or death were determined by univariate analysis and multiple logistic regression and are summarized in Tables 5⇓ and 6⇓. Independent preoperative risk factors for stroke or death by multivariate analysis were a history of angina, a history of CHF, and lack of preoperative antiplatelet medication. At least one cardiac complication developed in 8.9% of patients (26/291). The risk factors for cardiac complications, determined by univariate and multivariate analyses, are also summarized in Tables 5⇓ and 6⇓. Multiple logistic regression yielded age >75 years and a history of CHF as being significant independent variables associated with cardiac complications.
Appropriateness of CEA
Assessments of the appropriateness of CEA in surgical series or populations of surgical patients are rare. Winslow and colleagues3 examined the surgical indications of 1302 Medicare patients from three geographic areas who underwent CEA in 1981. Using appropriateness guidelines established by a multidisciplinary panel of experts, Winslow et al determined that 35% of these patients had surgery for appropriate reasons, 32% had equivocal surgical indications, and 32% underwent CEA inappropriately. Of this latter group of patients, 48% underwent surgery for minimal carotid stenosis, 11% were at excessively high preoperative risk, 9% had cerebrovascular symptoms contralateral to the operated side, and 6% received surgery for carotid occlusion. The authors concluded that CEA was substantially overused in the regions studied, and since the perioperative stroke or death rate was 9.8% in their study group, they felt that any benefit at all from CEA was questionable. Application of these same criteria to 107 cases of CEA in five Veterans Administration hospitals in 1981 revealed that 55% of surgeries were appropriate and 32% were equivocal.21 In that study, 13% of CEAs were clearly inappropriate, mostly because of surgery on patients with carotid occlusion and asymptomatic patients with moderate stenoses at higher surgical risk.
On a smaller scale, Asaph and colleagues22 found that in a retrospective community surgical audit of 243 patients from 1986 through 1987, 37% of CEA patients underwent surgery for <80% asymptomatic stenosis. After these results were publicized, a local surgical committee formulated the following appropriate indications for CEA: carotid lesions causing transient ischemic attack or reversible ischemic neurological deficit, “significant” carotid stenoses causing stroke with subsequent recovery, asymptomatic carotid stenoses >80%, or any other reason if supported by a disinterested third party. As a result, over the succeeding 21 months, there was a drop in the number of operations performed, the surgical complication rate, and the number of “inappropriate” operations for asymptomatic carotid disease.
Such concerns over the appropriateness of CEA, along with its unproven efficacy and possible excessive use,1 2 3 resulted in the initiation of a number of randomized controlled CEA trials, including NASCET,9 ECST,10 the symptomatic and asymptomatic American Veterans Affairs trials,11 12 and ACAS.13 The published results of these studies have supplied evidence to support different indications for CEA with correspondingly variable degrees of certainty, and they have also provided benchmarks for acceptable complication rates associated with each indication (Table 7⇓). Although NASCET and ECST measured carotid stenosis in different ways, the preliminary results of both indicate that CEA is much more effective than medical therapy in preventing stroke for severe symptomatic stenoses, providing that the perioperative stroke or death rate is <6%.18 Surgery for patients with symptomatic stenoses <70% was designated an uncertain indication for CEA in our analysis, since NASCET had not yet completed comparing CEA with medical therapy for such patients during our study period and manuscript preparation. Because NASCET specifically excluded patients with progressing or disabling neurological deficits and those with uncontrolled medical illness, and because previous studies have suggested that patients with these risk factors have a high risk of postoperative stroke,19 20 in our review patients with these features were considered inappropriate candidates for surgery.
Asymptomatic stenosis >60% (as determined by the NASCET method) was classified as an uncertain indication for CEA in our analysis. The only randomized controlled study to date indicating efficacy of CEA for stroke prevention in asymptomatic patients is ACAS, and the results of that study require careful interpretation.13 Although this trial demonstrated that the risk of stroke or death was significantly lowered in the surgical group, the absolute risk reduction with CEA was only slightly more than 1% per year projected over 5 years. As well, major strokes were not prevented, women did not appear to benefit from CEA, and the overall positive result for surgery was only realized with a very low 2.3% perioperative stroke or death rate. These concerns, coupled with the fact that the natural history of asymptomatic carotid stenosis is fairly benign with an annual stroke or death risk on the order of only 1% to 2%,23 24 suggest that the results of ACAS may be statistically but not clinically significant. It may be argued that before the ACAS results were first published in December 1994,25 halfway during our study period, there was even less compelling scientific evidence to suggest that asymptomatic stenosis was a good indication for CEA. Because the benefit of CEA for asymptomatic stenosis seems marginal and dependent on both the patients treated and a very low operative complication rate, we decided a priori in our analysis that this indication would be considered “uncertain” rather than unequivocally “appropriate.”
