Clinical Review & Education

Review

Acute Stroke Intervention

A Systematic Review
Shyam Prabhakaran, MD, MS; liana Ruff. MD; Richard A. Bernstein, MD, PhD

[Jj Supplemental content at

im p o r t a n c e

Acute ischemic stroke is a m ajor cause o f m orta lity and m orb id ity in th e United

jama.com

States. We review th e latest data and evidence supporting catheter-directed tre a tm e n t for
proxim al artery occlusion as an adjunct to intravenous throm bolysis in patients w ith acute
stroke.

o b j e c t iv e

To review th e pathophysiology o f acute brain ischemia and infarction and the

evidence supporting various stroke reperfusion treatm ents.

e v id e n c e

REVIEW Systematic literature search o f MEDLINE databases published between

January 1,1990, and February 11,2015, was perform ed to id e n tify studies addressing the role
o f throm bolysis and mechanical th ro m b e cto m y in acute stroke management. Studies
included randomized clinical trials, observational studies, guideline statements, and review
articles. Sixty-eight articles (N = 108 0 8 2 patients) were selected fo r review.

f in d in g s

Intravenous throm bolysis is the mainstay o f acute ischemic stroke management fo r

any patient w ith disabling deficits presenting w ith in 4.5 hours from sym ptom onset.
Randomized trials have dem onstrated th a t more patients return to having good function
(defined by being independent and having slight disability or less) when treated w ith in
4.5 hours afte r sym ptom onset w ith intravenous recom binant tissue plasminogen activator
(IV rtPA) therapy. Mechanical th ro m b e cto m y in select patients w ith acute ischemic stroke and
proxim al artery occlusions has dem onstrated substantial rates o f partial or com plete arterial
recanalization and im proved outcom es compared w ith IV rtPA or best medical trea tm e nt
alone in m ultiple randomized clinical trials. Regardless o f mode o f reperfusion, earlier
reperfusion is associated w ith b e tte r clinical outcomes.

CONCLUSIONS A N D

relevance

Author Affiliations: Department of

Neurology. Northwestern University

Intravenous rtPA remains the standard o f care fo r patients

Feinberg School o f Medicine,

Chicago, Illinois.

w ith m oderate to severe neurological deficits w ho present w ith in 4.5 hours o f sym ptom
onset. O utcomes fo r some patients w ith acute ischemic stroke and m oderate to severe

Corresponding Author: Shyam

neurological deficits due to proximal artery occlusion are im proved w ith endovascular
reperfusion therapy. Efforts to hasten reperfusion therapy, regardless o f the mode, should be
undertaken w ith in organized stroke systems o f care.

Prabhakaran, MD, MS, Department o f

Neurology, Northwestern University,
710 N Lake Shore Dr. Ste 1422,
Chicago, IL 60611 (shyam.prabhakaran
@northwestern.edu).

JAMA. 2015;313(14):1451-1462. doi:l0.100l/jama.2015.3058

Section Editor: Mary McGrae

McDermott, MD, Senior Editor.

troke is a leading cause o f disability in the United States and

(recanalization) and restoring cerebral blood flo w (reperfusion) as

th e fifth leading cause o f death.' An acute ischemic stroke

soon as possible to reduce tissue injury and im prove outcomes.

(AIS) occurs when an artery supplying the brain becomes oc

There is general consensus based on strong evidence th a t in pa

cluded, leading to th e death o f brain tissue and focal neurological

tients presenting w ith in 4.5 hours o f sym ptom onset, intravenous

deficits. An estimated 7 0 0 0 0 0 ischemic strokes occur in the United

recom binant tissue plasminogen activator (IV rtPA) therapy is ben

States each year, costing more than $70 billion to society.' The toll

eficial. However, many patients present w ith occlusion o f a large

exerted on individual patients and families by a devastating stroke

proximal artery beyond 4.5 hours or have contraindications to sys

is incalculable; most elderly patients fear a disabling stroke more than

tem ic throm bolysis (ie, recent major surgery or active bleeding). For

th e y fear death.2 Thus, im proving neurological outcom e a fte r an is

these reasons and because proxim al a rte ry occlusions are rela

chemic stroke is a m ajor societal p riority and has attracted intense

tively resistant to intravenous thrombolysis, catheter-based or intra

attention o f clinical and basic researchers, governm ent funding agen

arterial approaches to directly remove the clot and restore blood flo w

cies, and industry. Consistent w ith animal models o f ischemia, the

to the brain have been th e focus o f recent randomized clinical trials.

overarching goal o f AIS therapy is relieving th e arterial occlusion

We review acute stroke tre a tm e n t w ith an em phasis on intra-

jama.com

JAMA

April 14,2015

Volume 313, Number 14

1451

Clinical Review & Education Review

Acute Stroke Intervention: A Systematic Review

arterial treatm ents th a t have recently been shown to im prove o u t

comes fo r th e m ost severe o f patients w ith AIS.

Box 1. Common Terminology in Acute Ischemic Stroke

Management

Search M e th o d s a n d Results

Acute ischemic stroke (AIS): occlusion o f the brain, retina, or spi

Systematic literature search o f MEDLINE databases published be

nal cord supplying artery that results in focal tissue infarction and

tween January 1,1990, and February 11,2015, was performed to iden

corresponding sudden neurological deficits

tify studies addressing th e role o f th rom bolysis and mechanical

Ischemic core: the area o f irreversible severe ischemia w ith loss of

th ro m b e cto m y in acute stroke management. We searched Medical

oxygen and glucose supply and resultant depletion o f energy

Subject Headings (MeSH) term s in m ultiple com binations includ

stores, cellular necrosis, and cavitation

ing brain ischemia/drug therapy, stroke drug/therapy, tissue plas

Ischemic penumbra: the area surrounding the ischemic core, char

minogen activator, fibrinolytic agents, endovascular procedures,

acterized by moderate ischemia and cellular dysfunction but not

thrombectomy, time factors, emergency service, treatment out

cell death, which is potentially reversible w ith prom pt reperfusion

come, multicenter study, and randomized controlled trial. Studies re

porting outcom es from acute throm bolysis and mechanical th ro m
bectom y were included fo r review. The search was lim ited to human
studies w ith o u t language restrictions applied. This search was supple

Modified Rankin Scale (mRS): a scale used to measure disability

after stroke w ith ordinal scores from 0 to 6 w ith 0 indicating no
symptoms or disability: 1, symptoms but no disability; 2, slight dis
ability but requires no assistance; 3, moderate disability, requiring
some assistance w ith activities o f daily living but able to walk inde

m ented by review ing additional references from included studies.

pendently; 4, moderately severe disability and unable to walk or

Sixty-eight articles (N = 108 0 8 2 patients) w ere selected fo r re

care for bodily needs w ithout assistance; 5, severe disability, bed

view (eFigure in the Supplement).

ridden, and requiring constant care; and 6, dead

B a ckg ro un d

measure o f neurological dysfunction or deficit after stroke across

The central pathophysiological hypothesis underlying AIS therapy

multiple domains including motor, sensory, visual, and language

functions and ranging from 0 to 42 w ith a higher score indicating

National Institutes of Health Stroke Scale (NIHSS): a quantitative

is th a t a fte r a cerebral a rte ry becomes occluded, th e re is some

am ount o f hypoperfused brain tissue at risk fo r perm anent infarc
tion th a t could be salvaged by expeditious restoration o f blood flo w
(Box 1). Preventing this tissue-at-risk, known as th e ischemic pen
umbra, from progressingto irreversible infarction is the goal o f acute
reperfusion therapy (Figure 1). In contrast, th e ischemic core de
fines brain tissue th a t has already experienced irreversible damage
and, therefore, cannot be salvaged by reperfusion. The ischemic pen

more severe neurological deficit

Coil retrievers: thrombectom y devices that engage from distally to
proximally; wraps around the clot and is pulled back through the
guide catheter to remove the occluding thrombus
Aspiration devices: thrombectom y systems that use proximal suc
tion to remove the occluding thrombus through the guide catheter
Stent retrievers: throm bectom y devices that allow for immediate
restoration o f blood flow by stent expansion at the site o f occlu

umbra m odel predicts th a t earlier reperfusion leads to b e tter o u t

sion followed by entrapment o f the thrombus between the stent

comes in patients w ith AIS. Over the first fe w m inutes to hours af

and the vessel wall and thrombus extraction when the stent is re

ter an acute arterial occlusion, ischemic penumbral tissue progresses

moved through the guide catheter

to an infarct core, and the potential benefit o f restoring blood flo w

reduces over tim e. It is estimated th a t fo r every m inute an artery is
occluded duringan ischemicstroke, 2 million neurons die, which over

placebo, which evaluated 624 patients presenting w ith ischemic stroke

10 hours is equivalent to the expected neuronal loss occurring w ith

symptoms w ithin 3 hours o f symptom onset. Patients receiving IV rtPA

26 years o f normal aging.3

compared w ith placebo had an absolute16% increase in favorable o u t

Although it is n o t controversial th a t an ischemic penumbra ex

come (m odified Rankin Scale [mRS] 0-1) at 3 m onths (42.6% fo r IV

ists or that all penumbral tissue, by definition, in the absence o f tim ely

rtPA vs 26.6% fo r placebo, P < .01; number needed to treat to ben

reperfusion, is destined fo r irreversible infarction, distinguishing true

e fit [NNTB], 6) (Table 1). Although there was an increased risk fo r

penumbra from core infarct has been challenging. Another major

symptomatic brain hemorrhage from IV rtPA (6.4% for IV rtPA vs 0.6%

challenge is distinguishing true penumbra from regions o f th e brain

fo r placebo, P < .001), the benefits o f this treatm ent outweighed this

th a t are h y p o p e rfu s e d b u t n o t a t risk fo r in fa rc tio n (b e n ig n

risk. Consequently, the US Food and Drug Adm inistration (FDA) ap

oligemia).4 Indeed, recent trials have produced mixed results when

proved IV rtPA for treatm ent fo r patients w ith acute stroke who pre

applying this concept o f imaging-based selection fo r reperfusion

sent w ithin 3 hours o f sym ptom onset. A subsequent European ran

therapies in part due to inaccurate determ ination o f tru e penum

domized clinical trial o f 821 patients w ith moderate severity ischemic

bra and core infarct volumes.

