Two-Year Outcome After Endovascular Treatment For Acute Ischemic Stroke

The n e w e ng l a n d j o u r na l of m e dic i n e
Original Article
Two-Year Outcome after Endovascular

Treatment for Acute Ischemic Stroke
Lucie A. van den Berg, M.D., Marcel G.W. Dijkgraaf, Ph.D.,
Olvert A. Berkhemer, M.D., Ph.D., Puck S.S. Fransen, M.D.,
Debbie Beumer, M.D., Hester F. Lingsma, Ph.D.,
Charles B.L.M. Majoie, M.D., Ph.D., Diederik W.J. Dippel, M.D., Ph.D.,
Aad van der Lugt, M.D., Ph.D., Robert J. van Oostenbrugge, M.D., Ph.D.,
Wim H. van Zwam, M.D., Ph.D., and Yvo B.W.E.M. Roos, M.D., Ph.D.,
for the MR CLEAN Investigators*
A BS T R AC T
BACKGROUND
Several trials involving patients with acute ischemic stroke have shown better functional From the Departments of Neurology
outcomes with endovascular treatment than with conventional treatment at 90 days after (L.A.B., Y.B.W.E.M.R.) and Radiology
(O.A.B., C.B.L.M.M.) and the Clinical Re
initiation of treatment. However, results on long-term clinical outcomes are lacking. search Unit (M.G.W.D.), Academic Medi
cal Center, Amsterdam, the Departments
METHODS of Neurology (O.A.B., P.S.S.F., D.W.J.D.),
We assessed clinical outcomes 2 years after patients were randomly assigned to receive Radiology (O.A.B., A.L.), and Public Health
either endovascular treatment (intervention group) or conventional treatment (control (H.F.L.), Erasmus MC University Medical
Center, Rotterdam, and the Departments
group) for acute ischemic stroke. The primary outcome was the score on the modified of Neurology (D.B., R.J.O.) and Radiology
Rankin scale at 2 years; this scale measures functional outcome, with scores ranging (W.H.Z.), Maastricht University Medical
from 0 (no symptoms) to 6 (death). Secondary outcomes included all-cause mortality Center, Maastricht — all in the Nether
lands. Address reprint requests to Dr.
and the quality of life at 2 years, as measured by means of a health utility index that Roos at the Department of Neurology,
is based on the European Quality of Life–5 Dimensions questionnaire (scores range Academic Medical Center, P.O. Box 22660,
from −0.329 to 1, with higher scores indicating better health). 1100 DD Amsterdam, the Netherlands,
or at y.b.roos@amc.uva.nl.
RESULTS
* A complete list of investigators in the
Of the 500 patients who underwent randomization in the original trial, 2-year data for Multicenter Randomized Clinical Trial of
this extended follow-up trial were available for 391 patients (78.2%) and information Endovascular Treatment for Acute Is
on death was available for 459 patients (91.8%). The distribution of outcomes on the chemic Stroke in the Netherlands (MR
CLEAN) is provided in the Supplemen
modified Rankin scale favored endovascular treatment over conventional treatment tary Appendix, available at NEJM.org.
(adjusted common odds ratio, 1.68; 95% confidence interval [CI], 1.15 to 2.45;
N Engl J Med 2017;376:1341-9.
P = 0.007). There was no significant difference between the treatment groups in the DOI: 10.1056/NEJMoa1612136
percentage of patients who had an excellent outcome (i.e., a modified Rankin scale Copyright © 2017 Massachusetts Medical Society.
score of 0 or 1). The mean quality-of-life score was 0.48 among patients randomly as-
signed to endovascular treatment as compared with 0.38 among patients randomly
assigned to conventional treatment (mean difference, 0.10; 95% CI, 0.03 to 0.16;
P = 0.006). The cumulative 2-year mortality rate was 26.0% in the intervention group
and 31.0% in the control group (adjusted hazard ratio, 0.9; 95% CI, 0.6 to 1.2; P = 0.46).
CONCLUSIONS
In this extended follow-up trial, the beneficial effect of endovascular treatment on
functional outcome at 2 years in patients with acute ischemic stroke was similar to that
reported at 90 days in the original trial. (Funded by the Netherlands Organization for
Health Research and Development and others; MR CLEAN Current Controlled Trials
number, ISRCTN10888758, and Netherlands Trial Register number, NTR1804, and MR
CLEAN extended follow-up trial Netherlands Trial Register number, NTR5073.)
n engl j med 376;14 nejm.org April 6, 2017 1341

