Mishra, NK, Donnan, GA, Furlan, AJ, Hacke, W., and Lees, KR

12 Pages·2010·0.37 MB·English

Mishra, N.K., Donnan, G.A., Furlan, A.J., Hacke, W., and Lees,
K.R. (2010) Mismatch-based delayed thrombolysis: a meta-
analysis. Stroke, 41 (1). e25-e33. ISSN 0039-2499

http://eprints.gla.ac.uk/25902/

Deposited on: 18 January 2012

Enlighten – Research publications by members of the University of Glasgow
http://eprints.gla.ac.uk
Mismatch-Based Delayed Thrombolysis : A Meta-Analysis
Nishant K. Mishra, Gregory W. Albers, Stephen M. Davis, Geoffrey A. Donnan,
Anthony J. Furlan, Werner Hacke and Kennedy R. Lees
Stroke 2010, 41:e25-e33: originally published online November 19, 2009
doi: 10.1161/STROKEAHA.109.566869
Stroke is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX 72514
Copyright © 2009 American Heart Association. 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/41/1/e25
An erratum has been published regarding this article. Please see the attached page for:
http://stroke.ahajournals.org/http://stroke.ahajournals.org/content/41/4/e399.full.pdf
Subscriptions: Information about subscribing to Stroke is online at
http://stroke.ahajournals.org//subscriptions/

Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: [phone]. Fax:
410-528-8550. E-mail:
[email protected]

Reprints: Information about reprints can be found online at
http://www.lww.com/reprints
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
Mismatch-Based Delayed Thrombolysis
A Meta-Analysis
Nishant K. Mishra, MBBS; Gregory W. Albers, MD; Stephen M. Davis, MD, FRACP;
Geoffrey A. Donnan, MD, FRACP; Anthony J. Furlan, MD;
Werner Hacke, MD; Kennedy R. Lees, MD, FRCP
BackgroundandPurpose—Clinicalbenefitfromthrombolysisisreducedasstrokeonsettotreatmenttimeincreases.The
useof“mismatch”imagingtoidentifypatientsfordelayedtreatmenthasfacevalidityandhasbeenusedincaseseries
andclinicaltrials.Weundertookameta-analysisofrelevanttrialstoexaminewhetherpresentevidencesupportsdelayed
thrombolysis among patients selected according to mismatch criteria.
Methods—Wecollatedoutcomedataforpatientswhowereenrolledafter3hoursofstrokeonsetinthrombolysistrialsand
had mismatch on pretreatment imaging. We selected the trials on the basis of a systematic search of the Web of
Knowledge. We compared favorable outcome, reperfusion and/or recanalization, mortality, and symptomatic intrace-
rebralhemorrhagebetweenthethrombolyzedandnonthrombolyzedgroupsofpatientsandtheprobabilityofafavorable
outcomeamongpatientswithsuccessfulreperfusionandclinicalfindingsfor3to6versus6to9hoursfrompoststroke
onset.Resultsareexpressedasadjustedoddsratios(a-ORs)with95%CIs.Heterogeneitywasexploredbyteststatistics
for clinical heterogeneity, I2 (inconsistency), and L’Abbé plot.
Results—We identified articles describing the DIAS, DIAS II, DEDAS, DEFUSE, and EPITHET trials, giving a total of
502 mismatch patients thrombolyzed beyond 3 hours. The combined a-ORs for favorable outcomes were greater for
patients who had successful reperfusion (a-OR(cid:1)5.2; 95% CI, 3 to 9; I2(cid:1)0%). Favorable clinical outcome was not
significantlyimprovedbythrombolysis(a-OR(cid:1)1.3;95%CI,0.8to2.0;I2(cid:1)20.9%).Oddsforreperfusion/recanalization
were increased among patients who received thrombolytic therapy (a-OR(cid:1)3.0; 95% CI, 1.6 to 5.8; I2(cid:1)25.7%). The
combined data showed a significant increase in mortality after thrombolysis (a-OR(cid:1)2.4; 95% CI, 1.2 to 4.9; I2(cid:1)0%),
but this was not confirmed when we excluded data from desmoteplase doses that were abandoned in clinical
development (a-OR(cid:1)1.6; 95% CI, 0.7 to 3.7; I2(cid:1)0%). Symptomatic intracerebral hemorrhage was significantly
increasedafterthrombolysis(a-OR(cid:1)6.5;95%CI,1.2to35.4;I2(cid:1)0%)butnotsignificantafterexclusionofabandoned
doses of desmoteplase (a-OR(cid:1)5.4; 95% CI, 0.9 to 31.8; I2(cid:1)0%).
Conclusions—Delayed thrombolysis amongst patients selected according to mismatch imaging is associated with
increased reperfusion/recanalization. Recanalization/reperfusion is associated with improved outcomes. However,
delayed thrombolysis in mismatch patients was not confirmed to improve clinical outcome, although a useful clinical
benefitremainspossible.Thrombolysiscarriesasignificantriskofsymptomaticintracerebralhemorrhageandpossibly
increasedmortality.Criteriatodiagnosemismatcharestillevolving.Validationofthemismatchselectionparadigmis
required with a phase III trial. Pending these results, delayed treatment, even according to mismatch selection, cannot
be recommended as part of routine care. (Stroke. 2010;41:e25-e33.)
Key Words: thrombolysis (cid:1) mismatch (cid:1) perfusion (cid:1) desmoteplase
Thrombolysis is the principal therapy for acute stroke toward6hoursfromstrokeonset[(oddsratio[OR](cid:1)2.8;95%
patientsintheearlyhoursaftersymptomonset1–3buthas CI, 1.8 to 4.5) for 0 to 90 minutes, 1.6 (95% CI, 1.1 to 2.2)
ashorttreatmentwindow.Inameta-analysisofdataderived for91to180minutes,1.4(95%CI,1.1to1.9)for181to270
from 2775 patients (pooled from the ATLANTIS, ECASS, minutes, and 1.2 (95% CI, 0.9 to 1.5) for 271 to 360
andNINDStrials),therewasagraduallydiminishingbenefit minutes].4 Recently, the ECASS III trial (N(cid:1)821; treatment
ReceivedAugust29,2009;acceptedSeptember22,2009.
FromtheAcuteStrokeUnit(N.K.M.,K.R.L.),UniversityDepartmentofMedicineandTherapeutics,GardinerInstitute,WesternInfirmary,andFaculty
ofMedicine,UniversityofGlasgow,Glasgow,Scotland;StanfordStrokeCenter(G.W.A.),StanfordUniversityMedicalCenter,Stanford,Calif;Royal
MelbourneHospital(S.M.D.),UniversityofMelbourne,Parkville,andNationalStrokeResearchInstitute(G.A.D.),UniversityofMelbourne,Melbourne,
Australia;NeurologicalInstitute(A.J.F.),CaseMedicalCenter,CaseWesternReserveUniversitySchoolofMedicine,Cleveland,Ohio;andUniversity
ofHeidelberg(W.H.),Heidelberg,Germany.
