Special Report. Recommendations for Comprehensive Stroke Centers A Consensus Statement From the Brain Attack Coalition

Recommendations

for

Comprehensive

Stroke

Centers

A

Consensus Statement From

the Brain

Attack

Coalition

Mark

J.

Alberts,

MD;

Richard

E.

Latchaw,

MD;

Warren

R.

Selman,

MD;

Timothy

Shephard,

RN;

Mark

N.

Hadley,

MD;

Lawrence

M.

Brass,

MD;

Walter

Koroshetz,

MD;

John

R.

Marler,

MD;

John

Booss,

MD;

Richard

D.

Zorowitz,

MD;

Janet

B.

Croft,

PhD;

Ellen

Magnis,

MBA;

Diane

Mulligan;

Andrew

Jagoda,

MD;

Robert

O’Connor,

MD;

C.

Michael

Cawley,

MD;

J.J.

Connors,

MD;

Jean

A.

Rose-DeRenzy,

CN,

RN;

Marian

Emr;

Margo

Warren;

Michael

D.

Walker,

MD;

for

the

Brain

Attack

Coalition

BackgroundandPurpose—Todeveloprecommendationsfortheestablishmentofcomprehensivestrokecenterscapable

ofdeliveringthefullspectrumofcaretoseriouslyillpatientswithstrokeandcerebrovasculardisease.Recommenda­ tionsweredevelopedbymembersoftheBrainAttackCoalition(BAC),whichisamultidisciplinarygroupofmembers frommajorprofessionalorganizationsinvolvedwiththe careofpatientswithstrokeandcerebrovasculardisease.

SummaryofReview—Acomprehensiveliteraturesearchwasconductedfrom1966throughDecember2004usingMedline

andPubMed.Articles withinformationaboutclinicaltrials,meta-analyses,careguidelines,scientificguidelines,and other relevant clinical and research reports were examined and graded using established evidence-based medicine approachesfortherapeuticanddiagnosticmodalities.Evidencewasalsoobtainedfromaquestionnairesurveysentto leadersincerebrovasculardisease.MembersofBACreviewedliteraturerelatedtotheirfieldandgradedthescientific evidence on the various diagnostic and treatment modalitiesfor stroke. Input was obtained from the organizations representedbyBAC.BACmetonseveral occasionstorevieweach specificrecommendationandreach aconsensus aboutitsimportance inlightofothermedical,logistical,andfinancialfactors.

Conclusions—There are a number of key areas supported by evidence-based medicine that are important for a

comprehensivestrokecenteranditsabilitytodeliverthewidevarietyofspecializedcareneededbypatientswithserious cerebrovasculardisease.Theseareasinclude:(1)healthcarepersonnelwithspecificexpertiseinanumberofdisciplines, includingneurosurgeryandvascularneurology;(2)advancedneuroimagingcapabilitiessuchasMRIandvarioustypes of cerebral angiography; (3) surgical and endovascular techniques, including clipping and coiling of intracranial aneurysms, carotidendarterectomy, and intra-arterial thrombolytictherapy; and (4) other specific infrastructureand programmatic elements such as an intensive care unit and a stroke registry. Integration of these elements into a coordinated hospital-based program or system is likely to improve outcomes of patients with strokes and complex cerebrovasculardiseasewhorequirethe servicesof acomprehensivestrokecenter.(Stroke.2005;36:1597-1618.)

KeyWords:cerebrovasculardisorders _• cerebral hemorrhage _• healthcaresystems •patientcare _• universitymedical centers

S

troke is a common and serious disorder. Each year,

See

Editorial

Comment,

pg

1616

=750000 individualshaveaneworrecurrentstrokein

the United States.1 _{Hospitalizations attributable to stroke suggest that the care of patients hospitalized because of a} appear to be increasing, with 822000 per year in 1997 strokewillcontinuetobeasignificanthealthcareissueinto compared with 593000 per year in 1988.2 _{These figures theforeseeablefuture.}

ReceivedJanuary5,2005;acceptedFebruary8,2005.

FromtheDepartment ofRadiology (R.E.L.),UniversityofCalifornia atDavis,Sacramento; DepartmentofNeurosurgery (W.R.S.),University

Hospitals ofCleveland,Ohio; Neuroscience Consultants(T.S.),Richmond, Va;Department ofNeurosurgery(M.N.H.), UniversityofAlabama at

Birmingham;NeurologyService(L.B.),VAConnecticutHealthcareSystem,NewHaven;NeurologyService(W.K.),MassachusettsGeneralHospital,

Boston;NationalInstituteofNeurologicalDiseasesandStroke(J.R.M.,M.E.,M.W.,M.D.W.),Bethesda,Md;TheOfficeoftheNationalDirectorof

NeurologyoftheDepartmentofVeteransAffairs(J.B.),WestHaven,Conn;DepartmentofPhysicalMedicineandRehabilitation(R.Z.),Universityof

Pennsylvania,Philadelphia;MiamiCardiacandVascularInstitute(J.J.C.III),Florida;CentersforDiseaseControlandPrevention(J.C.),Atlanta,Ga;

AmericanStrokeAssociation(E.M.),Dallas,Tex;NationalStrokeAssociation(D.M.),Englewood,Colo;DepartmentofEmergencyMedicine(A.J.),Mt.

SinaiSchoolofMedicine,NewYork,NY;ChristianaCareHealthSystem(R.O.),Newark,Del;DepartmentofNeurosurgery(M.C.),EmoryUniversity,

Atlanta,Ga;andOSFSaintFrancisMedicalCenter(J.A.R.-D.),Peoria,Ill.

CorrespondencetoMarkJ.Alberts,MD,Director,StrokeProgram,NorthwesternUniversityMedicalSchool,710NLakeShoreDr,Room1420,

Chicago,[email protected]

©2005AmericanHeartAssociation,Inc.

Strokeisavailableathttp://www.strokeaha.org DOI:10.1161/01.STR.0000170622.07210.b4

In2000,the BrainAttackCoalition(BAC) discussedthe concept of stroke centers andproposed 2 types ofcenters: primaryandcomprehensive.3_{Aprimarystrokecenter(PSC)} has the necessary staffing, infrastructure, and programs to stabilize andtreat mostacute stroke patients.Details about specificelementsofaPSChavebeenpublishedpreviously.3 Effortsare nowunderwaytocredentialfacilitiesasPSCs.4 SeveraldozenhospitalshaveeitherbeencertifiedasPSCsor are going through a certification process. Although PSCs providestrokepatientswithhigh-qualitycare,somepatients with complex stroke types, severe deficits, or multiorgan disease may require and benefit from specialized care and technologicalresourcesnotavailable inatypicalPSC.Such patients often require advanced diagnostic and treatment proceduresdirectedbyspeciallytrainedphysiciansandother healthcareprofessionals.