stroke symptoms w ho were younger than 8 0 years and presenting

w ithin 3 to 4.5 hours o f sym ptom onset also showed the benefits o f

A c u te R e p e rfu sio n T he ra p y

IV rtPA, but the effect size was lower (mRS 0-1:52.4% fo r IV rtPA vs

Strategies to rapidly reperfuse brain tissue at risk o f infarction in

45.2% fo r placebo, P = .04; NNTB, 14).11

clude intravenous and intra-arterial adm inistration o f th ro m b o lytic

Analyzed in this dichotom ous manner fo r no or minimal disabil

drugs and the use o f various th ro m b e cto m y devices under angio

ity, th e results o f the NINDS and European Cooperative Acute Stroke

graphic and fluoroscopic guidance.5'9

Study III (ECASS-3) trials are compelling; however, when analyzed

fo r a sh ift tow ard im proved outcom es across the full range o f dis

1452

In tra v e n o u s T h ro m b o ly s is

ability (ie, ordinal sh ift in the mRS), IV rtPA is even more strongly as

In th e 1990s, the National Institute o f Neurological Disorders and

sociated w ith be n efit (NNTB: 3 in th e 0 -3 -h o u r w indow ; 7 in th e

Stroke (NINDS) sponsored 2 randomized clinical trials10 o f IV rtPA vs

3 -4 .5 -h o u r w in d o w ).12,13 These trials and fu rth e r c o n firm a to ry

JAMA April 14.2015 Volume 313, Number 14

jama.com

Review Clinical Review & Education

Acute Stroke Intervention: A Systematic Review

studies14'16 have established IV rtPA as a standard therapy for pa

younger than 80 years and without contraindications in some guide

tients with AIS within 3 hours from symptom onset. Although not
approved by the FDA for use in the 3- to 4.5-hour window, IV rtPA is

lines for stroke management (Box 2).17

A meta-analysis of IV rtPA trials inclusive of 2775 patients con

recommended for patients w ith moderately severe symptoms

firmed the time dependency of thrombolytic therapy with the fol-

Figure 1. Regions o f Cerebral Hypoperfusion Following Acute Ischemic Stroke

[~a ] Schematic representation of regions

of cerebral hypoperfusion

[~i~l MRl following acute stroke

Diffusion-weighted MRl

Perfusion-weighted MRl

91 Ischemic core
PH Ischemic penumbra
Benign oligemia

MRl indicates magnetic resonance imaging. A, Schematic representation of

material to estimate cerebral blood flow. The color scale represents mean

regions o f hypoperfused brain tissue following acute occlusion o f the middle

cerebral artery. The ischemic core is an area o f irreversible ischemia and cell
death; ischemic penumbra, potentially salvageable tissue w ith prompt

transit tim e o f a contrast bolus; blue indicates normal transit tim e and shades o f
green, yellow, orange, and red indicate delay in transit tim e (ischemia). The

reperfusion; benign oligemia, decreased perfusion but no infarction risk

regardless o f treatm ent. The infarct core can enlarge into the penumbra if

(black) in the PWI, indicative o f irreversible injury. The area w ith abnormal
transit tim e surrounding the core is considered the ischemic penumbra. These
images are from a 49-year-old patient who presented w ith sudden onset of

region o f the ischemic core as defined in the DWI shows areas o f no contrast

reperfusion is not successful. B, Top, Axial diffusion-weighted MRl (DWI)

showing a hyperintensity consistent w ith irreversible ischemia (ischemic core)

dysarthria and left hemiparesis. The MRl images were obtained following

in the deep perforating territory o f the right middle cerebral artery affecting the
caudate, internal capsule, and lentiform nucleus. Bottom, Axial

intravenous recombinant tissue plasminogen activator administered

approximately 50 minutes after symptom onset to assess eligibility for

perfusion-weighted MRl (PWI) at the same level as the DWI showed a much
larger area o f hypoperfusion. Perfusion-weighted imaging uses contrast

mechanical thrombectomy.

Table 1. Summary o f th e NINDS and ECASS-3 Trials

No. (%)
IV rtPA

Placebo

312

312

14

15

P Value

NINDS, 199510
No. enrolled
Median baseline NIHSS score3
Favorable 90-d outcome, mRS 0-1
Symptomatic intracranial hemorrhage11

Abbreviations: IV rtPA, intravenous

recombinant tissue plasminogen

133 (42.6)

83 (26.6)

<.01

20 (6.4)

2 (0.6)

<.01

activator; mRS, m odified Rankin

Scale; NIHSS, National Institutes of
Health Stroke Scale.
3 The NIHSS is a quantitative measure
o f neurological dysfunction after

ECASS-3, 2 00811
No. enrolled
Median baseline NIHSS score3
Favorable 90-d outcome, mRS 0-1
Symptomatic intracranial hemorrhage11

jama.com

418

403

10

219 (52.4)
33 (7.9)

182 (45.2)
1 4 (3 .5 )

stroke and ranges from 0 to 42 with

a higher score indicating more
severe neurological deficit.

.04

b National Institute o f Neurological

Disorders and Stroke definition.

<.01

JAMA

April 14,2015

Volume 313, Number 14

1453

Clinical Review & Education Review

Acute Stroke Intervention: A Systematic Review

Chemical Thrombolysis
Box 2. Common Signs o f Acute Stroke and Tests Used in Ischemic
Stroke Diagnostic Evaluation

The Prolyse in Acute Cerebral Thromboem bolism (PROACT) II trial25

Examination Findings Suggestive o f Acute Stroke

Aphasia

compared w ith heparin alone fo r 180 patients presenting w ith in 6

assessed th e efficacy o f intra-arterial recom binant prourokinase

(r-proUK) w ith heparin (2 0 0 0 -U bolus and 5 0 0 -U /h fo r 4 hours)

Hemiparesis
Hemisensory loss

hours a fte r sym ptom onset and w ith angiographically confirm ed

middle cerebral artery occlusions. No mechanical clot disruption was
perm itted in this trial, which makes th e results d ifficu lt to interpret

Hemineglect

in the m odern era. This study m et the prim ary end p o in t o f fu n c

Visual field deficit

tional independence (mRS 0-2) at 9 0 days (39.7% fo r r-proUI< vs

Gaze deviation and eye movement abnormalities

25.4% fo r control; OR, 2.13 [95% Cl, 1.02-4.42], P = .04; NNTB, 7).

Dysarthria

There was, however, a higher risk o f sym pto m a tic brain he m o r

Gait instability and incoordination

rhage in the r-proUK group (10.2% fo r r-proUI< vs 1.9% fo r control,

P = .06). Because th e benefits o f r-proUI< were marginal and offset

Diagnostic Tests to Establish Diagnosis and Cause o f Acute Stroke

Magnetic resonance imaging or computed tomography o f the
brain to evaluate for ischemia and exclude hemorrhage
Computed tomography or magnetic resonance angiogram to
evaluate for intracranial or extracranial stenosis or occlusion
Echocardiogram to evaluate for cardioembolic source
(ie, thrombus)
Telemetry and extended outpatient cardiac monitoring to evaluate
for arrhythmias (ie, atrial fibrillation)
Lipid panel to evaluate for hyperlipidemia
Hemoglobin Alc to evaluate for diabetes mellitus

partially by increased risk o f harm, r-proUK was not approved by the

FDA fo r intra-arterial throm bolysis in patients w ith AIS. Notably, a
subsequent secondary analysis from an analogous Japanese study
using urokinase in comparison w ith best medical care (control) later
supported the findings o f the PROACT II trial (mRS 0-1:42.1% for uro
kinase vs 22.8% fo r control, P = .045; sym ptom atic brain hem or
rhage: 8.8% fo r urokinase vs 1.8% fo r control, P = ,21).27
Mechanical Thrombectomy
Endovascular treatm ent for AIS has continued to evolve w ith the in
troduction o f catheter-based mechanical throm bectom y devices. The