The New England Journal of Medicine
Downloaded from nejm.org on April 5, 2017. For personal use only. No other uses without permission.
Copyright © 2017 Massachusetts Medical Society. All rights reserved.
W
e reported previously the 90-day and completeness of the available data from the
outcomes of a trial (Multicenter Ran- extended follow-up trial and for the fidelity of
domized Clinical Trial of Endovascu- the trial to the amended protocol.
lar Treatment for Acute Ischemic Stroke in the
Netherlands [MR CLEAN]) in which standard Patients and Trial Activities
treatment was compared with endovascular treat- Eligible patients had a score of 2 or higher on
ment, administered within 6 hours after the onset the National Institutes of Health Stroke Scale
of acute ischemic stroke caused by an intracranial (NIHSS) (scores range from 0 to 42, with higher
arterial occlusion of the anterior circulation.1 scores indicating more severe neurologic deficits)
Most patients in the intervention group were and could be treated within 6 hours after the on-
treated by mechanical thrombectomy with the set of stroke.8 The intracranial artery occlusion
use of retrievable stents. The trial showed that had to be confirmed by means of computed tomo-
functional recovery at 90 days was better with graphic angiography, magnetic resonance angiog-
the intervention than with standard treatment. raphy, or digital-subtraction angiography.
Subsequently, the beneficial effect of mechanical Funding for the extended follow-up trial be-
thrombectomy on 90-day outcomes was shown in came available in May 2013, when enrollment in
several other trials, as well as in a meta-analysis the MR CLEAN trial was approximately halfway
of individual patient data from these trials.2-6 to the projected number of patients to be includ-
Information regarding long-term outcomes may ed. Consequently, many patients had already com-
be useful for clinical practice and for health care pleted their 90-day follow-up assessment, and the
policy decisions. We report the results of clinical 2-year follow-up time point had passed for some
follow-up at 2 years after randomization among patients. After trial personnel consulted the Dutch
patients in the MR CLEAN trial. Municipal Personal Records Database to deter-
mine the vital status of patients who had com-
pleted the MR CLEAN trial, eligible patients who
Me thods
were alive were invited to take part in the ex-
Trial Design and Oversight tended follow-up trial. Patients received an invi-
MR CLEAN was a randomized, multicenter trial tation letter and were subsequently contacted by
in which endovascular treatment plus conven- telephone so that we could confirm their par-
tional care (intervention group) was compared ticipation and explain additional trial goals and
with conventional care alone (control group) in activities. Patients who wished to participate in
patients with acute ischemic stroke caused by a the extended follow-up trial provided informed
proximal intracranial arterial occlusion of the consent by telephone. For patients who declined
anterior circulation.1,7 Conventional care consist- to provide informed consent for the long-term
ed of care that represented the most appropriate follow-up, permission was obtained from the
medical management according to national and institutional review board to use information re-
international guidelines and could include intra- garding the patients’ vital status from the Dutch
venous administration of alteplase. Trial treatment Municipal Personal Records Database.
was open-label, and the evaluation of outcomes For the remaining patients who were continu-
was blinded. The protocol, which is available with ing in the MR CLEAN trial after May 2013, the
the full text of this article at NEJM.org, was longer duration of follow-up and additional trial
amended during the trial to include follow-up activities were part of the new informed consent
assessments up to 2 years after randomization. process described above. All patients (or their
An economic evaluation was prespecified as part primary caregiver if they were unable to respond)
of the amended protocol (results are not included who provided informed consent to participate in
here). Trial investigators at the coordinating cen- the extended follow-up were contacted by tele-
ter collected the data for this extended follow-up phone at 6 months, 1 year, 18 months, and 2 years.
trial. Members of the MR CLEAN executive com- One trial investigator, who was unaware of the
mittee designed the extended follow-up trial, treatment-group assignments, assessed functional
analyzed the data, prepared the manuscript, and outcome by means of a structured questionnaire,
made the decision to submit the manuscript for which was validated for assessment by telephone,
publication. The authors vouch for the accuracy to determine the patient’s modified Rankin scale
1342 n engl j med 376;14 nejm.org April 6, 2017

Endovascular Treatment for Acute Ischemic Stroke
score.9 Medical events that occurred between fol- Statistical Analysis