A comprehensive stroke center (CSC) is defined as a facility or system with the necessary personnel, infrastruc­ ture, expertise, and programs to diagnose and treat stroke patientswhorequireahighintensityofmedicalandsurgical care,specializedtests,orinterventionaltherapies.Thetypes ofpatients whomight useandbenefitfrom a CSCinclude (butarenotlimitedto)patientswithlargeischemicstrokesor hemorrhagicstrokes,thosewithstrokesfromunusualetiolo­ gies or requiring specialized testing or therapies, or those requiringmultispecialtymanagement.Additionalfunctionsof aCSCwouldbetoactasaresourcecenterforotherfacilities intheirregion,suchasPSCs. Thismightincludeproviding expertiseaboutmanagingparticularcases,offeringguidance fortriage ofpatients, makingdiagnostictests or treatments availabletopatients treatedinitially ataPSC,andbeingan educational resource for other hospitals and health care professionalsinacityorregion.

Inanefforttoprovideguidancetohealthcareprofession­ als, hospitals,andadministrators, BAC hasestablished rec­ ommendationsforthedevelopmentofaCSCorsystem.The purposesofthisarticlearetopresentthekeycomponentsof aCSCorsystemandoutlinehoweachelementofaCSCcan be metanddocumented.Theserecommendations shouldbe viewedwithsomeflexibilitysothatindividualfacilitiesand healthcaresystemsmaydevelop theirownCSCcriteria on thebasisoftheserecommendations,yetmodifiedtoaddress andmeetlocalpracticesandpreferences.Thedesignationof aPSCversusaCSCdoesnotimplyadifferenceinthequality ofcare,whichisexpectedtobehighatbothtypesofcenters. As this document iscirculated anddiscussed, BAC antici­ pates further refinements andimprovements that will meet the needs of patients, health care providers, hospitals, and healthcaresystems.

Methods

FiveprocesseswereusedforthedevelopmentoftheCSCrecommen­

dations: (1) a comprehensive literature review, (2) a questionnaire

survey of stroke thought leaders, (3) input from the professional

organizationsrepresented byBAC,(4)gradingofpublishedmedical

evidencefortreatmentsanddiagnostictests,and(5)groupconsensusof

theBACexecutivecommittee.The literaturereview wasconducted

usingtheMedlinedatabaseandPubMedfrom1966toDecember2004.

English language articles that focused on various tests, techniques,

expertise, or programs related to the care of stroke patients were

reviewedandevaluated.Meta-analyses,consensusstatements,practice

guidelines,andpositionarticleswerealsoreviewed.Somecomponents

suchaspersonnelandstaffingarenoteasily gradedusing

evidence-basedmedicineprotocols.Insuchcases,theothermethodslistedabove

wereusedtoformulaterecommendations.

MembersoftheexecutivecommitteeofBACwereaskedtoquery

theirparentorganizationsforguidanceabouttheessentialelements

for a CSC. Their specific recommendations were considered for

inclusionintothisdocument.Thequestionnairesurveyconsistedof

40questionsdealingwithvariouspotentialelementsofaCSC.Itwas

mailedto160stroke programdirectorsand otherphysicianswith

interestand expertise in stroke care (ie,vascular neurosurgeons,

neurologists,emergencydepartment[ED]physicians).Respondents

wereaskedtorankeachelementonascaleof1to5(leastimportant

tomost important)in terms ofimportance for thecare of stroke

patientsin aCSC andtoindicate whethertheirhospital hadeach

element.

Whereappropriate,standardevidence-basedmedicineassessment

criteriawereused tograderecommendationsforvarioustherapies

usedataCSC(Table1).5_{Fordiagnostictesting,weevaluatedthe}

evidenceusingcriteriadevelopedrecentlyforassessingtheutilityof

cerebralperfusiontechniques(withsomemodifications).6_Responses

from the questionnairecited above were used when appropriate.

Finally,BACexecutivecommitteemembersmetonseveralocca­

sionsinpersonandviateleconferencetoreviewandrefinethelistof

elements for a CSC and to develop suggestions for how each

componentcouldbedocumented.

Results

ThekeycomponentsofaCSCorsystemcanbedefinedin4 majorareas:(1)personnelwithspecificareasofexpertise,(2) specialized diagnosticandtreatmenttechniques,(3)facility infrastructure, and (4)other programmatic areas (Table 2). Theseelements are bestillustrated by the types ofpatients caredforinaCSC andthe medicalneeds ofthosepatients (Table3).Resultsofthenationalsurveyare includedinthe appendix.

Personneland ClinicalExpertise

A CSC should have the following personnel: (1) a center director, (2) neurologists and neurosurgeons, (3) surgeons withexpertiseperformingcarotidendarterectomy(CEA),(4) diagnosticradiologists,(5)physicianswithexpertiseininter­ ventionalendovascularneuroradiologyproceduresandtech­ niques (6) ED personnel and links to emergency medical services(EMS),(7)radiologytechnologists,(8)nursingstaff whoaretrainedinthe careofstrokepatients,(9)advanced practice nurses (APNs), (10) physicians with expertise in criticalcareorneurointensivecare,echocardiography,carotid ultrasound(U/S),andtranscranialDoppler(TCD),(11)phy­ siciansandtherapistswithtraininginrehabilitation,and(11) casemanagersandsocialworkers.

Strongleadershipisanimportantelementforthesuccess­ fulformationandoperationofaCSC.TheneedfortheCSC directortohaveasignificantamountoftrainingandexpertise in vascular neurology or neurosurgery is supported by the questionnairesurveyandBAC.TheCSCdirectormightbea neurologistorneurosurgeon,althoughother medicalprofes­ sionalscouldfulfillthisrole.Examplesofqualificationsfora CSC director include :2 of the following: (1) a board-certified neurologist or neurosurgeonwho hascompleted a strokefellowshiporvascularneurosurgeryfellowshiporhas equivalentexperience,(2)boardcertifiedinvascularneurol­ ogy,(3)afellowoftheStrokeCounciloftheAmericanHeart

TABLE1. ApproachtoGradingRecommendations*

LevelofEvidence Therapy/Treatment DiagnosticTest

I DatafromRCTwithsufficientstatisticalpowertomakefalse Evidencefromprospectivestudy(s)inabroadspectrumof

positive/negativefindingsunlikely;treatmentmaybe patients;goldstandardcomparisonswhenappropriate;high

FDA-approved accuracyrate

II DatafromRCT,butmayhavefalsepositivesornegatives;may Evidencefromprospectivestudyofanarrowpatientpopulationor

notbeFDA-approved,butRxiswidelyorcommonlyusedin well-designedretrospectivestudiesofabroadpopulation;

manymedicalcenters comparisonwithgoldstandardorotherreasonablevalidated

alternativetest

III Datafromnonrandomizedcohortstudies;Rxisusedinsome Evidencefromretrospectivestudiesinanarrowpatientpopulation

settingsbutnotwidelyadopted

IV Datafromnonrandomizedstudiesusinghistoricalcontrols Mostevidencefromcaseseriesorexpertopinionpanels

V Datafromanecdotalcaseseriesorseveralcasereports

StrengthofRecommendation

GradeA SupportedbylevelIevidence Establishedasuseful/predictiveforconditioninspecificpopulation

GradeB SupportedbylevelIIevidence Probablyuseful/predictiveforconditioninspecificpopulation

GradeC SupportedbylevelIII,IV,orVevidence Possibleuseful/predictiveforconditioninspecificpopulation

GradeD Datainadequateorconflicting;valueoftestunclearor

controversial

*Incasesinwhichthistypeofgradingisnotdirectlyapplicabletoaspecificrecommendation,theauthorsconsideredthebodyofavailableevidenceandpractice

standardstodeterminetheappropriategrading.