CorrespondencetoKennedyR.Lees,MD,FRCP,UniversityDepartmentofMedicineandTherapeutics,GardinerInstitute,WesternInfirmaryand
FacultyofMedicine,UniversityofGlasgow,Glasgow,G116NT,[email protected]
©2009AmericanHeartAssociation,Inc.
Strokeisavailableathttp://stroke.ahajournals.org DOI:10.1161/STROKEAHA.109.566869
Downloaded from http://stroke.ahajournals.orge/2 a5t GLASGOW UNIV LIB on January 18, 2012
e26 Stroke January 2010
vs placebo 1:1; median time for administration of alte- trials. For example, in DIAS II, the mismatch population was
plase(cid:1)3hours,59minutes)confirmedclinicalbenefitwithin identified on the basis of either CT perfusion or MRI perfusion,
4.5 hours of stroke onset. (OR(cid:1)1.34; 95% CI, 1.02 to 1.76; according to center preference. The determination of mismatch in
DEFUSE and EPITHET trials was based on postprocessed
P(cid:1)0.04).4However,thewider95%CIat6hours(0.9to1.5
perfusion-weightedimagingdatathatincludedcorrectionforarterial
for271to360minutesinthemeta-analysis4)havesuggested input and thresholding. In contrast, in the desmoteplase studies,
that there may still be patients able to benefit from mismatchwasdeterminedin“realtime,”withoutpostprocessing,by
thrombolysisevenbeyond4.5hours.Conversely,othersmay theinvestigatorusingthe“eyeball”technique.
be at increased risk from late treatment. The use of imaging
End Points
approachestoselectpatientswhohaveremainingsalvageable
End points of interest for our meta-analysis were comparisons
tissuefordelayedtreatmenthasbeenproposed,mostnotably
between thrombolyzed and nonthrombolyzed patients in (1) favor-
approaches that include magnetic resonance imaging (MRI) able outcome, (2) reperfusion and/or recanalization, (3) mortality,
perfusion/diffusion “mismatch.”5,6 Several trials have tested and (4) SICH. We also examined the rates of favorable versus
thrombolysis in patients selected after MRI; some centers unfavorable clinical outcome amongst successfully reperfused
patients.
have also incorporated mismatch imaging and delayed
thrombolysis into their routine clinical practice.7 We under-
Search
tookameta-analysisofdatainthepublicdomaintoexamine
We first searched the Web of Knowledge for 10 broad terms:
whether extension of the treatment window among patients “clinical trial*,” “prospective study,” “stroke trial*” “thrombolytic
selected according to the presence of mismatch can be agent,”“desmoteplase,”“tissueplasminogenactivator,”“recanaliza-
recommended for routine clinical practice. tion in stroke,” “ reperfusion therapy in stroke,” “penumbra in
stroke,”and“mismatchhypotheses.”Thenwerefinedoursearchby
combiningthesewithtermsthatunderlinethemismatchhypotheses
Methods
andthrombolysis.OurlastsearchwasundertakenonMarch1,2009.
Selection of Trials From a review of the title and abstract, we selected for further
examination all relevant articles describing the original findings of
We planned to include only relevant articles that described the
studiesthatusedthemismatchhypothesesandselectedpatientsfor
findings of studies that either undertook prospective enrollment of
thrombolysis despite delay beyond 3 hours of stroke onset. We
consecutive stroke patients with a mismatch profile suitable for
checked whether any later article or abstract offered supplemental
delayedthrombolysis(beyond3hoursofstrokeonset)orhadstudied
data. Once selected, each article was read completely and the
mismatch-based, delayed thrombolysis in a randomized controlled
relevantdataextracted.Wealsosearchedthebibliographyofeachof
design.Weexcludedcasereports,caseseries,andstudiesrestricted
thesearticlesforadditionalarticles.
tospecificanatomicbrainlocations.8Wedefinedthe(1)mismatch
profileasaperfusionvolumeatleast1.2timesthatoftheinfarctcore
Statistical Analysis
with use of the imaging methodology available at the specific trial
center, (2) symptomatic intracerebral hemorrhage (SICH) as a Forthismeta-analysis,weretrieved“estimate(s)ofeffect”fromthe
radiologically confirmed cerebral hemorrhage in association with abstract(s). When relevant data were missing, we searched the full
clinicalworseningafterthrombolytictherapy(within36hoursinthe text and any supplementary articles.21 Primarily, we wished to
caseoftherapywithrecombinanttissue-typeplasminogenactivator analyzedataderivedfromthepatientswithamismatchprofileonan
[rt-PA]and72hoursinthecaseoftherapywithdesmoteplase),(3) intention-to-treat basis, but when intention-to-treat data were un-
reperfusionand/orrecanalizationaccordingtotherespectivestudies’ available,weaccepted“perprotocol”dataanddescribedtheunder-
definitions,(4)favorableclinicaloutcomeasaNationalInstitutesof lying limitations. Our comparisons were mainly planned between
HealthStrokeScale(NIHSS)improvementof8pointsfrombaseline patients offered any dose of any thrombolytic agent and the corre-
orattainmentofanNIHSSscoreof0or1and/oramodifiedRankin spondingplacebo-treatedpatients.
Scalescoreof0or1,and(5)mortalityasdeath(RankinScalescore We performed subgroup analyses amongst patients who were
of6)inthe90daysafterthrombolytictherapy.Weconsideredrt-PA treated with thrombolytics at doses approved or still under clinical
and desmoteplase together because both are thrombolytic agents.9 investigation, ie, 90 (cid:1)g/kg desmoteplase or 0.9 mg/kg rt-PA.