low-up contacts were recorded during this inter- Analyses were based on the modified intention-
view. Furthermore, the patient or the primary to-treat principle. The treatment effect on the
caregiver was invited to complete the three- primary outcome at 2 years was calculated as an
level version of the European Quality of Life 5– adjusted common odds ratio for a better distribu-
Dimensions (EQ-5D-3L) questionnaire, for assess- tion of outcomes on the modified Rankin scale;
ment of quality of life.10 Consensus with respect this ratio was estimated with ordinal logistic
to final functional outcome on the modified regression. Dichotomized scores of the modified
Rankin scale was reached among three trial inves- Rankin scale at 2 years (0 or 1 versus 2 to 6, 0 to
tigators (all of whom were unaware of the treat- 2 versus 3 to 6, and 0 to 3 versus 4 to 6) were
ment-group assignments) on the basis of the analyzed with the use of logistic regression, with
structured questionnaire for the modified Rankin the odds ratio as the effect variable. All-cause
scale and of additional information available in mortality was assessed by means of the Kaplan–
the detailed reports of the telephone interviews. Meier method and a Cox proportional-hazards
model, with the hazard ratio as the effect vari-
Outcome Measures able. Data from all 500 patients who underwent
The primary outcome was the score on the mod- randomization in the original trial were included
ified Rankin scale at 2 years. The modified in the analysis of death from any cause. Addi-
Rankin scale is an ordinal scale that ranges tional information from the Dutch Municipal Per-
from 0 (no symptoms) to 6 (death).11,12 Second- sonal Records Database on vital status at 2 years
ary outcomes included categories scores of the was used for the outcome analysis, and data for
modified Rankin scale at 2 years (0 or 1 [excel- patients whose vital status was missing were
lent outcome], 0 to 2 [good outcome, indicating censored at the time of their withdrawal from
functional independence], and 0 to 3 [favorable the trial. For the analysis of quality of life, a utility
outcome]); death from any cause during the value for each observed EQ-5D-3L health status
2-year period after randomization; the first major profile was calculated with the use of an exist-
vascular event after the index stroke that occurred ing algorithm on the basis of valuations elicited
between 90 days and 2 years after randomiza- by time trade-off techniques applied to the gen-
tion; and quality of life at 2 years as assessed by eral Dutch population. An unstandardized re-
the EQ-5D-3L.10 The EQ-5D-3L consists of a de- gression parameter beta was estimated with the
scriptive system that assesses five dimensions of use of a multivariable linear regression model
quality of life: mobility, self-care, usual activi- and represented the difference between the two
ties, pain or discomfort, and anxiety or depression. treatment groups in the health utility score. As
Each dimension has three levels of severity: 1 (no in the original trial, all effect variables were ad-
problems), 2 (some problems), or 3 (extreme prob- justed for potential imbalances between the inter-
lems). Scores range from −0.329 (indicating the vention group and the control group in the fol-
worst health status [serious problems in all do- lowing prognostic variables at baseline: age;
mains]) to 1 (indicating the best health status stroke severity (as assessed by the NIHSS); time
[no problems at all]), with death assigned a value from stroke onset to randomization; status with
of zero.13 Major vascular events included fatal or respect to previous stroke, atrial fibrillation, and
nonfatal cardiac events, fatal or nonfatal recur- diabetes mellitus; and occlusion of the internal-
rent stroke, and fatal or nonfatal major periph- carotid-artery terminus (yes vs. no). To assess
eral arterial or thromboembolic events (details the difference between the two treatment groups
regarding the classification of major vascular in the occurrence of long-term major vascular
events are provided in the Supplementary Appen- events, a rate ratio was calculated on the basis of
dix, available at NEJM.org). All major vascular person-years at risk. Treatment-effect modifica-
events were confirmed by clinical reports or by tion was evaluated in the same prespecified sub-
the treating physician or the general practitioner. groups of patients as in the original trial (see the
Two trial investigators who were unaware of the Supplementary Appendix).
treatment-group assignments reviewed all events. To evaluate potential selection bias, we used
In the event of disagreement, classification was conventional statistics to compare patients whose
based on consensus. 2-year outcome data were missing with patients