In select patients, consider inflammatory markers, hypercoagu-

FDA has approved several mechanical throm bectom y devices to treat

lable workup, ultrasound o f the lower extremities, lumbar punc

ture, and blood cultures

AIS based on technical efficacy and safety reported from large m ulti
center case registries. These devices can successfully recanalize proxi
mal arterial occlusions w ith acceptable complication rates; in these

low ing adjusted odds ratios (ORs) o f good outcom e by trea tm e nt

studies, 7% to 19% o f patients experienced device- and procedure-

tim e w indow : 2.55 (95% Cl, 1.44-4.52) fo r 0 to 9 0 m inutes, 1.64

related complications such as device fracture, vessel perforation and

(95% Cl, 1.12-2.40) fo r 91 to 180 m inutes, and 1.34 (95% Cl, 1.06-

hemorrhage, and nontarget artery embolization (Figure 2).5'9

1.68) fo r 181 to 270 m inutes. There was no net benefit fo r patients

Mechanical throm bectom y devices are introduced into the fem

receiving IV rtPA beyond 4.5 hours. It also confirm ed a sim ilar risk

oral artery via guide catheters and advanced to th e affected artery

o f sym ptom atic brain hemorrhage to th a t seen in the NINDSIV rtPA

usingangiographicguidance.Am icrocatheterandguidewirearethen

tria l (5.2% fo r IV rtPA vs 1.0% fo r c o n tro l; OR, 5.37 [9 5 % Cl,

inserted into th e intracranial vessels beyond th e guide catheter and

3.2-9.0]).18

throm bectom y is performed w ith proximal balloon occlusion to pre

Outcomes are b e tte r in patients treated w ith IV rtPA than pla

ve n t distal em bolization during th e procedure. The approved de

cebo treatm ent fo r all stroke subtypes and across the range o f m od

vices have d iffe rin g mechanisms o f action: (1) a coil retriever de

erate to severe stroke severity.109'20 However, after IV rtPA alone,

vice th a t engages and wraps around the clot and then is pulled back

only 10% to 15% o f internal carotid artery occlusions and 25% to 50%

to the catheter to remove the throm bus; (2) an aspiration device that

o f proximal middle cerebral artery occlusions recanalizeand only 35%

uses proximal suction to remove throm bus; and (3) stent retriev

to 4 0 % o f p a tie n ts a ch ie ve g o o d o u tc o m e s (ie, fu n c tio n a l

ers th a t allow fo r im m ediate restoration o f blood flo w by stent ex

independence).21,22 These data suggest th a t proximal artery occlu

pansion at the site o f occlusion followed by entrapm ent o f the th ro m

sions (ie, m iddle cerebral artery and internal carotid artery) may be

bus between th e stent and the vessel wall and throm bus extraction

relatively resistant to IV rtPA alone.22'24 Because proximal artery oc

when th e stent is removed (Figure 2).

clusions account fo r o n e-third o f AIS, typ ica lly result in m ore se

vere strokes, and are associated w ith poor outcom es w ith o u t effec

Earlier Generation Devices

tiv e rep e rfu sio n,23'25'26 e ffo rts to im p ro ve recanalization rates

The coil retriever and aspiration devices were FDA-approved based

beyond w hat is possible w ith IV rtPA alone, either w ith alternate or

on single-group studies showing improved revascularization fo r a va

adjunctive m ethods, have been the focus o f several large random

riety o f proximal artery occlusions compared w ith the historical con

ized clinical trials.

tro l group from PROACT ll.5-8'9 However, clinical efficacy (ie, im

Intra-arterial Therapy

devices were not compared directly w ith oth er treatm ents or to a

Although catheter-based trea tm e nt o f large, proximal clots should

placebo group.

proving patient functional outcom es) was not proven because the

1454

im prove outcom es, early trials using first-generation approaches

Since then, tw o phase 3 randomized clinical trials evaluated the

failed to show clinical benefit despite successful recanalization rates

efficacy o f earlier generation endovasculartherapies (ie, coil retriever

(Table 2).

and aspiration devices and intra-arterial throm bolytics) in patients

JAMA April 14.2015 Volume 313, Number 14

jama.com

Review Clinical Review & Education

Acute Stroke Intervention: A Systematic Review

TO
C
TO ^
CO CD

ll T

2015

SWIF'

S
cn
Q- m

<

<

<

<

>
+

<
^

>

>

+
>

< 9

CD
>

5 1
3 5
CD LO
r o 1-H

<
Q_

+
> 00 >
LU CD

' l
lu r o

>

<

u-i

LU N -

o
<
1
O
Z ro
LU LO
1 1
X O
LU <N

TO
TO
t o T3
3 CD C
< z
TO
rH

c S
TO rvi

<
CL
t
>
+
>
lu

LO
ro
<
Q_
LO >
ro

<
cl

ro

>
+

<

<
CL
t!
>

<
CL
t
>

+
> l
LU

> i1
^
>
LU

lu

ro

E
O

2P S

CD

d)

So

ro

>

ai to

E c
<

ir o

C -~x
ro u

i in

. 1

b
^

JO

bo o

3
w -C
CD QJ

* L_> _ l ' 1
-2; r o "q CD

>- CD
^

> 82

> *

>
o >
D
iiilO
. i it
iii C
io
ii

> s
L U IN U

>
^
LU H

i
U

CU "sT

Q_
b
>

Q. w
t ! CD
> 2

+'
> u2
LJJ LD

+ I ^D
> cn
LU iI

>
w o^
LU CD U ' J

>
O rO
LU CO U

<

c
o
co

Q . Co
<u d "
O

L 20

<

LD
+1 r o - r \ i

Q. CL C
qj E .2
o

't

> d" o > I"- O >

ro

r o r\i
^ CM
CM

D-

LD

CL-

> <

>

i>

?
> L

a.

>

>>

r
>

o
- n
^ < CO
'- 'Q . ro

>

C ^
O
u fN

L0

bJO

3
8
c r T3

CNJ ^ CD
LD O- LO

m
O

_Q
rN

cn i

E 1
s

|
Z
>CO

^
CL
X
LU

TO

> <
LU g

>

>
LU

<

>

<
;
+1 t l
< >

"2

<

0
CJ H

lu q

*3
CM

r o ro
rH

00
ro

+1
<

a)
o '
^ - LO
^
r>
< 00
L U ^ d ---+in
t ! n
NR

ii ,i
co co

NR

CO

<

CJ

>

> P
LO
+ 1 ---O
<

<
CL
-H
->

d) TO

P
LO
'
O

r-s
TO
CL

TO
LO

CD
CD " o CO
j L m

<

S K

ro

ro "0
ro
<

i_i r o

<

l_

u Z

< c o O (N
(N O fN

ra
.3 5

E<

pv.
o
o
(N

NR

m
ir
1
W
S

<

S c d
LO 8 H

E 1/1

8!

S'} co
O

< >

S 2
I o
cu 2

<

cn kd

LU

cn

lo

<
t

^
u

O 00 V
CO LO ^

<

<

00
'1

0
O co CD
CO
1

- c n c cn
h cd o >
' CJ CD

<
b

> 9
< ro
t CD;

0 cr. 01
CO Cj" ^

LD

a.

00

00 CD
S 'P
CD CD
H= ^

r-. o

00
rH iH

0 "u
>
H-

rH p
T 'T

O
-C O
.. O
O CO ' CD

> '5
CD Qj
H= ^

CO
u r f

.^

o
CO
> 00 o o
d
+ lC
'3 - - b ar \ i CD

> rN o h

' L D _r Ir'-
1 y co
>~o

2 CD ^
l r - - > lti

> ^
-HI 5

v_,

O > 2 -5
COO .- T O )
CJ
> CD 3 o d
<D CD

u -

<=>

h - CD

> 00

TO <U

> .2

aj

CO

-2 co .E
E X y,

z 'a
JZ
u

0)

O) 1/1 O

tT:O _v:
.e p
CD
_

2
^

c? y
s TO
i *

<d

: .

cS

.3

cd

Q_ D_ I

CD

<
T.
8

'

~ -c .2 to
O^
3 TO

E 8

*5

rN
-a ^
>- ?

~
Cl

J5 E

xj

5 TO E
f

ro

T
OcTO
C .
CD I

o
o-TK
TO

s
i
c u

E .2-g
2

6 ~

jama.com

to

^-J5E
c

cd

Ed
CD 0
Ez
0 23 rN1
0 0

U
O
TO__
TO C
CD
D
CD
3 O
E ITO
^ i_ r

"to S o S
>
CL E ! E d

co u
o cn

^
E "8
.2 -2 o

> 0 - 0 .
d) CD CD
_n D

%18
<

4=

Ut CD

CD

*
CD

.1 I

> >

ai

ai

ai

TO

TO

ai

V)
S
I uc >u4- >u4"2

"2

"2

"E

^
U

^ CD
U "O

^ a

xi

-E

-E .E

co co co co

d5

TO

C
TO

TO
C
TO

TO

C
TO

TO
C
TO

JAMA April 14,2015 Volume 313, Number 14

1455

Clinical Review & Education

Review

Acute Stroke Intervention: A Systematic Review

Figure 2. E ndovascular T re a tm e n t o f A cu te Ischem ic S troke

[ a] Left MCA occlusion

fil Cerebral angiogram before (left) and after (right) mechanical

thrombectomy of a proximal artery occlusion in the left MCA

ICA indicates internal carotid artery,

MCA, middle cerebral artery. A,
Illustration o f a proximal artery
occlusion in the left MCA. B.
Anterior-posterior fluoroscopic
cerebral angiography showing
interrupted blood flow distal to a left

[cf] Mechanical thrombectomy devices

MCA occlusion and subsequent

reperfusion following mechanical

Coil retriever

throm bectom y using a stent retriever

device. The angiogram images are
from a 40-year-old patient who
presented w ith acute right
hemiparesis and expressive aphasia.
The patient received intravenous
recombinant tissue plasminogen
activator 1 hour and 15 minutes after
symptom onset w itho u t clinical
improvement. Mechanical
throm bectom y following
angiographic confirmation o f the
MCA occlusion was performed
Access through
femoral artery

