2 - Crest 2010

new england
The
journal of medicine
established in 1812 july 1, 2010 vol. 363 no. 1
Stenting versus Endarterectomy for Treatment

of Carotid-Artery Stenosis
Thomas G. Brott, M.D., Robert W. Hobson, II, M.D.,* George Howard, Dr.P.H., Gary S. Roubin, M.D., Ph.D.,
Wayne M. Clark, M.D., William Brooks, M.D., Ariane Mackey, M.D., Michael D. Hill, M.D., Pierre P. Leimgruber, M.D.,
Alice J. Sheffet, Ph.D., Virginia J. Howard, Ph.D., Wesley S. Moore, M.D., Jenifer H. Voeks, Ph.D.,
L. Nelson Hopkins, M.D., Donald E. Cutlip, M.D., David J. Cohen, M.D., Jeffrey J. Popma, M.D.,
Robert D. Ferguson, M.D., Stanley N. Cohen, M.D., Joseph L. Blackshear, M.D., Frank L. Silver, M.D.,
J.P. Mohr, M.D., Brajesh K. Lal, M.D., and James F. Meschia, M.D., for the CREST Investigators†
A BS T R AC T
BACKGROUND From Mayo Clinic, Jacksonville, FL (T.G.B.,

Carotid-artery stenting and carotid endarterectomy are both options for treating J.F.M., J.L.B.); University of Medicine and
Dentistry of New Jersey, Newark (T.G.B.,
carotid-artery stenosis, an important cause of stroke.
R.W.H., A.J.S.); University of Alabama at Bir-
METHODS mingham, Birmingham (G.H., V.J.H., J.H.V.);
Lenox Hill Hospital (G.S.R.) and Columbia
We randomly assigned patients with symptomatic or asymptomatic carotid stenosis University Medical Center (J.P.M.) — both
to undergo carotid-artery stenting or carotid endarterectomy. The primary compos- in New York; Oregon Health and Science
ite end point was stroke, myocardial infarction, or death from any cause during the University, Portland (W.M.C.); Central Bap-
tist Hospital, Lexington, KY (W.B.); Centre
periprocedural period or any ipsilateral stroke within 4 years after randomization. Hospitalier Affilié Hôpital de l’Enfant-Jésus,
Quebec City, QC (A.M.), University of Cal-
RESULTS gary, Calgary, AB (M.D.H.), and University of
For 2502 patients over a median follow-up period of 2.5 years, there was no significant Toronto, Toronto (F.L.S.) — all in Canada;
difference in the estimated 4-year rates of the primary end point between the stenting Spokane Cardiology Heart and Vascular
Health, Spokane, WA (P.P.L.); UCLA, Los
group and the endarterectomy group (7.2% and 6.8%, respectively; hazard ratio with Angeles (W.S.M.); University at Buffalo,
stenting, 1.11; 95% confidence interval, 0.81 to 1.51; P = 0.51). There was no differen- State University of New York, Buffalo
tial treatment effect with regard to the primary end point according to symptomatic (L.N.H.); Harvard Clinical Research Institute
(D.E.C.) and Beth Israel Deaconess Medical
status (P = 0.84) or sex (P = 0.34). The 4-year rate of stroke or death was 6.4% with stent- Center, Harvard Medical School (J.J.P.) —
ing and 4.7% with endarterectomy (hazard ratio, 1.50; P = 0.03); the rates among symp- both in Boston; Saint Luke’s Mid America
tomatic patients were 8.0% and 6.4% (hazard ratio, 1.37; P = 0.14), and the rates among Heart Institute, Kansas City, MO (D.J.C.);
MetroHealth Medical Center, Cleveland
asymptomatic patients were 4.5% and 2.7% (hazard ratio, 1.86; P = 0.07), respectively. (R.D.F.); Nevada Neuroscience Institute,
Periprocedural rates of individual components of the end points differed between the Las Vegas (S.N.C.); and University of Mary-
stenting group and the endarterectomy group: for death (0.7% vs. 0.3%, P = 0.18), for land Medical Center, Baltimore (B.K.L.). Ad-
dress reprint requests to Dr. Brott at Mayo
stroke (4.1% vs. 2.3%, P = 0.01), and for myocardial infarction (1.1% vs. 2.3%, P = 0.03). Clinic, Griffin 304, 4500 San Pablo Rd., Jack-
After this period, the incidences of ipsilateral stroke with stenting and with endar- sonville, FL 32224, or at brott.thomas@
terectomy were similarly low (2.0% and 2.4%, respectively; P = 0.85). mayo.edu.
*Deceased.
CONCLUSIONS †The Carotid Revascularization Endarter
Among patients with symptomatic or asymptomatic carotid stenosis, the risk of the ectomy vs. Stenting Trial (CREST) inves-
tigators and committee members are
composite primary outcome of stroke, myocardial infarction, or death did not differ listed in the Appendix.
significantly in the group undergoing carotid-artery stenting and the group undergo- This article (10.1056/NEJMoa0912321) was
ing carotid endarterectomy. During the periprocedural period, there was a higher risk published on May 26, 2010, and was last
of stroke with stenting and a higher risk of myocardial infarction with endarterec- updated on July 28, 2010, at NEJM.org.
tomy. (ClinicalTrials.gov number, NCT00004732.) N Engl J Med 2010;363:11-23.
Copyright © 2010 Massachusetts Medical Society.
n engl j med 363;1 nejm.org july 1, 2010 11

The New England Journal of Medicine
Downloaded from nejm.org at LOMA LINDA UNIV LIBRARY on October 25, 2023. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e
C
arotid-artery atherosclerosis is an means of a validated selection process8 document-
important cause of ischemic stroke.1 Ca- ing that they performed more than 12 procedures
rotid endarterectomy has been established per year and that the rates of complications and
as effective treatment for both symptomatic pa- death were less than 3% among asymptomatic pa-
tients and asymptomatic patients.2-4 Carotid-artery tients and less than 5% among symptomatic pa-
stenting is another option for treatment. The re- tients. The 224 interventionists were certified af-
sults of randomized trials comparing carotid-artery ter satisfactory evaluation of their endovascular
stenting and carotid endarterectomy for use in experience, carotid-stenting results, participation
symptomatic patients are conflicting.5-7 The pri- in hands-on training, and participation in a lead-
mary aim of the Carotid Revascularization Endar- in phase of training (see additional details in the
terectomy vs. Stenting Trial (CREST) was to com- Supplementary Appendix).9
pare the outcomes of carotid-artery stenting with
those of carotid endarterectomy among patients Selection of Study Patients
with symptomatic or asymptomatic extracranial Patients were considered to be symptomatic if they
carotid stenosis. had had a transient ischemic attack, amaurosis fu-
gax, or minor nondisabling stroke involving the
Me thods study carotid artery within 180 days before ran-
domization. Eligibility criteria were stenosis of 50%
Study Design or more on angiography, 70% or more on ultra-
CREST is a randomized, controlled trial with blind- sonography, or 70% or more on computed tomo-
ed end-point adjudication. Ethics review boards graphic angiography or magnetic resonance an-
at all participating centers approved the protocol. giography if the stenosis on ultrasonography was
All patients provided written informed consent. 50 to 69%. Eligibility was extended in 2005 to in-
The authors designed the study, gathered and ana- clude asymptomatic patients, for whom the crite-
lyzed the data, wrote the manuscript, made the ria were stenosis of 60% or more on angiography,
decision to publish the findings, vouch for the 70% or more on ultrasonography, or 80% or more
completeness and accuracy of the data, and attest on computed tomographic angiography or mag-
to the fidelity of the report to the study protocol. netic resonance angiography if the stenosis on ul-
Abbott Vascular Solutions (formerly Guidant) do- trasonography was 50 to 69%. Patients were ex-
nated the Accunet and Acculink systems to all cluded if they had had a previous stroke that was
CREST centers in Canada and to CREST centers in sufficiently severe to confound the assessment of
the United States that were at Veterans Affairs end points or if they had chronic atrial fibrilla-
sites. Abbott has a nonvoting seat on the CREST tion, paroxysmal atrial fibrillation that had oc-
Executive Committee and reviewed the final curred within the preceding 6 months or that
draft of the manuscript before submission. Ab- necessitated anticoagulation therapy, myocardial
bott assisted with CREST site monitoring and infarction within the previous 30 days, or unsta-
was responsible for all site monitoring of the ble angina. Additional eligibility criteria were clin-
Canadian centers. The full protocol and statisti- ical and anatomical suitability, before random-
cal analysis plan can be found in the Supple- ization, for management by means of either of the
mentary Appendix with the full text of this ar- study revascularization techniques. The full eli-
ticle at NEJM.org. gibility criteria have been published elsewhere.10
Centers and Investigators Randomization

