(Stroke. 1999;30:1745-1750.)
© 1999 American Heart Association, Inc.
Editorial |
From the Department of Neurologic Sciences, Rush Medical College, Chicago, Ill.
Correspondence to Philip B. Gorelick, MD, MPH, Center for Stroke Research, 1645 West Jackson, Suite 400, Chicago, IL 60612.
Key Words: carotid endarterectomy complications editorials stroke
| Introduction |
|---|
|
|
|---|
These pioneers1 2 3 4 5 helped to usher in the modern era of CEA. Early attempts at surgical correction of extracranial carotid artery occlusive disease led to the Joint Study of Extracranial Arterial Occlusion in 1959.6 The Joint Study was designed to assess the efficacy of carotid surgery and included 6535 patients (69% men). In 1961, 5 and later 13 of the centers agreed to a controlled trial whereby patients would be randomly assigned to surgical or nonsurgical therapy. The main finding of the trial was that there was no statistically significant difference in stroke or death between the 2 treatment groups. The study had shortcomings and was criticized for being inconclusive. The trial, however, was the only major randomized and controlled study until the 1980s.
The 1980s were punctuated by cautionary notes concerning CEA.7 8 9 Concern was fueled by startling reports of high complication rates,10 11 yet there was exponential growth in the performance of the procedure based largely on anecdotal evidence. It was estimated that only 35% of CEAs were being performed for appropriate indications.12 By the mid-1980s, dissemination of such sobering news may have led to a decline in the number of CEAs.13
CEA had become a topic of emotional debate and uncertainty since rigorous scientific investigation was wanting.14 15 Finally, in the 1990s the main results of large-scale clinical trials were published.16 17 18 19 20 21 22 23 24 25 At last, the debate would endor would it? The National Institute of Neurological Disorders and Stroke supported 3 of these major trials.17 18 22 23
What have we learned from these clinical trials? The main results
for symptomatic and asymptomatic patients are
summarized in Table 1![]()
. CEA is beneficial
in symptomatic patients with high-grade carotid
stenosis (70% to 99%). This was a main conclusion of the
North American Symptomatic Carotid
Endarterectomy Trial (NASCET) and the Medical
Research Council European Carotid Surgery Trial
(ECST).17 19 Furthermore, within the range of 70% to 99%
carotid diameter stenosis as defined by cerebral arteriography,
NASCET showed a stenosis-dependent effect for degree of carotid
stenosis (70% to 79%, 80% to 89%, and 90% to 99%) and
degree of risk reduction after CEA in a secondary
analysis.17 The final results of
ECST20 suggested that the cutoff point for major benefit
with CEA in arteriography-determined stenosis was approximately
80%, which correlated with approximately 60% stenosis in
NASCET (ECST and NASCET used different methods for measuring carotid
stenosis26 ; an arteriographic stenosis
equivalence scale is shown below Table 1
). ECST showed a
downward trend in the benefit of surgery for carotid stenosis
from a range of 90% to 100% to a range of 80% to
89%.20 Overall, men had more benefit than women, and
younger patients had more benefit over a narrow range of severe
stenosis than older patients.20 The Veterans
Affairs Cooperative Study21 was halted when results of
NASCET and ECST became available. The results of this study of 193
participants supported a stenosis-dependent effect that favored
CEA and showed the benefit of CEA when there was high-grade
stenosis (>70%).
|
|
Less severe carotid stenosis (0% to 29%) was studied in ECST.19 In this group the risk of ipsilateral ischemic stroke was low, even in the absence of surgery, and any benefits of CEA were outweighed by early operative risks.
What are the results when there is moderate stenosis (eg, 30%
to 69%)? In NASCET, when there was <50% stenosis, the
failure rate did not differ significantly in the surgically treated and
medically treated groups.18 Among those with 50% to 69%
stenosis, however, there was a "moderate" benefit of
ipsilateral stroke reduction for CEA over medical therapy alone during
a 5-year period. Risk factors for perioperative stroke
or death were contralateral carotid occlusion, left-sided carotid
disease, taking <650 mg aspirin per day, absence of a history of
myocardial infarction or angina, lesion present on CT or MRI that
was ipsilateral to the symptomatic stenosed artery, history
of diabetes mellitus, and diastolic blood pressure
>90 mm Hg. CEA was associated with greater long-term benefit in
men, those with a recent stroke, recent hemispheric symptoms, and those
taking
650 mg of aspirin per day. Prospective study of the
relationship between aspirin dose and CEA clarified results from post
hoc subgroup analysis in NASCET. The Aspirin
Carotid Endarterectomy (ACE) Trial showed that the
risk of stroke, myocardial infarction, and death within 30 days and 3
months of CEA was lower for patients taking 81 or 325 mg of aspirin
before surgery compared with those taking 650 or 1300 mg of aspirin per
day (written communication, H.J.M. Barnett, MD, 1999). Finally,
patients with
70% stenosis had a durable benefit from CEA at
8 years of follow-up.