Data from randomized controlled trials have afforded an opportunity to base determinations of CEA appropriateness on objective evidence of surgical efficacy. Previous studies evaluating the appropriateness of CEA may have been limited by biases in the opinions of what was considered appropriate surgery.26 27 28 29 Our analysis is also unique in that it links surgical appropriateness with standardized carotid stenosis remeasurements. Our choice of the NASCET method of measurement was based on its relatively high intraobserver and interobserver reliability30 and its use in most of the randomized controlled CEA trials that have provided the basis of our classification of appropriateness.9 11 12 13 31 The finding of our study, where reasonable but not strong agreement was found between the original radiologist’s interpretation and the remeasurement value, validates previous concerns that CEA may be performed on the basis of inaccurate or nonuniform stenosis measurements.32 However, the actual impact of this discrepancy on the management of our patients could not be determined, since it is not known whether each surgeon in our study based treatment on the radiologist’s report or on personal review and measurement of the angiogram. The fact that the most common discrepancy found was an overestimation of stenosis severity suggests that surgeons either were not consistently and accurately remeasuring angiograms with the NASCET method before making a decision regarding surgery or were making management decisions only partly based on the degree of stenosis.
We found that the sensitivity of ultrasound for assessing angiographic stenoses in our study population was fairly high, although the specificity of this imaging technique was only moderate. As well, the reliability of ultrasonography in our region may be suspect, since the positive predictive value of ultrasonography was only 50% for detecting severe stenoses ≥80%. In our series, carotid ultrasound tended to overestimate the actual severity of carotid stenosis. Other studies have shown ultrasonography to be less accurate in assessing mild and moderate stenoses than severe stenoses, with one study demonstrating test sensitivities ranging from 25% to 92%, depending on the degree of vessel narrowing.33 34 35
In our study, by combining the best indication for surgery as ascertained from patient records with the carotid stenosis remeasurement value, we found that 49% of patients who underwent CEA had an uncertain indication and that the majority of these patients had asymptomatic stenoses ≥60%. Inappropriate CEAs were found in 18% of our patient population or in almost 1 in every 5 patients overall. This group was made up largely of patients with <60% asymptomatic stenoses, of which 62% (23/37) had original angiographic readings that overestimated the NASCET remeasurement value. The remainder in this group underwent surgery in the face of neurological instability or high-risk medical comorbidity. An argument can be made, however, to consider neurologically unstable patients separately rather than as inappropriate surgical candidates a priori. Although unproven, it may be that some of these patients may in fact benefit overall from CEA, even with an elevated attendant perioperative complication rate, if their risk of stroke or death without surgery is greater than with CEA.
While efficacy of CEA for symptomatic stenoses <70% may yet be shown in the continuing part of NASCET, perhaps the most disquieting finding in our analysis was the large proportion of patients undergoing surgery for asymptomatic stenosis. Available evidence suggests that any modest benefit from surgery for asymptomatic patients is realized only with a very low perioperative complication rate, which may be difficult to achieve outside the research setting.
Complications of CEA
A recent literature review of studies published since 1980 that examined the results of CEA performed for symptomatic carotid stenosis found a mean stroke or death rate of 5.6% (95% CI, 4.4 to 6.9).36 For the symptomatic subgroup of patients in our series, the 5.2% stroke or death rate was similar to this and the 5.8% rate found in NASCET.9 However, the 5.1% postoperative stroke or death rate in our asymptomatic patients was higher than the 3.4% rate found in a review of asymptomatic series37 and significantly more than the corresponding 1.2% rate found in ACAS (Fisher’s exact test, P=.01).13 The comparatively greater complication rate for our asymptomatic patients probably negated any overall benefit from CEA for this patient subgroup in our region. This observed higher frequency of complications is unexplained, but it possibly may be related to technical performance or patient selection or may simply be a realistic overall portrayal of CEA as it is practiced in a community-wide setting. Higher complication rates in community-based studies, compared with randomized controlled trials, are not unexpected. In randomized controlled trials, artificial constraints such as stringent inclusion and exclusion criteria, generally excellent patient care, and close patient follow-up may result in better outcomes than in routine clinical practice.