1456

approximately 3 hours after symptom

onset. C, Types o f mechanical
throm bectom y devices. Vascular
access to the cerebral vasculature is
generally via the femoral artery

w ith AIS. The Interventional Management o f Stroke (IMS) III trial com

Based on a successful p ilo t tria l c o m p a rin g in tra -a rte ria l

pared standard-dose IV rtPA w ith a com bination o f low-dose IV rtPA

therapy w ith intravenous throm bolysis,34 the Intra-arterial vs Sys

and intra-arterial rtPA or mechanical th ro m b e cto m y (79% o f the

te m ic Throm bolysis fo r Acute Ischemic Stroke (SYNTHESIS EXP)

study patients received intra-arterial rtPA; 45% had clot removal w ith

study random ized patients w ith ischemic stroke into 2 groups o f

coil retriever and aspiration devices; and only in 1% were stent re

181 patients each: 1 group received standard IV rtPA and th e oth er

trievers used).22 Patients were enrolled a fte r baseline neuroim ag

received m echanical th ro m b e c to m y o r in tra -a rte ria l th e ra p y

ing excluded intracranial hemorrhage. No preprocedure vascular

w ith in 4.5 hours o f sym ptom onset. O f th e patients undergoing

imaging selection was perform ed in 46.6% o f patients, which led to

intra-arterial therapy, m ost w ere trea te d w ith rtPA infusion and

the inclusion o f 89 patients (21%) w ith o u t proxim al artery occlu

m ic ro g u id e w ire th ro m b u s fra g m e n ta tio n (6 0 % ), 31% w e re

sion in the intra-arterial treatm ent group for intention-to-treat analy

treated w ith th ro m b e cto m y devices, and 13% were treated w ith

sis. A fte r enrolling 656 patients over 6 years, th e trial was stopped

s te n t r e tr ie v e rs . No b e n e fit o f e n d o v a s c u la r th e ra p y was

fo r fu tility. There was no significant difference in long-term fu n c

observed in th is tria l as no d ifference in prim a ry outcom e, alive

tional outcom e between groups: (mRS 0-2 at 9 0 days between the

w ith o u t disability at 9 0 days, was found (30.4% fo r endovascular

groups; 4 0 .8 % fo r com bined therapy vs 38.7% fo r IV rtPA; abso

vs 34.8% fo r IV rtPA,

lute difference, 1.5% [95% Cl, -6.1% to 9.1%]) and no significant d if

safety d iffe re n ce s w ere noted b etw een th e 2 groups (s y m p to

ference in m ortality (19.1%for combined therapy vs 21.6% fo r IV rtPA,

m atic intracranial hemorrhage, 5.5% fo r endovascular vs 5.5% for

P = .52) or sym pto m a tic brain hem orrhage (6.2% fo r com bined

IV rtPA,

therapy vs 5.9% fo r IV rtPA, P = .83).22

rtPA,

JAMA

April 14,2015

Volume 313, Number 14

P = .37; OR, 0 .8 2 [9 5 % Cl, 0.53-1.27]). No

P = .99; m orta lity, 7.7% fo r endovascular vs 6.1% fo r IV

P = .53).28
jama.com

Review Clinical Review & Education

Acute Stroke Intervention: A Systematic Review

A fourth trial, SOLITAIRE with Intention for Thrombectomy as

Stent Retrievers
Two studies directly compared the newer stent retriever devices with

Primary Endovascular Treatment (SWIFT PRIME),33 also stopped

the earlier coil retriever devices and observed improved recanali

early after enrolling 196 patients due to efficacy o f endovascular

zation, reduced mortality, and better functional outcomes with the

treatment. This study found that in rtPA-treated patients able to un

stent retriever devices6,7 These studies established that stent re

dergo catheter intervention within 6 hours with anterior circula

trievers are superior to coil retrievers. Though these trials included

tion occlusion, thrombectomy was superior to IV rtPA alone (mRS

some patients who had received IV rtPA therapy initially in both treat

0-2:60.2% for intra-arterial treatment vs 35.5% for IV rtPA, P < .01;

ment groups, there was no direct comparison with a control group

NNTB, 4).

that received IV rtPA alone or no acute reperfusion treatment at all.

Recently, 4 randomized clinical trials of stent retrievers against

Imaging-Based Patient Selection

medical treatment have established the benefits o f endovascular

In parallel with advances in device technology, new neuroimaging

treatm ent in patients w ith proximal artery occlusions.30'33 The

techniques have been developed to evaluate the location of the ar

Multicenter Randomized Clinical Trial o f Endovascular Treatment

terial occlusion and the extent of the penumbra, collateral blood flow

for Acute Ischemic Stroke (MR CLEAN)30 phase 3 randomized clini

status, and core infarct areas and thereby to improve patient selec

cal trial provided the first evidence that mechanical thrombectomy

tion for endovascular therapy. In addition to, or instead of, clinical

within 6 hours o f symptom onset improved 90-day clinical out

criteria, these tools have been adopted in recent trials to identify pa

comes compared w ith standard medical treatm ent, in which

tients who will most benefit from acute reperfusion and exclude pa

90.6% received IV rtPA w ithin 4.5 hours. Acute stroke patients

tients in whom reperfusion would be futile or dangerous 35-37 How

with confirmed proximal artery occlusions and treated at 1 o f 16

ever, although CT or magnetic resonance angiography can identify

stroke centers in Holland were randomized to standard medical

management alone (n = 267) or standard medical management fol

proximal artery occlusions with high accuracy, the ideal imaging mo

lowed by intra-arterial (predominantly stent retriever) treatment

for penumbra identification (ie, bolus transit times, blood flow, and

(n = 233). Functional results were better for patients who under

volume measurements), and the potential for adverse conse

went intra-arterial treatm ent (mRS 0-2: 32.6% for intra-arterial

quences of acquiring such imaging (ie, contrast injury and time de

P < .01; NNTB, 8). Although there

were no hemorrhagic safety concerns in MR CLEAN (symptomatic

lays) remain controversial. Because the penumbra is an area o f the

brain with reversible ischemia, imaging techniques that provide high-

intracranial hemorrhage, 7.7% for intra-arterial treatment vs 6.4%

quality information about the penumbra and its extent compared

for IV rtPA; 30-day mortality, 18.4% for intra-arterial treatment vs

with the irreversibly infracted core should help guide stroke therapy.
The first of several perfusion imaging studies, the Diffusion and

treatm ent vs 19.1% in IV rtPA;

18.9% for IV rtP A ), endovascular treatment increased the risk of

dality (CT vs magnetic resonance perfusion) and the optimal marker

new ischemic stroke within 90 days (5.6% for intra-arterial treat

Perfusion Imaging Evaluation for Understanding Stroke Evolution

P < .001) presumably due to procedure-

Study-2 (DEFUSE-2),38 was an uncontrolled, prospective cohort

related embolization into other unaffected cerebral vasculature

study of 99 patients with acute stroke. Patients with a favorable pen-

(noted in 8.6% of treated patients).

umbral pattern (defined as penumbra to infarct core ratios >1.8) had

ment vs 0.4% for IV rtPA,

In the Endovascular Treatment for Small Core and Proximal Oc

better 90-day outcomes with successful reperfusion compared with

clusion Ischemic Stroke (ESCAPE) trial,31316 patients presenting

no reperfusion (mRS 0-2:56.5% for reperfusion vs 31.3% for no rep

within 12 hours from symptom onset at 22 global sites and in whom

erfusion,

proximal artery occlusions were identified using computed tomog

P = .04), whereas no benefit was seen in patients w ith

raphy (CT) angiography were randomized to best medical therapy

out a favorable penumbra pattern (mRS 0-2: 25.0% for reperfu

sion vs 22.2% for no reperfusion, P >.99). In contrast, the MR RESCUE

alone (78% IV rtPA in the control group) or with adjunctive intra

study,29 a phase 2b, open-label, randomized clinical trial w ith a

arterial treatment (of which 86% were stent retrievers). The trial was
stopped early due to efficacy of endovascular therapy (mRS 0-2:
53.0% for intra-arterial treatm ent vs 29.3% for control, P < .01;

blinded outcome o f 118 patients with AIS found no benefit of me

chanical thrombectomy in patients with favorable penumbral pat
terns (20.6% for mechanical thrombectomy vs 26.5% for medical

NNTB, 4). Of note, improved workflow and enhanced patient se

management, P = .78). These 2 studies differed in that (1) DEFUSE-2

lection using rapid-acquisition CT imaging was emphasized through

used only magnetic resonance imaging-based selection whereas MR

out with median time from symptom onset to reperfusion of 241 min

RESCUE included CT perfusion in 20% of analyzed patients, intro

utes such that only 15.5% of patients were treated beyond 6 hours.

ducing heterogeneity into the analysis of penumbra determina

A third recently published study, the Australian Extending the

tion; (2) DEFUSE-2 defined the penumbra to core ratio as greater

than 1.8 with a maximum core infarct volume o f 70 mL whereas MR

Time for Thrombolysis in Emergency Neurological DeficitsIntraArterial (EXTEND-IA) trial,32 used CT perfusion imaging, analogous