Two large-scale trials of CEA for asymptomatic carotid
stenosis have been published. The Asymptomatic
Carotid Atherosclerosis Study (ACAS) showed a modest
benefit projected over 5 years that favored surgery in those with
60% stenosis.22 Risk of ipsilateral stroke and
any perioperative stroke or death was less for CEA by
an estimated average annual absolute difference of approximately 1.2%
(P=0.004). In subgroup analysis, men had a 66%
5-year reduction in stroke risk (95% CI, 0.36 to 0.82), whereas women
had an event rate reduction of only 17% (95% CI, -0.96 to 0.65). The
difference between men and women, however, was not statistically
significant (P=0.10). Possibly, women did not derive benefit
from CEA because they had more perioperative
complications (3.6% versus 1.7%). In addition, there was a trend for
greater risk reduction for CEA in younger patients (P=0.50)
but no stenosis-dependent effect by degree of carotid
stenosis (60% to 69%, 70% to 79%, 80% to 99%).
Finally, there was no statistically significant difference in favor of
CEA for reduction of major ipsilateral stroke or any
perioperative major stroke or death
(P=0.12). Overall, there were relatively small numbers of
events in the various subgroups under study, which limits statistical
power to draw more firm conclusions about subgroup results.
Another trial, the Veterans Affairs Cooperative Study, included
patients with asymptomatic carotid stenosis of
50% diagnosed by arteriography.24 Ipsilateral
neurological end points were reduced when stroke was combined with
transient ischemic attack (TIA) and transient monocular
blindness. There was no significant difference between groups, however,
in an analysis of all stroke and death. Overall mortality,
including that of the postoperative period, was largely due to
coronary events.
The Mayo Asymptomatic Carotid Endarterectomy Study23 was halted prematurely because there was an excess of myocardial infarction in the surgical group. The Carotid Artery Stenosis with Asymptomatic Narrowing: Operation Versus Aspirin (CASANOVA) study25 included 410 patients with 50% to 90% stenosis. The investigators concluded that CEA was not beneficial; however, interpretation of data were confounded by a complex study design and exclusion of patients with >90% carotid stenosis.
In this issue of Stroke, Paciaroni et al27 and Ferguson et al28 share with us results of final primary data from NASCET. These articles focus on 2 key topics: medical complications of CEA and a detailed review of surgical results. Medical complications that occurred within 30 days after CEA were reported in 1415 patients with symptomatic internal carotid artery stenosis (30% to 99%) and were compared with those in 1433 patients who received medical care alone. Overall, medical complications occurred in 8.1% who underwent CEA. Most were of short duration (69.7%), and cardiovascular disorders were common (14 myocardial infarction, 22 arrhythmia, 14 congestive heart failure, 19 angina pectoris, 20 hypertension, 24 hypotension, and 2 sudden death). There were 5 deaths; 3 were from myocardial infarction and 2 were sudden. In the medical treatment arm, complications occurred with approximately one third the frequency (3.4% versus 10.0%). Predictors of medical complications in the CEA group were history of myocardial infarction or angina and hypertension. Overall, there was a low rate of serious perioperative medical complications, and perioperative myocardial infarction (1.0%) was uncommon. It must be kept in mind, however, that patients with recent history of myocardial infarction, unstable angina pectoris, or recent congestive heart failure were excluded from NASCET.
The NASCET surgical results study detailed rates of
perioperative stroke and death at 30 days, stroke
severity at 90 days, variables that influenced
perioperative risk, and durability of
CEA.28 Among 1415 patients, there were 92
perioperative outcome events, for an overall rate of
6.5% (death, 1.0%, 1.1%; disabling stroke, 1.8%, 0.9%; and
nondisabling stroke, 3.7%, 4.5% at 30 and 90 days, respectively).
Approximately one third of the perioperative events
occurred intraoperatively, and most strokes were attributed to
thromboembolism. Baseline predictors of surgical risk were hemispheric
TIA as qualifying event (versus retinal TIA), left-sided procedure,
contralateral carotid occlusion, ipsilateral ischemic lesion on
CT scan, and irregular or ulcerated ipsilateral plaque. History of
coronary artery disease with prior cardiac procedure was a
protective factor. Other baseline variables that did not increase
or decrease perioperative risk were age (<65 versus
65 years); sex; hemispheric TIA versus stroke; surgery within 30 days
of the qualifying event; hyperlipidemia; intraoperative
monitoring, a shunt, or both; and geographic location or type of
surgeon (neurosurgeon versus vascular surgeon). These 2
studies27 28 sharpen our focus for
perioperative medical and surgical complications
associated with CEA and their predictors.