We enrolled patients at 108 centers in the United Eligible patients were randomly assigned, with the
States and 9 in Canada. Centers were required to use of a Web-based system, to undergo either ca-
have a team consisting of a neurologist, an inter- rotid-artery stenting or carotid endarterectomy.
ventionist, a surgeon, and a research coordinator. Randomization was based on a permuted-block
Patients could not be randomly assigned to a treat- design (with random block sizes of 2, 4, or 6),
ment group until the operators performing carotid- was stratified according to center and symptom-
artery stenting and carotid endarterectomy had atic status, and was performed after the patient,
been certified. Certification was achieved by 477 surgeon, and interventionist could arrange for a
surgeons, whose clinical results were audited by date for the procedure within 2 weeks.
12 n engl j med 363;1 nejm.org july 1, 2010

Stenting vs. Endarterectomy for Carotid-Artery Stenosis
2522 Patients underwent randomization
1271 Were assigned to CAS 1251 Were assigned to CEA

1144 Underwent assigned intervention 1194 Underwent assigned intervention
127 Did not undergo assigned intervention 57 Did not undergo assigned intervention
8 Had CAS attempted, but underwent CEA 13 Underwent CAS
5 Had severe vascular tortuosity 5 Declined assigned treatment
or inappropriate anatomy 6 Had other medical reason
3 Had other vessel characteristics 2 Had unknown reason
65 Underwent CEA 44 Did not undergo CAS or CEA
18 Had severe vascular tortuosity 21 Declined assigned treatment
or inappropriate anatomy 10 Had other medical reason
13 Had other vessel characteristics 4 Had stenosis <50% (if symptomatic)
7 Had occlusion or “string sign” >1 cm or <60% (if asymptomatic)
in length 3 Had unfavorable lesion characteristics
6 Declined assigned treatment 2 Had occlusion or “string sign” >1 cm
2 Had stenosis <50% (if symptomatic) in length
or <60% (if asymptomatic) 4 Had unknown reason
5 Had other medical reason
14 Had unknown reason
54 Did not undergo CAS or CEA
4 Had severe vascular tortuosity or
inappropriate anatomy
1 Had abnormal angiographic findings
6 Had occlusion or “string sign” >1 cm
in length
16 Declined assigned treatment
2 Had renal impairment
20 Had stenosis <50% (if symptomatic)
or <60% (if asymptomatic)
5 Had unknown reason
1262 Had data included in analysis 1240 Had data included in analysis
9 Had data excluded from analysis 11 Had data excluded from analysis
owing to scientific misconduct owing to scientific misconduct
69 Discontinued follow-up 111 Discontinued follow-up

36 Withdrew consent 64 Withdrew consent
33 Were lost to follow-up 47 Were lost to follow-up
Figure 1. Randomization and Follow-up of the Study Patients.

Patients were assessed for eligibility before randomization, but the number of patients assessed is not available,
because screening logs were not maintained. CAS denotes carotid-artery stenting, and CEA carotid endarterectomy.
Stenting and Endarterectomy was scheduled for within 48 hours after random-
Carotid-artery stenting and carotid endarterectomy ization, 650 mg of aspirin and 450 mg of clopid
were performed according to published guide- ogrel were given 4 or more hours before the pro-
lines.9,11,12 For carotid-artery stenting, the proto- cedure. After the procedure, patients received one
col specified use of the RX Acculink stent and, or two 325-mg doses of aspirin daily for 30 days
whenever feasible, the RX Accunet embolic-pro- and either clopidogrel, 75 mg daily, or ticlopidine,
tection device. At least 48 hours before carotid- 250 mg twice daily, for 4 weeks. The continuation
artery stenting, patients received aspirin, at a dose of antiplatelet therapy for more than 4 weeks after
of 325 mg twice daily, and clopidogrel at a dose the procedure was recommended for all patients
of 75 mg twice daily. When carotid-artery stenting who had undergone carotid-artery stenting. At

least 48 hours before carotid endarterectomy, pa- NIHSS is a 15-item neurologic-impairment scale,
tients received 325 mg of aspirin daily and con- with possible scores ranging from 0 (no deficit)
tinued to receive that dose for a year or more. Al- to 42 (quadriplegia and coma). The modified
ternatives to this regimen were ticlopidine given Rankin scale is a disability scale on which scores
at a dose of 250 mg twice daily, clopidogrel at a can range from 0 (no symptoms) to 6 (death).
dose of 75 mg daily, aspirin at a dose of 81 mg The TIA–Stroke Questionnaire consists of a ques-
daily, or aspirin and extended-release dipyridamole tion about whether there is a history of TIA, one
twice daily (Table 1 in the Supplementary Appen- about whether there is a history of stroke, and six
dix). Patients undergoing either study procedure questions about whether there was a sudden on-
received medical therapy that was consistent with set of any of various focal neurologic symptoms
the current standard of care, including treatment consistent with TIA or stroke.
of hypertension and hyperlipidemia. Cardiac-enzyme levels were measured before
the study procedure and 6 to 8 hours after the
Follow-up Assessments of End Points procedure. Electrocardiography (ECG) was per-
Neurologic evaluation was performed at baseline formed before stenting or endarterectomy, as well
and 18 to 54 hours after the study procedure, as 6 to 48 hours and 1 month afterward. Carotid
1 month afterward, and every 6 months thereafter. ultrasonography was performed before the study
The evaluation consisted of the use of the National procedure; 1, 6, and 12 months afterward; and
Institutes of Health (NIH) Stroke Scale (NIHSS),13 annually thereafter.10 A follow-up telephone inter-
the modified Rankin scale, and the Transient view, including administration of the TIA–Stroke
Ischemic Attack (TIA)–Stroke Questionnaire.14 The Questionnaire,14 was conducted at 3 months and
Table 1. Baseline Characteristics of the Study Population, According to Treatment Group.*
Carotid-Artery Stenting Carotid Endarterectomy

Characteristic (N = 1262) (N = 1240)
Age (yr) 68.9±9.0 69.2±8.7
Male sex (% of patients) 63.9 66.4
White race (% of patients)† 92.9 93.5
Asymptomatic arteries (% of patients) 47.1 47.3
Risk factors (% of patients)
Hypertension 85.8 86.1
Diabetes 30.6 30.4
Dyslipidemia‡ 82.9 85.8
Current smoker 26.4 26.1
Previous cardiovascular disease 42.4 45.0
Previous coronary-artery bypass 19.9 21.5
Blood pressure (mm Hg)
Systolic 141.6±20.2 141.2±20.5
Diastolic 74.0±11.6 73.9±11.5
Percent stenosis at randomization
Moderate (<70%) 13.1 14.9
Severe (≥70%) 86.9 85.1
Stenosis characteristics (% of patients)
Left carotid artery treated 50.6 52.3
Contralateral occlusion 2.7 3.2
Time from randomization to treatment (no. of days)
Median 6 7
Interquartile range 9 9

Table 1. (Continued.)
Carotid-Artery Stenting Carotid Endarterectomy

Characteristic (N = 1262) (N = 1240)
Procedural characteristics
Target-lesion length (mm) 17.8±8.5 —
Total length of stented segment (mm) 34.4±7.3 —
Balloon angioplasty before stenting (% of patients) 67.7 —
Embolic protection (% of patients) 96.1 —
Medical treatment (% of patients)
Antiplatelet therapy 48 hr before stenting 97.7 —
During procedure
Heparin 86.4 —
Bivalirudin 13.6 —
Vasopressors 29.9 —
After procedure
Any antiplatelet therapy 99.0 —
Aspirin plus either clopidogrel or ticlopidine for 4 wk 87.9
General anesthesia (% of patients) — 90.0
Surgical technique (% of patients)
Patch — 62.4
Shunt — 56.7
Medical treatment (% of patients)
Aspirin 48 hr before endarterectomy — 92.1
Vasopressors during endarterectomy — 61.0
Antiplatelet therapy after endarterectomy — 91.1
* Plus–minus values are means ±SD. Dashes indicate not applicable.