whose 2-year outcome data were available with the control group). A total of 26 patients were
respect to the main prognostic variables, treat- lost to follow-up. In total, 391 of the 500 patients
ment assignment, and 90-day functional outcome. (78.2%) had 2-year follow-up data and were in-
In addition, a sensitivity analysis was performed cluded in the primary analysis for functional
with the use of a regression-based multiple impu- outcome (194 in the intervention group and 197
tation to account for missing data on the func- in the control group) (Fig. S1 in the Supplemen-
tional outcome at 2 years. A detailed description tary Appendix).
of the multiple imputation method is provided in Demographic characteristics, risk factors for
the Supplementary Appendix. a poor outcome, clinical risk factors for stroke,
Two-sided P values of 0.05 or less were con- and the use of intravenous alteplase treatment
sidered to indicate statistical significance for all for the 391 patients included in the analysis of
statistical tests. Because our trial had a single the primary outcome were evenly distributed
primary outcome, no adjustments for multiple between the two treatment groups at baseline
tests were made. All analyses were performed with (Table 1). The median age of the patients was 66
the use of SPSS software, version 24.0 (IBM). years and 58.6% were men. The median NIHSS
score was 18 (interquartile range, 14 to 22).
R e sult s
Primary Outcome
Trial Population Among the 391 patients with available data on
In the original trial, 502 patients underwent ran- the modified Rankin scale at 2 years after ran-
domization at 16 Dutch hospitals from Decem- domization, the adjusted common odds ratio was
ber 2010 through March 2014 (Table S1 in the 1.68 (95% confidence interval [CI], 1.15 to 2.45;
Supplementary Appendix). Two patients, whose P = 0.007) for a better distribution of outcomes
representatives withdrew consent immediately on the modified Rankin scale with endovascular
after randomization and assignment to the con- treatment than with conventional treatment
trol group, could not be included in the intention- (Fig. 1). The median and interquartile range of
to-treat analysis of the original trial. A total of the scores on the modified Rankin scale for the
233 patients (46.6%) were randomly assigned to two treatment groups and the effect size favoring
the intervention group and 267 (53.4%) to the endovascular treatment are shown in Table 2.
control group. Endovascular treatment was never
initiated in 17 of the 233 patients (7.3%) who Secondary Outcomes
had been assigned to the intervention group, Dichotomized Scores on the Modified Rankin Scale
and 1 patient received endovascular treatment Patients in the intervention group were more
after being assigned to the control group.1 likely than patients in the control group to have
At the time of the inception of the extended a good outcome (i.e., a modified Rankin scale
follow-up trial, 332 patients were already in- score of 0 to 2) (37.1% vs. 23.9%; adjusted odds
cluded in the original trial and an additional 168 ratio, 2.21; 95% CI, 1.30 to 3.73; P = 0.003) and
patients would be enrolled and prospectively fol- to have a favorable outcome (i.e., a modified
lowed for up to 2 years after randomization. Of Rankin scale score of 0 to 3) (55.2% vs. 40.6%;
the 332 patients already enrolled in the original adjusted odds ratio, 2.13; 95% CI, 1.30 to 3.43;
trial, 14 (4.2%) had passed the 2-year follow-up P = 0.003). Among patients with an excellent out-
time point, and 87 (40 in the intervention group come (i.e., a modified Rankin scale score of 0 or
and 47 in the control group) died before the ex- 1), no significant difference between the treat-
tended follow-up period began. Of the remain- ment groups was observed (7.2% in the interven-
ing 231 patients who were included in the origi- tion group and 6.1% in the control group; ad-
nal trial, those who did not provide informed justed odds ratio, 1.22; 95% CI, 0.53 to 2.84;
consent to participate in the extended follow-up P = 0.64) (Table 2).
trial initially were reinvited to participate: 61 pa-
tients declined and 8 patients withdrew consent All-Cause Mortality
during the follow-up period (of these 69 patients, Information regarding vital status at 2 years was
18 were in the intervention group and 51 were in available for 459 of the original 500 patients who

Table 1. Baseline Characteristics of the 391 Patients Included in the Analysis of the Primary Outcome, According to
Treatment Assignment.*
Intervention Control
Characteristic (N = 194) (N = 197)
Age — yr
Median 65.9 65.5
Interquartile range 55.8−76.2 56.6−76.6
Male sex — no. (%) 111 (57.2) 118 (59.9)
NIHSS score†
Median 17 18
Interquartile range 14−21 14−22
Location of stroke in left hemisphere — no. (%) 96 (49.5) 116 (58.9)
History of ischemic stroke — no. (%) 24 (12.4) 23 (11.7)
Atrial fibrillation — no. (%) 53 (27.3) 44 (22.3)
Diabetes mellitus — no. (%) 29 (14.9) 25 (12.7)
Prestroke modified Rankin scale score — no. (%)‡
0 155 (79.9) 156 (79.2)
1 19 (9.8) 24 (12.2)
2 10 (5.2) 8 (4.1)
>2 10 (5.2) 9 (4.6)
Systolic blood pressure — mm Hg 147.5±27.0 144.8±23.7
Treatment with IV alteplase — no. (%) 168 (86.6) 182 (92.4)
Time from stroke onset to start of IV alteplase — min
Median 85 85
ASPECTS§
Median 9 9
Intracranial arterial occlusion — no. (%)
Intracranial ICA 1 (0.5) 1 (0.5)
ICA with involvement of the M1 middle cerebral artery segment 49 (25.3) 54 (27.4)
M1 middle cerebral artery segment 130 (67.0) 124 (62.9)
M2 middle cerebral artery segment 13 (6.7) 16 (8.1)
A1 or A2 anterior cerebral artery segment 1 (0.5) 2 (1.0)
Extracranial ICA occlusion — no. (%)¶ 60 (30.9) 51 (25.9)
Time from stroke onset to randomization — min
Median 205 190
Time from stroke onset to groin puncture — min
Median 263 NA
Interquartile range 210−307
* Plus−minus values are means ±SD. There were no significant between-group differences with respect to any of the vari
ables listed in this table. ICA denotes internal carotid artery (intracranial segment), IV intravenous, and NA not applicable.
† Scores on the National Institutes of Health Stroke Scale (NIHSS) range from 0 to 42, with higher scores indicating
more severe neurologic deficits.
‡ Scores on the modified Rankin scale of functional disability range from 0 (no symptoms) to 6 (death). A score of 0 or 1 in
dicates an excellent outcome; 0 to 2 a good outcome, indicating functional independence; and 0 to 3 a favorable outcome.
§ The Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is a measure of the extent of stroke. Scores
range from 0 to 10, with higher scores indicating fewer early ischemic changes. Scores were not available for four patients
assigned to the control group: noncontrast computed tomography was not performed in one patient, and three patients
had strokes in the territory of the anterior cerebral artery.
¶ Extracranial ICA occlusions were reported by local investigators.