RESCUE used a smaller penumbra to core ratio of greater than 1.4

to the Mechanical Retrieval and Recanalization o f Stroke Clots Using

and larger maximum core infarct volume of 90 mL. These method

Embolectomy (MR RESCUE) trial, to randomize patients with favor

ological differences may have led to greater rates of futile reperfu

able mismatch patterns to IV rtPA alone vs IV rtPA plus stent re

triever mechanical thrombectomy within 4.5 hours of symptom on

sion in the MR RESCUE study compared with the DEFUSE-2 study.

set. Following MRCLEAN, this trial too was stopped early fortechnical
efficacy in favor of endovascular treatment after enrollingjust 35 pa

The recent EXTEND-IA trial32 used a similar algorithm as the MR

RESCUE investigators but at earlier time points (<4.5 hours) and was
able to confirm the benefit of reperfusion therapy on improvingclini-

tients in each group, and also found benefit on the lead secondary
clinical efficacy outcome (mRS 0-2:71.4% for IV rtPA plus stent re

cal outcomes with similar effect sizes as seen in DEFUSE-2.

triever vs 40.0% for IV rtPA; P < .01; NNTB, 3).

the recent randomized clinical trials: (1) a pragmatic, simple ap-

jama.com

Varied approaches for patient selection have been adopted in

JAMA April 14,2015 Volume 313, Number 14

1457

Clinical Review & Education Review

Acute Stroke Intervention: A Systematic Review

Figure 3. Rate o f R eperfusion and Tim e to R eperfusion C om pared W ith P ercentage o f G ood O u tco m e s in th e 6 Trials C om p a rin g E ndovascular
T re a tm e n t to M edical T re a tm e n t A lon e

1a ] Substantial reperfusion rates

[T ] Time to reperfusion

Patients With Good Clinical Outcome at 90 d (mRS 0-2), %

mRS indicates m odified Rankin Scale; TICI, thrombolysis in cerebral infarction. The dotted lines indicate 95% CIs.

proach o f id entifying proxim al artery occlusions by CT angiography

simple, and accurate, the prim acy o f tim e in AIS managem ent may

and enrolling based on tim e w in d o w alone as done in MR CLEAN

supplant penum bral imaging to a lesser role such as th e evaluation

(<6 hours); (2) assessment o f early infarct signs (ie, core infarct) using

o f atypical presentations or in delayed trea tm e nt w indow s (ie, >6

noncontrast head CT and tim e w in d o w as done in some patients in

hours or stroke upon awakening).

ESCAPE (<12 hours) and some patients in SWIFT PRIME (<6 hours);
(3) additional CT angiography assessment o f collateral blood sup

Endovascular Trial Comparisons

ply as done in some patients in ESCAPE; and (4) penum bra imaging

The Interventional Management o f Stroke III (IMS III), SYNTHESIS

using CT or magnetic resonance perfusion imaging along w ith an

EXP, and MR RESCUE trials tested various first-generation strate

giography to confirm the occluded artery w ith in 4.5 hours as done

gies fo r intra-arterial trea tm e nt o f proximal artery occlusions. Three

in EXTEND-IA and some patients in SWIFT PRIME. W hether one ap

key factors differentiate the earlier trials from the m ost recent ones:

proach is superior to another is not clear, but all 3 have now been

(1) rates o f substantial reperfusion; (2) tim e to reperfusion; (3) se

shown in randomized clinical trials to select patients w h o benefit

lection based on confirm ed proximal arterial occlusion.

from adjunctive intra-arterial therapy. The num ber excluded on the

First, use o f less effective, first-generation th ro m b e cto m y de

basis o f the enhanced selection process should be carefully exam

vices may have neutralized th e potential benefit o f endovascular

ined. In th e EXTEND-IA study, only 25% o f otherw ise eligible pa

therapy in th e earlier trials. In fact, SYNTHESIS EXP used clot frag

tie n ts were excluded on the basis o f the perfusion-im aging selec

m entation w ith intra-arterial rtPA infusion in 6 0 % o f treated pa

tio n criteria. Overall, only 70 o f 819 patients (8.5% ) treated w ith

tients, a practice th a t is n o t com m on in th e United States. In con

IV rtPA and screened fo r th e study were randomized, w ith absence

trast, th e MR CLEAN, ESCAPE, EXTEND-IA, and SWIFT PRIME trials

o f large artery occlusion being the most common reason for not quali

predom inantly used stent retrievers.30'32 Rates o f substantial rep

fying fo r enrollm ent.

erfusion (ie, Thrombolysis In Cerebral Ischemia [TICI] grades 2b or

Further advances in th e accurate and reliable m easurement o f

brain ischemia are also needed. As stated in a recent review article

1458

3) were lower in IMS III (4 0 % ) and MR RESCUE (27%) when com

pared w ith the stent retriever trials (58% -88% ).

regarding perfusion im aging selection, it may be p rudent to "im

Second, tim e to reperfusion was lo w e r in recent trials co m

prove the science before changing clinical practice."39 The lack o f op

pared w ith earlier efforts. Although tim es to groin puncture were

tim al consensus thresholds o f blood flo w and volum e and valida

n o t significantly d iffe re n t (3.1-3.5 hours) from IMS III, tim es to rep

tio n o f these on d iffe re n t scanner types and m odalities still hamper

erfusion were lower in ESCAPE (4 hours), EXTEND-IA (4.1 hours), and

its use in clinical practice. Accurate im aging parameters need to be

SWIFT PRIME (4.2 hours) than in IMS III (5.4 hours). Therefore, an

standardized and each m ethod tested against each other. It also re

o ther advantage o f th e stent retrievers may be th e ir ability to re

mains unknow n w h e th er th e tim e th a t elapses to acquire and ana

duce tim e from groin puncture to reperfusion. Figure 3A shows a

lyze this advanced "penum bral" imaging (ie, up to 3 0 minutes) ne

scatterplot o f substantial reperfusion rates (TICI 2b or 3) vs pro p o r

g a te s a n y e ffic a c y a d v a n ta g e c o n fe rre d by th e a d d itio n a l

tion w ith good outcom es (mRS 0-2 at 9 0 days) and Figure 3B shows

pathophysiological inform ation fo r optimal endovascular patient se

onset to reperfusion tim es vs proportion w ith good outcom es in the

lection th a t th e imaging study m ight provide. As a result, some in

6 com pleted endovascular trials o f mechanical throm bectom y.

vestigators as in th e ESCAPE and MR CLEAN trials have recom

T hird , un like IMS III and SYNTHESIS EXP, th e MR CLEAN,

mended using a sim plified approach th a t assesses degree o f early

ESCAPE, EXTEND-IA, and SWIFT PRIME trials required confirm a

infarct changes on initial head CT and co n firm a to ry CT angiogra

tio n o f proximal artery occlusion on baseline CT angiography fo r en

phy alone fo r endovascular triag e .40'42 Unless it is made quick.

rollment, leadingto a more homogeneous cohort and one more likely

JAMA

April 14.2015

Volume 313, Number 14

jama.com

Review Clinical Review & Education

Acute Stroke Intervention: A Systematic Review

rect transport o f stroke patients w ith high suspicion o f proximal ar

Box 3. Patient Inform ation

te ry occlusion to comprehensive stroke centers, instead o f prim ary

or nonstroke centers.

JAMA has patient information available for stroke:

1. Torpy JM, Burke AE, Glass RM. Hemorrhagic stroke [JAMA Patient
Page], JAMA. 2010;303(22):2312.
2. Pluta RM, Lynm C, Golub RM. Stroke imaging [JAMA Patient Page].
JAMA. 2011;306(11):1277.
3. Jin J. Warning signs o f a stroke. JAMA. 2014;311(16):1704.

Prehospital and hospital in itia tive s to avoid delays result in

demonstrable reductions in stroke onset-to-treatm ent tim es.49'52
Prior to these efforts, the m ajority o f treated patients in the United
States did not receive guideline-recommended17 IV rtPA w ithin 6 0
m inutes o f hospital arrival.53 Im plem entation o f prehospital and
emergency departm ent best practices in participating US hospitals
in th e A m erican S troke A sso cia tio n 's Target: S troke in itia tiv e
reduced median door-to-needle tim es to less than 6 0 m inutes.54

to show be n efit from endovascular therapy. Several also included

Internationally, similar efforts have shown even better results w ith

more selective imaging-based approaches (see above) th a t may have

sustainable door-to-needle tim es averaging 2 0 to 3 0 m inutes in

reduced the likelihood o f fu tile reperfusion.

Helsinki, Finland, and M elbourne, Australia.50,51 Three key best

Enhanced selection to enrich the trial population in favor o f en

practices lead to th e m ajor im provem ent: (1) p re n o tifica tio n by

dovascular therapy is a potential criticism o f 1 o f the 4 trials. In the

ambulances o f em ergency departm ents to enhance stroke team

MR CLEAN trial, although th e re was an absolute benefit in favor o f

p re p a re dn e ss; (2) d ire c t-to -C T im a ging fro m th e a m b u la n ce

intra-arterial therapy, patients treated w ith standard management

bypassing the emergency departm ents initially; and (3) premixing

(90 .6 % o f w hom received IV rtPA) fared worse than th a t observed

and delive ring rtPA im m e d ia te ly a fte r a CT scan excluded brain

in prior studies (achieved mRS 0-2:19.1% fo r th e M R CLEAN trial vs

hemorrhage. In Berlin, an experim ent using ambulances staffed by

3 5 % -4 0 % in o th er studies w ith IV rtPA alone).21,22 The prolonged

stroke experts and fitte d w ith CT scanners advanced the concept

tim e from IV rtPA to intra-arterial groin puncture (172 m inutes) sug

o f "ambulysis" (throm bolysis in th e ambulance) p rior to hospital

gests th a t despite rapid onset to IV rtPA trea tm e nt tim es (median,

arrival w ith 32% o f patients receiving IV rtPA in the first 6 0 minutes

8 6 minutes), these patients may have been more resistant to IV rtPA,

o f ischem ia w ith am bulysis com pared w ith 5% w ith standard

as evidenced by having persistent proximal arterial occlusion nearly

hospital-based trea tm e nt (P < ,01).55,56 Earlier throm bolysis trans

3 hours later, and selected on th a t basis. However, we cannot con

lates into reduced sym ptom atic intracranial hemorrhage and m or

firm this hypothesis because th e numbers o f to ta l screened and ex

ta lity and better functional outcomes follow ing a stroke.57

cluded patients are n o t presented in th e publication.