RCTindicatesrandomizedcontrolledtrial;Rx,therapy.

Association(AHA)(4)aclinicianwhodiagnosesandtreats :50 patients with cerebrovascular disease annually; (5) a clinicianwith :10peer-reviewed publicationsdealingwith cerebrovasculardisease,(6)aclinicianwith:12continuing medicaleducation(CME)creditseachyearinareasdirectly related to cerebrovascular disease,and (7)other criteria as determinedbythelocalhealthcaresystem.

Thecenterdirectororhis/herdesigneeshouldbeavailable 24 hours per day, 7 days per week (24/7) to provide leadership and deal with difficult medical, logistical, and administrativeissues.It isexpected that inmost cases,the center director would be involved in the assessment of patients and provide consultative advice to other treating physicians.Itisrecommendedthat:1otherphysicianswith expertisein cerebrovasculardisease alsobeon staffsothat continuous24/7coveragecanbeassured.ACSCshouldhave :1neurologists(preferablywith fellowshiptraininginvas­ cularneurology).Publishedobservationalstudieshaveshown thatstrokepatientscaredforbyneurologistshaveimproved outcomes compared with care by other physicians (level IIIC).7,8 _{Such physicians should be available within 20} minutestoansweremergencycallsbyphoneandbeavailable in-house within 45 minutes if needed. The need for a neurosurgeonisdiscussedbelow.

Manypatients caredforinaCSCwill havehemorrhagic strokes and require care in an intensive care unit (ICU). Physicianswith training in criticalcaremedicine or neuro­ critical care should be part of the CSC to manage these patientsintheICUorneuroscienceICU.9,10_{Suchpersonnel} wouldtypicallybeaboard-eligibleorboard-certifiedneurol­ ogist, neurosurgeon, anesthesiologist, or internist who has completed either a critical care fellowship or neurocritical carefellowship.Itisrecommendedthattheseclinicianscare for:20 patientswith acutestrokesper yearandattend:4

hours per year of CME activities (or similar educational programs)relatedtoorfocusedoncerebrovasculardisease.

Althoughit isdifficultto quantify the qualityofnursing care, the consensus of BAC and other practitioners isthat high-quality nursing care is a key factor in determining patientoutcomesafterastroke.Themajorityofnursescaring forstrokepatientsinanICU,strokeunit,andwardshouldbe registerednurses.AllnursesinaCSCshouldbefamiliarwith standardneurologicassessmentsandscales,strokeprotocols, caremaps,ongoingresearch projects, andnewpatientcare techniques related to stroke. Nurses caring primarily for strokepatientsshouldattendtraining sessionssponsoredby the CSC(ie,inservices, seminars,specialized lectures):3 timesperyear.Suchnursesshouldparticipatein:10hours of continuing education units (CEUs) activities (or other educationalprograms)annuallythatarerelatedtoorfocused oncerebrovasculardisease.Eachnurseshouldhaveafilethat documentshis/herparticipationin the aboveactivities.Itis suggestedthateachCSCnurse(strokeunitorICU)attend:1 nationalorregionalmeetingeveryotheryearthatfocuseson someaspectofcerebrovasculardisease.

An APN is a vital team member involved in several importantaspectsofaCSCsuchaspatientcare,caremaps, research activities, stroke registries, educational programs, and quality assurance.11 _{The designation of APN could} includeanursepractitioner,master’s-preparedclinicalnurse specialist, or American Board of Neuroscience Nurses– certifiednurse.ItisrecommendedthataCSChave:1APN (or similar personnel) to implement and coordinate the programsoutlinedbelow.Thisrecommendationissupported byBACaswellasthesurveyresults.

ItisvitalthattheCSCstaffbefullyintegratedwithEMS personnel andED staff.EMS and EDpersonnel shouldbe very familiar with the diagnosis and treatment of patients with cerebrovascular disease. Several studies have

docu-TABLE2. ComponentsofaCSC

Recommendation(grade) Optional

Personnelwithexpertiseinthefollowingareas

Vascularneurology Neuroscienceintensivecare

Vascularneurosurgery Nursingdirectorforstrokeprogram

APN

Vascularsurgery

Diagnosticradiology/neuroradiology

Interventional/endovascularphysician(s)

Criticalcaremedicine

Physicalmedicineandrehabilitation

Rehabilitationtherapy(physical,occupational,speechtherapy)

Staffstrokenurse(s)

RT

Swallowingassessment

Diagnostictechniques

MRIwithdiffusion(IA) MRperfusion(IIB)

MRA/MRV(IA) CTperfusion(IIIC)

CTA(IA) XenonCT(IIIC)

Digitalcerebralangiography(IA) SPECT(IIIC)

TCD(IA) PET(IIB)

CarotidduplexU/S(IA)

Transesophagealecho(IA)

Surgicalandinterventionaltherapies

CEA(IA)

Clippingofintracranialaneurysm(IA) Stenting/angioplastyofextracranialvessels(IIB)*

Placementofventriculostomy(IA) Stenting/angioplastyofintracranialvessels(IIIC)*

Hematomaremoval/draining(IIB–VC)

Placementofintracranialpressuretransducer(VC)

EndovascularablationofIAs/AVMs(IA)

IAreperfusiontherapy(IIB)

EndovascularRxofvasospasm(IIIC)

Infrastructure

Strokeunit†(IA)

ICU Strokeclinic

Operatingroomstaffed24/7 Airambulance

Interventionalservicescoverage24/7 NeuroscienceICU

Strokeregistry(IIIC)

Educational/researchprograms

Communityeducation(IA) Clinicalresearch

Communityprevention(IA) Laboratoryresearch

Professionaleducation Fellowshipprogram

Patienteducation Presentationsatnationalmeetings

*Althoughthese therapiesarecurrentlynotsupportedbygradeIA evidence,theymaybeusefulforselected

patientsinsomeclinicalsettings.Therefore,aCSCthatdoesnotofferthesetherapiesshouldhaveanestablished

referralmechanismandprotocolto sendappropriatepatients toanotherfacilitythatdoesofferthesetherapies;

†strokeunitmaybepartofanICU.