† Race was self-reported.
‡ P = 0.05 for the difference in the baseline rate of dyslipidemia between the two groups.
every 6 months thereafter. General health status 30 days after the procedure. When the procedure
was assessed at baseline, at 2 weeks, 1 month after was not performed within 30 days after random-
the procedure, and 1 year after randomization, ization, the periprocedural period was defined as
with the use of the Medical Outcomes Study 36- the period from randomization through 36 days
Item Short-Form Health Survey (SF-36), which after randomization. Study committees unaware
evaluates eight dimensions of health, with scores of the treatment assignments adjudicated stroke
for each ranging from 0 to 100, and higher scores and myocardial infarction.
indicating better health status. The SF-36 provides Stroke was defined as an acute neurologic event
summary scales for overall physical and mental with focal symptoms and signs, lasting for 24
health, with norm-based standardization of the hours or more, that were consistent with focal
scores to a mean of 50 and a standard deviation cerebral ischemia. The adjudication process was
of 10.15,16 initiated after a clinically significant neurologic
The primary end point was the composite of event, any positive response on the TIA–Stroke
any stroke, myocardial infarction, or death dur- Questionnaire, or an increase by 2 points or more
ing the periprocedural period or ipsilateral stroke in the NIHSS score. Stroke was defined as major
within 4 years after randomization. When the pro- stroke on the basis of clinical data or if the NIHSS
cedure was performed within 30 days after ran- score was 9 or higher 90 days after the procedure.
domization, the periprocedural period was de- Myocardial infarction was defined by a creatine
fined as the period from randomization through kinase MB or troponin level that was twice the up-

Table 2. Primary End Point, Components of the Primary End Point, and Other Events, According to Treatment Group.*
End Point Periprocedural Period

Absolute Treatment Hazard Ratio for
Effect of CAS vs. CEA CAS vs. CEA
CAS (N = 1262) CEA (N = 1240) (95% CI) (95% CI) P Value
no. of patients (% ±SE) percentage points
Death 9 (0.7±0.2) 4 (0.3±0.2) 0.4 (−0.2 to 1.0) 2.25 (0.69 to 7.30)† 0.18†
Stroke
Any 52 (4.1±0.6) 29 (2.3±0.4) 1.8 (0.4 to 3.2) 1.79 (1.14 to 2.82) 0.01
Major ipsilateral 11 (0.9±0.3) 4 (0.3±0.2) 0.5 (−0.1 to 1.2) 2.67 (0.85 to 8.40) 0.09
Major nonipsilateral‡ 0 4 (0.3±0.2) NA NA NA
Minor ipsilateral 37 (2.9±0.5) 17 (1.4±0.3) 1.6 (0.4 to 2.7) 2.16 (1.22 to 3.83) 0.009
Minor nonipsilateral 4 (0.3±0.2) 4 (0.3±0.2) 0.0 (−0.4 to 0.4) 1.02 (0.25 to 4.07) 0.98†
Myocardial infarction 14 (1.1±0.3) 28 (2.3±0.4) −1.1 (−2.2 to −0.1) 0.50 (0.26 to 0.94) 0.03
Any periprocedural stroke or postprocedural 52 (4.1±0.6) 29 (2.3±0.4) 1.8 (0.4 to 3.2) 1.79 (1.14 to 2.82) 0.01
ipsilateral stroke
Major stroke 11 (0.9±0.3) 8 (0.6±0.2) 0.2 (−0.5 to 0.9) 1.35 (0.54 to 3.36) 0.52
Minor stroke 41 (3.2±0.5) 21 (1.7±0.4) 1.6 (0.3 to 2.8) 1.95 (1.15 to 3.30) 0.01
Any periprocedural stroke or death or post- 55 (4.4±0.6) 29 (2.3±0.4) 2.0 (0.6 to 3.4) 1.90 (1.21 to 2.98) 0.005
procedural ipsilateral stroke
Primary end point (any periprocedural stroke, 66 (5.2±0.6) 56 (4.5±0.6) 0.7 (−1.0 to 2.4) 1.18 (0.82 to 1.68) 0.38
myocardial infarction, or death or
postprocedural ipsilateral stroke)
* The periprocedural period was defined, according to the study protocol, as the 30-day period after the procedure (for all patients who under-
went the assigned procedure within 30 days after randomization) or the 36-day period after randomization (for all patients who did not un-
dergo the assigned procedure within 30 days after randomization). Because the periprocedural period was relatively short, which minimized
the need for censoring, event proportions and the absolute differences in event proportions were calculated as the percentage of patients
with events. For the 4-year study period, proportions reflecting the absolute efficacy of carotid-artery stenting (CAS) over that of carotid end
arterectomy (CEA) were based on Kaplan–Meier survival estimates at the end of the 4 years. Hazard ratios for the periprocedural period
were based on data for all patients, censored at the end of the periprocedural period. All hazard ratios were adjusted for age, symptomatic
status, and sex. P values were calculated on the basis of the significance of the hazard ratio (per study protocol). For death, stroke, and
myocardial infarction end points, patients may have had more than one event (e.g., fatal stroke was counted as both a death and a stroke,
and patients may have had an ipsilateral stroke followed by a nonipsilateral stroke).
† Because of the small number of events, a univariate proportional-hazards model was used to estimate the hazard ratio for death during the
periprocedural period, the P value for minor nonipsilateral stroke during the periprocedural period, and the P value for major nonipsilateral
stroke during the 4-year study period.
‡ Absolute treatment effect, hazard ratio, and P value for major nonipsilateral stroke were not available (NA) because of the small number of
events, resulting in unreliable estimates.
per limit of the normal range or higher according end point of less than 0.54 or more than 1.49 with
to the center’s laboratory, in addition to either stenting as compared with endarterectomy, ap-
chest pain or symptoms consistent with ischemia proximating an absolute difference of 1.2 percent-
or ECG evidence of ischemia, including new ST- age points per year in the rate of the primary end
segment depression or elevation of more than point between the two treatment groups. Inten-
1 mm in two or more contiguous leads according tion-to-treat survival analysis was used, and Kap
to the core laboratory.17 lan–Meier survival curves were plotted. Two in-
terim analyses were performed with the use of
Statistical Analyses O’Brien–Fleming boundaries,18 the first after ap-
Analyses were aimed at testing for superiority. The proximately one fifth of the patients had been re-
null hypothesis was that the two study treatments cruited, and the second after approximately half
are equivalent; the alternative hypothesis was that the patients had been recruited. Multiple-imputa-
the treatments differ. A sample size of 2500 pa- tion techniques19 were used to assess bias from
tients was selected to provide a statistical power differential rates of withdrawal from the study in
of 90% to detect a hazard ratio for the primary the two groups.

4-Yr Study Period (Including Periprocedural Period)

Absolute Treatment Hazard Ratio for
Effect of CAS vs. CEA CAS vs. CEA
CAS (N = 1262) CEA (N = 1240) (95% CI) (95% CI) P Value
no. of patients (% ±SE) percentage points
94 (11.3±1.2) 83 (12.6±1.5) −1.3 (−5.1 to 2.5) 1.12 (0.83 to 1.51) 0.45
105 (10.2±1.1) 75 (7.9±1.0) 2.3 (−0.6 to 5.2) 1.40 (1.04 to 1.89) 0.03
16 (1.4±0.3) 6 (0.5±0.2) 0.8 (0.1 to 1.6) 2.56 (1.00 to 6.54) 0.05
6 (0.9±0.4) 8 (0.8±0.3) 0.1 (−0.9 to 1.1) 0.73 (0.25 to 2.11) 0.56†
52 (4.5±0.6) 36 (3.5±0.6) 1.0 (−0.7 to 2.7) 1.43 (0.94 to 2.19) 0.10
33 (4.0±0.8) 29 (3.8±0.9) 0.2 (−2.1 to 2.4) 1.11 (0.67 to 1.82) 0.69
72 (6.2±0.7) 50 (4.7±0.7) 1.5 (−0.4 to 3.4) 1.44 (1.00 to 2.06) 0.049
16 (1.4±0.3) 10 (0.8±0.3) 0.6 (−0.2 to 1.4) 1.55 (0.70 to 3.42) 0.28