Theydifferinsomefeatures:desmoteplaselacksthesecondkringle Comparisons (summary estimates) are expressed as ORs and their
site in its molecular structure, does not need to be cleaved by 95%CIs.Whereasweappliedbothfixed(inverse-varianceweight-
plasmin,isactiveinitssingle-chainform,hasreducedneurotoxicity, ingmethod)andrandom(adjustedOR[a-OR]21)methodstocalcu-
and has limited passage through the blood-brain barrier. Desmote- latethesummaryestimate,wereportedonlythefindingsofthefixed
plase has a theoretical advantage over rt-PA because the former is methodbuthaveindicatedtheinstanceswheretheresultsdiverged.
almostnonfunctionalwhenfibrinisabsent.9–15Alteplaseisalready We assessed the heterogeneity with the test statistics for heteroge-
a proven therapy for treating stroke patients within the early hours neityandI2forinconsistencysupportedbyexaminationofL’Abbé
after stroke onset (NINDS16 and ECASS III3).17 Doses that have plots.
acceptable safety and efficacy have been identified.18–20 Both Ouranalysisincludeddataderivedfromthosepatientswhowere
desmoteplase and alteplase remain investigational for delayed selected(orcouldhavebeenselected)onthebasisoftheirmismatch
thrombolysis.However,weundertookasensitivityanalysesforany profile.Toassesswhetherfavorableoutcomes(clinicaloutcomesat
differentialeffectbetweendesmoteplaseandalteplase. day 90) were more common amongst patients who had successful
UntiltheDIASIIstudy,identificationoftheischemicpenumbra reperfusion, we retrieved data on 242 patients for whom the
was based on the mismatch between MRI perfusion-weighted reperfusion findings were available (the DIAS I trial, N(cid:1)9720; the
imaging and diffusion-weighted imaging.18 For the first time, the DEDAStrial,N(cid:1)3419;theEPITHETtrial,N(cid:1)77[“goodneurolog-
DIASIIinvestigatorswerepermittedtoselectpatientsonthebasis ical outcome” for patients with {n(cid:1)30} and without {n(cid:1)47}
ofvisualinspectionofthemismatchonperfusioncomputedtomog- reperfusion in mismatch patients only]22; and the DEFUSE trial,
raphy (CT) images as an alternative to MRI perfusion studies, N(cid:1)34,inmismatchpatientswith[n(cid:1)18]andwithout[n(cid:1)16]early
dependingonthelocalexpertiseoftheimagingcenter.Weincluded reperfusion23). Corresponding information was not reported in the
datafromeithermethodasreportedintheDIASIIpublication.18 DIAS II trial.18 Similarly, to answer whether a favorable clinical
We included all trials that defined the mismatch profile as the outcome occurred more frequently in the thrombolyzed group of
perfusion volume being 1.2 times the infarct core. We placed no patients, information on 410 patients was available (DIAS I,
restrictiononthemannerinwhichperfusionwasmeasuredinthese N(cid:1)10220; DIAS II, N(cid:1)18618; DEDAS, N(cid:1)3719; and EPITHET,
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
Mishra et al Meta-Analysis of Mismatch-Based Delayed Thrombolysis e27
N(cid:1)85; mismatch patients with and without good neurological a
outcome in the thrombolysis group, n(cid:1)42, and the placebo group, DIAS I 4.08 (1.29, 15.22)
n(cid:1)4322)forthosepatientswhoreceivedanythrombolyticagentat
any dosage. Next, to answer whether reperfusion or recanalization
occurred more frequently amongst those who were thrombolyzed, DEDAS 0.88 (0.14, 6.85)
weretrieveddataon211patientswhoreceivedthrombolytictherapy
at any dose (DIAS I, 97 patients20; DEDAS, intention to treat 37
patients19 and target population 23 patients; and EPITHET, 77 EPITHET 3.68 (1.27, 10.83)
patients22). To assess mortality between thrombolyzed and non-
thrombolyzed patients, we extracted data on 410 patients (DIAS I,
102patients20;DIASII,186patients18;DEDAS,37patients19;and combined [fixed] 3.04 (1.59, 5.82)
EPITHET, 85 mismatch only patients22). To assess SICH between
thrombolyzed and nonthrombolyzed patients, we extracted data on 0.1 0.2 0.5 1 2 5 10 100
odds ratio (95% confidence interval)
405 patients (DIAS I, 102 patients20; DIAS II, 186 patients18;
DEDAS,37patients19;andEPITHET,80mismatchpatientsonly22). b
DIAS I 3.68 (0.73, 19.04)
OwingtomathematicaldifficultiesinvolvedincalculatingORwhen
thenumeratoriszero,wecombinedtheDEDASdatawithDIASI
dataformortalityanalysis. DEDAS 0.37 (0.02, 4.68)
Weundertooksensitivity(subgroup)analysesinwhichwecom-
paredthedataafterexcludingthedataforthosewhoreceiveddoses
EPITHET 3.68 (1.27, 10.83)
ofdesmoteplasethatwereabandonedforfurtherevaluation.Wealso
analyzed differences in clinical outcome between the patients who
werethrombolyzedwithin3to6hoursofstrokeonsetversusthose combined [fixed] 2.65 (1.28, 5.45)
whowerethrombolyzedbeyond6hours.Finally,wecomparedand
0.01 0.10.2 0.5 1 2 5 10 100
contrastedtheattributesofthestudiesandassessedtheirqualityon
the basis of the manner in which patients were enrolled and the odds ratio (95% confidence interval)
resultingbaselinecharacteristics. c
DIAS I 3.68 (0.73, 19.04)
Results
DEDAS 0.37 (0.02, 4.68)
Literature Search
The literature search led to 13 citations on the DEFUSE
EPITHET 3.68 (1.27, 10.83)
trial (10 articles)23–32, 2 on the DEDAS trial (1 article19),
6 on the DIAS trial20,30, 9 on the EPITHET trial (8 combined [random] 2.27 (0.71, 7.27)
articles20,22,33–36),and2onDIASII(1article).37Information
0.01 0.10.20.51 2 5 10 100
on 502 patients was obtained from the 5 main articles
odds ratio (95% confidence interval)
describingtherelevanttrials(DIAS,104patients20;DIASII,
Figure1.Didreperfusionorrecanalizationoccurmorefre-
186patients18;DEDAS,37patients19;DEFUSE,74patients23;
quentlyinpatientswhowerethrombolyzed?Findingsareshown
and EPITHET, 101 patients22), and the data corresponding to fromthefixed-methodanalysisofcombineddata(a)afterexclu-
patients with a mismatch profile were retrieved for subsequent sionofabandoneddosesbyfixed(b)andrandom-method(c)
analyses.
analysis.
received thrombolytic therapy in: DIAS I20 (OR(cid:1)4.1; 95%
Comparative Analysis of the Mismatch Trials
CI, 1.3 to 15.2) and EPITHET (OR(cid:1)3.7; 95% CI, 1.3 to
We compared the attributes that differed between trials to
10.8). Odds were nonsignificant in the DEDAS trial19
highlighttheunderlyingheterogeneityinthemannerinwhich
(OR(cid:1)0.9; 95% CI, 0.1 to 6.9). The combined data gave a
the selected trials were conducted (Supplemental Table I
greateradjustedoddsforreperfusion/recanalizationforthe
available online at http://stroke.ahajournals.org). DIAS II18
patients who had thrombolytic therapy at any dosage
enrolled the least severely affected stroke patients (median
NIHSSscore(cid:1)9)andEPITHET,22themostseverelyaffected (a-OR(cid:1)3.0; 95% CI, 1.6 to 5.8; P(cid:2)0.05, P for heteroge-
(medianNIHSSscore(cid:1)14inthetreatmentarmand10inthe neity(cid:1)0.26, and I2(cid:1)25.7%; Figure 1a).