curred during 239 person-years of follow-up in

Modified Rankin Scale Score
0 1 2 3 4 5 6 the intervention group (incidence rate, 0.02 per
No symptoms Death year), and the remaining three events occurred
A Modified Rankin Scale Scores at 90 Days during 235 person-years of follow-up in the con-
3.0 trol group (incidence rate, 0.01 per year) (rate
Intervention
9.0 21.0 18.0 22.0 6.0 21.0
ratio for major cerebrovascular events, 1.64; 95%
(N=233) CI, 0.40 to 6.78; P = 0.50) (Table S3 in the Supple-
mentary Appendix).
Control
6.0 13.0 16.0 30.0 12.0 22.0
(N=267) Quality of Life
1.0
At 2 years, patients in the intervention group
0 20 40 60 80 100
had a higher quality of life, as assessed with the
Patients (%)
EQ-5D-3L instrument, than did patients in the
B Modified Rankin Scale Scores at 2 Years control group (mean health utility score, 0.48 vs.
2.6 0.38). The treatment effect size, which was esti-
Intervention
4.6 29.9 18.0 6.2 8.2 30.4
mated with the use of multiple linear regression,
(N=194) was 0.10 (95% CI, 0.03 to 0.16; P = 0.006) in favor
of endovascular treatment (Table 2). The difference
Control
5.1 17.8 16.8 10.2 10.7 38.6 in the treatment effect between the two groups
(N=197)
was attributable mainly to the EQ-5D-3L dimen-
1.0
sions of “mobility,” “self-care,” and “usual activi-
0 20 40 60 80 100
ties” (Fig. S2 in the Supplementary Appendix).
Patients (%)
Figure 1. Modified Rankin Scale Scores at 90 Days and 2 Years According to

Sensitivity Analysis
Treatment Assignment. Of the 109 patients whose outcome data were
Shown are the distribution of scores on the modified Rankin scale at 90 days missing at 2 years, 6 patients died between 90
(Panel A) and at 2 years (Panel B). Scores range from 0 to 6, with 0 indicating days and 2 years after randomization and were
no symptoms, 1 no clinically significant disability, 2 slight disability (patient assigned a score of 6 on the modified Rankin
is able to look after own affairs without assistance but is unable to carry
scale at 2 years for an initial sensitivity analysis.
out all previous activities), 3 moderate disability (patient requires some
help but is able to walk unassisted), 4 moderately severe disability (patient The missing outcome data for the remaining 103
is unable to attend to bodily needs without assistance and is unable to walk patients were imputed with the use of multiple
unassisted), 5 severe disability (patient requires constant nursing care and imputation. The pooled effect on the primary
attention), and 6 death. The adjusted common odds ratio was 1.68 (95% CI, outcome after multiple imputation (adjusted com-
1.15 to 2.45; P = 0.007), which indicated a distribution of the scores on the
mon odds ratio, 1.59; 95% CI, 1.08 to 2.35;
modified Rankin scale in favor of endovascular treatment as compared with
conventional treatment. P = 0.02) was similar to the result of the main
analysis of the primary outcome (adjusted com-
mon odds ratio, 1.68; 95% CI, 1.15 to 2.45;
underwent randomization (91.8%). The cumula- P = 0.007). The subgroup of patients whose out-
tive 2-year rate of death was 26.0% in the inter- come data were missing had a higher incidence
vention group and 31.0% in the control group of atrial fibrillation at baseline than the subgroup
(adjusted hazard ratio for death, 0.9; 95% CI, 0.6 of patients who had outcome data at 2 years
to 1.2; P = 0.46) (Fig. 2). (35.9% vs. 26.4%, P = 0.02), were more likely to
have been randomly assigned to the control group
Major Vascular Events (62.1% vs. 48.9%, P = 0.05), had a longer median
Eight major vascular events were reported be- time from onset of stroke to randomization (218
tween 90 days and 2 years after randomization minutes vs. 195 minutes, P = 0.003), and were
(the maximum duration of follow-up per patient more likely to have poor functional outcomes
was 1.75 years [i.e., 21 months, which represents (i.e., a modified Rankin scale score of 4 or 5,
the duration of time between the 90-day and indicating moderately severe or severe disability)
2-year time points for all patients included in the at 90 days (57.3% vs. 30%, P = 0.005) (Table S2 in
extended follow-up trial]). Five of the events oc- the Supplementary Appendix).