Though tim e benchmarks have been recom m ended fo r intra

A n o th e r p o ten tia l concern is th a t th e halted trials are m ore

arterial therapy 58,59 these have been less well-studied compared

prone to exaggerated e ffe cts due to reduced sam ple size.43,44

w ith IV rtPA process im provem ent. Several studies have observed

Because th e 3 m ost recent trials were term inated early fo llo w in g

th a t perform ance o f additional im aging studies, use o f combined

th e publication o f MR CLEAN after m eeting prespecified stopping

endovascular and intravenous approaches, and interhospital trans

rules, one needs to consider th e ir findings w ith caution w ith regard

fe r add significant delays to intra-arterial tre a tm e n t and worsen

to treatm ent benefit magnitude. Indeed, the largest effect size was

outcom es.47,60' 63 Indeed, observational studies and th e recently

noted in the smallest, most selective trial, EXTEND-IA (n = 70; OR,

com pleted clinical trials have shown th a t th e odds o f good o u t

3.8 [95% Cl, 1.4-10.0]), fo llo w e d next by SWIFT PRIME (n = 196;

come follow ing endovascular therapy decline rapidly w ith each 30-

OR, 2.7 [9 5 % Cl. 1.53-4.95]), and then ESCAPE (n = 316; OR, 1.8

to 6 0 -m in u te delay in tim e to tre a tm e n t.61' 65 The successful

[95% Cl, 1.4-2.4]). However, this gradient also reflects the intensity

ESCAPE, EXTEND-IA, and SWIFT PRIME trials emphasized work-

o f imaging selection in the 3 trials, so it may not be related to early

flo w efficiency and speed w ith goals fo r first CT to groin puncture

cessation. Because these trials w ere replicated independently in

(picture to puncture) o f less than 6 0 minutes, a practice reinforced

separate health systems around th e globe, it guards against the

by quality im provem ent initiatives and site performance m on ito r

p o ssib ility th a t th e d ire c tio n o f e ffe c t in fa vo r o f endovascular

ing visits.

therapy was by chance alone.45

Lastly, the costs associated w ith these interventions and the in

frastructure th a t supports them need to be balanced against the ac

Future Directions

tual patient benefits in term s o f population health. Health systems

The use o f intra-arterial therapy fo r stroke has remained low (l% -2%

w ill need to consider changes in cu rre n t public education cam

o f all patients w ith AIS) and along w ith 5% to 7% o f patients w ith

paigns to emphasize early arrival and use o f emergency medical ser

AIS w h o are treated w ith IV rtPA, approxim ately 10% o f patients are

vices, prehospital triage and hospital bypass policies to d ive rt pa

afforded reperfusion treatm ents in th e United States.15,46 Because

tients to stroke centers capable o f providing intra-arterial therapy,

successful stroke treatm ent requires rapid intervention, greater em

and stream lined protocols fo r im aging selection in th e emergency

phasis on im proving prehospital systems and processes fo r rapid ac

departm ent to effectively screen and tre a t eligible patients.

cess and optim izing delivery o f reperfusion therapies should be a

public health priority. Insufficient com m unity awareness o f stroke

Evidence-Based Guidelines

signsand sym ptom s result in delayed use o f emergency medical ser

Current American Heart Association and American Stroke Associa

vices and hospital arrival.47,48 Educating the public about stroke signs

tion guidelines17 recom m end th a t IV rtPA be administered to all eli

and s ym pto m s and th e need fo r rapid tre a tm e n t is im p o rta n t

gible patients (Table 3) as quickly as possible (door-to-needle tim e

(Box 3). W ith th e developm ent o f organized strokes systems o f care,

should be less than 6 0 minutes) in th e 0 - to 3-hour w in d o w (Class I,

regional policies w ill also need to consider prehospital criteria fo r di-

level o f evidence [LOE] A), in th e 3- to 4.5-hour w in d o w (Class I,

jama.com

JAMA April 14,2015 Volume 313, Number 14

1459

Clinical Review & Education Review

Acute Stroke Intervention: A Systematic Review

Table 3. Eligibility Criteria for Intravenous and Endovascular Reperfusion Therapy in Acute Ischemic Stroke

IV rtPA eligibility

Intra-arterial thrombolysis
elig ib ility3

Mechanical thrombectomy e lig ib ility

Clinical diagnosis

Ischemic stroke causing moderate to Major large artery ischemic stroke

severe disabling neurological deficit; causing severe disabling neurological
NIHSS score <25 for 3- to 4.5-hr
deficit; NIHSS score 4-30 except for
window
isolate aphasia or hemianopia

Major large artery ischemic stroke causing

severe disabling neurological deficit; NIHSS
score 8-30c; CTA confirmation of MCA
and/or ICA occlusion; small core by NCCT
(ASPECTS >5); optional: mismatch by
perfusion imaging; good collateral grade

Timing of administration

Known time of onset or last known

well 4.5 h before IV rtPA
administration

Planned intervention can be

performed within 6 h of patient's last
known well

Planned intervention can be performed

w ithin 8 h of patient's last known well

Age, y

>18 for 0- to 3-hr window; 18-80

for 3- to 4.5-hr window

18-85

18-85

Eligibility for IV rtPA

Ineligible to receive IV rtPA

Ineligibility or failure to respond to IV rtPA

Prestroke mRS >2; life expectancy >6 moe

Exclusion Criteria
Head CT findings

Evidence of hemorrhage or other

large mass for 0- to 3-hr window;
> one-third MCA territory infact for
3 - to 4.5-hr window

Evidence of hemorrhage or other

large mass or involvement of
> one-third MCA territory

Evidence of hemorrhage or other large

mass or involvement > one-third MCA
territory (>100 mL volume)

History of nontraumatic intracranial

hemorrhage

Yes

Yes

Yes

Symptoms suggestive of subarachnoid

hemorrhage even with normal head CT

Yes

Yes

Yes

Blood pressure

>185/110 mm Hg or requiring
aggressive measures to lower blood
pressure to below these lim its are
needed

>180/100 mm Hg on 3 occasions at
least 10 min apart or requiring IV
medications

>185/110 mm Hg or requiring aggressive

measures to lower blood pressure to below
these lim its are needed

Arteriovenous malformation, aneurysm,

or intracranial neoplasm

Yes

Yes

Yes

Major surgery

Surgery with unacceptable bleeding

risk within 14 d

Surgery, trauma w ith internal injuries, Surgery, trauma, or biopsy of parenchymal

or lumbar puncture w ithin 30 d
organ with unacceptable bleeding risk
within 30 d

Major head trauma, intracranial/spinal

surgery, or prior stroke (symptoms lasting
>24 h) w ithin 3 mo

Yes

Yes; history of stroke (symptoms

lasting >24 h) within 6 weeks'

Yes

Hemorrhagic diathesis

Yes

Yes

Yes; coagulation factor deficiency

Vitamin l< antagonist with

INR>1.7 (current use of direct thrombin
inhibitors or direct factor Xa inhibitors)

Yes for 0- to 3-hr window; any oral

anticoagulant use w ithin 24 h is a
contraindication regardless of INR
for 3 - to 4.5-hr window

Yes

Vitamin K antagonist with INR >3

Heparin

Received w ithin 48 h, resulting in

abnormally elevated a PTT greater
than the upper lim it of normal

Heparin with PTT 1.5 over upper lim it Heparin with PTT >2 upper lim it
of normal
of normal

Platelets/pl

<100000

<100 000

Evidence of major active internal bleeding

Yes

Yes

Arterial puncture at noncompressible site

in previous 7 d

Yes

Blood glucose, mg/dL

<50; history of prior stroke and

diabetes mellitus is an exclusion
for 3 - to 4.5-hr window

<50, >400

Seizure at the time of symptom onset

Yes

Yes

Presumed septic embolus, or suspicion

of bacterial endocarditis

Yes

Yes

<30000f

<50, >400

Yes

Severe contrast allergy

Yes

Yes

Excessive tortuosity of the vessel

precluding device delivery or >50%
stenosis of vessel proximal to target lesion

Yes

Yes

Abbreviations: ASPECTS, Alberta Stroke Program Early Computed Tomography

Score; CT, computed tomography,- CTA, computed tomography angiography;
NIHSS, National Institutes o f Health Stroke Scale; ICA, internal carotid artery; INR,
international normalized ratio; IV rtPA, intravenous recombinant tissue
plasminogen activator; MCA, middle cerebral artery; mRS, modified Rankin Scale;
NCCT, noncontrast computed tomography; PTT, partial thromboplastin time.
SI conversion factor: To convert blood glucose to mmol/L, m ultiply by 0.0555.
3 Proposed intra-arterial thrombolysis criteria adapted from eligibility criteria for
PROACTII study.
b Proposed mechanical embolectomy criteria based on SWIFT, MERCI,

c Difference o f NIHSS score ranges for eligibility across the trials: MERCI, higher
than 8; TREV02,7 to 29; SWIFT, 8 to 30; IMS III, higher than 9 (changed to >8
later in trial); SWIFT PRIME, 8 to 29; MR CLEAN, higher than 2; EXTEND-IA
and ESCAPE, no NIHSS criteria used (imaging criteria only).
d Difference o f age ranges for eligibility across the trials: SWIFT, 22 to 85 years;
TREV02,18 to 85 years; IMS III, 18 to 82 years; SWIFT PRIME, 18 to 8 0 years;
MR CLEAN, ESCAPE, and EXTEND-IA, older than 18 years (no upper limit).
e Life expectancy: M ERCI, more than 3 months; TREV02: more than 6 months.
1 MR CLEAN: less than 4 0 0 0 0 platelets/pL; severe head trauma less than 4
weeks w ith intended intra-arterial thrombolysis.