56 (4.8±0.6) 40 (3.8±0.6) 1.0 (−0.8 to 2.7) 1.39 (0.93 to 2.09) 0.11
75 (6.4±0.7) 50 (4.7±0.7) 1.7 (−0.2 to 3.7) 1.50 (1.05 to 2.15) 0.03
85 (7.2±0.8) 76 (6.8±0.8) 0.4 (−1.7 to 2.6) 1.11 (0.81 to 1.51) 0.51
Secondary aims included estimating the mod- endarterectomy, 64 (5.1%) withdrew consent, 13
ification of the treatment effect by symptomatic (1.0%) underwent carotid-artery stenting, and 47
status, sex, and age, which were assessed through (3.8%) were lost to follow-up.
inclusion of the interaction terms in the propor- Quality-control and site-monitoring activities
tional-hazards models (as a single indicator vari- resulted in the detection of irregular data from one
able for sex and symptomatic status and a linear center. The principal investigator and the Office
term for age). The analyses of age were planned of Research Integrity of the Department of Health
before data analysis began but were not described and Human Services were notified and subse-
in the study protocol. Longitudinal random-effect quently determined that some data were fabricat-
growth-curve models20 were used to evaluate the ed. All data from this center (which had enrolled
effect of periprocedural events on health status 9 patients undergoing carotid-artery stenting and
at 1 year, as assessed with the use of the SF-36 11 undergoing carotid endarterectomy) were ex-
physical and mental health scales. These models cluded before any analyses were performed, result-
were adjusted for symptomatic status, sex, age, ing in a cohort of 2502 patients for all analyses.
and baseline health status. Dyslipidemia was more common among pa-
tients in the endarterectomy group than among
R e sult s those in the stenting group (85.8% vs. 82.9%,
P = 0.048), both groups had high rates of vascular
Study Population and Treatments risk factors, and more than 80% of patients had
From December 2000 through July 2008, a total severe stenosis (Table 1). Baseline characteristics
of 2522 patients were randomly assigned to one are reported according to symptomatic status in
of the two treatments (Fig. 1). After randomiza- Tables 2 and 3 in the Supplementary Appendix.
tion, among the 1271 patients randomly assigned The median time from randomization to the
to undergo carotid-artery stenting, 36 (2.8%) with- procedure was 6 days for carotid-artery stenting
drew consent, 73 (5.7%) underwent carotid end and 7 days for carotid endarterectomy. Stenting
arterectomy, and 33 (2.6%) were lost to follow- was performed with embolic protection in 96.1%
up; among the 1251 patients assigned to carotid of patients assigned to the stenting group, and

endarterectomy was performed with the use of tween the stenting group and the endarterectomy
general anesthesia in 90.0% of patients assigned group among symptomatic patients (6.7% vs. 5.4%;
to the endarterectomy group. The median duration hazard ratio for stenting, 1.26; 95% CI, 0.81 to
of follow-up was 2.5 years. During that time, the 1.96) or among asymptomatic patients (3.5% vs.
level or prevalence of selected risk factors remained 3.6%; hazard ratio, 1.02; 95% CI, 0.55 to 1.86) (Ta-
similar between the two treatment groups, except ble 3). Cranial-nerve palsies were less frequent dur-
for current smoking, the prevalence of which was ing the periprocedural period with carotid-artery
similar at baseline (26.4% with stenting and 26.1% stenting (0.3%, vs. 4.7% with carotid endarterec-
with endarterectomy) but for which differences tomy; hazard ratio, 0.07; 95% CI, 0.02 to 0.18).
developed during follow-up (21.8% with stenting Additional serious adverse events are detailed in
vs. 13.8% with endarterectomy, P = 0.03) (Table 4 Table 6 in the Supplementary Appendix. The 4-year
in the Supplementary Appendix). rate of stroke or death was 6.4% in the stenting
group as compared with 4.7% in the endarterec-
Primary End Point tomy group (hazard ratio, 1.50; 95% CI, 1.05 to
There was no significant difference in the estimat- 2.15; P = 0.03); the respective rates were 8.0% and
ed 4-year rates of the primary end point between 6.4% among symptomatic patients (hazard ratio,
carotid-artery stenting and carotid endarterecto- 1.37; 95% CI, 0.90 to 2.09; P = 0.14) and 4.5% and
my (7.2% and 6.8%, respectively; hazard ratio for 2.7% among asymptomatic patients (hazard ratio,
stenting, 1.11; 95% confidence interval [CI], 0.81 1.86; 95% CI, 0.95 to 3.66; P = 0.07).
to 1.51; P = 0.51) (Table 2 and Fig. 2A). Findings
from the multiple-imputation analysis suggested Post Hoc Analyses
that the withdrawal of patients in each group did Longitudinal growth-curve models were used to
not introduce bias. Of the end-point events, 13 estimate the effect of periprocedural stroke and
strokes were fatal (7 in the stenting group and myocardial infarction on health status at 1 year
6 in the endarterectomy group), and 1 myocardial (Fig. 2 and 3 in the Supplementary Appendix). Ma-
infarction was fatal (in the endarterectomy group). jor stroke and minor stroke were found to have
During the periprocedural period, the incidence an effect on physical health at 1 year, according
of the primary end point was similar with carotid- to the SF-36 physical component scale (mean ef-
artery stenting and carotid endarterectomy (5.2 fect estimates, −15.8 points [95% CI, −25.1 to −6.4]
and 4.5%, respectively; hazard ratio for stenting, and −4.5 points [95% CI, −7.9 to −1.2], respectively),
1.18; 95% CI, 0.82 to 1.68; P = 0.38), although the whereas the effect of periprocedural myocardial
rates of the individual end points differed between infarction was less certain (mean effect estimate,
the stenting group and the endarterectomy group −3.0 points [95% CI, −7.1 to 1.1]). Minor stroke had
(death, 0.7% vs. 0.3%; P = 0.18; stroke, 4.1% vs. a significant effect on mental health at 1 year, as
2.3%; P = 0.01; myocardial infarction, 1.1% vs. 2.3%; measured on the SF-36 mental component scale
P = 0.03) (Table 2, and Fig. 1 in the Supplementary (mean effect estimate, –3.4 points [95% CI, −6.3
Appendix). After the periprocedural period, the in- to −0.5]). The likelihood of the primary end point
cidence of ipsilateral stroke was similarly low with was not significantly affected by the medical spe-
carotid-artery stenting and with carotid endarter cialty of the interventionist performing the carotid-
ectomy (2.0% and 2.4%, respectively; P = 0.85). artery stenting (P = 0.51) (Table 7 in the Supplemen-
Prespecified analyses did not show a modifica- tary Appendix).9
tion of the treatment effect by symptomatic sta-
tus (P = 0.84) or by sex (P = 0.34) (Table 5 in the Discussion
Supplementary Appendix). However, an interaction
between age and treatment efficacy was detected Our CREST results indicate that carotid-artery
(P = 0.02) (Fig. 2B and 2C), with a crossover at an stenting and carotid endarterectomy were associ-
age of approximately 70 years; carotid-artery stent-ated with similar rates of the primary composite
ing tended to show greater efficacy at younger outcome — periprocedural stroke, myocardial in-
ages, and carotid endarterectomy at older ages. farction, or death and subsequent ipsilateral stroke
— among men and women with either symptom-
Prespecified Secondary Analyses atic or asymptomatic carotid stenosis. However,
During the periprocedural period, rates of the pri- the incidence of periprocedural stroke was lower
mary end point did not differ significantly be- in the endarterectomy group than in the stenting

group, whereas the incidence of periprocedural

myocardial infarction was lower in the stenting A
100
group. These countervailing effects during the
Freedom from Primary End Point (%)

CEA
periprocedural period resulted in similar rates of
the primary outcomes because the rates of events 90
CAS
after the periprocedural period were similar in

the two groups. Although previous studies have
80
indicated that both stroke and myocardial infarc-
tion are associated with substantial morbidity and
mortality,21,22 quality-of-life analyses among sur- 70
vivors at 1 year in our trial indicate that stroke
had a greater adverse effect on a broad range of 60
health-status domains than did myocardial in- 0
farction. 0 1 2 3 4
The selection of patients for either carotid- Year of Follow-up
artery stenting or carotid endarterectomy may No. at Risk
CAS 1262 1100 787 460 162
require attention to age, with younger patients CEA 1240 1099 770 430 145
having a slightly better outcome with carotid-
artery stenting and older patients having a better B
outcome with carotid endarterectomy.23 The as- 4
P=0.02
sociation between older age and increased risk
Hazard Ratio for Primary
of adverse events after carotid-artery stenting

End Point with CAS
3
was seen in our lead-in cohort,9 the Stent-Pro-
tected Angioplasty versus Carotid Endarterecto- 2
CEA better
my (SPACE) trial (Current Controlled Trials num-
ber, ISRCTN57874028),6 and the International 1
Carotid Stenting Study (ICSS; ISRCTN25337470).24
CAS better
An effect of age on differences between carotid- 0
artery stenting and carotid endarterectomy was 40 50 60 70 80 90
found in the SPACE trial25 as well as in our study. Age (yr)