Rxindicatestherapy.

mentedtheimportanceoftheEMSsystemandEDpersonnel focusoncerebrovasculardisease.Ideally,theEDphysicians forthe rapid identificationand transportationof strokeand shouldbe board certified. They shouldmeet with the CSC patients and the initiation of therapy.12–16 _{EMS and ED directorat leastsemiannuallyandreviewcareissues.Other} personnelshouldattendinitialandongoingeducationalpro- aspectsof the integrationof the ED/EMS personnel witha grams (ie, in services, CME programs, grand rounds) that strokecenterarereviewedinthePSCrecommendations.3

TABLE3. UseofCSCComponentsinVariousPatientPopulations

Ischemic

Stroke ICH SAH

Personnel Vascularneurologist X X X Neurosurgeon X X X Intensivist Asneeded X X Vascularsurgeon X Endovascularspecialist X X X Caresetting Strokeunit X X X ICU X X X Neuroimaging MRI/MRA,DWI X X X MRV X X X Digitalangiography X X X Carotidultrasound X TCD X X TEE X Endovasculartherapy Aneurysmablation X AVMembolization X X

Angioplastyforvasospasm X

Stent/angioplastyforatherothrombosis X

Reperfusiontechniques X

Surgery

Ventriculostomy X X X

Intracranialpressuretransducer X X X

Hemicraniectomy X X

Hematomaremoval X X

Aneurysmclipping X

CEA X

Brainbiopsy X X

Rehabilitationassessments andtreatments (physical ther­ apy,occupationaltherapy,andspeechtherapy)areanimpor­ tant component of acute care and long-term recovery and should begin soon after the patient is admitted and stabi­ lized.17,18 _{Below is a more complete discussion of the} personnel recommendations for rehabilitation. Expertise in assessingswallowingfunction isan important elementofa CSCbecauseofhighratesofdysphagiainstrokepatients(up to 50%) and a risk of aspiration pneumonia.19,20 _These assessmentsareoftenperformedbyaspeciallytrainedspeech therapistorotolaryngologist,althoughnursesandotherscan performsomeswallowingevaluations.21_{Casemanagersand} social workers who have experience dealing with stroke patients and their families/caregivers are an invaluable re­ source.ItisrecommendedthataCSChave:1casemanager orsocialworker onstafftoprovidecoverageforpatientsin needofhis/herservices.

MuchofwhatdistinguishesaCSCfromotherfacilitiesis expertiseandinfrastructurein3keyareas:diagnosticradiol­ ogy,endovasculartherapy,andsurgery.Theseareasarevital

in the management of patients with large ischemicstrokes and hemorrhagic strokes, and they are discussed below in detail.Thereisaseparatesectionthatreviewsrecommenda­ tionsforrehabilitation.

DiagnosticImaging:Techniques andPersonnel

Patients in a CSC need accurate imaging of the brain and related vasculature and physiological evaluation regarding theeffectsofcerebralischemiaandhemorrhage.Appropriate computedtomographicresourcesareaprerequisiteforbeing a PSC, the recommendations for which are not repeated.3 Thissection detailsthe recommendations forotherimaging andrelatedtechniques.

MRIandRelatedTechniques

The contrast resolutionof MRI issignificantly higher than computed tomography (CT), making it far more sensitive thanCTfordetectingthe oftensubtleabnormalitiesseen in earlycerebralischemiaandotherconditions.22,23,24_Numerous studieshaveclearlydemonstratedthesuperiorityofMRIfor detectingacuteischemia(especiallyintheposteriorfossa)as well as other processes that can present with stroke-like symptoms (grade IA).5,25–28 _{Basic MRI at a CSC must be} availableona24/7basis,evenifpersonnelarecalledinfrom home.AnMRIshouldbecompletedwithin2hoursofthetest beingorderedata CSC.

Diffusion-weightedMRI (DWI) isvery sensitivefor de­ tectingcerebral ischemiawithin minutes after its onset,far exceeding any other imaging method available today.29,30 Calculationoftheapparentdiffusioncoefficientisimportant to confirm that a diffusion abnormality is attributable to ischemia.31_{DWI detects}>_{90%ofacuteischemiclesionsin} thebrain.27,32–34_{Patternsofstrokeseenwith DWImayalso} provide important information about stroke mechanism.35 Although results of DWI may not affect outcome, it is a valuablediagnostictoolandshouldbepartoftheevaluation ofpatientswithanacuteischemicstroke(gradeIA).36,37,38_It shouldbeperformedaspartofastandardMRI,withthesame timerequirements.

Magnetic resonance (MR) perfusion provides valuable information aboutblood flow in specific brain regions and vascularterritories.Itcanbeusefulindeterminingthesizeof a perfusiondeficit andidentifyingbrain tissue that maybe ischemic but not infarcted.39 _{MR perfusion defects may} correlate with clinical outcomes.40 _{However, the clinical} utility of MR perfusion for guiding therapy or affecting outcomehasnotbeendocumentedbylargeprospectivetrials; therefore, it is considered an optional element of a CSC (gradeIIB).6,37,38

MR angiography (MRA)isan effective andnoninvasive technique for visualizing abnormalities of the extracranial andtheintracranialcerebralcirculation.Theoverallsensitiv­ ityandspecificityofMRAforextracranialcarotiddiseaseis 82% to 86% and 98%, respectively.24,41 _{MRA is more} sensitivethanU/Salonefordiagnosinghigh-gradeextracra­ nialcarotidstenosis.42_{Its accuracyfordetecting significant} high-gradeextracranial vasculardiseaseinsomecases(par­ ticularly elliptic centric contrast MRA) approaches that of catheter-baseddigitalangiography(CA),consideredtobethe

“goldstandard,”withasensitivityof97%andaspecificityof 95%(gradeIA).43_{Theuseofintravenouslyinjectedcontrast} materialhasfurtherincreasedtheaccuracy,spatialresolution, andreproducibilityofMRA.43–47_{Theaccuracyandreproduc­} ibility ofMRA fordetecting intracranialstenoses isless.48 ThetimeframefordoingMRAissimilartothatforabrain MRI.

MRA can also be usefulfor detecting intracranial aneu­ rysms.Theaccuracyofthistechniquedependsonthesizeof theaneurysm,thefieldstrengthofthemagnet,andthetypeof MRAsequence used.24_{Forintracranialaneurysms}>_5mm, nonenhanced3D time-of-flightMRA performedon a 1.5-T systemhas an accuracyof >85% relativeto CA, although accuracy approaches 100% with increasing aneurysm size (gradeIIB).49–51

MRvenography (MRV)isa safe,rapid,andnoninvasive techniquetodiagnosecerebralvenousthrombosis(CVT).52_A positive MRV caneliminate the need for invasive cerebral angiographyinmanycasesofCVT,althoughfalsepositives canoccur.53–55_{SoftwareforperformingMRVisavailableon} allcurrentscannersystems.BecauseoftheabilityofMRVto noninvasively diagnose CVT and the wide availability of MRV,itisarecommendedtechniqueforaCSC(gradeIIB).