Penumbra Pivotal Stroke Trial, TREV02, IMS III, MR CLEAN, ESCAPE,

EXEND-IA, and SWIFT PRIME trials.

1 460

JAMA

April 14,2015

Volume 313, Number 14

jama.com

Acute Stroke Intervention: A Systematic Review

Review Clinical Review & Education

LOE B), and even if considering other adjunctive therapies (Class I,

o f symptom onset. Outcomes for some patients with acute ische

LOE A). Regardless o f treatment mode, efforts should be made to

mic stroke and moderate to severe neurological deficits due to

reduce and avoid delays to reperfusion therapy as earlier treat

proxim al arte ry occlusion are im proved w ith endovascular

ment leads to better clinical outcomes (Class I, LOE A).

reperfusion therapy. Efforts to hasten reperfusion therapy, regard

Current guidelines also recommend intra-arterial thromboly

sis with rtPA in carefully selected (Table 3) patients with middle ce

less of the mode, should be undertaken within organized stroke sys

tems of care.

rebral artery occlusion within 6 hours o f symptom onset (Class I,

LOE B) based on the Middle Cerebral Artery Embolism Local Fibri
nolytic Intervention Trial (MELT) and the Prolyse in Acute Cerebral

Bottom-Line Clinical Messages

Thromboembolism II (PROACTII) trial. Although the guidelines pub

lished in 2013 recommend stent retrievers over earlier generation
coil retrievers (Class I, LOE A), they provide much weaker recom
mendations for the clinical efficacy of mechanical thrombectomy

Intravenous rtPA is the standard treatm ent for patients with

acute stroke who are moderately or severely disabled presenting
within 4.5 hours of symptom onset.
Based on multiple randomized clinical trials, intra-arterial therapy

(Class lla, LOE B) as they do not include the MR CLEAN, ESCAPE,

using stent retrievers improves recanalization of proximal artery

EXTEND-IA, and SWIFT PRIME trial results published in 2015.

occlusions and clinical outcomes beyond that possible with IV

rtPA or supportive care alone.
Efforts should be made to hasten reperfusion therapy in orga
nized systems of stroke care.

Conclusions

Public education about the importance of stroke symptom rec

ognition and use of emergency medical services remains

Intravenous rtPA remains the standard of care for patients with mod
erate to severe neurological deficits who present within 4.5 hours

ARTICLE INFORMATION

paramount.

6. Nogueira RG, Lutsep HL, Gupta R, et al. Trevovs

Merci retrievers for throm bectom y

16. Wahlgren N. Ahmed N, Davalos A, et al.

revascularisation o f large vessel occlusions in acute

Thrombolysis w ith alteplase fo r acute ischaemic

stroke in the Safe Implementation o f Thrombolysis

study and take responsibility fo r the integrity o f the

data and the accuracy o f the data analysis. Drs

ischaemic stroke (TREVO 2). Lancet. 2012:380

(98491:1231-1240.

in Stroke-Monitoring Study (SITS-MOST). Lancet.

2007;369(9558):275-282.

Prabhakaran and Ruff contributed equally to this

w ork as co-first authors.

7. Saver JL.Jahan R, Levy El, etal. Solitaire flow

17. Jauch EC, Saver JL, Adams HP Jr, et al.

Author Contributions: Drs Prabhakaran, Bernstein,

and Ruff had full access to all o f the data in the

Study concept and design: All authors.

Acquisition, analysis, or interpretation o f data:
Prabhakaran, Ruff.
Drafting o f the manuscript: All authors.
Critical revision o f the manuscript for im portant

restoration device vs the Merci retriever in patients

w ith acute ischaemic stroke (SWIFT). Lancet. 2012;
380(98491:1241-1249.
8. Smith WS, Sung G, Saver J, et al. Mechanical
throm bectom y fo r acute ischemic stroke. Stroke.
2008:39(41:1205-1212.

Guidelines for the early management o f patients

w ith acute ischemic stroke. Stroke. 2013;44(3):870947.
18. Lees KR, Bluhmki E, von Kummer R, et al. Time
to treatm ent w ith intravenous alteplase and

intellectual content: All authors.

Administrative, technical, or material support: Ruff,

9. Smith WS, Sung G, Starkman S, et al. Safety and

outcome in stroke. Lancet. 2010;375(9727):16951703.

Bernstein.
Study supervision: Prabhakaran.

efficacy o f mechanical embolectomy in acute

ischemic stroke. Stroke. 2005;36(7):1432-1438.

19. Ingall TJ, O'Fallon WM, Asplund K, et al.

Findings from the reanalysis o f the NINDS tissue

Conflict of Interest Disclosures: All authors have

10. Tissue plasminogen activator fo r acute

ischemic stroke. N Engl J Med. 1995:333(241:15811587.

plasminogen activator for acute ischemic stroke

treatm ent trial. Stroke. 2004;35(10):2418-2424.

completed and submitted the ICMJE Form for

Disclosure o f Potential Conflicts o f Interest. Dr
Prabhakaran reports receiving grant funding from
the National Institutes o f Health. No other
disclosures were reported.
Submissions: We encourage authors to submit
papers fo r consideration as a Review. Please

11. Hacke W, Kaste M, Bluhmki E, et al.

Thrombolysis w ith alteplase 3 to 4.5 hours after

20. Generalized efficacy o f t-PA fo r acute stroke.

Stroke. 1997:28(111:2119-2125.
21. Gonzalez RG, Furie KL, Goldmacher GV, et al.

acute ischemic stroke. N Engl J Med. 2008:359(13):

1317-1329.

Good outcome rate o f 35% in IV tPA-treated

patients w ith computed tomography

12. Saver JL. Number needed to treat estimates

angiography-confirmed severe anterior circulation

occlusive stroke. Stroke. 2013:44(111:3109-3113.

contact Mary McGrae McDermott, MD, at

incorporating effects over the entire range of

m dm 608@ northwestern.edu.

clinical outcomes. Arch Neurol. 2004:61(71:10661070.

22. Broderick JP, Palesch YY, Demchuk AM. et al.

13. Saver JL, GornbeinJ, Grotta J, etal. Number

alone for stroke. N Engl J Med. 2013:368(101:893-

1. Mozaffarian D. Benjamin EJ. Go AS, et al. Heart

needed to treat to benefit and to harm for

903.

disease and stroke statistics-2015 update.

23. Jansen 0, von Kummer R, Forsting M. Hacke W,

Circulation. 20l5;131(4):e29-e322.

intravenous tissue plasminogen activator therapy in

the 3- to 4.5-hour window. Stroke. 2009 :4 0 (7):

2. Solomon NA, Glick HA, Russo CJ, Lee J,

2433-2437.

Schulman KA. Patient preferences fo r stroke

outcomes. Stroke. 1994:25(9):1721-1725.

14. Ahmed N, Wahlgren N. Grand M, et al.

Implementation and outcome o f thrombolysis w ith

3. Saver JL. Time is brain-quantified. Stroke. 2 00 6 :

alteplase 3-4.5 h after an acute stroke. Lancet Neurol.

2010:9(91:866-874.

REFERENCES

37(0:263-266.
4. Kidwell CS, Alger JR, Saver JL. Beyond
mismatch. Stroke. 2003:34(10:2729-2735.
5. Penumbra Pivotal Stroke Trial Investigators. The
penumbra pivotal stroke trial. Stroke. 2 0 0 9 :4 0
(81:2761-2768.

jama.com

15. Schwamm LH, Ali SF, Reeves MJ, et al. Temporal

trends in patient characteristics and treatm ent w ith
intravenous thrombolysis among acute ischemic
stroke patients at Get W ith The Guidelines-Stroke
hospitals. C/rc Cardiovasc Qual Outcomes. 2013;6
(51:543-549.

Endovascular therapy after intravenous t-PA vs t-PA

Sartor K. Throm bolytic therapy in acute occlusion of

the intracranial internal carotid artery bifurcation.
AJNRAmJNeuroradiol. 1995:16(101:1977-1986.
24. Wolpert SM, Bruckmann H, Greenlee R,
Wechsler L, Pessin MS, del Zoppo GJ.
Neuroradiologic evaluation o f patients w ith acute
stroke treated w ith recombinant tissue
plasminogen activator. AJNRAm J Neuroradiol.
1993:14(11:3-13.
25. Furlan A, Higashida R, Wechsler L, et al.
Intra-arterial prourokinase fo r acute ischemic
stroke. JAMA. 1999:282(211:2003-2011.