Mechanisms underlying the increased risk with
carotid-artery stenting in very elderly patients C
600
probably include vascular tortuosity and severe CEA
No. of Patients
vascular calcification.23 CAS

400 232 268
The periprocedural outcomes for carotid-artery 188
241
stenting and carotid endarterectomy reported here 200 117
16 261 264 88 24
are the best reported from a randomized, carotid- 6
4 18 60 143 182 205
100 28
57
revascularization study that incorporated pre- and 0
5
50
55
60
65
70
75
80
85
postprocedural medical, neurologic, ECG, and 5

≤4
>8
–
–
46
51
56
61
66
71
76
81
biomarker evaluations (Table 3). The rate of stroke Age Group (yr)
or death among our symptomatic patients after
carotid-artery stenting (6.0%) was lower than the Figure 2. Primary End Point, According to Treatment Group.
corresponding rates in the SPACE trial (6.8%, not The primary end point was a composite of stroke, myocardial infarction,
including nonipsilateral stroke), the Endarterec- or death from any cause during the periprocedural period or ipsilateral
tomy versus Angioplasty in Patients with Symp- stroke within 4 years after randomization. Panel A shows the Kaplan–Meier
curves for patients undergoing carotid-artery stenting (CAS) and those un-
tomatic Severe Carotid Stenosis (EVA-3S) trial
dergoing carotid endarterectomy (CEA) in whom the primary end point did
(ClinicalTrials.gov number, NCT00190398) (9.6%), not occur, according to year of follow-up. Panel B shows the hazard ratios
and ICSS (7.4%). The rate of stroke or death for the primary end point, as calculated for the CAS group versus the CEA
among our symptomatic patients after carotid end- group, according to age at the time of the procedure. The hazard ratios
arterectomy (3.2%) was also lower than the cor- were estimated from the proportional-hazards model with adjustment for
sex and symptomatic status. Dashed lines indicate the 95% confidence in-
responding percentage in SPACE (6.3%) and was
tervals. Panel C shows the numbers of patients in each age group.
similar to the corresponding percentage in EVA-
3S (3.9%) as well as that in ICSS (3.4%); in ICSS,

20
Table 3. Primary End Point and Its Individual Components among the 1181 Asymptomatic and the 1321 Symptomatic Patients, According to Treatment Group.*
End Point Periprocedural Period 4-Yr Study Period (Including Periprocedural Period)
Absolute Treatment Hazard Ratio for Absolute Treatment Hazard Ratio for
Effect of CAS vs. CEA CAS vs. CEA Effect of CAS vs. CEA CAS vs. CEA
CAS CEA (95% CI) (95% CI) P Value CAS CEA (95% CI) (95% CI) P Value
no. of patients (% ±SE) percentage points no. of patients (% ±SE) percentage points
Myocardial infarction
Asymptomatic patients 7 (1.2±0.4) 13 (2.2±0.6) −1.0 (−2.5 to 0.4) 0.55 (0.22 to 1.38) 0.20
Symptomatic patients 7 (1.0±0.4) 15 (2.3±0.6) −1.2 (−2.6 to 0.1) 0.45 (0.18 to 1.11) 0.08
The
Any periprocedural stroke or

postprocedural ipsilateral
stroke
Asymptomatic patients 15 (2.5±0.6) 8 (1.4±0.5) 1.2 (−0.4 to 2.7) 1.88 (0.79 to 4.42) 0.15 24 (4.5±0.9) 13 (2.7±0.8) 1.9 (−0.5 to 4.3) 1.86 (0.95 to 3.66) 0.07
Symptomatic patients 37 (5.5±0.9) 21 (3.2±0.7) 2.3 (0.1 to 4.5) 1.74 (1.02 to 2.98) 0.04 48 (7.6±1.1) 37 (6.4±1.1) 1.2 (−1.8 to 4.1) 1.29 (0.84 to 1.98) 0.25
Any periprocedural stroke or
n engl j med 363;1

death or postprocedural
ipsilateral stroke
Symptomatic patients 40 (6.0±0.9) 21 (3.2±0.7) 2.8 (0.5 to 5.0) 1.89 (1.11 to 3.21) 0.02 51 (8.0±1.1) 37 (6.4±1.1) 1.6 (−1.4 to 4.6) 1.37 (0.90 to 2.09) 0.14
nejm.org
n e w e ng l a n d j o u r na l
Primary end point (any peripro-

of
cedural stroke, myocar-

dial infarction, or death
or postprocedural ipsi-
july 1, 2010
lateral stroke)

m e dic i n e
Symptomatic patients 45 (6.7±1.0) 35 (5.4±0.9) 1.4 (−1.2 to 3.9) 1.26 (0.81 to 1.96) 0.30 55 (8.6±1.1) 50 (8.4±1.2) 0.2 (−3.0 to 3.4) 1.08 (0.74 to 1.59) 0.69
* The 1181 asymptomatic patients consisted of 594 patients in the carotid-artery stenting (CAS) group and 587 in the carotid endarterectomy (CEA) group. The 1321 symptomatic pa-
tients consisted of 668 in the CAS group and 653 in the CEA group. The periprocedural period was defined, according to the study protocol, as the 30-day period after the procedure
(for all patients who underwent the assigned procedure within 30 days after randomization) or the 36-day period after randomization (for all patients who did not undergo the assigned
procedure within 30 days after randomization). Because the periprocedural period is relatively short, which minimizes the need for censoring, event proportions and the absolute differ-
ences in event proportions were calculated as the percentage of patients with events. For the 4-year study period, proportions reflecting the absolute efficacy of CAS over that of CEA
were based on Kaplan–Meier survival estimates at the end of the 4 years. Hazard ratios for the periprocedural period were based on data for all patients, censored at the end of the
periprocedural period. All hazard ratios were adjusted for age, symptomatic status, and sex. P values were calculated on the basis of the significance of the hazard ratio (per study pro-
tocol). For death, stroke, and myocardial infarction end points, patients could have had more than one event (e.g., fatal stroke was counted as both a death and a stroke, and patients
may have had an ipsilateral stroke followed by a nonipsilateral stroke).
the protocol did not require formal examination had the potential to compromise the statistical
until 30 days after the procedure. Among asymp- power. However, that lower event rate was offset
tomatic patients, the rate of stroke or death in by the higher number of events associated with
the carotid-artery stenting group in our trial the extended enrollment and follow-up periods.
(2.5%) was similar to that in the Asymptomatic The use of medical therapy alone was not stud-
Carotid Atherosclerosis Study (ACAS) (2.3%, ied, sharply limiting our ability to generalize the
excluding patients older than 79 years)8 and was results. For example, the rates of ipsilateral and
lower than that in the Asymptomatic Carotid contralateral stroke were similar in the two groups
Surgery Trial (ACST; ISRCTN26156392) (3.1%),26 after the periprocedural period. For asymptom-
and the rate of stroke and death in our carotid- atic patients, this similarity could indicate that
endarterectomy group (1.4%) was lower than that revascularization is durable or that asymptomatic
in ACAS and ACST. carotid stenosis is benign if treated medically.
The improved periprocedural outcomes in In conclusion, carotid revascularization per-
CREST as compared with previous trials may re- formed by highly qualified surgeons and inter-
flect the effective surgeon credentialing, assimi- ventionists is effective and safe. Stroke was more
lation of evolving endovascular technology, and likely after carotid-artery stenting. Myocardial in-
rigorous training and credentialing of carotid- farction was more likely after carotid endarterec-
artery stenting operators.9 The differential results tomy, but the effect on the quality of life was less
for myocardial infarction and stroke offer oppor- than the effect of stroke. Younger patients had
tunities for improvement. To reduce the risk of slightly fewer events after carotid-artery stenting
stroke after carotid-artery stenting, improvements than after carotid endarterectomy; older patients
in training and technique, embolic protection and had fewer events after carotid endarterectomy. The
stent design, and patient selection (especially low absolute risk of recurrent stroke suggests that
among patients older than 70 years of age) hold both carotid-artery stenting and carotid endar
promise.23,27 To reduce the risk of myocardial in- terectomy are clinically durable and may also
farction after carotid endarterectomy, more de- reflect advances in medical therapy.
tailed preoperative cardiovascular evaluation and The content of this article is solely the responsibility of the au-
the use of dual antiplatelet therapy, statins,28 car- thors and does not necessarily represent the official views of the
dioprotective pharmacotherapy,29 or local anes- NINDS or the NIH.
Supported by the National Institute of Neurological Disorders
thesia could be investigated. and Stroke (NINDS) and the NIH (R01 NS 038384) and supple-
The clinical durability of carotid-artery stent- mental funding from Abbott Vascular Solutions (formerly
ing and carotid endarterectomy is important. The Guidant), including donations of Accunet and Acculink systems,
equivalent to approximately 15% of the total study cost, to CREST
rates of ipsilateral stroke during our follow-up centers in Canada and to CREST centers in the United States that
period — 2.0% with carotid-artery stenting and were at Veterans Affairs sites.
2.4% with carotid endarterectomy — are similar Dr. Roubin reports receiving consulting fees, honoraria, pay-
ments for the development of educational presentations, and
to the rates in the SPACE trial and EVA-3S, sug- travel reimbursement from Abbott Vascular, and royalties from
gesting excellent durability for up to 4 years. Be- Abbott Vascular and Cook; Dr. Mackey reports serving on the
cause the life expectancy of our average-aged pa- Board of Directors of the Canadian Stroke Consortium; Dr. Hill
reports receiving consulting fees from Vernalis Group, hono-
tient is 15 years after the procedure,30 outcomes raria from Hoffmann–La Roche Canada, Stem Cell Therapeu-
are being assessed in CREST out to 10 years. tics, Portola Therapeutics, Sanofi-Aventis Canada, Bristol-Myers
The CREST study does have limitations. The Squibb, and Merck Canada; grant support from Hoffmann–
La Roche Canada, Bayer Canada, and Merck Canada; payment for
certification requirements were important for pa- the development of educational presentations from Boehringer-
tient safety, but they limit the generalizability of Ingelheim Canada; stock options from Calgary Scientific; and
the results and conclusions to similarly qualified travel reimbursement from GE Canada Healthcare; Dr. Sheffet
reports being a member of the Society for Clinical Trials; Dr.
operators. One interventional carotid-artery stent- Hopkins reports serving on the boards of AccessClosure and Mi-
ing system was used. Improvements were incor- crus; receiving consulting fees from Abbott Vascular, Boston Sci-
porated into that system as technology evolved. entific, Cordis, Micrus, and W.L. Gore Endovascular; providing
expert testimony for Brown and Tarantino; Bollinger, Ruberry
Although the use of one system facilitated stan- and Garvey; Kevin Ricotta; Hamilton, Altman, Canale and Dillon;
dardization, external validity may have been af- Martin, Clearwater and Bell; and Bunnell, Woulfe, Kirschbaum,
fected by our prohibition on the use of other stent Keller, McIntyre, Gregoire and Klein; receiving honoraria from
Lenox Hill Hospital, Accumetrics, Society for Cardiac Angiogra-
systems. The addition of asymptomatic patients phy and Interventions (SCAI), Penn State College of Medicine, the
and the anticipated lower event rate for that group Vascular Interventional Advances (VIVA) Conference, Northwest-