CatheterAngiography(GradeIA)

Digital subtraction angiography (DSA) represents the gold standard for the detection and characterization of cerebral aneurysms,arteriovenousmalformations(AVMs), andarte­ riovenous fistulae (AVFs), and for measuring the exact degree of stenosis in extracranial and intracranial arteries (grade IA).56 –59_{Itistheprocedure ofchoiceforevaluating} the third- and fourth-order intracranial branchesto makea diagnosis of a central nervous system (CNS) vasculitis.24 Single-plane systems suffice for diagnostic uses, although biplanesystemsprovideashorterexaminationtimeandfewer injections. Because of the emergent nature of some of the stroketypesdiscussedabove,cerebralangiographymustbe available ata CSC ona 24/7 basis,with supportpersonnel available to comein fromhome fora procedurewithin 60 minutesofbeingcalled.ACSCmustdemonstrateaperipro­ cedurestrokeanddeathrate of<1%andanoverallserious complicationrateof�2%forCA.60

CTAngiography(GradeIA)

CT angiography (CTA) is a noninvasive technique that is veryusefulforrapidlyimagingthelargevesselsintheneck andmanyfirst- andsecond-orderarteriesinthebrain.CTA can detect vascular stenoses, acute emboli, and cerebral aneurysms with a high degree of sensitivity and specifici­ ty.61– 63 _{Thespatial resolution ofCTA issuperiorto MRA,} and a “string sign” may be detected more accurately than even DSA because of its cross-sectional image acquisition andabilitytodetectminuteamountsofcontrastmaterial.64_In general, CTA has sensitivities and specificities of 80% to 100%fordetectinghigh-gradeextracraniallesions.64 – 66_CTA hasasensitivityof53%to100%andaspecificityof87%to 100% for detecting intracranial aneurysms. For aneurysms :7mm,CTAhasa sensitivityof95% andaspecificity of 98.9%.67_{Mostrecentstudieshavereported sensitivitiesand} specificitiesforCTAof>90%to95%whencomparedwith

digital angiography for the detection of aneurysms.68–71 _In somecases,aCTAcandetectananeurysmmissedbyCA.62 CTAcannotprovidethesamedetailedcerebralhemodynamic data provided by CA, nor can it accurately image small cerebralvessels. However,because of the significant flexi­ bility and accuracy of CTA, particularly for patients who cannot undergoan MRA or a conventional cerebralangio­ gram, and its noninvasive nature, it is a recommended elementforaCSC(gradeIA).Itispossiblethatinthenear future,CTAmightreplaceCAformanyindications.72

ExtracranialUltrasonography(GradeIA)

CarotidU/Sisrelativelyinexpensive,verysafe,andcanbe usedto noninvasivelyscreen fordiseaseandfollow known diseaseintheextracranialcarotidandvertebralarteries.Itcan beusedinpatientsunabletoreceivecontrastdyesorinwhom anMRAiscontraindicated(pacemaker,metalimplants,etc). ThesensitivityandspecificityofcarotidU/Scanbeashigh as85%to90%forhemodynamicallysignificantlesionsatthe carotidbifurcation,althoughitislesssensitivefordiseasein the vertebral arteries.24,41,73 _{Because of its easeof use and} accuracy, it is recommended that a CSC have extracranial U/Sanddemonstrateacceptableproficiencyusingguidelines establishedby the Intersocietal Committee forthe Accredi­ tationofVascularLaboratories(ICAVL)orasimilarcreden­ tialingorganization.

TranscranialDoppler (GradeIA)

TCD is a safe, noninvasive, and low-cost technique for imagingthelarge intracranialvesselsatthe skullbase.Itis usedinpatientswithacutecerebralischemiaforthedetection ofintracranialstenosisandocclusionsandforthedetectionof vasospasm in patients with neurological deterioration after subarachnoid hemorrhage (SAH).74 –77 _{For the detection of} vasospasm,TCDhasasensitivityof80%andaspecificityof 95%comparedwithCA.78,79_{OtherstudieshaveshownTCD} tobeusefulformonitoringrecanalizationafterthrombolytic therapy.80_{Basedonitsaccuracyandimportanceinmonitor­} ingpatientswithSAH,TCDisarecommendedelementofa CSC.81_{AswithcarotidU/S,theTCDlaboratoryshouldtrack} theirresultsandseek certificationfromICAVLorasimilar organization.

TransthoracicandTransesophagealEchocardiography (GradeIA)

Becauseasignificantpercentageofstrokesareofcardioem­ bolic origin, cardiac imaging is an important test in most strokepatients.5_{Practiceguidelinessupportcardiacimaging} incasesoftransientischemicattackandstroke.82_Transtho­ racicechocardiography(TTE)isaroutinetestusedtoimage theheartforthepresenceofclots,valvularabnormalities,and thedeterminationofleftventricularfunctionandwallmotion abnormalities.83_{Transesophagealechocardiography(TEE)is} ahighlysensitivetestfordetectingseveralcardiacandaortic lesionsthatmaycauseischemicstrokes,includingthrombiin the left atrium, masses on the mitral and aortic valves, a patent foramen ovale, intra-atrial septal aneurysm, and atherothromoboticlesionsinthe aorticarch.84 – 89_Numerous studieshave proven the increased sensitivity ofTEE com­ paredwith TTE inpatients with ischemicstrokes.90,91 _TTE

andTEEmustbe performed andinterpretedby technicians andcardiologistswithtraininginthesetechniques.92

TestsofCerebralBloodFlowandMetabolism

Thereareavarietyofmethodologiescurrentlyavailablethat assess cerebral blood flow, including MR perfusion, CT perfusion, single-photon emission CT, positron emission tomography(PET),andxenonCT.6_{PETprovidesdataabout} cerebralbloodflow,brainmetabolism,anddegreeofische­ mia.Itmaybeusefulinsomecasesforguidingacutetherapy (gradeIIIC).93,94_{However,sophisticatedhardwareisrequired} to detect and measure these isotopes, and their production requiresexpensiveinfrastructure. Allofthesetestsare non­ invasive. There isno compelling datathat these tests alter managementoroutcomesin mostpatients.Somecannotbe doneonan emergentbasis.Theymaybemostusefulatthe presenttimeaspartofresearchprotocols.

DiagnosticRadiologyPersonnel

ACSC musthavephysiciansavailable toevaluate imaging studies 24/7. Although it is preferable that the attending physicianbeafellowship-trained neuroradiologist,veryfew institutionshaveanin-housefellowship-trainedneuroradiolo­ gistona24/7basis,althoughmanyhaveageneraldiagnostic radiologistavailablein-house24/7.Forurgentneuroimaging studies,physiciansexperiencedininterpreting headCTand brain MRI studies must be available to read these scans within 20 minutesof theircompletion.The proliferation of telecommunications systems for the rapid assessment of diagnosticimages makesquite feasibletherequirementthat anemergencyCTscan orMRIisevaluatedbyaneuroradi­ ologist, general diagnostic radiologist, or other suitably trained physician in a variety of care settings within 20 minutesofscanacquisition.95,96

Because of the need for the performance of a CT scan within 25 minutes, there must be an in-house technologist capableofperformingaCTscanandanyCT-based studies. TheAmericanSocietyofRadiologicalTechnologists,Amer­ ican Registry of Radiological Technologists, and the Joint ReviewCommittee onEducationinRadiologicTechnology allhaverequirementsforthetraining,testing,andcertifica­ tionoftechnologistsperformingalltypesofimagingstudies. ACSCmusthave:1certifiedradiologytechnologisttrained inCTtechniquesin-houseona24/7basis.