Several groups have described potential risk factors for surgical complications after CEA.8 38 39 40 41 In 1975, Sundt et al19 provided evidence that patients with certain cardiovascular and neurological risk factors were at highest risk for postoperative myocardial infarction and stroke. Some of these same risk factors have been confirmed statistically in a retrospective analysis of 1160 CEA patients.20 The authors of this latter study found seven predictors that can stratify candidates for CEA at higher risks of stroke and cardiac complications: age over 75 years, symptomatic (versus asymptomatic) stenosis, severe hypertension, CEA in preparation for coronary artery surgery, a history of angina, intraluminal thrombus apparent on angiography, and intracranial arterial stenosis. Our analysis of a smaller group of patients identified a history of angina, a history of CHF, and lack of antiplatelet agent use before surgery as preoperative independent variables associated with stroke or death.
Although risk factors for cardiac complications alone after CEA have been less commonly studied in the literature, patients undergoing CEA frequently harbor occult or symptomatic cardiac disease.42 43 44 45 46 47 48 49 Our present study identified age over 75 years and a history of CHF as independent preoperative variables predictive of cardiac complications after CEA. The detection of these risk factors may allow the preoperative identification of high-risk patients who may then be given more intensive perioperative monitoring or treatment.50
The findings of this study emphasize the usefulness of a regional analysis in uncovering local clinical concerns. We determined that CEA may be overused in our region because of the substantial number of operations for inappropriate and uncertain indications. In particular, the observed high number of complications suggests restricting the use of CEA for our asymptomatic patients. We are currently undertaking a local prospective quality-control program to address the issues identified in this study. However, it may be reasonable to assume that these concerns are not isolated to our own region only. Because estimates have placed the proportion of operations for asymptomatic carotid disease at 25% to 50% of all CEAs performed in the United States,51 our disturbing findings regarding the misuse of CEA and elevated complication rates may reflect similar problems in other communities. It deserves mention that surgical audits such as reported here are time-consuming and expensive endeavors, and they require the interpretive skills of individuals familiar with cerebrovascular disease. As additional information from randomized controlled trials becomes available,52 thus more fully defining the indications for CEA, appropriateness classifications can be revised, new audits undertaken, and previous audits reanalyzed.
In summary, the combined rate of uncertain and inappropriate surgeries was high at 67% in our healthcare region, with half of all patients undergoing CEA for uncertain indications. Inappropriate surgery was seen in about 1 in 5 patients in this analysis and was related mainly to overestimation of mild and moderate asymptomatic carotid stenoses, but it was also due to the use of CEA in the presence of significant neurological and medical instability in some instances. The uniform use of a standard method of angiographic measurement of carotid stenosis may reduce the rate of surgeries for uncertain and inappropriate indications. A substantial proportion of our study population was asymptomatic, and any benefit of surgery for these patients was overshadowed by a significant risk of postoperative stroke or death. The preoperative identification of high-risk patients through risk factor analysis may offer an opportunity to reduce complication rates. Independent local audits of surgical performance are recommended to identify areas of concern at a regional level.
Selected Abbreviations and Acronyms
|ACAS||=||Asymptomatic Carotid Atherosclerosis Study|
|CHF||=||congestive heart failure|
|ECST||=||European Carotid Surgery Trial|
|ICA||=||internal carotid artery|
|NASCET||=||North American Symptomatic Carotid Endarterectomy Trial|
- Received November 18, 1996.
- Revision received February 25, 1997.
- Accepted March 3, 1997.
- Copyright © 1997 by American Heart Association
Warlow C. Carotid endarterectomy: does it work? Stroke. 1984;15:1068-1076.
Barnett HJ, Plum F, Walton JN. Carotid endarterectomy: an expression of concern. Stroke. 1984;15:941-943.
Easton JD, Sherman DG. Stroke and mortality rate in carotid endarterectomy: 228 consecutive operations. Stroke. 1977;8:565-568.
Brott T, Thalinger K. The practice of carotid endarterectomy in a large metropolitan area. Stroke. 1984;15:950-955.
Dyken ML, Pokras R. The performance of endarterectomy for disease of the extracranial arteries of the head. Stroke. 1984;15:948-950.
Fode NC, Sundt TM Jr, Robertson JT, Peerless SJ, Shields CB. Multicenter retrospective review of results and complications of carotid endarterectomy in 1981. Stroke. 1986;17:370-376.
Mayberg MR, Wilson SE, Yatsu F, Weiss DG, Messina L, Hershey LA, Colling C, Eskridge J, Deykin D, Winn HR, for the Veterans Affairs Cooperative Studies Program 309 Trialist Group. Carotid endarterectomy and prevention of cerebral ischemia in symptomatic carotid stenosis. JAMA. 1991;266:3289-3294.