ern University, Boston Scientific, the Advances in Cardiac and Company Cleviprex Partnership Summit, the Cardiovascular Re-
Endovascular Therapies (ACE) Conference, Cleveland Clinic, the search Technologies Conference, the Neurosurgical Society of
International Symposium on Endovascular Therapy (ISET) Con- America Annual Meeting, Strategic Medical Seminars; and being
ference, Memorial Healthcare Systems, and Canisius College; an employee of University at Buffalo Neurosurgery; Dr. Cutlip re-
payment for the development of educational presentations and ports receiving grant support from Cordis, travel reimbursement
speakers’ fees from Abbott Vascular; stock options from Access- from Abbott Vascular, and additional funding for the Clinical
Closure, Boston Scientific, and Micrus; royalties from Informa Events Committee from Boston Scientific; Dr. David Cohen re-
UK, Cordis, Taylor and Francis Books, and Elsevier; grant support ports serving on the boards of Cordis and Medtronic and receiv-
from Toshiba; and travel reimbursement from Leipzig Interven- ing consulting fees from Medtronic, grant support from Abbott
tional Course (LINC), Micrus, AccessClosure, SCAI, Vascular Vascular, Boston Scientific, Eli Lilly–Daiichi Sankyo, and Bristol-
Disease Conference, the Cardiovascular Research Foundation Myers Squibb–Sanofi, payment for development of educational
(CRF) Conference, All that Jazz, the EuroPCR Conference, Ce- presentations from Eli Lilly, and travel reimbursement from Eli
dars–Sinai, the Multidisciplinary European Endovascular Therapy Lilly, Medtronic, and Cordis; Dr. Popma reports receiving consult-
(MEET) Conference, the American Association of Neurological ing fees from Boston Scientific, Abbott Vascular, Cordis, Medtron-
Surgeons (AANS), the VIVA Conference, the Congress of Neuro- ic, Ablomed, AstraZeneca and Bristol-Myers Squibb; and grant
surgeons Conference, the LINC Conference, the Transcatheter support from Cordis, Medtronic, Boston Scientific, Abbott Vascu-
Cardiovascular Therapeutics (TCT) Conference, the Latin Ameri- lar, and Abiomed; and Dr. Stanley Cohen reports receiving hono-
can Neurosurgery (CLAN) Conference, the Vascular and Endovas- raria and payment for the development of educational presenta-
cular Issues, Techniques, and Horizons (VEITH) Conference, the tions from the Boehringer-Ingelheim Speakers’ Bureau. No other
International Course on Carotid Angioplasty and Other Cerebro- potential conflict of interest relevant to this article was reported.
vascular Interventions (ICCA) Conference, the ISET Conference, Disclosure forms provided by the authors are available with
the Innovation in Intervention: i2 Summit 2009 Research Founda- the full text of this article at NEJM.org.
tion, NeuroVASX, W.L. Gore Endovascular, the Western Neuro- We thank the patients for their participation, their families,
logical Society, the Neurological Society of Australia Meeting, the and the original principal investigator, Robert W. Hobson, II,
Alta Bates Summer Stroke Conference, Toshiba, the Medicines M.D. (deceased), for his leadership and courage.
Appendix
The CREST investigators and committee members were as follows — Principal investigators (in order of decreasing number of patients
who were randomly assigned to a treatment group): W. Clark, Oregon Health and Science University, Portland; W. Brooks, Central
Baptist Hospital, Lexington, KY; A. Mackey, Centre Hospitalier Affilié Universitaire de Québec–Hôpital de l’Enfant-Jésus, Quebec City,
QC, Canada; M. Hill, A. Buchan, University of Calgary–Foothills Medical Centre, Calgary, AB, Canada; P. Leimgruber, Deaconess
Medical Center–Northwest Cardiovascular Research Institute, Spokane, WA; V. Mantese, St. John’s Mercy Medical Center, St. Louis; C.
Timaran, University of Texas Southwestern Medical School, Dallas; L.N. Hopkins, State University of New York, Buffalo; D. Chiu,
Methodist Hospital, Houston; R. Begg, Tri-State Medical Group, Beaver, PA; Z. Jamil, St. Michael’s Medical Center, Newark, NJ; R. Hye,
Kaiser Permanente San Diego Medical Center, San Diego, CA; B. Demaerschalk, Mayo Clinic, Scottsdale, AZ; O.W. Brown, Beaumont
Hospital, Royal Oak, MI; G.S. Roubin, S. Iyer, Lenox Hill Hospital, New York; D. Heck, Forsyth Medical Center, Winston-Salem, NC;
R. Farb, W. Montanera, S.K. Lee, Toronto Western Hospital, Toronto; I. Altafullah, North Memorial Medical Center, Golden Valley,
MN; G. Ansel, MidWest Cardiology Research Foundation, Columbus, OH; A. Sam II, Vascular Specialty Associates, Baton Rouge, LA;
N. Gonzales, M. Campbell, J. Choi, University of Texas Medical School, Houston; P. Soukas, St. Elizabeth’s Medical Center, Boston; L.
Wechsler, University of Pittsburgh Medical Center–Shadyside Hospital and Presbyterian Hospital, Pittsburgh; D. Clair, K. Ouriel, Cleve-
land Clinic, Cleveland; M. Reisman, W. Gray, Swedish Medical Center, Seattle; J. Eidt, University of Arkansas for Medical Sciences–
Central Arkansas Veterans Healthcare System, Little Rock; S. Orlow, Northern Indiana Research Alliance, Fort Wayne; J. Burke, S. So-
renson, P.J. Casterella, Intermountain Medical Center, Salt Lake City; M. Malas, K. Murphy, Johns Hopkins Medical Institutions, Balti-
more; M. Rinaldi, Sanger Heart and Vascular Institute, Charlotte, NC; K. Rosenfield, Massachusetts General Hospital, Boston; C.
Sternbergh III, R. Felberg, Ochsner Health System, New Orleans; R. McCann, T. Smith, Duke University Medical Center, Durham, NC;
C. O’Mara, Mississippi Baptist Medical Center, Jackson; A. Hakaim, Mayo Clinic, Jacksonville, FL; B. Katzen, Baptist Cardiac and Vas-
cular Institute, Miami; R. Spetzler, Barrow Neurological Institute, Phoenix, AZ; A. Pucillo, Westchester Medical Center, Valhalla, NY; J.
Elmore, Geisinger Medical Center, Danville, PA; W. Jordan, University of Alabama at Birmingham, Birmingham; D. Lew, Leesburg
Regional Medical Center, Leesburg, FL; R. Powell, Dartmouth Hitchcock Medical Center, Lebanon, NH; R. Bulas, the Christ Hospital,
Cincinnati; B. Kluck, Lehigh Valley Hospital, Allentown, PA; J. Rapp, San Francisco Veterans Affairs Medical Center, San Francisco; G.
Mishkel, Prairie Education and Research Cooperative, Springfield, IL; F. Weaver, University of Southern California, Los Angeles; M.
Nazzal, University of Toledo Medical Center, Toledo, OH; C. Narins, University of Rochester Medical Center, Rochester, NY; R. Molnar,
Michigan Vascular Research Center, Flint; M. Eskandari, Northwestern Memorial Hospital, Chicago; H. Aronow, Michigan Heart and
Vascular Institute, Ann Arbor; F. Shawl, Washington Adventist Hospital, Takoma Park, MD; R. Rosenwasser, Thomas Jefferson Univer-
sity, Philadelphia; H. Chastain, Parkview Research Center, Fort Wayne, IN; M. Foster, Baptist Hospital West, Knoxville, TN; R. Raabe,
Providence Medical Research Center, Spokane, WA; D. Pelz, London Health Sciences Centre, London, ON, Canada; G. Stotts, Ottawa
Hospital, Ottawa; H. Cloft, Mayo Clinic, Rochester, MN; L. Heller, W. Knopf, St. Joseph’s Hospital of Atlanta, Atlanta; B. Reddy, Pied-
mont Hospital–Fuqua Heart Center, Atlanta; K. Hodgson, Southern Illinois School of Medicine, Springfield; K. Fraser, Order of St.
Francis Saint Francis Medical Center, Peoria, IL; W. Gray, J.P. Mohr, Columbia University Medical Center–New York Presbyterian Hos-
pital, New York; A. Shepard, D. Reddy, Henry Ford Hospital, Detroit; W. Montanera, St. Michael’s Hospital, Toronto; W. Moore, UCLA,
Los Angeles; E. Chaikof, B. Stern, Emory University Hospital–Atlanta Veterans Affairs Medical Center, Atlanta; S. Johnson, Cardiovas-
cular Research Foundation–Cardiovascular Associates, Elk Grove, IL; R. Zelman, Cape Cod Research Institute, Hyannis, MA; A. Evans,
University of Virginia Health System, Charlottesville; D. Burkart, St. Joseph Medical Center, Kansas City, MO; D. Bandyk, University of
South Florida, Tampa; P. Karanjia, Marshfield Clinic, Marshfield, WI; J. Zarghami, Providence Hospital and Medical Centers, Southfield,
MI; A. Arthur, J. Barr, Baptist Memorial Hospital, Memphis, TN; M. Mehta, Vascular Interventional Project, Albany, NY; A. Comerota,
Jobst Vascular Center, Toledo, OH; K. Natarajan, K. Krol, St. Vincent Hospital, Indianapolis; J. Howington, St. Joseph/Candler Health
System, Savannah, GA; D. Selchen, Trillium Health Centre, Mississauga, ON, Canada; M. Schermerhorn, F.W. LoGerfo, Beth Israel
Deaconess Medical Center, Boston; S. Laster, St. Luke’s Hospital, Kansas City, MO; M. Sanz, St. Patrick Hospital, Missoula, MT; E.
Lopez del Valle, Morton Plant Hospital, Clearwater, FL; J. Andriole, Orlando Regional Healthcare, Orlando, FL; A. Ringer, University of
Cincinnati, Cincinnati; J. Martin, Anne Arundel Medical Center, Annapolis, MD; R. Guzman, University of Manitoba, Winnipeg, MB,