Table 2. Primary and Secondary Outcomes and Treatment Effects.*
Intervention Control Effect Unadjusted Value Adjusted Value

Outcome (N = 194) (N = 197) Variable (95% CI) (95% CI)† P Value‡
Primary outcome
Modified Rankin scale score at 3 (2−6) 4 (3−6) Common odds 1.63 (1.14−2.32) 1.68 (1.15−2.45) 0.007
2 years — median (IQR) ratio
Secondary outcomes
Modified Rankin scale score of 14 (7.2) 12 (6.1) Odds ratio 1.20 (0.54−2.66) 1.22 (0.53−2.84) 0.64
0 or 1 at 2 years — no. (%)
0−2 at 2 years — no. (%)
0−3 at 2 years — no. (%)
EQ-5D-3L score at 2 years§ 0.48±0.40 0.38±0.39 Beta 0.10 (0.02−0.18) 0.10 (0.03−0.16) 0.006
* Plus−minus values are means ±SD. IQR denotes interquartile range.

† Values were adjusted for age, NIHSS at baseline, time to randomization, previous stroke, atrial fibrillation, diabetes mellitus, and presence
of ICA terminus occlusion.
‡ P values are for the comparisons of the adjusted values of each effect variable, as determined byordinal logistic regression analysis for the
primary outcome, by logistic regression analysis for the dichotomizations of the modified Rankin scale, and by multivariable linear regres
sion analysis for the European Quality of Life 5–Dimensions (EQ-5D-3L) score.
§ The three-level version of the EQ-5D-3L questionnaire is a standardized instrument for the measurement of health status. The health status
profiles on five domains (mobility, self-care, daily activities, pain or discomfort, and anxiety or depression), each differentiating three levels
of severity (no problems, some problems, or extreme problems), were converted to a score between −0.329 and 1, with higher scores indi
cating better health and with death assigned a value of zero. The EQ-5D-3L score was missing for two patients.
Subgroup Analyses (32.6%). Although the above results are similar,

No significant interactions (effect modifications) notable differences were observed between the two
were observed between the prespecified subgroups, time points. First, during the extended follow-up
which were defined according to baseline char- period, the mortality rate was lower with endo-
acteristics, and treatment at 2 years. The treat- vascular intervention than with conventional treat-
ment effect remained consistent in all prespeci- ment, although this difference was not statisti-
fied subgroups (Fig. S3 in the Supplementary cally significant, whereas at 90 days, the risk of
Appendix). However, some subgroups were small, death was similar in the two groups. Second,
which resulted in wide confidence intervals. the percentages of patients with modified Rankin
scale scores of 0 or 1 at 2 years were lower than
the percentages at 90 days in both groups. A pos-
Discussion
sible explanation for the lower rate of modified
The results of the extended follow-up evaluation Rankin scale scores of 0 or 1 at 2 years is that
of the MR CLEAN trial showed that endovascu- patients with stroke undergo a period of reha-
lar treatment in patients with acute ischemic bilitation in the first months after the event,
stroke resulted in functional recovery, as mea- during which the effect of the stroke on their
sured on the modified Rankin scale, that was daily activities may not be fully apparent. There-
similar to the originally reported results at 90 days. after, when patients are living at home with less
The odds ratio for better scores on the modified assistance, small changes in functional ability
Rankin scale in the endovascular group than in the may become apparent that could move patients
conventional treatment group was 1.67 at 90 days, from a modified Rankin scale score of 0 (no
as compared with an odds ratio of 1.68 at 2 years. symptoms) to a score of 1 (no significant dis-
The percentage of patients in the intervention ability), or to a score of 2 (slight disability).
group who were functionally independent (i.e., a Other studies of reperfusion therapy for ische
modified Rankin scale score of 0 to 2) at 2 years mic stroke have suggested that the effects of
(37.1%) was also similar to the results at 90 days intravenous alteplase and of endovascular treat-