We repeated our analysis after excluding desmoteplase
placebo arm). Median baseline NIHSS scores were 11.5 and
doses that were abandoned for clinical development; the
12, respectively, in the DEFUSE23 and DIAS I20 trials. We
subanalysis restricted to 90 (cid:1)g/kg desmoteplase or rt-PA
alsocomparedthetimesincestrokeonsetuntilthrombolysis
gave an a-OR(cid:1)2.65 and a 95% CI of 1.3 to 5.5 (P(cid:1)0.007
(OTT), and we assessed qualitatively the proportion of
fixedmethod;Figure1b)andana-OR(cid:1)2.28anda95%CI
patients treated in each trial after 4.5 hours (Supplemental
of 0.7 to 7.3 (P(cid:1)0.17 random method; Figure 1c) (P for
Table II, available online at http://stroke.ahajournals.org).
clinical heterogeneity(cid:1)0.13, and I2(cid:1)50.5%). We also
Detailed analysis of OTT could not be undertaken without
examined the underlying heterogeneity by L’Abbé plot
raw data.
(Figures 2a and 2b).
Findings From Statistical Analyses
Are Favorable Outcomes More Common in Patients Who
Did Reperfusion or Recanalization Occur More Underwent Reperfusion?
Frequently in Patients Who Were Thrombolyzed? Theindividualoddsforafavorableclinicaloutcomeinthe4
The data from 211 patients showed greater individual odds studies reporting this end point were greater in patients who
for reperfusion and/or recanalization amongst those who underwent reperfusion compared with those who did not
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
e28 Stroke January 2010
Figure2.Didreperfusionorrecanalization
occurmorefrequentlyinpatientswho
werethrombolyzed?L’Abbéplotexamin-
ing(a)thecompletedatasetand(b)the
abandoneddosesexcludedforheteroge-
neity.Thecirclesizedenotesthesample
size;DIAS,graycircles;DEDASopen
circles;andEPITHET,blackcircles.
(DIASI20OR(cid:1)3.4;95%CI,1.3to8.8;DEDAS19OR(cid:1)9.6; the thrombolysis group of patients: DIAS I20 OR(cid:1)2.2; 95%
95% CI, 1.5 to 64.6; EPITHET22 OR(cid:1)7.2; 95% CI, 2.3 to CI, 0.7 to 7.4; DEDAS19 OR(cid:1)2.4; 95% CI, 0.4 to 28.0;
23.2;andDEFUSE23OR(cid:1)5.4;95%CI,0.94to38.1).Forall EPITHET22 OR(cid:1)1.7; 95% CI, 0.7 to 4.4; and DIAS II18
trialscombined,thea-ORsweregreaterforpatientswhohad OR(cid:1)0.8; 95% CI, 0.4 to 1.6. The combined data analysis
successful reperfusion compared with those who did not failed to show a significant benefit (a-OR(cid:1)1.28; 95% CI,
(a-OR(cid:1)5.2; 95% CI, 3 to 9.1; P for clinical heterogene- 0.84 to 1.97; P for clinical heterogeneity(cid:1)0.28; I2(cid:1)20.9%;
ity(cid:1)0.60; I2(cid:1)0%; Figure 3a). Figure4a).AfterexclusionofDIASIIdata,a-ORwas1.96,
In a sensitivity analyses in which DEFUSE23 trial data 95% CI was 1.06 to 3.63, and for clinical heterogeneity, I2
were excluded (as DEFUSE,23 unlike others, was a nonran- was 0% and P was 0.89 (Figure 4b).
domized,prospectivelyconductedstudy),thea-ORremained We repeated our analysis after excluding desmoteplase
greater among patients with successful reperfusion (a- dosesthatwereabandonedforclinicaldevelopment:with90
OR(cid:1)5.2;95%CI,2.8to9.5;P(cid:1)0.00;heterogeneitystatistics (cid:1)g/kg desmoteplase and rt-PA 0.9 mg/kg data alone, we
P(cid:1)0.4; I2(cid:1)0%; Figure 3b). found a-OR(cid:1)1.4; 95% CI, 0.9 to 2.3, P(cid:1)0.16; for clinical
heterogeneity,P(cid:1)0.56andI2(cid:1)0%.AfterexclusionofDIAS
Did a Favorable Clinical Outcome Occur More II data, OR(cid:1)1.88; 95% CI, 0.95 to 3.72, and heterogeneity
Frequently in the Thrombolyzed Group of Patients?
With the exception of DIAS II,18 all trials reported nonsig-
nificantly improved odds of a favorable clinical outcome in a
DIAS I 2.21 (0.74, 7.44)
DIAS II 0.83 (0.43, 1.61)
a
DIAS I 3.39 (1.31, 8.84) DEDAS 2.44 (0.35, 28.02)
EPITHET 1.69 (0.65, 4.39)
DEDAS 9.56 (1.54, 64.55)
combined [fixed] 1.28 (0.84, 1.97)
0.2 0.5 1 2 5 10 100
EPITHET 7.19 (2.31, 23.21) odds ratio (95% confidence interval)
b
DEFUSE 5.42 (0.94, 38.12) DIAS I 2.21 (0.74, 7.44)
DEDAS 2.44 (0.35, 28.02)
combined [fixed] 5.19 (2.95, 9.13)
EPITHET 1.69 (0.65, 4.39)
0.5 1 2 5 10 100
odds ratio (95% confidence interval)
combined [fixed] 1.96 (1.06, 3.63)
b
DIAS I 3.39 (1.31, 8.84) 0.2 0.5 1 2 5 10 100
odds ratio (95% confidence interval)
c
DIAS I 3.06 (0.64, 14.75)
DEDAS 9.56 (1.54, 64.55)
DIAS II 1.06 (0.48, 2.30)
EPITHET 7.19 (2.31, 23.21) DEDAS 1.20 (0.12, 17.00)
EPITHET 1.69 (0.65, 4.39)
combined [fixed] 5.15 (2.81, 9.45) combined [fixed] 1.43 (0.87, 2.33)
0.1 0.2 0.5 1 2 5 10 100
1 2 5 10 100 odds ratio (95% confidence interval)
odds ratio (95% confidence interval)
Figure4.Didafavorableclinicaloutcomeoccurmorefre-
Figure3.Arefavorableoutcomesmorecommoninpatients quentlyinthethrombolyzedgroupofpatients?Findingsare
whounderwentreperfusion?Findingsareshownfromthe shownfromthefixed-methodanalysisofcombineddata(a),
fixed-methodanalysisofcombineddata(a)andafterexcluding afterexclusionofDIASIIdata(b),andafterexclusionofaban-
DEFUSEdata(b). doneddoses(c).