Canada; P. Teal, Vancouver General Hospital, Vancouver, BC, Canada; F. Hellinger, Florida Hospital, Orlando; G. Petrossian, R.B. Ber-
roya, St. Francis Hospital, Roslyn, NY; M. Bates, Charleston Area Medical Center, Charleston, WV; J. Mills, University of Arizona College
of Medicine, Tucson; M. Golden, R. Fairman, Hospital of the University of Pennsylvania, Philadelphia; A. Mansour, Butterworth Hos-
pital, Grand Rapids, MI; A. MacBeth, H. Madyoon, St. Joseph’s Medical Center, Stockton, CA; L.A. Iannone, Iowa Heart Center, Des
Moines; K. Hansen, Wake Forest University Health Sciences, Winston-Salem, NC; J. Biller, Loyola University Medical Center, Maywood,
IL; J. Shuck, Christiana Care Health Services, Newark, DE; P. Gobin, Weill Cornell Medical College of New York Presbyterian Hospital,
New York; K. Dauterman, Rogue Valley Medical Center, Medford, OR; J. Melton, Oklahoma Foundation for Cardiovascular Research,
Oklahoma City; D. Benckart, Allegheny General Hospital, Pittsburgh; W. Lesley, Scott and White Memorial Hospital, Temple, TX; M.
Belkin, Brigham and Women’s Hospital, Boston; T. Bajwa, St. Luke’s Medical Center, Milwaukee; S. Myla, Hoag Memorial Hospital,
Newport Beach, CA; J. Snell, Rush University Medical Center, Chicago; H. Shownkeen, Central DuPage Hospital, Winfield, IL; A. Abou-
Chebl, University of Louisville, Louisville, KY; Executive Committee: T.G. Brott (chair), R.W. Hobson, II (deceased), K.W. Beach, D.J.
Cohen, J. Cordell, D.E. Cutlip, R.D. Ferguson, L.N. Hopkins, G. Howard, V.J. Howard, L. Keene, B.K. Lal, J.F. Meschia, J.P. Mohr, W.S.
Moore, J. Popma, G.S. Roubin, A.J. Sheffet, F.L. Silver; Stroke Adjudication Committee: S.N. Cohen (chair), J. Biller, J.P. Broderick, S.
Chaturvedi, M.A. Kalafut, E.J. Skalabrin; Myocardial Infarction Adjudication Committee: J.L. Blackshear (chair), S.P. Glasser, R.J. Prin-
eas; Data and Safety Monitoring Board: J. Marsh (chair), R. Higashida, G. Lamas, B. Tilley, M. Walker, and G.R. Cutter (2001–2002).
REFERENCES
1. Petty GW, Brown RD Jr, Whisnant JP, 10. Sheffet AJ, Roubin G, Howard G, et al. diovascular events in patients with tran-
Sicks JD, O’Fallon WM, Wiebers DO. Design of the Carotid Revascularization sient ischaemic attack or minor ischaemic
Ischemic stroke subtypes: a population- Endarterectomy vs. Stenting Trial (CREST). stroke: comparison between arterial and
based study of incidence and risk factors. Int J Stroke 2010;5:40-6. cardiac source of the index event. J Neurol
Stroke 1999;30:2513-6. 11. Hobson RW II. CREST (Carotid Revas- Neurosurg Psychiatry 2008;79:895-9.
2. Goldstein LB, Adams R, Alberts MJ, et cularization Endarterectomy versus Stent 22. Landesberg G, Shatz V, Akopnik I, et al.
al. Primary prevention of ischemic stroke: Trial): background, design, and current Association of cardiac troponin, CK-MB,
a guideline from the American Heart As- status. Semin Vasc Surg 2000;13:139-43. and postoperative myocardial ischemia
sociation/American Stroke Association 12. Moore WS, Vescera CL, Robertson JT, with long-term survival after major vascular
Stroke Council: cosponsored by the Ath- Baker WH, Howard VJ, Toole JF. Selection surgery. J Am Coll Cardiol 2003;42:1547-
erosclerotic Peripheral Vascular Disease process for surgeons in the Asymptomatic 54.
Interdisciplinary Working Group; Cardio- Carotid Atherosclerosis Study. Stroke 23. Chiam PT, Roubin GS, Iyer SS, et al.
vascular Nursing Council; Clinical Cardiol- 1991;22:1353-7. Carotid artery stenting in elderly patients:
ogy Council; Nutrition, Physical Activity, 13. Brott T, Adams HP Jr, Olinger CP, et importance of case selection. Catheter Car-
and Metabolism Council; and the Quality al. Measurements of acute cerebral infarc- diovasc Interv 2008;72:318-24.
of Care and Outcomes Research Interdisci- tion: a clinical examination scale. Stroke 24. Ederle J, Dobson J, Featherstone RL, et
plinary Working Group: The American 1989;20:864-70. al. Carotid artery stenting compared with
Academy of Neurology affirms the value of 14. Meschia JF, Brott TG, Chukwudelun- endarterectomy in patients with symptom-
this guideline. Stroke 2006;37:1583-633. zu FE, et al. Verifying the stroke-free phe- atic carotid stenosis (International Carotid
[Erratum, Stroke 2007;38:207.] notype by structured telephone interview. Stenting Study): an interim analysis of a
3. Adams RJ, Albers G, Alberts MJ, et al. Stroke 2000;31:1076-80. randomised controlled trial. Lancet 2010;
Update to the AHA/ASA recommendations 15. Ware JE Jr, Sherbourne CD. The MOS 375:985-97.
for the prevention of stroke in patients 36-Item Short-Form Health Survey (SF-36). 25. Stingele R, Berger J, Alfke K, et al.
with stroke and transient ischemic attack. I. Conceptual framework and item selec- Clinical and angiographic risk factors for
Stroke 2008;39:1647-52. tion. Med Care 1992;30:473-83. stroke and death within 30 days after ca-
4. Ederle J, Featherstone RL, Brown MM. 16. Ware JE Jr, Kosinski M, Bayliss MS, rotid endarterectomy and stent-protected
Percutaneous transluminal angioplasty and McHomey CA, Rogers WH, Raczek A. angioplasty: a subanalysis of the SPACE
stenting for carotid artery stenosis. Coch Comparison of methods for the scoring study. Lancet Neurol 2008;7:216-22.
rane Database Syst Rev 2007;4:CD000515. and statistical analysis of SF-36 health pro- 26. Halliday A, Mansfield A, Marro J, et al.
5. Yadav JS, Wholey MH, Kuntz RE, et al. file and summary measures: summary of Prevention of disabling and fatal strokes
Protected carotid-artery stenting versus results from the Medical Outcomes Study. by successful carotid endarterectomy in
enda rterectomy in high-risk patients. Med Care 1995;33:Suppl:AS264-AS279. patients without recent neurological symp-
N Engl J Med 2004;351:1493-501. 17. Rautaharju PM, MacInnis PJ, Warren toms: randomised controlled trial. Lancet
6. Ringleb PA, Allenberg J, Bruckmann H, JW, Wolf HK, Rykers PM, Calhoun HP. 2004;363:1491-502.
et al. 30 Day results from the SPACE trial Methodology of ECG interpretation in the 27. Roubin GS, Iyer S, Halkin A, Vitek J,
of stent-protected angioplasty versus ca- Dalhousie Program: NOVACODE ECG clas- Brennan C. Realizing the potential of ca-
rotid endarterectomy in symptomatic pa- sification procedures for clinical trials and rotid artery stenting: proposed paradigms
tients: a randomised non-inferiority trial. population health surveys. Methods Inf for patient selection and procedural tech-
Lancet 2006;368:1239-47. [Erratum, Lancet Med 1990;29:362-74. nique. Circulation 2006;113:2021-30.
2006;368:1238.] 18. O’Brien PC, Fleming TR. A multiple 28. Schouten O, Boersma E, Hoeks SE, et
7. Mas JL, Chatellier G, Beyssen B, et al. testing procedure for clinical trials. Bio- al. Fluvastatin and perioperative events in
Endarterectomy versus stenting in patients metrics 1979;35:549-56. patients undergoing vascular surgery.
with symptomatic severe carotid stenosis. 19. Taylor JMG, Cooper KL, Wei JT, Sarma N Engl J Med 2009;361:980-9.
N Engl J Med 2006;355:1660-71. AV, Raghunathan TE, Heeringa SG. Use of 29. Landesberg G, Beattie WS, Mosseri M,
8. Executive Committee for the Asymp- multiple imputation to correct for nonre- Jaffe AS, Alpert JS. Perioperative myocar-
tomatic Carotid Atherosclerosis Study. End sponse bias in a survey of urologic symp- dial infarction. Circulation 2009;119:2936-
arterectomy for asymptomatic carotid ar- toms among African-American men. Am J 44.
tery stenosis. JAMA 1995;273:1421-8. Epidemiol 2002;156:774-82. 30. Arias E, Curtin L, Wei R, Anderson R.
9. Hopkins LN, Rougin GS, Chakhtoura 20. Jennrich RI, Schluchter MD. Unbal- United States decennial life tables for
EY, et al. The Carotid Revascularization anced repeated-measures models with 1999–2001: United States life tables. Natl
Endarterectomy versus Stenting Trial: cre- structured covariance matrices. Biomet- Vital Stat Rep 2008;57:1-36.
dentialing of interventionalists and final rics 1986;42:805-20. Copyright © 2010 Massachusetts Medical Society.
results of lead-in phase. J Stroke Cerebro- 21. van Wijk I, Koudstaal P, Kappelle LJ, et
vasc Dis 2010;19:153-62. al. Long-term occurrence of death and car-