Therequirementthata CSCwillperformMRIstudieson a24/7basismeansthataqualifiedMRtechnologistmustbe available(butnotnecessarilyin-house)ona24/7basis.The technologistmaytakecallsfromhomeaslongashe/shecan be at the hospital within 1 hour ofbeing paged.A similar requirementappliestotechnologists andtechniciansneeded to performa cerebralangiogram. U/S andvarious cerebral perfusionstudiesarecommonlyelective,andtheavailability oftechnologiststoperformthemwillvaryamonginstitutions.

EndovascularTherapy:Proceduresand Personnel

Endovascular techniques and devices are being used with increasingfrequencyforthe treatment ofa varietyofcere­ brovascular diseases. These include ablation of cerebral aneurysms, angioplasty and stenting of occlusive lesions, intracranial angioplasty for vasospasm, intra-arterial (IA)

thrombolysisforacutestroke,andembolizationofAVMsand AVFs.

The endovascular ablation of aneurysms is a safe and effectivealternativeto surgicalclippinginselectedpatients (gradeIA).Publishedmulticentertrialresultsandguidelines supporttheuseofendovasculartherapyinsuchpatients.97,98 A multinational trial of endovascular treatment using the Guglielmidetachablecoil(GDC)versussurgicalclippingof ruptured intracranial aneurysms found a 7% absolute risk reductionofdeathordependencyinpatientstreatedwiththe GDC compared with surgery.99 _{These results may not be} extrapolated to all patients with all types of aneurysms. Complete aneurysm ablation may be less common with endovascularcoilingthanwithclipping,andtheremaybea higherrateofearlyrebleeding.100_{Thelong-termdurabilityof} endovascular ablation versus surgical clipping remains un­ clear.100_{Someaneurysmsappeartobebettertreatedwithan} endovascular approach and others with surgical therapy. Therefore,aCSCisrequiredtohavethecapabilitytoperform microsurgicalneurovascularclippingandneuroendovascular coiling.

Vasospasm is a frequent and deadly complication of an SAH.101_{Medicalmanagementsuchashemodynamictherapy} often fails to reverse the clinical effects of the vaso­ spasm.102,103 _{Catheter-directed intracerebral IA infusion of} vasodilatorsisanimportanttherapeuticoptionusedroutinely in some cases of vasospasm with mixed results (grade IIIC).104–108 _{Intracranial angioplasty for vasospasm has a} success rate of >90% in correcting the angiographically visiblevasospasm,withclinicalimprovementin60%to80% ofpatients (grade IIIC)106,109,110 _{anda complication rate of} 2% to 4%.111_{Although angioplasty for vasospasmhas not} been subjected to rigorous clinical study, it is considered very effective and is a standard therapy for severe vaso­ spasm.110 –112_{Becausetheothertherapeuticoptionsforsymp­} tomatic vasospasm are limited and often ineffective, the abilitytoperformintracranialangioplastyorIAinfusionsof vasodilatorsisrecommendedforaCSC.IfaCSCistempo­ rarily unable to offer this therapy, it is recommended that protocols be developed for the rapid transfer of patients needingthesetreatments toanearbyfacilitythat doesoffer thistherapy.

IAthrombolysisinvolvestheuseofadvancedangiographic techniques for the placement of a microcatheter into a cerebralvessel for the infusion ofa thrombolyticdrug. IA thrombolyticshave increasedefficacycompared with intra­ venouslyticsfordissolvingthrombiwithinthelargearteries at the skullbase, althoughit carries a 10%to 18% risk of symptomatic intracerebral hemorrhage (ICH) in some cas­ es.113–116 _{The use ofIA lytic agents might extendthe time} window for therapy beyond the 3-hour requirement for intravenous thrombolysis.114 _{One prospective, randomized} trialofIApro-urokinaseshoweda15%absoluteincreasein goodneurologic outcomesand a 10% rate of symptomatic ICH.115_{Othersmallercaseserieshavealsofoundsignificant} benefits for IA thrombolysis in stroke patients with large arteryocclusions.117–119

There iscurrently no fibrinolytic agent with a Food and DrugAdministration(FDA)labelindicationforIA

adminis-trationforthe treatmentofacuteischemicstroke.However, therehasbeenextensiveexperiencewiththistechnique,itis commonly used at many medical centers, and it isrecom­ mendedinthecurrentAHAAdvancedCardiacLifeSupport handbook.120–122 _{Based on all of these factors and the} consensusofBAC,IAlyticsareconsideredarecommended componentofaCSC(gradeIIB).Complicationratesshould bemonitoredclosely.Mechanical thrombectomytechniques forthecerebralcirculationarealsobeingdevelopedthatuse a variety of devices such as microcatheters, snares, clot retrievers, and balloons123–126_{(grade VC). A clot retrieval} device recently received FDA approval, although clinical experienceislimited.127_{Intrasinus lytic agentsmayalsobe} efficaciousintreatingselected casesofCVT, althoughthis manner128–132 therapy has not been studied in a rigorous

(gradeVC).

Carotidangioplastyandstenting(CAS)maybeanoption for the treatment of selected patients with symptomatic or asymptomaticcarotidarterystenosis.Overthepast10years, thetechnicalsuccessrateforCAShasrisento>97%,andthe complication rates have fallen.133,134 _{However, there is a} paucityof data fromprospective, randomized studies com­ paring the efficacy and safety of CAS to CEA or to best current medical therapy Onerandomized study of 220 pa­ tients with symptomatic carotid artery stenosis found the 1-yearstrokeandvasculardeathratetobehigherinthestent group versus CEA group (10.4% versus 4.6%), although these differences were not statistically significant.135 _The Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS) compared endarterectomy with angio­ plasty(25%ofpatientsalsoreceivedastent)inpatientswith carotidorvertebralarterystenosis.136_{Overall,therewereno} significantdifferencesinmajoroutcomessuchasstrokeand death. Data from a large unrandomized registry found the 1-year stroke anddeath rate to be 11% in the stent group, abouthalfofwhomhadasymptomaticlesions.137_Astudyof high surgical–risk patients found the 30-day complication rateofcarotidstentingtobe halfthatofCEAwhenusinga distalprotectiondevice(6%versus 12%).138