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
Mishra et al Meta-Analysis of Mismatch-Based Delayed Thrombolysis e29
a b
Thrombolysis % Thrombolysis %
100 100
80 80
60 3 60
1 3
1
40 4 40
2 2
Figure5.Didafavorableclinicaloutcome
occurmorefrequentlyinthethrombo-
20 20
lyzedgroupofpatients?L’Abbéplot
examiningheterogeneityintheanalysis(a)
0 0 forcompletedata,(b)forDIASIIdata
0 20 40 60 80 100 0 20 40 60 80 100 excluded,(c)forcompletedatabutaban-
c Placebo % d Placebo % doneddosesexcluded,and(d)forDIASII
Thrombolysis % dataandabandoned-dosedataexcluded.
100 Thrombolysis % Thesizeofthesquaredenotesthesam-
100 plesize.1indicatesDEDAS;2,DIASI;3,
EPITHET;and4,DIASII(blackrectangle).
80
80
60 3 60 3
2 2
40 4 40
1
1
20 20
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Placebo % Placebo %
statisticsI2(cid:1)0%andP(cid:1)0.69(Figure4c).L’Abbéplotswere 0.5 to infinity; and EPITHET OR(cid:1)152.6; 95% CI, 15.9 to
examined for underlying heterogeneity in these analyses infinity; but the combined odds for SICH were significantly
(Figure5).Undersensitivityanalysis,nodifferentialeffectof greater for the group that underwent thrombolytic therapy
desmoteplase versus alteplase was found, with the ratio of (a-OR(cid:1)24.7; 95% CI, 5.2 to 118.2; heterogeneity statistics
OR(cid:1)0.7 (95% CI, 0.24 to 1.92; P(cid:1)0.46). I2(cid:1)35.4% and P(cid:1)0.2; Figure 7a). After we combined data
fromDEDASwithDIASI,thefindingsremainednonsignif-
Was There a Greater Probability of Mortality icant for the individual odds (DIAS I(cid:3)DEDAS OR(cid:1)7.1;
in Thrombolyzed Compared With
95%CI,0.7toinfinity)butweresignificantforthecombined
Nonthrombolyzed Patients?
Here,theindividualoddsformortalitywerenonsignificantin analysis (a-OR(cid:1)6.5; 95% CI, 1.2 to 35.4, and for clinical
the thrombolysis group: DIAS II18 OR(cid:1)2.4; 95% CI, 0.7 to heterogeneity, P(cid:1)1.0 and I2(cid:1)0%; Figure 7b).
10.1; DIAS I OR(cid:1)3.6; 95% CI, 0.5 to 161.3; EPITHET22 Repeating the analysis by excluding the data associated
OR(cid:1)2.7;95%CI,0.8to10.9;andDEDAS19OR(cid:1)0.5;95% with abandoned thrombolytic doses, the findings were non-
CI, 0.0 to 34.9. The combined data analysis found a signifi- significant for both individual odds (DIAS I(cid:3)DEDAS
cant increase in mortality in the thrombolysis group of OR(cid:1)3.7;95%CI,0.03toinfinity;DIASIIOR(cid:1)5.7;95%CI,
patients compared with the placebo group (a-OR(cid:1)2.4; 95% 0.2 to infinity; and EPITHET OR(cid:1)6.5; 95% CI, 0.4 to
CI, 1.2 to 4.9; P(cid:1)0.02; P for heterogeneity(cid:1)0.67; and infinity)andincombinationa-OR(cid:1)5.4;95%CI,0.9to31.8;
I2(cid:1)0%; Figure 6a). P for heterogeneity(cid:1)0.97; and I2(cid:1)0% (Figure 7c) but at-
Repeating our analysis after excluding data from the tained marginal significance of the adjusted odds derived
abandoneddesmoteplasedoses,ie,restrictingtheanalysisto by considering the DIAS I and DEDAS data separately
patients treated with 90 (cid:1)g/kg desmoteplase or 0.9 mg/kg (a-OR(cid:1)6.0; 95% CI, 1.00 to 35.8; heterogeneity statistics
rt-PA, we found a-OR(cid:1)1.6; 95% CI, 0.7 to 3.7; P(cid:1)0.28; P P(cid:1)1.00 and I2(cid:1)0%). There were no SICH occurrences in
for heterogeneity(cid:1)0.56; and I2(cid:1)0% (Figure 6b). Under the placebo arms, and therefore, a sensitivity analysis to
sensitivity analysis, no differential effect of desmoteplase assess any differential effect of desmoteplase versus alte-
versusalteplasewasfound,withtheOR(cid:1)0.8(95%CI,0.2to plase could not be undertaken.
3.5; P(cid:1)0.8).
Were There Better Clinical Findings (Outcomes or
Was There a Greater Probability of SICH Reperfusion) When Treatment Was Commenced Within 3
in Thrombolyzed Compared With to 6 Hours Versus 6 to 9 Hours?