What possible effect has recent publication of these major CEA trials had on the rate of CEA? As mentioned previously, the substantial increase in use of CEA between 1980 and 1985 was followed by a marked fall between 1985 and 1988 and a plateau effect until approximately 1992.29 30 31 32 33 Rates of CEA, however, increased sharply after publication of ECST and NASCET in 1991.29 30 There have also been reports of substantial increase in the number of CEAs since the ACAS Clinical Advisory in September 1994.34 35 Asymptomatic patients constitute approximately 40% to 50% of those who undergo CEA.34 36 37 Women and blacks have considerably lower rates.29
With the performance of CEA on the rise again, we are reminded of the potential for inappropriate use36 37 38 39 and community complication rates40 41 42 43 44 45 that may exceed acceptable levels established in recent clinical trials.16 17 18 19 20 21 22 23 24 Factors such as perceived availability of cerebral angiography and physician specialty may influence the frequency of use of CEA.46 Lack of use or inappropriate use may be partially addressed through targeted educational programs. Of concern, however, are past reports of substantially higher perioperative complication rates in the community than reported in recent trials.40 Combined morbidity and mortality that exceed the expected 3% for patients with asymptomatic carotid stenosis and 6% for patients with symptomatic stenosis could eliminate the benefit gained from surgery.40 This has led to a call to action for mandatory audits of surgical complications.38 40 Physicians who refer patients for CEA must be knowledgeable of acceptable surgical morbidity and mortality rates. Factual, objective information on complications should be recorded for individual surgeons, who can then share these results with their patients and local physicians.
Surgical skill is critical to the success of CEA. High-volume
surgeons and hospitals seem to have better results.42 43
Furthermore, as shown in this issue of Stroke by the NASCET
collaborators27 28 and by others,47 48
complications of CEA may be predicted by clinical data that may be used
to stratify operative risk.47 48 Risk factors for
postoperative complications in asymptomatic carotid
stenosis, for example, may include age
75 years, history of
congestive heart failure, combined CEA and coronary surgery,
and sex (women).49
So then, where do we draw the line? I am convinced that medical therapy (ie, antiplatelet agents and cardiovascular risk factor control/prevention50 ) is the treatment of choice for those with symptomatic (TIA or minor stroke) lower-grade carotid stenosis (<50% by NASCET criteria). For high-grade symptomatic carotid stenosis (70% to 99% by NASCET criteria), CEA plus antiplatelet therapy and cardiovascular risk factor control/prevention is the treatment of choice. Those with moderate-grade symptomatic carotid stenosis (50% to 69% by NASCET criteria) walk a finer line. Absolute risk reductions for CEA in this case are more modest than for high-grade carotid stenosis. Thus, as major surgical perioperative complication rates exceed 6%, any benefit gained from surgery is soon lost. Chassin30 estimates that for every increase in 2 percentage points in the CEA complication rate above 6% to 7%, the 5-year benefit will be reduced by 20%.This is in contradistinction to high-grade carotid stenosis (70% to 99%), in which a modest increase in surgical morbidity and mortality may not negate benefit from CEA. In addition, before committing to CEA for symptomatic carotid stenosis, one must consider the patient's risk factor profile and comorbid conditions. As shown by the NASCET investigators, such factors as contralateral carotid occlusion, appropriate lesion on CT scan, history of diabetes mellitus, and elevated diastolic blood pressure (>90 mm Hg) may double the perioperative risk of stroke or death, whereas men and those with recent stroke or hemispheric symptoms may derive greater long-term benefit from CEA.18 These criteria may be useful to decide whether CEA should be recommended for those with 50% to 69% symptomatic stenosis.
The patient with
60% asymptomatic carotid
stenosis (NASCET criteria) walks the finest line. There is
little room for additional major surgical morbidity and mortality
beyond the 3% standard, in view of the modest benefit of surgery. An
increase of 2 percentage points in the complication rate could reduce
the surgical benefit by
30%30 and render CEA
ineffective and possibly harmful. If CEA for asymptomatic
carotid stenosis is performed at higher than the accepted 3%
surgical morbidity and mortality rate, as could potentially occur in
the community on the basis of past studies,51 52 the
public health impact in terms of dollars spent and morbidity and
mortality could be substantial. Thus, with approximately one half of an
estimated 130 000 CEAs30 performed in
asymptomatic patients, wholesale application of CEA in less
than optimal surgical circumstances could have adverse implications.
Concerns over reproducibility of low clinical trial surgical morbidity
by community surgeons and the ACAS finding that CEA may not reduce the
risk of major stroke have led to nonuniform support of CEA in
asymptomatic patients.52 There is a need to
replicate the
60% asymptomatic carotid stenosis
cut point50 and to have additional information about CEA
in women, especially perioperative complication rates.