(NIHSS score ≥20) (32.5% vs. 18.6%, P = 0.04).16,17

0.5 In the REVASCAT trial (Endovascular Revascular-
ization With Solitaire Device Versus Best Medi-
cal Therapy in Anterior Circulation Stroke Within
8 Hours), the common odds ratio for the distri-
0.4 bution of scores on the modified Rankin scale
at 1 year was 1.80 (95% CI, 1.10 to 2.99) in favor
of endovascular revascularization, which was
similar to the finding at 90 days (common odds
Cumulative Incidence of Death
Control
ratio, 1.71; 95% CI, 1.05 to 2.81).18 Although all
0.3
these trials were smaller and had a shorter follow-
up period than our trial, their results regarding
Intervention longer-term functional outcomes are consistent
with ours.
0.2 We observed few major vascular events be-
tween 90 days and 2 years after randomization
(1.6% in 474 person-years at risk). A possible
explanation for this low rate is that patients with
0.1 large-vessel extracranial or intracranial occlusion
could have had either embolic strokes (artery-to-
artery or cardiogenic) or extracranial or intracra-
nial atherosclerotic occlusion of vessels, and these
conditions are treated or prevented more effec-
0.0
0 100 200 300 400 500 600 700 800 tively than are other causes of ischemic stroke.19-21
Days since Randomization Our trial has several limitations. The trial
No. at Risk
was powered to detect an effect at 90 days and
Intervention 233 175 159 157 155 153 153 153 did not take into account loss to follow-up dur-
Control 267 200 183 177 171 166 166 165 ing the 2-year follow-up period. Furthermore, pa-
tients whose outcome data were missing at 2 years
Figure 2. Kaplan–Meier Estimates of All-Cause Mortality during the 2-Year
had worse clinical characteristics and worse
Follow-up Period after Randomization.
functional outcomes at 90 days and were more
The adjusted hazard ratio for death in the intervention group, as compared
with the control group, was 0.9 (95% CI, 0.6 to 1.2; P = 0.46). likely to have been assigned to the control group,
and therefore, selection bias may have been intro-
duced. However, a sensitivity analysis in which
ment do not diminish over time.14-18 The National missing outcomes were imputed by means of
Institute of Neurological Disorders and Stroke model-based multiple imputation showed results
Recombinant Tissue Plasminogen Activator Stroke that were similar to those of the main analysis
Study showed that patients treated with intrave- of the primary outcome, which suggests a limited
nous alteplase were 30% more likely than patients effect of bias.
in the placebo group to have an excellent outcome In conclusion, the beneficial effect of endo-
(i.e., a modified Rankin scale score of 0 or 1) at vascular treatment in patients with acute ischemic
1 year, which was similar to the 90-day results stroke caused by a proximal intracranial occlusion
of the MR CLEAN trial.15 Furthermore, in the of the anterior circulation was sustained during
Interventional Management of Stroke Phase III the course of at least 2 years.
Trial, the percentage of patients with a good out-
Supported by the Netherlands Organization for Health Re-
come (i.e., a modified Rankin scale score of 0 to 2) search and Development. The MR CLEAN trial was partly sup-
at 1 year was higher among patients who reported by the Dutch Heart Foundation and through unrestricted
ceived endovascular treatment than among pa- grants from AngioCare BV, Covidien/EV3, MEDAC/LAMEPRO,
Stryker, and Penumbra.
tients who received intravenous alteplase alone Disclosure forms provided by the authors are available with
in a subgroup of patients with severe stroke the full text of this article at NEJM.org.