ThereisgeneralagreementthatCASmaybeanacceptable treatmentoptionin patientsthoughttobe at highrisk fora CEA(ie,restenosis afterCEA,radiationfibrosis, fibromus­ culardysplasia,surgicallyinaccessiblestenosis,contralateral carotid disease, and significant cardiac or pulmonary dis­ ease).139,140_{Stentingmayalsobeconsideredinpatientswith} arterialdissectionthatisunresponsivetomedicaltherapyor in whom treatment with anticoagulation is contraindicat­ ed.141,142 _{Based on all of this information, stenting of ex­} tracranial carotid arteries for atherothrombotic disease is gradeIIBandisconsideredanoptionalelementofaCSC(see below). The National Institutes ofHealth (NIH)–sponsored multicenterCREST (Carotid Revascularization:Endarterec­ tomy versus Stenting) trial that is now under way will hopefullydeterminethe relativesafetyandefficacyofCAS comparedwithCEAinpatientswithaveragesurgicalriskand symptomaticextracranialcarotidstenosis.143

BAC recommendsthat forpatientswith average surgical risk,suchasthosewhowouldhavequalifiedforenrollment intheNorthAmericanSymptomaticCarotidEndarterectomy

Trial(NASCET)andtheAsymptomaticCarotidAtheroscle­ rosisStudy(ACAS),CASshouldbe performedaspartofa randomized clinical trial such as CREST or under a local institutionalreviewboard–approvedinvestigationalprogram. CASplacementshouldonlybeperformedbyanindividualor team with training and expertise in cerebral angiography, cerebrovascularpathophysiology,hemodynamics,andneuro­ vascularinterventions.144,145

Angioplastyand stenting ofstenotic lesionsin the intra-cranial circulation (including vertebral– basilar territory) is anotherareaofgreatinterestbecauseofthepooroutcomein patients who fail medical therapy.146 _{However, there is a} paucityofdatafromrandomizedcontrolledtrialstoproperly evaluatethistreatmentapproach.147_{Somestudieshavefound} 30-day complications rates of 5% to 30% and 12-month stroke and death rates of 28% to 40% for intracranial angioplasty/stenting.148 –151_{Incomparison,theWarfarin–As­} pirin Symptomatic Intracranial Disease (WASID) study foundan annual strokerate of15% forpatients with intra-cranialsymptomaticlesionswith>50%stenosis.152

TheStentingofSymptomaticLesionsoftheVertebraland IntracranialArteries(SSYLVIA)trialreported a strokeand death rate of 6.6% in 30 days and a 12-month stroke and deathrateof13.2%.153_{TheSSYLVIAtrialresultedinFDA} approval for a specific angioplasty balloon and stent for intracranial atherosclerotic stenosis. The 2 largest reported seriesof long-termstrokepreventionforintracranialangio­ plasty orstenting demonstratedvery low long-term stroke/ deathratesof<3.5%,with>300patientyearsoffollow-up in1seriesand70patientyearsintheother.151,154_Becauseof the lack of data from large, prospective randomized trials, extracranialandintracranialangioplasty/stentingforcerebro­ vascular disease isconsideredan optionalcomponent fora CSC, althoughthere areselected casesin whichsuchtech­ niquesmaybeofvalue(gradeIVC).Ifacenterdoesofferthis procedure, it is recommended that cases be entered into a registrytotrackoutcomes.ItisrecommendedthatifaCSC doesnotofferextracranialandintracranialangioplasty/stent­ ing,ithas availablea referralarrangement tosend selected patientstoanotherfacilitythatdoesoffertheseinterventions. Some AVMs cannot be easily treated with conventional surgery alone. Preoperative embolization maydecrease the flowsufficientlysothatsurgicalresectionispossible.155,156_In othercases, embolizationmayreduce the sizeof the AVM sufficientlytoallowfocusedirradiation.157,158_{Rarely,embo­} lization alone may be curative.159 _{Some AVFs are treated} solelywithendovascularablationtechniques,whereasothers mightbetreatedwithsurgeryalone. Occasionally,acombi­ nationofmethodsmustbeused.160,161_{Thesetechniqueshave} notbeenstudiedinrigorous,randomizedclinicaltrials(grade IIIC).162,163

Aneuroendovascularspecialist(eg,endovascularsurgical neuroradiologist)isrecommendedasanecessarycomponent of a CSC. An individualwith such expertise iscapable of performing extracranial and intracranial angioplasty and stentingforatherosclerosisorvasospasmaswellasperform­ ing emergency catheter-directed IA stroke therapy. The endovasculartreatmentofpatientswithcerebralaneurysms, AVMs,andAVFs requiresthese specialized skills.98_These

� neuroendovascularproceduresaretechnicallyandcognitively

demandingandshouldonlybeperformedbyphysicianswith formal and specific training (or equivalent experience) in neurointerventionaltherapy,working incoordinationwitha multidisciplinary team.144,145 _{Specific pathways and guide­} linesforthetrainingandcredentialingofaneuroendovascular specialistandforcarotidstentinghavebeenpublished.164–167_In allcases,BACrecommendsthattheneuroendovascularspecial­ ists receive specific formal training and accrue significant experienceinaprocedurebecausepaststudieshaveshownthat moreexperienceandanincreasingnumberofproceduresreduce complicationrates.168

Itisrecommendedthat amultidisciplinary teamevaluate patientsbeforeandaftersomeoftheendovascularprocedures outlined above to discuss treatment options and assess for complications during and after the intervention.169 _{This is} important because of the emerging and changing role of endovascularapproachestodiseasetreatmentbutshouldalso beappliedtosomevascular surgicalprocedures.Othervital teammembersincludenursesandtechnologistswithtraining inendovascularprocedures.Aregistryshouldbeestablished to track treatments, outcomes, and complications. Yearly comparisonsshouldbemadebetweentheCSCcomplication rateandratesreportedfromnationalsurveysandguidelines. For all of the endovascular andsurgical procedures per­ formed at a CSC, the number, indications, and outcomes should be recorded and available for review. A quality assuranceprocessshouldconfirmthatproceduresandthera­ piesareperformedforappropriateindications,withratesof success and complications that meet acceptable standards. Thecommitteeshoulddefinealistofappropriateindicators thatwouldtriggerautomaticchartreview.Whencasereviews find significant deviations in the standards of care, the committeeshouldrecommend correctiveactionthroughap­ propriatemethods.