Nonthrombolyzed Patients? Limited data were available to examine OTT, and neither
The individual odds for SICH were nonsignificant: DIAS I DIAS I20 nor DIAS II individually suggested significantly
OR(cid:1)7.9;95%CI,0.7toinfinity;DIASIIOR(cid:1)5.9;95%CI, greaterodds(DIASIOR(cid:1)1.07;95%CI,0.4to2.9;P(cid:1)0.9;
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
e30 Stroke January 2010
a
DIAS I 3.55 (0.45, 161.28) DIAS I 7.86 (0.74, ∞)
DIAS II 5.89 (0.47, ∞)
DIAS II 2.37 (0.72, 10.07)
DEDAS * (excluded)
DEDAS 0.52 (0.02, 34.92)
EPITHET 152.59 (15.86,∞)
combined [fixed] 24.69 (5.16, 118.17)
EPITHET 2.70 (0.75, 10.88)
0.2 0.5 1 2 5 10 100 1000
odds ratio (95% confidence interval)
combined [fixed] 2.37 (1.15, 4.87) b
DIAS I & DEDAS 7.06 (0.68, ∞)
0.01 0.10.20.5 1 2 5 10 100 1000
odds ratio (95% confidence interval)
DIAS II 5.89 (0.47,∞)
DIAS I 1.86 (0.02, 151.29)
EPITHET 6.48 (0.36, ∞)
DIAS II 0.82 (0.11, 5.09) combined [fixed] 6.49 (1.19, 35.41)
0.2 0.5 1 2 5 10 100
odds ratio (95% confidence interval)
DEDAS 0.54 (0.01, 48.25)
c
DIAS I & DEDAS 3.74 (0.03, ∞))
EPITHET 2.69 (0.76, 10.88)
DIAS II 5.72 (0.21, ∞)
combined [fixed] 1.61 (0.71, 3.65)
EPITHET 6.48 (0.36, ∞)
0.001 0.01 0.10.20.512 510 100 1000
odds ratio (95% confidence interval)
Figure6.Wasthereagreaterprobabilityofmortalityinthrom- combined [fixed] 5.37 (0.91, 31.82)
bolyzedpatientscomparedwiththosenotthrombolyzed?Find-
ingsareshownfromthefixed-methodanalysisofcombined 0.01 0.10.2 0.5 1 2 5 10 100
data(a)andafterexclusionoftheabandoned-dosedata(b). odds ratio (95% confidence interval)
Figure7.WasthereagreaterprobabilityofSICHinthrombolyzed
DIASIIOR(cid:1)0.8;95%CI,0.4to1.8;P(cid:1)0.7).Withthedata patientscomparedwiththosenotthrombolyzed?Findingsare
frombothtrialscombined,thea-OR(cid:1)0.9;95%CI,0.5to1.7, shownfromthefixed-methodanalysisforallstudiescombinedbut
withDEDASdataexcluded(a),DEDAScombinedwithDIASIdata
and P(cid:1)0.8 (Figure 8). (b),andafterexclusionoftheabandoned-dosedata(c).
Analysis of Mortality
mismatch hypothesis to select and thrombolyze patients
InDIASI,1placeboand2desmoteplasedeathsoccurreddue
despite delays beyond 3 hours. Five trials, DIAS I,20 DIAS
to cardiac causes. In the DIAS II trial, only 1 of 3 deaths in
II,18 EPITHET,22 DEFUSE,23 and DEDAS,19 were available
the90(cid:1)g/kggroupand3of14deathsinthe125(cid:1)g/kggroup
for inclusion. Our results indicate that reperfusion/recanali-
were considered related to the trial medication. In the
zationismorecommonwiththrombolysiswhenalldosesare
DEDAStrial,thesoledeathinthe90(cid:1)g/kggroupwasdueto
considered together, but the significance was lost with the
aspiration pneumonia, whereas that in the 125 (cid:1)g/kg groups
exclusion of data for abandoned doses, which reduced the
wasduetoevolvingneurologicdeteriorationofaleftmiddle
power of our analysis through effects on sample size. Fur-
cerebral artery infarct, leading to pneumonia.
thermore, a favorable clinical outcome was more common
Discussion
We undertook a meta-analysis of all previous studies that
DIAS I 1.07 (0.35, 3.22)
evaluated the principle of physiologic selection for delayed
thrombolysis, based on the presence of potentially viable
tissue in the ischemic penumbra.38,39 These trials used the DIAS II 0.84 (0.37, 1.92)
mismatch hypothesis with either MRI (perfusion/diffusion
mismatch) or CT (perfusion/cerebral blood volume mismatch)
as a signature of the putative penumbra.19,20,22,24,25,40–43 Apart combined [fixed] 0.92 (0.51, 1.67)
fromtherecentDIASIItrial,18thesetrialshadsupportedthe
physiologicbasisofthemismatchconcept.Thedisappointing 0.2 0.5 1 2 5
odds ratio (95% confidence interval)
findings of the DIAS II trial have been attributed to limita-
Figure8.Weretherebetterclinicalfindings(outcomesorreper-
tionsofthestudyandtochance.37Totestforconsistency,we fusion)whentreatmentwascommencedwithin3to6hoursvs
undertook a meta-analysis of the studies that studied the 6to9hours?
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
Mishra et al Meta-Analysis of Mismatch-Based Delayed Thrombolysis e31
amongstpatientswithsuccessfulreperfusionoftheischemic potential benefit among mismatch patients may be time
parenchyma, despite delays beyond 3 hours from stroke dependent, but it appears unlikely that this is sufficient to
onset.Thisconclusionwasnotinfluencedbyinclusionofthe explainalleffects.NowthattheECASSIIIresultshavebeen
nonrandomizedDEFUSEtrial23data.TheDIASIItrial18did presented, another meta-analysis of individual patient data
not report reperfusion findings. fromthetrialsstudiedhereinshouldbeundertakentoassess
However,wedidnotfindevidencethatafavorableclinical clinical and radiologic outcomes for patients who were
outcomewassignificantlyimprovedinthegroupthatunder- thrombolyzedbeyond4.5hoursofstrokeonset.Similarly,an
went thrombolysis. Neither did we find a significant benefit additional analysis comparing outcomes in patients with
when we excluded doses of desmoteplase that were aban- mismatchversusthosewithoutmismatchisdesirablebutwas
doned for clinical development. The CI around our estimate beyond the scope of our meta-analysis.
of effect remains wide and would be consistent with a Our meta-analysis included data from 5 different
doubling of odds for a favorable outcome, although in this trials,18–20,22–23ofwhichDEFUSE23couldbeconsideredonly
respect, DIAS II suggests that the likely upper limit may be intheanalysisofafavorableclinicaloutcomeamongpatients
1.6.Evenso,oddsof1.6remaingreaterthanthoseachieved with reperfusion versus no reperfusion. DIAS II18 did not
in unselected patients treated with rt-PA in the ECASS III report reperfusion findings and had to be excluded where
trial3 and have been regarded as sufficient to influence these data were needed. The L’Abbé plot46,47 suggested that
national and European stroke treatment guidelines (SIGN44 DIAS II18 contributed to the heterogeneity in the combined
and ESO45). analysis of favorable outcomes in all thrombolyzed patients,
Latetreatment,evenamongstselectedpatients,maycarry and the DEDAS trial contributed to the heterogeneity in the
somerisk.Wefoundamarginallysignificantincreaseinthe analysis of reperfusion and recanalization in patients throm-
oddsofdeathamongalltreatedpatients,withapointestimate bolyzedwiththeabandoneddosesexcluded.Bothsourcesof
of 2.4. When we restricted the analysis to 0.9 mg/kg rt-PA heterogeneity appeared to affect the results by virtue of the
andtothedoseofdesmoteplasethatremainsunderdevelop- effects of sample size on the power of a study.
ment(90(cid:1)g/kg),theORformortalityfellto1.6andtherisk We know that the number needed to treat to achieve an
was nonsignificant. Higher doses of desmoteplase were enhancedfavorableoutcomewithalteplasemaybeasfewas
clearlylinkedtoexcessiveSICHandwereabandonedforthis 7within3hours,butthisnumberhasrisenby3to4.5hours
reason. Our analysis did not take into account the attributed to (cid:4)14.3 When treatment with alteplase is started within 6
causeofdeath.ManydeathsinDIASIIandEPITHETwere hours OTT, the number needed to treat rises to 25.48 Hence,
considered unrelated to treatment. The attribution may be our challenge is to identify those patients most likely to
important for understanding the mechanism of effect, but benefitfromdelayedthrombolysis.TheuseofeitherMRIto
caution is required when drawing conclusions from subjec- identifyperfusion/diffusionmismatchoraCT-basedalterna-
tiveassessmentssuchasthese.Treatmentfailurecancontrib- tive is attractive. It is clear from our data that delayed
ute to late death, just as unrecognized excitotoxic damage thrombolysisamongpatientsselectedaccordingtomismatch
mayrepresentapotentialmechanism.Regardless,ifmortality imaging is associated with increased reperfusion/recanaliza-
is increased, this may be mediated via hemorrhagic tion and that recanalization/reperfusion is associated with
transformation. improved outcomes. At present, although the data remain
Despite a lack of significance in the individual odds for consistent with improved functional outcome from delayed
SICH in patients given thrombolytic therapy, the a-OR thrombolysisamongmismatchpatients,astatisticallysignif-
indicated a statistically significant increase in SICH after icantbenefitonfunctionaloutcomeshasnotbeenconfirmed.
delayed thrombolysis. Similarly, an increased risk of SICH Although our pooled results suggest that mortality may be
has long been recognized for time-based t-PA in the estab- higher, the retention of excessive doses of desmoteplase in
lished clinical windows, but this is offset by the improved the analysis is likely to lead to overestimation of any risk.
clinicaloutcomesintreatedpatients.Afterexclusionofdoses Wenotethatexistingmethodsfordefiningmismatchmay
of desmoteplase that were abandoned for clinical develop- be optimized in the future, resulting in greater power of the
ment,20 the adjusted odds for SICH again lost significance. mismatch-based thrombolysis studies. For example, we con-
Cautionisrequiredininterpretingtheseposthocsubgroup sidered 1.2 as the cutoff for defining a mismatch profile.
analyses. Although the inclusion of data from all doses may However, a post hoc analysis of the DEFUSE study has
giveafalselypessimisticviewoftherisk/benefitprofileafter recently shown that the highest sensitivity and specificity
mismatch-based thrombolysis, post hoc exclusion of doses occurred at a mismatch ratio of 2.6, suggesting that the
thatwereabandonedinclinicaldevelopmentisadata-driven previous studies were probably underpowered and lacked a
decisionandraisesstatisticalconcernsofbiasthatcanonlybe sufficiently rigorous definition for the mismatch ratio.27
assuagedbyfurtherprospectivetrials.Wefoundnoevidence Furthermore, the 2-second threshold for Tmax is likely also
thatrelativelyearlier(3-to6-hour)versuslater(6-to9-hour) suboptimal,asaposthocanalysesofDEFUSEdatashoweda
treatment influenced our findings. This is particularly rele- significantly better correlation between infarct growth and
vant,becauseECASSIIIhasrecentlyshownthatunselected penumbra salvage volume for perfusion-weighted imaging
patients benefit from alteplase given within 4.5 hours of lesions defined by Tmax (cid:5)6 seconds.29 The EPITHET
stroke onset, and a small proportion of patients in the investigators reported similar findings.33 It is now clear that
mismatch trials would now be considered eligible for such bothtrialsincludedsignificantvolumesofbenignoligemiain
treatment.Wecannotexcludethepossibilitythatsomeofthe theirmismatchassessments.Recently,automatedonlineanal-
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012
e32 Stroke January 2010
ysisofMRmismatchhasbeendescribedthatfacilitatesrapid R,WahlgrenN,ToniD.Thrombolysiswithalteplase3to4.5hoursafter
selection of patients for delayed treatment. In summary, acuteischemicstroke.NEnglJMed.2008;359:1317–1329.
4. HackeW,DonnanG,FieschiC,KasteM,vonKummerR,BroderickJP,
continuedrefinementinthedefinitionsofdifferentperfusion
BrottT,FrankelM,GrottaJC,HaleyECJr,KwiatkowskiT,LevineSR,
parameters may result in a better choice of the best measure LewandowskiC,LuM,LydenP,MarlerJR,PatelS,TilleyBC,Albers
ofperfusion(Tmax,timetopeak,meantransittime,cerebral G, Bluhmki E, Wilhelm M, Hamilton S. Association of outcome with
blood volume, or cerebral blood flow) and correction for early stroke treatment: pooled analysis of ATLANTIS, ECASS, and
NINDSrt-PAstroketrials.Lancet.2004;363:768–774.
arterial input functions.
5. Barber PA, Darby DG, Desmond PM, Yang Q, Gerraty RP, Jolley D,
Thus, the definitions used in the trials published to date DonnanGA,TressBM,DavisSM.Predictionofstrokeoutcomewith
have been generous and have included many patients who echoplanarperfusion-anddiffusion-weightedMRI.Neurology.1998;51:
418–426.
hadlimitedpenumbraltissueandlimitedprospectsofclinical
6. JansenO,SchellingerP,FiebachJ,HackeW,SartorK.Earlyrecanal-
improvement in response to thrombolysis. The recently isation in acute ischaemic stroke saves tissue at risk defined by MRI.
formed STIR collaboration is initiating a detailed examina- Lancet.1999;353:2036–2037.
tion of this topic. The diversity of mismatch definitions and 7. SchellingerPD,JansenO,FiebachJB,PohlersO,RysselH,HeilandS,
SteinerT,HackeW,SartorK.FeasibilityandpracticalityofMRimaging
large number of investigators involved in these studies
ofstrokeinthemanagementofhyperacutecerebralischemia.AJNRAmJ
weakenconclusionsaboutthepotentialvalueofmismatchin Neuroradiol.2000;21:1184–1189.
the future clinical management of patients with stroke. 8. OstremJL,SaverJL,AlgerJR,StarkmanS,LearyMC,DuckwilerG,
Jahan R, Vespa P, Villablanca JP, Gobin YP, Vinuela F, Kidwell CS.
However,theseweaknessesdonotextendtoourconclusions
Acutebasilararteryocclusion:diffusion-perfusionMRIcharacterization
aboutthestatusofexistingevidenceforuseofthrombolysis
of tissue salvage in patients receiving intra-arterial stroke therapies.
amongmismatchpatients:patientswereselectedaccordingto Stroke.2004;35:e30–e34.
the best intentions of the investigators under protocols that 9. Dafer RM, Biller J. Desmoteplase in the treatment of acute ischemic
stroke.ExpertRevNeurother.2007;7:333–337.
werestateoftheartwhenwritten,althoughtheyhavealready
10. Debens T. Technology evaluation: desmoteplase, PAION/Forest. Curr
been superseded. Prospective phase III trials are required to OpinMolTher.2004;6:567–575.
test whether thrombolysis is associated with a favorable 11. MeretojaA,TatlisumakT.Novelthrombolyticdrugs:willtheymakea
risk/benefitratiowhenusedundermodifiedcircumstances.In difference in the treatment of ischaemic stroke? CNS Drugs. 2008;22:
619–629.
Australia, the EXTEND trial, which will use a phase III 12. Schleuning WD. Vampire bat plasminogen activator DSPA-(cid:2)-1 (des-
design and randomization of patients 4.5 to 9 hours after moteplase):athrombolyticdrugoptimizedbynaturalselection.Haemo-
strokeonsettoalteplaseorplaceboandautomatedmismatch stasis.2001;31:118–122.
13. PaciaroniM,MedeirosE,BogousslavskyJ.Desmoteplase.ExpertOpin
selection, will test this hypothesis. Meanwhile, although the
BiolTher.2009;9:773–778.
concept of selection of patients based on individual patho-
14. Bergum PW, Gardell SJ. Vampire bat salivary plasminogen activator
physiology rather than a rigid time window remains attrac- exhibits a strict and fastidious requirement for polymeric fibrin as its
tive, delayed treatment according to mismatch selection cofactor,unlikehumantissue-typeplasminogenactivator:akineticanal-
ysis.JBiolChem.1992;267:17726–17731.
cannotberecommendedaspartofroutinecareuntilorunless
15. Liberatore GT, Samson A, Bladin C, Schleuning WD, Medcalf RL.
further trials show benefit. Vampire bat salivary plasminogen activator (desmoteplase): a unique
fibrinolytic enzyme that does not promote neurodegeneration. Stroke.
Disclosures 2003;34:537–543.
16. TheNationalInstituteofNeurologicalDisordersandStrokert-PAStroke
N.K.M. is supported by a University of Glasgow scholarship.
Study Group. Tissue plasminogen activator for acute ischemic stroke.
G.W.A. was the principal investigator of the DEFUSE trial, is a
NEnglJMed.1995;333:1581–1587.
consultant to Genentech and to Lundbeck, and was cochair for the
17. WahlgrenN,AhmedN,DavalosA,HackeW,MillanM,MuirK,Roine
steeringcommitteeforDIASI–IV.S.M.D.wascoprincipalinvesti- RO,ToniD,LeesKR.Thrombolysiswithalteplase3–4.5hafteracute
gator of the EPITHET trial, is on the advisory board of Servier ischaemicstroke(SITS-ISTR):anobservationalstudy.Lancet.2008;372:
Australia,andhasreceivedhonorariafromBoehringerIngelheimfor 1303–1309.
lectures.G.A.D.wascoprincipalinvestigatoroftheEPITHETtrial, 18. HackeW,FurlanAJ,Al-RawiY,DavalosA,FiebachJB,GruberF,Kaste
isamemberofadvisoryboardsforServierAustraliaandBoehringer M, Lipka LJ, Pedraza S, Ringleb PA, Rowley HA, Schneider D,
Ingelheim,andhasreceivedhonorariafrombothcompanies.A.J.F. Schwamm LH, Leal JS, Sohngen M, Teal PA, Wilhelm-Ogunbiyi K,
is a consultant to Paion and to Forest Laboratories. W.H. was Wintermark M, Warach S. Intravenous desmoteplase in patients with
chairman of the steering committee of DIAS and cochair of the acute ischaemic stroke selected by MRI perfusion-diffusion weighted
steering committees of DEDAS and DIAS II trials, sponsored by imagingorperfusionCT(DIAS-2):aprospective,randomised,double-
Paion and Forest, and received honoraria for his activities in the blind,placebo-controlledstudy.LancetNeurol.2009;8:141–150.
conduct and development of the trial. K.R.L. was chairman of the 19. FurlanAJ,EydingD,AlbersGW,Al-RawiY,LeesKR,RowleyHA,
datamonitoringcommitteesforDIASI–IV,DEDAS,andECASSIII SacharaC,SoehngenM,WarachS,HackeW.DoseEscalationofDes-
moteplaseforAcuteIschemicStroke(DEDAS):evidenceofsafetyand
trialsofthrombolysisinacuteischemicstroke,sponsoredbyPaion,
efficacy3to9hoursafterstrokeonset.Stroke.2006;37:1227–1231.
ForestLaboratories,Lundbeck,andBoehringerIngelheim.
20. HackeW,AlbersG,Al-RawiY,BogousslavskyJ,DavalosA,Eliasziw
M,FischerM,FurlanA,KasteM,LeesKR,SoehngenM,WarachS.The
References
Desmoteplase in Acute Ischemic Stroke Trial (DIAS): a phase II
1. Wahlgren N, Ahmed N, Davalos A, Ford GA, Grond M, Hacke W, MRI-based9-hourwindowacutestrokethrombolysistrialwithintrave-
Hennerici MG, Kaste M, Kuelkens S, Larrue V, Lees KR, Roine RO, nousdesmoteplase.Stroke.2005;36:66–73.
SoinneL,ToniD,VanhoorenG.Thrombolysiswithalteplaseforacute 21. PetittiDB,ed.Meta-Analysis,DecisionAnalysis,andCost-Effectiveness
ischaemicstrokeintheSafeImplementationofThrombolysisinStroke- Analysis:MethodsforQuantitativeSynthesisinMedicine.2ndEd.New
MonitoringStudy(SITS-MOST):anobservationalstudy.Lancet.2007; York;OxfordUniversityPress:2000.
369:275–282. 22. DavisSM,DonnanGA,ParsonsMW,LeviC,ButcherKS,PeetersA,
2. BrottT.Thrombolysisforstroke.ArchNeurol.1996;53:1305–1306. BarberPA,BladinC,DeSilvaDA,ByrnesG,ChalkJB,FinkJN,Kimber
3. Hacke W, Kaste M, Bluhmki E, Brozman M, Davalos A, Guidetti D, TE,SchultzD,HandPJ,FrayneJ,HankeyG,MuirK,GerratyR,Tress
LarrueV,LeesKR,MedeghriZ,MachnigT,SchneiderD,vonKummer BM,DesmondPM.Effectsofalteplasebeyond3hafterstrokeinthe
Downloaded from http://stroke.ahajournals.org/ at GLASGOW UNIV LIB on January 18, 2012