We await results of the Asymptomatic Carotid Surgery Trial
(ACST) for confirmatory evidence.53
Mass screening for high-grade asymptomatic carotid artery stenosis is not recommended because it is not cost-effective.52 54 The prevalence of higher grades of asymptomatic carotid stenosis is low in the population at large, screening tests may lead to false-positive and false-negative results, and conventional cerebral angiography is associated with risks. Furthermore, the application of noninvasive blood flow measures to supplant conventional cerebral angiography remains a focus of debate that is unsettled.55 56 I continue to recommend cerebral angiography for patients who are being evaluated for CEA at centers with competent and experienced angiographers. Cerebral angiography remains the standard by which we determine degree of carotid stenosis for CEA. Further improvement in noninvasive carotid blood flow and imaging techniques to overcome inaccuracies and supplant conventional angiography with its inherent risks is a welcome advance that will lead to appropriate use of CEA with lower morbidity and mortality.
| Summary |
|---|
|
|
|---|
|
CEA is likely to continue to stir controversy and debate. Some of this will abate as we dynamically build on and refine important CEA guidelines59 and continue to perform important CEA studies that may prove contrary to hypothetical beliefs.60 With open minds, we will be able to advance the science and practice of CEA.
| Footnotes |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
V. L. Wright, W. Olan, B. Dick, H. Yu, N. Alberts-Grill, L. L. Latour, and A. E. Baird Assessment of CE-MRA for the rapid detection of supra-aortic vascular disease Neurology, July 12, 2005; 65(1): 27 - 32. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. E. Hyde, A. J. Fox, I. Gulka, P. Kalapos, D. H. Lee, D. M. Pelz, and D. W. Holdsworth Internal Carotid Artery Stenosis Measurement: Comparison of 3D Computed Rotational Angiography and Conventional Digital Subtraction Angiography Stroke, December 1, 2004; 35(12): 2776 - 2781. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. O'Rourke, N. Dean, N. Akhtar, and A. Shuaib Current and future concepts in stroke prevention Can. Med. Assoc. J., March 30, 2004; 170(7): 1123 - 1133. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. Linfante, J. A. Hirsch, M. Selim, G. Schlaug, L. R. Caplan, and A. S. Reddy Safety of Latest-Generation Self-expanding Stents in Patients With NASCET-Ineligible Severe Symptomatic Extracranial Internal Carotid Artery Stenosis Arch Neurol, January 1, 2004; 61(1): 39 - 43. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. M. Rothwell, S. A. Gutnikov, and C. P. Warlow Reanalysis of the Final Results of the European Carotid Surgery Trial Stroke, February 1, 2003; 34(2): 514 - 523. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Barth and C. Bassetti Patient Selection for Carotid Endarterectomy: How Far Is Risk Modeling Applicable to the Individual? Stroke, February 1, 2003; 34(2): 524 - 527. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. A. Mansour Carotid Endarterectomy in Elderly Patients--Invited Critique Arch Surg, November 1, 2002; 137(11): 1288 - 1288. [Full Text] [PDF] |
||||
![]() |
P. B. Gorelick Stroke Prevention Therapy Beyond Antithrombotics: Unifying Mechanisms in Ischemic Stroke Pathogenesis and Implications for Therapy: An Invited Review Stroke, March 1, 2002; 33(3): 862 - 875. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. J. Powers Oral Anticoagulant Therapy for the Prevention of Stroke N. Engl. J. Med., November 15, 2001; 345(20): 1493 - 1495. [Full Text] [PDF] |
||||
![]() |
V. M. Rao, L. Parker, D. C. Levin, J. Sunshine, and G. Bushee Use Trends and Geographic Variation in Neuroimaging: Nationwide Medicare Data for 1993 and 1998 AJNR Am. J. Neuroradiol., October 1, 2001; 22(9): 1643 - 1649. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. M. Malek, R. T. Higashida, C. C. Phatouros, T. E. Lempert, P. M. Meyers, W. S. Smith, C. F. Dowd, and V. V. Halbach Stent Angioplasty for Cervical Carotid Artery Stenosis in High-Risk Symptomatic NASCET-Ineligible Patients Stroke, December 1, 2000; 31(12): 3029 - 3033. [Abstract] [Full Text] [PDF] |
||||
![]() |
Carotid angioplasty and stenting: current status Can. Med. Assoc. J., May 1, 2000; 162(10): 1451 - 1454. |
||||
![]() |
Complications of Carotid Endarterectomy Journal Watch (General), October 15, 1999; 1999(1015): 7 - 7. [Full Text] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Stroke Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 1999 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |