Mechanical Thrombectomy for Acute Ischemic Stroke

Final Results of the Multi MERCI Trial

Wade S. Smith, MD, PhD; Gene Sung, MD, MPH; Jeffrey Saver, MD; Ronald Budzik, MD;
Gary Duckwiler, MD; David S. Liebeskind, MD; Helmi L. Lutsep, MD; Marilyn M. Rymer, MD;
Randall T. Higashida, MD; Sidney Starkman, MD; Y. Pierre Gobin, MD;
for the Multi MERCI Investigators

Background and PurposeEndovascular mechanical thrombectomy may be used during acute ischemic stroke due to
large vessel intracranial occlusion. First-generation MERCI devices achieved recanalization rates of 48% and, when
coupled with intraarterial thrombolytic drugs, recanalization rates of 60% have been reported. Enhancements in
embolectomy device design may improve recanalization rates.
MethodsMulti MERCI was an international, multicenter, prospective, single-arm trial of thrombectomy in patients with
large vessel stroke treated within 8 hours of symptom onset. Patients with persistent large vessel occlusion after IV tissue
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016

plasminogen activator treatment were included. Once the newer generation (L5 Retriever) device became available,
investigators were instructed to use the L5 Retriever to open vessels and could subsequently use older generation devices
and/or intraarterial tissue plasminogen activator. Primary outcome was recanalization of the target vessel.
ResultsOne hundred sixty-four patients received thrombectomy and 131 were initially treated with the L5 Retriever.
Mean ageSD was 6816 years, and baseline median (interquartile range) National Institutes of Health Stroke Scale
score was 19 (15 to 23). Treatment with the L5 Retriever resulted in successful recanalization in 75 of 131 (57.3%)
treatable vessels and in 91 of 131 (69.5%) after adjunctive therapy (intraarterial tissue plasminogen activator,
mechanical). Overall, favorable clinical outcomes (modified Rankin Scale 0 to 2) occurred in 36% and mortality was
34%; both outcomes were significantly related to vascular recanalization. Symptomatic intracerebral hemorrhage
occurred in 16 patients (9.8%); 4 (2.4%) of these were parenchymal hematoma type II. Clinically significant procedural
complications occurred in 9 (5.5%) patients.
ConclusionsHigher rates of recanalization were associated with a newer generation thrombectomy device compared with
first-generation devices, but these differences did not achieve statistical significance. Mortality trended lower and the
proportion of good clinical outcomes trended higher, consistent with better recanalization. (Stroke. 2008;39:1205-1212.)
Key Words: acute stroke fibrinolytic thrombectomy

I schemic stroke caused by large vessel occlusions (middle

cerebral, basilar artery, and carotid terminus) is particularly
morbid1 and neurological outcome is dependent on timely
recanalization. Thrombus within vessels of this size is rela-
tively resistant to dissolution from plasminogen activators
delivered intravenously2 4 providing a reason to pursue direct
endovascular techniques to open vessels. Intraarterial (IA)
plasminogen activators delivered directly to the clot have
been shown to significantly restore perfusion and improve
clinical outcomes for middle cerebral artery (MCA) occlu-
sions5 but may be less effective for carotid terminus or basilar
artery occlusions.6 Endovascular mechanical thrombectomy
can restore vascular patency of these vessels between 41%
and 54% of the time,79 providing an alternative or synergistic
method to restore blood flow. Because clinical outcome is
improved with better recanalization rates,8 11 further im-
provements in recanalization rates are desirable.
The Multi MERCI trial was designed in part to test the
performance of a newer generation thrombectomy device
designed to attain better grasp of the thrombus compared with
first-generation devices and in part to further evaluate the
safety and efficacy of combining IV tissue plasminogen
activator (tPA) with mechanical thrombectomy. Interim data
regarding the safety of combining IV tPA with thrombectomy

Received June 19, 2007; final revision received August 15, 2007; accepted September 11, 2007.
From the Departments of Neurology (W.S.S.) and Radiology (R.T.H.), University of California, San Francisco, San Francisco, Calif; the Department
of Neurology (G.S.), University of Southern California, Los Angeles, Calif; the Departments of Neurology (J.S., D.S.L.), Radiology (G.D.), and
Emergency Medicine (S.S.), University of California, Los Angeles, Los Angeles, Calif; Riverside Methodist Hospital (R.B.), Columbus, Ohio; Oregon
Health Sciences (H.L.L.), Portland, Ore; St Lukes Hospital (M.M.R.), Kansas City, Kan; and the Department of Radiology (Y.P.G.), New York
Presbyterian HospitalCornell, New York, NY.
Correspondence to Wade S. Smith, MD, PhD, Department of Neurology, University of California, San Francisco, 505 Parnassus Avenue, San
Francisco, CA 94143-0114. E-mail smithw@neurology.ucsf.edu
2008 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STROKEAHA.107.497115

1205
 1206 Stroke April 2008
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016
was reported previously9 and the final results are reported
here.
Materials and Methods
Multi MERCI was an international, multicenter, single-arm trial
using a family of thrombectomy devices (Merci Retriever X5, X6
and L5 Retriever) to restore cerebral perfusion (Figure 1). Treatment
was initiated within 8 hours of stroke symptom onset. The trial had
3 broad aims: (1) to gain greater experience with the Merci Retriever
first-generation devices (X5 and X6) in patients ineligible for IV
tPA, supplementing the data obtained in the MERCI trial8; (2) to
explore the safety and technical efficacy of using the Merci Retriever
in patients treated with IV tPA who failed to recanalize promptly;
and (3) to obtain safety and technical efficacy data on a second-
generation thrombectomy device (L5 Retriever) once these became
available for trial investigation. The X5 and X6 models were cleared
for commercial use in August 2004, and the L5 Retriever was used
under an US Food and Drug Administration-approved Investiga-
tional Device Exemption as part of this trial. The trial enrolled
patients at 15 sites, including 2 Canadian sites, shown in the
Appendix, and the trial protocol was approved by each local
Institutional Review Board. The study was supervised by an inde-
pendent data safety monitoring board (DSMB).
Patient eligibility in the IV tPA ineligible arm of Multi MERCI
was the same as the MERCI trial.8 Eligibility in the IV tPA-treated
arm was the same as in the IV tPA-ineligible arm except that patients
who had received tPA within 3 hours of onset under US Food and
Drug Administration-labeled indications could be enrolled if tPA
failed to open the intracranial large vessel as proved by conventional
angiography. Specifically, patients were eligible who met all of the
following criteria: age 18 years, signs and symptoms of acute
stroke, National Institutes of Health Stroke Scale (NIHSS) score 8,
and stroke symptom duration under 8 hours. After cerebral angiog-
raphy, eligible patients had to have occlusion of a treatable vessel.
Treatable vessels were defined as the intracranial vertebral artery,
basilar artery, intracranial carotid artery, internal carotid artery
terminus, or the MCA first division (M1) or second division (M2).
The patient was defined as enrolled once the balloon guide catheter
was placed in the vasculature.
Patients were ineligible for the study if any of the following were
true: informed consent was not obtained (and approval for waiver of
explicit consent for emergency circumstances had not been obtained
at the study site), current pregnancy, serum glucose 50 mg/dL,
excessive tortuosity of cervical vessels precluding device delivery/
deployment, known hemorrhagic diathesis, known coagulation fac-
tor deficiency, oral anticoagulation treatment with international
Figure 1. Drawing of the L5 device (top) and
X6 device (bottom) engaging a clot. The L5
device differs from the X5/X6 family of devices
by having a series of monofilaments that attach
proximal and distal to the helical nitinol coils.
The balloon guide catheter is placed more
proximal (not shown) to arrest flow in the vas-
cular segment as the blue microcatheter and
device are pulled proximally to remove the clot.
normalized ratio 3.0, use of heparin within 48 hours and a
prothrombin time 2 times normal, platelet count 30 000/L,
history of severe allergy to contrast media, sustained systolic blood
pressure 185 mm Hg or diastolic blood pressure 110 mm Hg
despite treatment, CT scan revealing significant mass effect with
midline shift, greater than 50% stenosis of the artery proximal to the
target vessel, or life expectancy under 3 months.
All patients underwent conventional cerebral angiography. The
embolectomy procedure has been described previously.9 Successful
recanalization was defined as achieving Thrombolysis In Myocardial
Infarction12 (TIMI) II or III flow in all treatable vessels. Successful
recanalization for the MCA required all M1 and M2 segments to be
at least TIMI II; for internal carotid artery terminus lesions, the
internal carotid artery, M1, and both M2 branches needed to be at
least TIMI II; and for the posterior circulation, both the vertebral and
basilar arteries needed to be at least TIMI II to be considered
recanalized. TIMI scoring of angiography was scored by the local
investigator who was not blinded to clinical outcome. If the treatable
vessel was not opened to at least TIMI II flow with a maximum of
6 passes with the device, it was considered a treatment failure for the
device. Intraarterial fibrinolysis (up to 24 mg tPA) was allowed in
cases of treatment failure with the device or to treat distal embolus
not accessible to the device after successful proximal embolectomy.
Use of glycoprotein (GP IIb/IIIa) antagonists and alternate mechan-
ical thrombectomy procedures were prohibited. Aspirin but not IV
heparin was allowed in the first 24 hours after the procedure among
patients who had not received fibrinolytics.
Patient demographics, medical history, vital signs, and routine
laboratory values were documented on standardized clinical report
forms. The NIHSS and modified Rankin scores (mRS) were obtained
at baseline and at 30 and 90 days. CT or MRI brain imaging was
performed at baseline, 24 hours, and at any time there was a decline
in patient neurological status.
The prespecified primary efficacy end point was L5 Retriever
device recanalization rate 44%, which represents 1 SD below the
mean of MERCI postdevice recanalization rate (48.2%4.2%)8 as a
test of noninferiority to the X5/X6 device. Postdevice recanalization
was defined as TIMI grades II and III flow assessed immediately
posttreatment with the device. Final recanalization was assessed
following any and all procedures, including administration of IA
plasminogen activators.
The major secondary end point was safety defined as the major
device-related serious adverse event rate 10%, which represents a
5% increase of this rate beyond the 5% reported rate in MERCI.8
Procedure-related adverse events were defined as vascular perfora-
tion, intramural arterial dissection, or embolization of a previously
 Smith et al
Table 1. Patient Demographics, Baseline Stroke Score, and
Site of Vascular Occlusion
No. of patients 164
Age, meanSD, years 68.116.0
Female, % 57%
Baseline NIHSS, median (interquartile range) 19 (1523)
Baseline mRS0 86%
Baseline mRS, meanSD 0.340.95
Site of vascular occlusion, % (n)
ICA/ICA terminal bifurcation 32% (52)
Middle cerebral artery 60% (98)
Vertebral/basilar/P1 8% (14)
ICA indicates internal carotid artery.
uninvolved territory, symptomatic hemorrhage adjudicated as
procedure-related, and access site complications requiring surgery or
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016
transfusion. Clinically significant procedural complications were
defined as a procedure complication with decline in NIHSS of 4 or
death or groin complication requiring surgery or blood transfusion.
Symptomatic intracranial hemorrhage was defined as a 4 or more
point decline in the NIHSS score within 24 hours with any blood
products identified on 24-hour head CT/MRI scan (petechial bleed-
ing, hematoma, or subarachnoid hemorrhage) or any intracranial
hemorrhage in which no further NIHSS scores were available
beyond baseline and the patient died. All 24-hour CT/MRI scans
were reviewed at a central core laboratory and adjudicated as
subarachnoid hemorrhage or as ECASS hemorrhagic infarction type
I or II or ECASS parenchymal hematoma types I and II.13 All
symptomatic intracerebral hemorrhages were reviewed by the
DSMB and adjudicated as to whether they were related to the
procedure; in cases in which IV or IA plasminogen activator was
given, the DSMB classified any hemorrhage as procedure-related.
Asymptomatic hemorrhage was defined as evidence of any blood on
the 24-hour CT or MRI scan with no more than a 3-point decline in
the NIHSS score.
Other secondary outcomes included clinical outcome, as measured
by the mRS at 90 days, 90-day mortality, and safety dichotomized by
use or no use of IV tPA. Good neurological outcome was defined as
mRS 2.
Statistical Analysis
Primary outcomes are reported based on patients who had the
Retriever deployed. Statistical tests used to determine the signifi-
cance of differences in variables are listed in the data tables and
within the text where relevant. All analyses were performed by a
biostatistician using SAS for Windows, version 8.2 (SAS Institute,
Inc, Cary, NC).
Results
Overall, 1088 patients were screened, 177 patients were
enrolled, and the device was deployed in 164 patients. Patient
demographics and baseline characteristics are shown in Table
1. Patient enrollment began January 20, 2004. The trial was
placed on hold May 2, 2005, by the DSMB because of a
question of safety regarding intracranial hemorrhages and
protocol violations. After review, the DSMB allowed the trial
to continue with stricter control on protocol violations.9 Final
enrollment occurred in July 2006. Thirteen patients did not
have the device deployed for the following reasons: vessel
tortuosity (n4), clot not penetrable with the microcatheter
(n1), spontaneous recanalization (n2), distal clot migra-
tion to the M3 segment (n1), and presence of significant
Final Results of the Multi MERCI Trial 1207
Table 2. Major Outcomes
Post Retriever recanalization, % (95% CI) 55 (4763)
Final recanalization, % (95% CI) 68 (6175)
Symptom onset to groin puncture, median hours 4.3 (3.25.3)
(IQR)
Procedure duration, median hours (IQR) 1.6 (1.22.3)
Attempts to remove clot, meanSD 2.91.6
IV tPA pretreatment, % 29
Procedural complications, % (95% CI) 9.8 (5.214.3)
Clinically significant procedure complication, % 5.5 (2.09.0)
(95% CI)
Device-related serious adverse event rate, % 2.4 (0.14.8)
(95% CI)
Good outcome (mRS 2) at 90 days, % (95% CI) 36 (2944)
NIHSS improvement of 10 points or 0 score at 26 (1933)
24 hours, % (95% CI)
Mortality at 90 days, % (95% CI) 34 (2641)
Symptomatic intracranial hemorrhage, % (95% CI) 9.8 (5.214.3)
Symptomatic PH-2 hemorrhage, % (95% CI) 2.4 (0.14.8)
IQR indicates interquartile range.
carotid stenosis (n5). All results are analyzed for the 164
patient cohort in whom the device was deployed.
All patients underwent conventional angiography and were
enrolled in the trial based on complete occlusion of the
intracranial vessels shown in Table 1. Overall recanalization
after device treatment alone and device followed by adjuvant
IA therapy is shown in Table 2. Postdevice recanalization was
55%, which exceeded 44%, thus meeting the prespecified
primary end point of the study (see Methods). The device-
related serious adverse events rate was 2.4%, which is below
the prespecified adverse event rate of 10%, thus meeting the
secondary prespecified end point. Clinically significant pro-
cedure complications occurred in 9 (5.5%) patients and did
not occur more often with any particular device. There was
one L5 device fracture on withdrawal of the device into the
balloon guide catheter and the fractured end was retrieved
with a snare; the patient had successful recanalization with no
adverse outcome and had a mRS score of 0 at 90 days. There
were 19 off-protocol mechanical interventions to address clot,
including 10 patients treated with snares or other foreign
body retrievers and 9 with balloons at the thrombus site; 11
patients received GP IIB/IIIa antagonists; 14 received IA
thrombolytics before or between passes and 10 had a proxi-
mal stenosis treated with a stent before embolectomy. These
protocol violations typically occurred as multiples within
individual patients; see the discussion in part I of this trial.9
At 90 days, 36% of patients had a favorable neurological
outcome (mRS 0 to 2) and 34% had died. Prestroke mRS
were 0 in 86% of patients, but 23 patients had mRS greater
than 0 on entry (9 had mRS1, one had mRS2, 8 had
mRS3, and 5 had mRS4). In patients with a baseline mRS
of 0, favorable neurological outcome (mRS 0 to 2) was seen
in 39.4% and 32% died. Favorable outcome was often
apparent at 24 hours; for patients who had an NIHSS
determined at 24 hours (146 cases), 32% (46 of 146) had an
8 or more point decrease in NIHSS or achieved a 0 score,
 1208 Stroke April 2008

Table 3. NIHSS, Recanalization, and Outcomes by Site of Vascular Occlusion for L5 Retriever Patients and
All Patients
Site of Occlusion

Posterior ICA-T MCA M1 MCA M2

N8 L5 N41 L5 N67 L5 N15 L5
Device N14 All N52 All N77 All N21 All P
Age mean, years L5 60 69 70 70 NS
All 62 67 70 68 NS
Baseline median NIHSS L5 17 21 17 18 0.013
All 19 21 17 16 0.027
Post Retriever recanalization, % L5 88 59 48 80 0.024
All 71 52 48 76 NS
Final recanalization, % L5 100 71 60 93 0.003
All 86 65 61 91 0.017
Favorable outcome (mRS 2), % L5 38 33 37 50 NS
All 29 33 36 50 NS
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016

Mortality, % L5 38 48 29 14 NS
All 43 45 29 15 NS
Symptomatic hemorrhage, % L5 13 9.8 10 6.7 NS
All 29 7.7 9.1 4.8 NS
Symptomatic PH-2 hemorrhage, % L5 0 2.4 4.5 0 NS
All 0 1.9 3.9 0 NS
P values are for differences in values across vessel categories; NS, P0.05; P values were calculated as follows: likelihood ratio
chi-squared tests for categorical variables except symptomatic PH-2 hemorrhage rate where Fisher exact test was used; analysis of
variance for age, and Kruskal-Wallis test for NIHSS scores.
ICA-T indicates internal carotid artery terminus.

whereas 26% (38 of 146) exhibited a 10 or more point
decrease or achieved a 0 score.
Among cases in which the L5 Retriever was deployed,
primary recanalization was achieved in 75 of 131 cases
(57.3%) as compared with recanalization in 15 of 33 (45.5%)
of cases in which older generation X5/X6 devices were the
first device deployed (P0.25). In patients receiving postad-
junctive IA therapies, recanalization was achieved in 91 of
131 patients (69.5%) in whom the L5 Retriever was used as
compared with 21 of 33 (63.6%) patients treated with the X5
or X6 only (P0.54).
Recanalization by target vessel is shown in Table 3. There
were significant differences in recanalization by vessel both
for L5 postdevice recanalization and for final recanalization
for both L5 and all patients. Mortality and proportion of good
neurological outcomes were not significantly different across
vessel categories. However, a significant difference in mor-
tality and proportion of patients with good neurological
outcome was observed based on vessel recanalization (Table
4). Furthermore, there were more patients with mRS 0, 1, and
2 scores and less with mRS 3, 4 and 5 scores in those who
recanalized (Figure 2, P0.001). Nearly half of recanalized
patients had a good neurological outcome and there were
fewer disabled survivors among those with recanalization;
therefore, patients did not appear to be saved only to live
disabled.
Forty-eight patients (29.3%) received IV tPA before an-
giography; as reported previously,9 no differences in the rates
of intracranial hemorrhage or clinically significant procedure
complications were seen between those patient treated with
IV tPA and those who were not (Table 5) suggesting that
pretreatment with IV tPA does not raise a concern of safety.
The significant difference in time of onset between those
receiving IV tPA and those who did not reflects the under-
lying eligibility for using IV tPA within 3 hours of symptom
onset. The significant difference in rates of IA thrombolytic
use between those patients who had successful vascular
recanalization with the device and those who did not reflects

Table 4. Neurological Outcomes at 90 Days

Overall
Overall Overall Recanalized Not Recanalized Relative Risk
Assessment (N164) (N112) (N52) (95% CI) P Value
Favorable outcome,* % (95% CI) 36 (2944) 49 (4059) 9.6 (1.618) 5.1 (2.212) 0.001
Mortality at 90 days, % (95% CI) 34 (2641) 25 (1733) 52 (3866) 0.48 (0.310.73) 0.001
*mRS 2.
P value is for ad hoc testing of the difference in outcome rates between the recanalized and not recanalized groups using a
two-tailed Fisher exact test.
 Smith et al Final Results of the Multi MERCI Trial 1209

Figure 2. Clinical outcome at 90 days

by mRS.

differential investigator use of adjuvant thrombolytic for P0.60). The rate of symptomatic PH-2 hemorrhages in
cases that failed to recanalize with the device. patients receiving no tPA was 1.3% (one of 76). Asymptom-
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016

Symptomatic hemorrhages were observed in 16 (9.8%)
patients and representative images of each patient are shown
in Figure 3. Each hemorrhage was classified as to hemorrhage
type and adjudicated by a core laboratory neuroradiologist;
disagreement between the site reading and this core labora-
tory interpretation was adjudicated by a second neuroradiolo-
gist. Only 4 of the hemorrhages (2.4%) were classified as
PH-2, suggesting that 12 of the 16 patients may have declined
not from the intracranial blood, but from cytotoxic edema
generated from the initial infarction. Three of the 4 PH-2
patients had hemorrhages adjudicated as procedure compli-
cations: one received IA tPA (Figure 3K), one was due to
guidewire perforation of the MCA vessel during IA tPA
infusion (Figure 3L), and one received IV tPA (Figure 3M).
Another patient (Figure 3N) had a symptomatic PH-2 hem-
orrhage that was not considered procedure-related because no
thrombolytics were administered and no apparent problem
during the procedure was reported by the investigator. The
rate of symptomatic hemorrhage in patients treated without
any tPA was 7.9% (6 of 76), whereas the rate of symptomatic
hemorrhages in patients given any tPA was 11% (10 of 88;
atic hemorrhages were observed in 30.5% of all patients.
Discussion
The final results of the Multi MERCI trial reveal that the
studys prespecified primary end point of recanalization was
achieved. The study found an increased rate of recanalization
of intracranial vessels using a newer generation thrombec-
tomy device with a higher rate (although not statistically
significant) of device-only recanalization among L5
Retriever-treated patients in Multi MERCI than among X5/
X6-treated patients in MERCI. In addition, these results
support the preliminary conclusions of Multi MERCI, part I
that pretreatment with IV tPA followed by mechanical
thrombectomy has an acceptable safety profile. Among pa-
tients who experienced recanalization, there was a 2-fold
survival advantage and a significantly higher proportion of
patients lived without significant disability.
Given the large separation between outcomes in recana-
lized and nonrecanalized patients, the stringent definition of
recanalization used in the Multi MERCI trial appears to be a
clinically meaningful one. Because there were no contempo-

Table 5. Use of IV or IA Thrombolytics

IV tPA No IV tPA IA Lytic No IA Lytic
Result (N48) (N116) P Value* (N57) (N107) P Value
Recanalization post Retriever, % 58 53 0.61 33 66 0.001
Recanalization post Adjuvant, % 73 66 0.46 68 68 0.99
Symptom onset to arterial puncture, hours 3.9 4.6 0.031 3.7 4.8 0.001
IA lytic use, % 35 34 0.99 100 0
mRS 2 at 90 days, % 38 35 0.72 32 39 0.49
Mortality at 90 days, % 28 36 0.36 43 29 0.08
Intracranial hemorrhage
Symptomatic ICH, % 10 9.5 0.99 14 7.5 0.27
Symptomatic PH-2, % 2.1 2.6 0.99 3.5 1.9 0.61
Clinically significant procedure complications, % 4.2 6.0 0.99 12 1.9 0.009
*Difference between IV and no IV tPA groups.
Difference between IA and no IA groups. Fisher exact test for all tests of significance, except symptom onset to arterial puncture in which
analysis of variance was used.
ICH indicates intracerebral hemorrhage.
 1210 Stroke April 2008
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016
raneous controls, the trial is unable to conclusively show that
thrombectomy actually improves stroke outcomes; a random-
ized trial would help answer that question. However, trial
results do strongly suggest that a treatment effect exists
despite the lack of contemporaneous controls. First, the
remarkable 27% absolute difference in mortality between
recanalizers and nonrecanalizers is consistent with the known
pathophysiology of stroke whereby timely recanalization of
ischemic brain mitigates tissue death and leads to better
survival and better outcomes of patients because less brain is
injured. The only other comparable effect of any stroke
therapy on mortality is hemicraniectomy for cytotoxic cere-
bral edema after ischemic stroke in which a 49% difference in
mortality was reported.14 It is unlikely that the mortality rate
of nonrecanalizers could have been inflated by iatrogenic
complications administered to patients whose vessels failed
to open with the procedure because the 27% difference in
mortality between groups is far higher than our blinded
adjudication of clinically significant procedural complica-
tions (5.5%). Second, the 68% final recanalization rate found
here is likely superior to that achieved with supportive
medical therapy. Based on the PROACT-II placebo arm that
showed a spontaneous recanalization rate of the MCA stem or
M2 branches of 18% at 2 hours after initial angiography, and
cerebral angiography-based studies of vascular patency of
25% to 30% within 6 hours of presumed large vessel stroke
onset,15 20% is a reasonable estimate for the spontaneous
recanalization rate of vessels of the size treated in the MERCI
trials. The reported recanalization rate of 68% in Multi
MERCI for all cerebral vessels, not just MCA vessels, is
similar to that reported in PROACT-II for MCA or M2
branches only (66%). Given that clot burdens in the internal
carotid artery terminus and basilar artery can be substantially
higher, and therefore less likely to be recanalized with
Figure 3. Representative images of all
16 symptomatic intracranial hemor-
rhages defined as any blood on the
postprocedure brain imaging study in a
patient with a 4 or more point decline in
NIHSS score. Images are arranged by
hemorrhage classification. (E and L) were
also classified as having subarachnoid
blood. The hemorrhage in image (P)
occurred outside of the area of ischemia
and therefore is not classified by the
ECASS method.13
thrombolytics, our data suggest that the device provides an
advantage over IA thrombolytic therapy alone for all large
vessel occlusions. Third, timely recanalization of cerebral
vessels is a highly significant predictor of good clinical
outcome as has been reported in several series1,8,11 and in a
recent meta-analysis of earlier published series.10 Finally,
26% of patients treated in Multi MERCI had dramatic clinical
improvement within 24 hours. For patients treated with IV
tPA within 3 hours, the rate of dramatic improvement within
24 hours is 22% to 32%.16,17 Because the patients in Multi
MERCI were treated up to 8 hours after stroke, and the
baseline stroke severity is markedly higher, the 26% rate seen
in Multi MERCI suggests a treatment benefit.
Multi MERCI also met its secondary prespecified safety
end point: device-related serious adverse events rate of 2.4%.
Clinically significant procedural complications were fewer in
this trial compared with the MERCI trial (5.5% versus 7.1%,
Pnot significant), although the absence of groin complica-
tions alone in Multi MERCI accounted for this numeric
difference. Only one device fracture occurred with the L5
device and this resulted in no clinical consequence. One
strength of both studies is that these estimates of procedural
risk are based on all treated patients within the trial; no patient
was excluded as a run-in patient, which is typical of other
human device trials. Investigators were trained with phan-
toms before device use but were not required to have treated
patients outside of the trial first before being able to enroll.
Therefore, the results reported here are likely representative
of how the device performs outside of investigational use.
Symptomatic intracranial hemorrhages were observed in
9.8% of cases as adjudicated by a radiology core laboratory
and reviewed by the independent DSMB. However, inspec-
tion of Figure 3 shows that only 4 of the 16 hemorrhages were
classified as PH-2, representing hematoma with significant
 Smith et al
mass effect on neighboring brain.13 Most of the hemorrhages
are hemorrhagic transformation of ischemic infarcts, which is
seen more often after vascular recanalization and has been
associated with more favorable neurological outcomes18 or no
clinical adverse effect.19 The DSMB classified a patient as
having a symptomatic hemorrhage if the patient declined by
4 or more NIHSS points and had any blood products on the
24-hour CT scan. This conservatism was proscribed to err on
the side of safety to ensure that no potential symptomatic
hemorrhages would be missed. If one considers, however,
only the symptomatic hemorrhages that were sizable enough
to cause mass effect, only 4 hemorrhages (2.4%) met these
criteria. This definition of symptomatic intracranial hemor-
rhage was used in a European registry of IV tPA-treated
patients20 because it was felt to be more representative of the
true clinically meaningful intracranial hemorrhage rate.
This trial has several limitations. Because the trial included
adults with no upper age limit, nor an upper limit to NIHSS,
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016
and 14% of treated patient had prestroke mRS scores exceed-
ing 0, the overall mortality results are difficult to compare
with other published stroke trials that exclude elderly patients
and those with high stroke severity and typically enroll only
patients with no prestroke disability. As previously reported,
the outcomes for patients in MERCI who were PROACT-II-
eligible was comparable to the PROACT-II outcomes8; there-
fore, although the mortality rates are relatively high in both
trials, this likely represents the overall stroke severity of the
patients enrolled. Finally, although all CT scans were re-
viewed by a blinded central reviewer, the conventional
angiograms were interpreted by the site investigator.
Mechanical embolectomy is undergoing further analysis in
the MR RESCUE and IMS-III trials. Both trials are random-
ized and each will provide important efficacy data with
primary clinical outcomes. At present, mechanical embolec-
tomy is a useful technique for restoring blood in patients with
large vessel acute ischemic stroke, especially in those who are
ineligible for thrombolytics or in those who have failed
thrombolytic therapy.
Appendix
Investigators
International Principal Investigator: Wade S. Smith, MD, PhD,
University of California, San Francisco. Data Safety Monitoring
Board: Chair: Gene Sung, MD, MPH, University of Southern
California. Biostatistician: Phil Hormel, MS. Members: Tim W.
Malisch, MD, Alexian Brothers Medical Center; Steven Rudolph,
MD, Maimonides Medical Center; and Arun Amar, MD, Stanford
University. Imaging Core Laboratory: Paul Kim, MD, University of
Southern California. Biostatistician: Phil Hormel, MS. Writing
Committee: Ronald Budzik, MD; Gary Duckwiler, MD; Donald Frei,
MD; Y. Pierre Gobin, MD; Thomas Grobelny, MD; Randall T.
Higashida; Frank Hellinger, MD; Dan Huddle, MD, MD; Chelsea
Kidwell, MD; Walter Koroshetz, MD; David S. Liebeskind, MD;
Helmi L. Lutsep, MD; Michael Marks, MD; Gary Nesbit, MD;
Marilyn M. Rymer, MD; Jeffrey Saver, MD; Isaac E. Silverman,
MD; Wade S. Smith, MD, PhD; Sidney Starkman, MD; and Gene
Sung, MD, MPH. Site Principal Investigator (PI), Coinvestigators,
and Study Coordinators in order of number of patients treated (N):
St Lukes Hospital: (50) Co-PIs: Naveed Akhtar, MD, and Thomas
Grobelny, MD; Annette Allen, RN; Steven Arkin, MD; Irene
Bettinger, MD; Christine Boutwell, MD; Charlene Grau, RN; Bar-
bara Gruenenfelder, RN; Marilyn Rymer, MD; Michael Schwartz-
Final Results of the Multi MERCI Trial 1211
man, MD; and Charles Weinstein, MD. Riverside Methodist Hos-
pital: (32) PI: Ronald Budzik, MD; Erik Arce, MD; Albert
Berarducci, MD; Tom Davis, MD; Mark Dean, MD; Eric Dolen;
Geoffrey Eubank, MD; Jim Fulop, MD; Xiamei Gao-Hickman,
MD; John Lippert, MD; William Mayr, MD; J. Kevin McGraw, MD;
Paula Meyers, RN; Peter Pema, MD; and Robert Wyatt, MD.
Oregon Stroke Center: (21) PI: Helmi Lutsep, MD; Stanley
Barnwell, MD; Wayne Clark, MD; Barbara Dugan, RN; Robert
Egan, MD; Todd Kuether, MD; Ted Lowenkopf, MD; Gary Nesbit,
MD; Elizabeth North, MD; Bryan Peterson, MD; John Roll, MD;
and Lisa Yanase, MD. The Stroke Center at Hartford Hospital:
(14) PI: Isaac Silverman, MD; Martha Ahlquist, LPN, CCRP; Dawn
Beland, MSN; Joao Gomes, MD; Stephen Ohki, MD; and Gary
Speigel, MD. University of California at Los Angeles Medical
Center (12): PI: Sidney Starkman, MD; Latisha Ali; Brian Buck,
MD; Dennis Chute, MD; Gary Duckwiler, MD; Judy Guzy, RN;
Reza Jahan, MD; Doojin Kim, MD; David S. Liebeskind, MD;
Victor Marder, MD; Bruce Ovbiagele, MD; Venkatakrishna Rajajee,
MD; Lucas Restrepo, MD; Nerses Sanossian, MD; Jeffrey Saver,
MD; Scott Selco, MD; Samir Shah, MD; Maria Shukman, RN;
Satoshi Tateshima, MD; Amytis Towfighi, MD; Paul Vespa, MD; J.
Pablo Villablanca, MD; Harry Vinters, MD; and Fernando Vinuela,
MD. Swedish (Denver) Medical Center: (9) Co-PIs: Don Frei, MD,
and Dan Huddle, MD; Richard Bellon, MD; Christopher Finale, MD;
Carol Greenwald, MD; and Don Smith, MD. Florida Hospital
Neuroscience Institute: (8) PI: Frank Hellinger, MD; Laura Bill-
anovic, RN; and Susan Mitchell, RN. NY Presbyterian Hospital
Cornell: (4) PI: Alan Segal, MD; Y. Pierre Gobin, MD; Jeffrey Katz,
MD; Igor Ougrets, MD; Howard Riina, MD; and Kimberly Salvag-
gio, NP. University of Calgary, Foothills Hospital (4): PI: Michael
Hill, MD; Philip Barker, MD; Andrew Demchuk, MD; Imanuel
Dzialowski; Karyn Fischer, RN, MD; William Hu; Mark Hudon,
MD; Will Morrish, MD; Suresh Subramanian, MD; Tim Watson,
MD; and John Wong, MD. NY Presbyterian HospitalColumbia:
(3) PI: John Pile-Spellman, MD; Sean Lavine, MD; Philip Meyers,
MD; and Leslie Schmidt, NP. Georgetown University: (3) PI:
Vance Watson, MD; John DeSimone, MD; Timea Hodics, MD;
Theresa Kowal, RN; Farid Parham, MD; Susan Sutten, MPH; and
Manual Yepes, MD. Stanford University Medical Center: (2) PI:
Michael Marks, MD; Gregory Albers, MD; James Castle, MD; Huy
Do, MD; Amie Hsia, MD; Mahesh Jayerman, MD; Marten Lansberg,
MD; Mary Marcellus, RN; David Tong, MD; Chitra Venkatsubma-
ran, MD; and Christine Wijman, MD. University of Alberta,
Edmonton: (2) PI: Ashfaq Shuaib, MD; Robert Ashforth, MD;
Derek Emery, MD; Faraz Al-Hussain, MD; Muhammad Hussain,
MD; Thomas Jeerakathil, MD; Kurshid Khan, MD; Mikael Murtao-
ghu, MD; Nazir Rizvi, MD; Maher Saqqur, MD; James Scozzafava,
MD; Brenda Scwindt, RN; Muzaffar Siddiqui, MD; and Khalida
Tariq, MD. Baptist Memorial Clinical Research Center: PI: John
Barr, MD; Paul Broadbent, MD; Sanat Dixit, MD; Grace Miller; and
Stephen D. Morris, MD. University of Pittsburgh Medical Center:
PI: Tudor Jovin, MD; Max Hammer, MD; Michael Horowitz, MD;
Vivek Reddy, MD; Tibetha Santucci, RN; Ken Uchiro, MD; Nirav
Vora, MD; and Lawrence Wechsler, MD.
Source of Funding
This study was funded by Concentric Medical, Inc.
Disclosures
W.S.S. and J.S. received significant research support from Boehr-
inger Ingelheim. R.B. received significant honoraria from Concentric
Mediral. W.S.S., G.D., and Y.P.G. report significant ownership
interests in Concentric Medical. Y.P.G. is the inventor of the MERCI
Retriever. G.Y.S. and H.L.L. have nothing to disclose.
References
1. Smith WS, Tsao JW, Billings ME, Johnston SC, Hemphill JC III,
Bonovich DC, Dillon WP. Prognostic significance of angiographically
confirmed large vessel intracranial occlusion in patients presenting with
acute brain ischemia. Neurocrit Care. 2006;4:14 17.
 1212 Stroke April 2008
2. Tissue plasminogen activator for acute ischemic stroke. The National
Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group.
N Engl J Med. 1995;333:15811587.
3. del Zoppo GJ, Poeck K, Pessin MS, Wolpert SM, Furlan AJ, Ferbert A,
Alberts MJ, Zivin JA, Wechsler L, Busse O, et al. Recombinant tissue
plasminogen activator in acute thrombotic and embolic stroke. Ann
Neurol. 1992;32:78 86.
4. Saqqur M, Uchino K, Demchuk AM, Molina CA, Garami Z, Calleja S,
Akhtar N, Orouk FO, Salam A, Shuaib A, Alexandrov AV. Site of arterial
occlusion identified by transcranial Doppler predicts the response to
intravenous thrombolysis for stroke. Stroke. 2007;38:948 954.
5. Furlan A, Higashida R, Wechsler L, Gent M, Rowley H, Kase C, Pessin
M, Ahuja A, Callahan F, Clark WM, Silver F, Rivera F. Intra-arterial
prourokinase for acute ischemic stroke. The PROACT II study: a ran-
domized controlled trial. Prolyse in acute cerebral thromboembolism.
JAMA. 1999;282:20032011.
6. IMS Trial Investigators. Combined intravenous and intra-arterial recan-
alization for acute ischemic stroke: The Interventional Management of
Stroke Study. Stroke. 2004;35:904 911.
7. Berlis A, Lutsep H, Barnwell S, Norbash A, Wechsler L, Jungreis CA,
Woolfenden A, Redekop G, Hartmann M, Schumacher M. Mechanical
thrombolysis in acute ischemic stroke with endovascular photoacoustic
recanalization. Stroke. 2004;35:11121116.
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016
8. Smith WS, Sung G, Starkman S, Saver JL, Kidwell CS, Gobin YP, Lutsep
HL, Nesbit GM, Grobelny T, Rymer MM, Silverman IE, Higashida RT,
Budzik RF, Marks MP. Safety and efficacy of mechanical embolectomy
in acute ischemic stroke: results of the MERCI trial. Stroke. 2005;36:
14321438.
9. Smith WS. Safety of mechanical thrombectomy and intravenous tissue
plasminogen activator in acute ischemic stroke. Results of the Multi
Mechanical Embolus Removal in Cerebral Ischemia (MERCI) trial,
part I. AJNR Am J Neuroradiol. 2006;27:11771182.
10. Rha JH, Saver JL. The impact of recanalization on ischemic stroke
outcome: a meta-analysis. Stroke. 2007;38:967973.
11. Smith WS. Intra-arterial thrombolytic therapy for acute basilar occlusion:
pro. Stroke. 2007;38:701703.
12. Williams DO, Borer J, Braunwald E, Chesebro JH, Cohen LS, Dalen J,
Dodge HT, Francis CK, Knatterud G, Ludbrook P, et al. Intravenous
recombinant tissue-type plasminogen activator in patients with acute
myocardial infarction: a report from the NHLBI Thrombolysis in Myo-
cardial Infarction Trial. Circulation. 1986;73:338 346.
13. Hacke W, Kaste M, Fieschi C, von Kummer R, Davalos A, Meier D,
Larrue V, Bluhmki E, Davis S, Donnan G, Schneider D, Diez-Tejedor E,
Trouillas P. Randomised double-blind placebo-controlled trial of
thrombolytic therapy with intravenous alteplase in acute ischaemic stroke
(ECASS II). Second EuropeanAustralasian Acute Stroke Study investi-
gators. Lancet. 1998;352:12451251.
14. Vahedi K, Hofmeijer J, Juettler E, Vicaut E, George B, Algra A, Amelink
GJ, Schmiedeck P, Schwab S, Rothwell PM, Bousser MG, van der Worp
HB, Hacke W. Early decompressive surgery in malignant infarction of the
middle cerebral artery: a pooled analysis of three randomised controlled
trials. Lancet Neurol. 2007;6:215222.
15. Kassem-Moussa H, Graffagnino C. Nonocclusion and spontaneous re-
canalization rates in acute ischemic stroke: a review of cerebral
angiography studies. Arch Neurol. 2002;59:1870 1873.
16. Brown DL, Johnston KC, Wagner DP, Haley EC Jr. Predicting major
neurological improvement with intravenous recombinant tissue plasmin-
ogen activator treatment of stroke. Stroke. 2004;35:147150.
17. Felberg RA, Okon NJ, El-Mitwalli A, Burgin WS, Grotta JC, Alexandrov
AV. Early dramatic recovery during intravenous tissue plasminogen ac-
tivator infusion: clinical pattern and outcome in acute middle cerebral
artery stroke. Stroke. 2002;33:13011307.
18. Molina CA, Alvarez-Sabin J, Montaner J, Abilleira S, Arenillas JF,
Coscojuela P, Romero F, Codina A. Thrombolysis-related hemorrhagic
infarction: a marker of early reperfusion, reduced infarct size, and
improved outcome in patients with proximal middle cerebral artery
occlusion. Stroke. 2002;33:15511556.
19. Trouillas P, von Kummer R. Classification and pathogenesis of cerebral
hemorrhages after thrombolysis in ischemic stroke. Stroke. 2006;37:
556 561.
20. Wahlgren N, Ahmed N, Davalos A, Ford GA, Grond M, Hacke W,
Hennerici MG, Kaste M, Kuelkens S, Larrue V, Lees KR, Roine RO,
Soinne L, Toni D, Vanhooren G. Thrombolysis with alteplase for acute
ischaemic stroke in the Safe Implementation of Thrombolysis in Stroke-
MOnitoring STudy (SITS-MOST): an observational study. Lancet.
2007;369:275282.
 Mechanical Thrombectomy for Acute Ischemic Stroke: Final Results of the Multi MERCI
Trial
Wade S. Smith, Gene Sung, Jeffrey Saver, Ronald Budzik, Gary Duckwiler, David S.
Liebeskind, Helmi L. Lutsep, Marilyn M. Rymer, Randall T. Higashida, Sidney Starkman and
Y. Pierre Gobin
Downloaded from http://stroke.ahajournals.org/ by guest on August 19, 2016

for the Multi MERCI Investigators

Stroke. 2008;39:1205-1212; originally published online February 28, 2008;

doi: 10.1161/STROKEAHA.107.497115
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 2008 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/39/4/1205

An erratum has been published regarding this article. Please see the attached page for:
http://stroke.ahajournals.org/content/43/10/e109.full.pdf

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Stroke is online at:

http://stroke.ahajournals.org//subscriptions/
 Correction

The article, Mechanical Thrombectomy for Acute Ischemic Stroke Final Results of the Multi
MERCI Trial by Smith et al (2008;39:12051212) included an error in the Appendix. Dr Fawaz
Al-hussains name and affiliation were incorrect. The correct information appears below as well
as in the current online version.

The investigators name should appear as Fawaz Al-hussain, MD, King Saud University.

(Stroke. 2012;43:e109.)
2012 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STR.0b013e318273da81

e109