JAMA

April 14,2015

Volume 313, Number 14

1461

Clinical Review & Education Review

26. Alexandrov AV, Grotta JC. Arterial reocclusion

41. Yoo AJ, Zaidat 0 0 , Chaudhry ZA, et al. Impact

55. Ebinger M, W inter B, Wendt M, et al. Effect o f

in stroke patients treated w ith intravenous tissue

plasminogen activator. Neurology. 2002:59(6):
862-867.

o f pretreatment noncontrast CT Alberta Stroke

the use o f ambulance-based thrombolysis on tim e

to thrombolysis in acute ischemic stroke. JAMA.

27. Ogawa A, Mori E, Minematsu K. et al.

746-751.

56. Ebinger M, Kunz A, Wendt M, et al. Effects o f

Randomized trial o f intra-arterial infusion o f

42. Barber PA, Demchuk AM, Zhang J, Buchan AM.

golden hour thrombolysis. JAMA. 2015;72(1):25-30.

urokinase w ithin 6 hours o f middle cerebral artery

stroke. Stroke. 2007;38(10):2633-2639.

Validity and reliability o f a quantitative computed

tomography score in predicting outcome of

28. Ciccone A, Valvassori L. Endovascular

hyperacute stroke before throm bolytic therapy.

treatm ent fo r acute ischemic stroke. N Engl J Med.

2013;368(25):2433-2434.

Lancet. 2000;355(9216):1670-1674.

43. Montori VM, Devereaux PJ, Adhikari NK, et al.

29. Kidwell CS. Jahan R, Gornbein J, et al. A trial o f

Randomized trials stopped early for benefit. JAMA.

imaging selection and endovascular treatm ent for

ischemic stroke. N Engl J Med. 2013;368(10):914-923.

2005;294(17):2203-2209.

Program Early CT Score on clinical outcome after

intra-arterial stroke therapy. Stroke. 2014;45(3):

44. Bassler D, Briel M, Montori VM, et al. Stopping

2014;311(16):1622-1631.

57. Saver JL, Fonarow GC, Smith EE, et al. Time to

treatm ent w ith intravenous tissue plasminogen
activator and outcome from acute ischemic stroke.
JAMA. 2013;309(23):2480-2488.

58. Sacks D, Black CM, Cognard C, et al.

Multisociety consensus quality improvement
guidelines fo r intra-arterial, catheter-directed

A randomized trial o f intra-arterial treatm ent for

randomized trials early fo r benefit and estimation of

treatm ent effects. JAMA. 2010;303(12):1180-1187.

treatm ent o f acute ischemic stroke, from the

American Society o f Neuroradiology, Canadian
Interventional Radiology Association,

30. Berkhemer OA, Fransen PS, Beumer D, et al.

acute ischemic stroke. N Engl J Med. 2015;372(1):11-

45. loannidis JP. Contradicted and initially stronger

Cardiovascular and Interventional Radiological

20.

effects in highly cited clinical research. JAMA.

31. Goyal M, Demchuk AM. Menon BK, et al.

2005;294(2):218-228.

Society o f Europe, Society for Cardiovascular

Angiography and Interventions, Society o f

Randomized assessment o f rapid endovascular

46. Menon BK, Saver JL, Goyal M, et al. Trends in

treatm ent o f ischemic stroke. N Engl J Med. 2015:

372(11):1019-1030.

endovascular therapy and clinical outcomes w ithin

the nationwide Get With The Guidelines-Stroke

32. Campbell BC, Mitchell PJ, KleinigTJ, etal.

Registry. Stroke. 2015;46(4):989-995.

Endovascular therapy fo r ischemic stroke w ith

perfusion-imaging selection. N Engl J Med. 2015;

47. Asif KS, Lazzaro MA, Zaidat 0. Identifying

Interventional Radiology, Society o f

Neurolnterventional Surgery, European Society of
Minimally Invasive Neurological Therapy, and
Society o f Vascular and Interventional Neurology.
Catheter Cardiovasc Interv. 2013;82(2):52-68.

59. Zaidat 0 0 , Yoo AJ, Khatri P, et al.

372(11):1009-1018.

delays to mechanical throm bectom y fo r acute

stroke. J NeurointervSurg. 2014;6(7):505-510.

33. Saver JL, Goyal M, Bonafe A, e t al. SOLITAIRE FR

48. Faiz KW, Sundseth A, Thommessen B, Ranning

ischemic stroke. Stroke. 2O13;44(9):2650-2663.

w ith the Intention for Thrombectomy as Primary

Endovascular Treatment for Acute Ischemic Stroke.

OM. Prehospital delay in acute stroke and TIA.

Emerg MedJ. 2013;30(8):669-674.

60. Sun CH, Nogueira RG, Glenn BA, et al. "Picture

Paper presented at: International Stroke Conference:

February 2015; Nashville, TN.

49. Ford AL, Williams JA, Spencer M, et al.

34. Ciccone A, Valvassori L, Ponzio M, et al.

Reducing door-to-needle times using Toyota's lean

manufacturing principles and value stream analysis.

Intra-arterial or intravenous thrombolysis fo r acute

Stroke. 2012;43(l2):3395-3398.

ischemic stroke? J NeurointervSurg. 2010;2(l):74-79.

Recommendations on angiographic
revascularization grading standards fo r acute

to puncture." Circulation. 2013;127(10):1139-1148.

61. Sun CH, Ribo M, Goyal M, et al.

Door-to-puncture: a practical metric for capturing

50. Meretoja A, Strbian D, Mustanoja S, Tatlisumak

and enhancing system processes associated with

endovascular stroke care, preliminary results from
the rapid reperfusion registry. J Am Heart Assoc.

35. Furlan AJ, Eyding D, Albers GW, et al. Dose

T, Lindsberg PJ, Kaste M. Reducing in-hospital delay

2014;3(2):e000859.

Escalation o f Desmoteplase fo r Acute Ischemic

to 2 0 minutes in stroke thrombolysis. Neurology.

Stroke (DEDAS). Stroke. 2006;37(5):1227-1231.

2012;79(4):306-313.

36. Hacke W, Albers G, Al-Rawi Y, et al. The

51. Meretoja A, Weir L, Ugalde M, et al. Helsinki

Desmoteplase in Acute Ischemic Stroke Trial (DIAS).

Stroke. 2005;36(l):66-73.

model cut stroke thrombolysis delays to 25 minutes

in Melbourne in only 4 months. Neurology. 2013:81
(12):1071-1076.

37. Albers GW, Thijs VN, Wechsler L, et al. Magnetic

resonance imaging profiles predict clinical response
to early reperfusion. Ann Neurol. 2 00 6 ;6 0 (5 ):5 0 8 517.
38. LansbergMG, Straka M, K em pS ,etal. MRI
profile and response to endovascular reperfusion
after stroke (DEFUSE 2). Lancet Neurol. 2012:11(10):
860-867.

39. Goyal M, Menon BK, Derdeyn CP. Perfusion

imaging in acute ischemic stroke. Radiology. 2013;
266(1):16-21.

52. Ruff IM, Ali SF, Goldstein JN, et al. Improving

door-to-needle times. Stroke. 2014;45(2):504-508.
53. Fonarow GC, Smith EE, Saver JL, et al.
Timeliness o f tissue-type plasminogen activator
therapy in acute ischemic stroke-, patient
characteristics, hospital factors, and outcomes
associated w ith door-to-needle times w ithin 6 0
minutes. Circulation. 2011;123(7):750-758.

54. Fonarow GC, Zhao X, Smith EE, et al.

62. Goyal M. Almekhlafi MA, Fan L, et al. Evaluation

o f interval times from onset to reperfusion in
patients undergoing endovascular therapy in the
Interventional Management o f Stroke III trial.
Circulation. 2014;130(3):265-272.

63. Menon BK, Almekhlafi MA, Pereira VM, et al.

Optimal w orkflow and process-based performance
measures for endovascular therapy in acute
ischemic stroke: analysis o f the SOLITAIRE FR
throm bectom y for acute revascularization study.
Stroke. 2014;45(7):2024-2029.

64. Khatri P, Yeatts SD, Mazighi M, et al. Time to

angiographic reperfusion and clinical outcome after
acute ischaemic stroke. Lancet Neurol. 2014:13(6):
567-574.

Door-to-needle times for tissue plasminogen

activator administration and clinical outcomes in

65. Mazighi M, Chaudhry SA, Ribo M, et al. Impact

baseline CT scan appearance and tim e to

acute ischemic stroke before and after a quality

Circulation. 2013;127(19):1980-1985.

recanalization on clinical outcomes in endovascular

improvement initiative. JAMA. 2014;311(16):16321640.

40. Goyal M, Menon BK, Coutts SB, et al. Effect o f

throm bectom y o f acute ischemic strokes. Stroke.

2011;42(1):93-97.

1462

Acute Stroke Intervention: A Systematic Review

JAMA April 14,2015 Volume 313, Number 14

o f onset-to-reperfusion tim e on stroke mortality.

jama.com

Copyright of JAMA: Journal of the American Medical Association is the property of

American Medical Association and its content may not be copied or emailed to multiple sites
or posted to a listserv without the copyright holder's express written permission. However,
users may print, download, or email articles for individual use.