References
1. Berkhemer OA, Fransen PSS, Beumer D, tion: a clinical examination scale. Stroke 16. Broderick JP, Palesch YY, Demchuk
et al. A randomized trial of intraarterial 1989;20:864-70. AM, et al. Endovascular therapy after in-
treatment for acute ischemic stroke. N Engl 9. Wilson JTL, Hareendran A, Grant M, travenous t-PA versus t-PA alone for stroke.
J Med 2015;372:11-20. et al. Improving the assessment of out- N Engl J Med 2013;368:893-903.
2. Goyal M, Demchuk AM, Menon BK, comes in stroke: use of a structured inter- 17. Palesch YY, Yeatts SD, Tomsick TA,
et al. Randomized assessment of rapid view to assign grades on the modified et al. Twelve-month clinical and quality-
endovascular treatment of ischemic stroke. Rankin Scale. Stroke 2002;33:2243-6. of-life outcomes in the Interventional Man-
N Engl J Med 2015;372:1019-30. 10. EuroQol Group. EuroQol — a new agement of Stroke III Trial. Stroke 2015;
3. Saver JL, Goyal M, Bonafe A, et al. facility for the measurement of health- 46:1321-7.
Stent-retriever thrombectomy after intra- related quality of life. Health Policy 1990; 18. Davalos A, Cobo E, Chamorro A, et al.
venous t-PA vs. t-PA alone in stroke. N Engl 16:199-208. Randomized trial of revascularization with
J Med 2015;372:2285-95. 11. Banks JL, Marotta CA. Outcomes Solitaire FR device versus best medical
4. Campbell BCV, Mitchell PJ, Kleinig TJ, validity and reliability of the modified therapy in the treatment of acute stroke
et al. Endovascular therapy for ischemic Rankin scale: implications for stroke due to anterior circulation large vessel oc-
stroke with perfusion-imaging selection. clinical trials: a literature review and syn- clusion presenting within eight hours of
N Engl J Med 2015;372:1009-18. thesis. Stroke 2007;38:1091-6. symptom onset (REVASCAT Trial) — fi-
5. Jovin TG, Chamorro A, Cobo E, et al. 12. van Swieten JC, Koudstaal PJ, Visser nal results at 12 months. Presented at the
Thrombectomy within 8 hours after symp- MC, Schouten HJ, van Gijn J. Interobserver International Stroke Conference, Los An-
tom onset in ischemic stroke. N Engl J Med agreement for the assessment of handi- geles, February 17–19, 2016. abstract.
2015;372:2296-306. cap in stroke patients. Stroke 1988;19: 19. Dabus G, Linfante I. The natural his-
6. Goyal M, Menon BK, van Zwam WH, 604-7. tory of acute ischemic stroke due to intra-
et al. Endovascular thrombectomy after 13. Lamers LM, McDonnell J, Stalmeier cranial large-vessel occlusion: what do we
large-vessel ischaemic stroke: a meta- PFM, Krabbe PFM, Busschbach JJV. The know? Tech Vasc Interv Radiol 2012;15:
analysis of individual patient data from Dutch tariff: results and arguments for an 2-4.
five randomised trials. Lancet 2016;387: effective design for national EQ-5D valua- 20. Rerkasem K, Rothwell PM. Carotid
1723-31. tion studies. Health Econ 2006;15:1121-32. endarterectomy for symptomatic carotid
7. Fransen PSS, Beumer D, Berkhemer 14. Schmitz ML, Simonsen CZ, Hundborg stenosis. Cochrane Database Syst Rev
OA, et al. MR CLEAN, a multicenter ran- H, et al. Acute ischemic stroke and long- 2011;4:CD001081.
domized clinical trial of endovascular term outcome after thrombolysis: nation- 21. Saxena R, Koudstaal PJ. Anticoagu-
treatment for acute ischemic stroke in the wide propensity score-matched follow-up lants for preventing stroke in patients
Netherlands: study protocol for a ran- study. Stroke 2014;45:3070-2. with nonrheumatic atrial fibrillation and
domized controlled trial. Trials 2014;15: 15. Kwiatkowski TG, Libman RB, Frankel a history of stroke or transient ischaemic
343. M, et al. Effects of tissue plasminogen ac- attack. Cochrane Database Syst Rev 2004;
8. Brott T, Adams HP Jr, Olinger CP, et al. tivator for acute ischemic stroke at one 1:CD000185.
Measurements of acute cerebral infarc- year. N Engl J Med 1999;340:1781-7. Copyright © 2017 Massachusetts Medical Society.
specialties and topics at nejm.org

Specialty pages at the Journal’s website (NEJM.org) feature articles in cardiology,
endocrinology, genetics, infectious disease, nephrology, pediatrics, and many other
medical specialties. These pages, along with collections of articles on clinical and
nonclinical topics, offer links to interactive and multimedia content and feature
recently published articles as well as material from the NEJM archive (1812–1989).


Two-Year Outcome After Endovascular Treatment For Acute Ischemic Stroke

Uploaded by

Copyright:

Available Formats

Two-Year Outcome After Endovascular Treatment For Acute Ischemic Stroke

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

Two-Year Outcome After Endovascular Treatment For Acute Ischemic Stroke

Uploaded by

Copyright:

Available Formats

The n e w e ng l a n d j o u r na l of m e dic i n e

Two-Year Outcome after Endovascular

n engl j med 376;14 nejm.org April 6, 2017 1341

1342 n engl j med 376;14 nejm.org April 6, 2017

The New England Journal of Medicine

score.9 Medical events that occurred between fol- Statistical Analysis

n engl j med 376;14 nejm.org April 6, 2017 1343

1344 n engl j med 376;14 nejm.org April 6, 2017

The New England Journal of Medicine

n engl j med 376;14 nejm.org April 6, 2017 1345

curred during 239 person-years of follow-up in

Figure 1. Modified Rankin Scale Scores at 90 Days and 2 Years According to

1346 n engl j med 376;14 nejm.org April 6, 2017

The New England Journal of Medicine

Table 2. Primary and Secondary Outcomes and Treatment Effects.*

Intervention Control Effect Unadjusted Value Adjusted Value

* Plus−minus values are means ±SD. IQR denotes interquartile range.

Subgroup Analyses (32.6%). Although the above results are similar,

n engl j med 376;14 nejm.org April 6, 2017 1347

(NIHSS score ≥20) (32.5% vs. 18.6%, P = 0.04).16,17

1348 n engl j med 376;14 nejm.org April 6, 2017

The New England Journal of Medicine

specialties and topics at nejm.org

n engl j med 376;14 nejm.org April 6, 2017 1349

You might also like

* Plus−minus values are means ±SD. IQR denotes interquartile range.