Gillum RF. Epidemiology of carotid endarterectomy and cerebral angiography in the United States. Stroke. 1995;26:1724-1728.
North American Symptomatic Carotid Endarterectomy Trial (NASCET) Steering Committee. North American Symptomatic Carotid Endarterectomy Trial: methods, patient characteristics, and progress. Stroke. 1991;22:711-720.
Streifler JY, Eliasziw M, Fox AJ, Benavente OR, Hachinski VC, Ferguson GG, Barnett HJ. Angiographic detection of carotid plaque ulceration: comparison with surgical observations in a multicenter study—North American Symptomatic Carotid Endarterectomy Trial. Stroke. 1994;25:1130-1132.
Moore WS, Barnett HJ, Beebe HG, Bernstein EF, Brener BJ, Brott T, Caplan LR, Day A, Goldstone J, Hobson RW II, Kempczinski RF, Matchar DB, Mayberg MR, Nicolaides AN, Norris JW, Ricotta JJ, Robertson JT, Rutherford RB, Thomas D, Toole JF, Trout HH III, Wiebers DO. Guidelines for carotid endarterectomy: a multidisciplinary consensus statement from the Ad Hoc Committee, American Heart Association. Stroke. 1995;26:188-201.
Findlay JM, Tucker WS, Ferguson GG, Holness RO, Wallace MC, Wong JH. Guidelines for the use of carotid endarterectomy: current recommendations from the Canadian Neurosurgical Society. Can Med Assoc J. In press.
McCrory DC, Goldstein LB, Samsa GP, Oddone EZ, Landsman PB, Moore WS, Matchar DB. Predicting complications of carotid endarterectomy. Stroke. 1993;24:1285-1291.
Clinical advisory: carotid endarterectomy for patients with asymptomatic internal carotid artery stenosis. Stroke. 1994;25:2523-2524. Special report.
Eliasziw M, Smith RF, Singh N, Holdsworth DW, Fox AJ, Barnett HJ. Further comments on the measurement of carotid stenosis from angiograms: North American Symptomatic Carotid Endarterectomy Trial (NASCET) Group. Stroke. 1994;25:2445-2449.
CASANOVA Study Group. Carotid surgery versus medical therapy in asymptomatic carotid stenosis. Stroke. 1991;22:1229-1235.
Barnett HJM, Taylor DW. Clinical trials in stroke prevention: persisting uncertainties, firm answers. Neurology. 1993;43:2163-2166.
Eliasziw M, Rankin RN, Fox AJ, Haynes RB, Barnett HJ. Accuracy and prognostic consequences of ultrasonography in identifying severe carotid artery stenosis: North American Symptomatic Carotid Endarterectomy Trial (NASCET) Group. Stroke. 1995;26:1747-1752.
Rothwell PM, Slattery J, Warlow CP. A systematic review of the risks of stroke and death due to endarterectomy for symptomatic carotid stenosis. Stroke. 1996;27:260-265.
Rothwell PM, Slattery J, Warlow CP. A systematic comparison of the risks of stroke and death due to carotid endarterectomy for symptomatic and asymptomatic stenosis. Stroke. 1996;27:266-269.
Goldstein LB, McCrory DC, Landsman PB, Samsa GP, Ancukiewicz M, Oddone EZ, Matchar DB. Multicenter review of preoperative risk factors for carotid endarterectomy in patients with ipsilateral symptoms. Stroke. 1994;25:1116-1121.
Chimowitz MI, Weiss DG, Cohen SL, Starling MR, Hobson RW II, for Veterans Affairs Cooperative Study Group 167. Cardiac prognosis of patients with carotid stenosis and no history of coronary artery disease. Stroke. 1994;25:759-765.
Urbinati S, Di Pasquale G, Andreoli A, Lusa AM, Carini G, Grazi P, Labanti G, Passarelli P, Corbelli C, Pinelli G. Preoperative noninvasive coronary risk stratification in candidates for carotid endarterectomy. Stroke. 1994;25:2022-2027.
Barnett HJ, Eliasziw M, Meldrum HE, Taylor DW. Do the facts and figures warrant a 10-fold increase in the performance of carotid endarterectomy on asymptomatic patients? Neurology. 1996;46:603-608.
Barnett HJ, Warlow CP. Carotid endarterectomy and the measurement of stenosis. Stroke. 1993;24:1281-1284 (erratum published in Stroke. 1994;25:758).