2 - Crest 2010

Uploaded by

Copyright:

Available Formats

2 - Crest 2010

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2 - Crest 2010

Uploaded by

Copyright:

Available Formats

new england

Stenting versus Endarterectomy for Treatment

BACKGROUND From Mayo Clinic, Jacksonville, FL (T.G.B.,

n engl j med 363;1 nejm.org july 1, 2010 11

Centers and Investigators Randomization

12 n engl j med 363;1 nejm.org july 1, 2010

The New England Journal of Medicine

2522 Patients underwent randomization

1271 Were assigned to CAS 1251 Were assigned to CEA

69 Discontinued follow-up 111 Discontinued follow-up

Figure 1. Randomization and Follow-up of the Study Patients.

n engl j med 363;1 nejm.org july 1, 2010 13

Table 1. Baseline Characteristics of the Study Population, According to Treatment Group.*

Carotid-Artery Stenting Carotid Endarterectomy

14 n engl j med 363;1 nejm.org july 1, 2010

The New England Journal of Medicine

Carotid-Artery Stenting Carotid Endarterectomy

* Plus–minus values are means ±SD. Dashes indicate not applicable.

n engl j med 363;1 nejm.org july 1, 2010 15

End Point Periprocedural Period

16 n engl j med 363;1 nejm.org july 1, 2010

The New England Journal of Medicine

4-Yr Study Period (Including Periprocedural Period)

72 (6.2±0.7) 50 (4.7±0.7) 1.5 (−0.4 to 3.4) 1.44 (1.00 to 2.06) 0.049

16 (1.4±0.3) 10 (0.8±0.3) 0.6 (−0.2 to 1.4) 1.55 (0.70 to 3.42) 0.28

85 (7.2±0.8) 76 (6.8±0.8) 0.4 (−1.7 to 2.6) 1.11 (0.81 to 1.51) 0.51

n engl j med 363;1 nejm.org july 1, 2010 17

The New England Journal of Medicine

group, whereas the incidence of periprocedural

Freedom from Primary End Point (%)

after the periprocedural period were similar in

of adverse events after carotid-artery stenting

found in the SPACE trial25 as well as in our study. Age (yr)

vascular calcification.23 CAS

postprocedural medical, neurologic, ECG, and 5

n engl j med 363;1 nejm.org july 1, 2010 19

Any periprocedural stroke or

n engl j med 363;1

Primary end point (any peripro-

The New England Journal of Medicine

cedural stroke, myocar-

Copyright © 2010 Massachusetts Medical Society. All rights reserved.

n engl j med 363;1 nejm.org july 1, 2010 21

22 n engl j med 363;1 nejm.org july 1, 2010

The New England Journal of Medicine

You might also like