NeurosurgeryandVascular Surgery

Somepatients withan ICH,SAH,large ischemicstroke, or significantcarotiddiseasewillrequiresurgicalinterventions (Table 3). Operative proceduresmayinclude evacuationof intracerebralhematoma,clippingofanintracranialaneurysm, excision of an AVM, placement of a ventriculostomy for drainageofcerebrospinal fluid,a CEA, decompressivecra­ niotomy,andother procedures.97,156,170 –172_{Themedical evi­} dence insupport of these recommendations variesbetween levelsIIBandVCdependingonthespecificclinicalscenario, especiallyinthecaseofICH.170,173,174_{Nonetheless,thereare} severalsurgicalproceduresimportantforthemanagementof strokepatientsthatonlyaneurosurgeoncanperformsuchas insertion of a ventriculostomy, clipping of an intracranial aneurysm,excisionofanAVM,orremovalofanintracere­ bral hematoma. Therefore, neurosurgicalexpertise mustbe availableinaCSCona24/7basis.Theattendingneurosur­ geons at a CSC should have expertise and experience in cerebrovascularsurgery.Atalltimes,theremustbepersonnel in-house(orabletobeatthehospitalwithin30minutes)who arecapableofperformingemergentneurosurgicalprocedures and treatinglife-threatening intracranial conditions such as increased intracranial pressure, mass effect from a hemor­

rhage,etc.Suchpersonnelmayincluderesidentsorfellowsin anapprovedneurosurgery residencytraining program.Such trainees must have attending-level back-up, available on a 24/7 basis, by a board-eligible or board-certifiedneurosur­ geon.Writtenneurosurgicalcallschedulesmustbeavailable intheEDandhospitalwardsofaCSC.

A CSCmust haveindividuals with expertise andexperi­ ence in microsurgery for aneurysm clipping and surgical excision of AVMs. A study of in-hospital deaths after craniotomies performed for unrupturedaneurysms between 1987and1993inNewYorkstatehospitalsrevealeda53% decreasein mortalityrate inthe21 hospitalsinwhich>10 craniotomiesperyearwereperformed,comparedwiththe89 hospitals in which 10 craniotomies per year were per­ formed (5.3% versus 11.2%mortality rate,respectively.)175 The range for mortality and morbidity rates for surgical clippingofunrupturedintracranialaneurysmsisquitebroad, varying from 0% to 7% for death, and 4% to 15.3% for complications.176_{Ameta-analysisof2460patientsreporteda} mortalityrateof2.6%andamorbidityrateof10.9%.177_An international study of clipping of unruptured aneurysms reportsa1-yearmortalityof3.2%,a5.8%rateofmoderateor severe disability, and a 6% rate of isolated cognitive impairment.178

ForpatientswithSAHattributabletorupturedintracranial aneurysms,hospital volumeisalsostronglyassociatedwith outcomes.AstudyinNewYorkstatefounda43%reduction in mortality in hospitals that operated on >30 SAH/IA patientsper year.175_{Anotherstudy foundthat hospitalsthat} caredfor:21patientsperyearwithaneurysmalSAHhada significantly lower mortality rate (32% versus 49%) and reduced rate of adverseoutcomes (56%versus 76%)com­ paredwithhospitalswithlowervolumes.179_{Arecentstudyof} 16399 hospitalizations for SAH in 18 states found that hospitals caring for :19 patients per year with SAH had reduced mortality compared with lower volume centers.180 Anotherrecentstudybasedondatafrom1995to1999found that hospitals with very low patient volumes for cerebral aneurysm clipping had higher mortality rates than very high-volume hospitals for emergency and elective surgeries.181

Based on these data, it is recommended that for an institutiontobeconsideredaCSC,theinstitutionshouldcare for:20SAHpatientsperyearandshouldaccomplish>10 craniotomiesperyearforaneurysmclipping(gradeIA).This doesnotpreclude lower-volumecentersfromhavingexcel­ lentoutcomes,nordoesitguaranteethathigh-volumecenters willachieveexcellentresults.182,183_{Eachcentershouldmon­} itoritsperioperativecomplicationratesandoveralloutcomes forcomparisonwithnationalbenchmarksaftercorrectingfor variouscomorbidities.Eachneurosurgeonshouldparticipate in:10suchcasesperyear.Theperioperativemortalityrate for aneurysm clipping at a CSC should be documented, reviewed, andcompared with published outcomes. Forthe treatment of AVMs, individual expertise is of paramount importance in treating these complex lesions. The CSC shouldhavethe capability andexpertise toprovidethe full spectrumof treatmentoptions required forthe treatmentof

�

TABLE4. ComplicationRatesforCEA Infrastructure

Asymptomatic Symptomatic EMS,ED,Referral,andTriage

Perioperativemortality* <0.2% <0.7%

Perioperativestrokeanddeath* _<3% _<6%

*Perioperativereferstowithin30daysofsurgery(seetextforreferences).

AVMs, includingmicrosurgical excision, endovascular em­ bolization,andstereotacticradiosurgery.156

CEA can reduce the risk of stroke in patients with symptomaticandasymptomaticcarotidarterystenosis(grade IA).184 –186 _{A CEA is typically performed by a vascular} surgeon or a neurosurgeon. Documentation of expertise in thisoperation is criticalto ensuring its efficacy. Published recommendations for perioperative complication rates for CEAshouldbeusedasbenchmarks(Table4).171,187_Results of CSC neurosurgeons and vascular surgeons should be auditedonayearlybasis,andtheresultsofarollingaverage

>3years shouldbe compared with published outcomeand complicationrates.

Revascularization procedures and microvascular tech­ niques are important options for the treatment of some ischemic disorders, and in the management of complex intracranial aneurysms. A large, prospective, randomized studyof EC/IC by-pass found that thissurgery was not of benefitforthemanagementofcarotidocclusion/stenosis,or middle cerebral artery occlusion/stenosis (grade I A).188 Other neurosurgical procedures may have benefit in the management ofselect patients. For example,microsurgical bypass procedures are an important treatment option fora selectbutdiversegroupofotherdisorders,includingmoya­ moya disease, aneurysms requiring sacrifice of the parent vessel, and tumors needing vessel occlusion (grade IV C).189–194,199,200–202 _{There has been renewed interest in these} proceduresforpatientswithsymptomaticcarotidocclusion.191 Recentstudieshaveindicatedthataselectgroupofpatientswith carotidocclusioninwhomthereisadocumentedalterationof cerebrovascularreserve, such asshownby anincrease inthe oxygenextractionfraction,haveanincreasedriskofsubsequent stroke.196–198_{AnNIH-sponsoredmulticenterstudyiscurrently} underwaytodetermineifsuperficialtemporaltomiddlecerebral arterybypass iseffectiveinreducingtheriskofstrokeinthis select group ofpatients (http://www.clinicaltrials.gov/ct/show/ NCT00029146).195_{Basedonalloftheseconsiderations,andthe} limited treatment options for some of these patients, it is recommended thata CSC have theability and equipment to performrevascularizationproceduresandmicrovascularsurgery (grade III C–IV C). Perioperative complications and graft patencyratesshouldbetrackedprospectively.

Neurosurgeonsareneededforprocedurestodiagnoseand treatincreasedintracranial pressure,includingplacementof anintracranialpressuretransducer,placementofaventricu­ lostomy, and performance ofa decompressivecraniectomy (grade IIB).170,202–207 _{At a CSC, an attending neurosurgeon} (orsenior-levelresident)mustbeavailablewithin30minutes forsuchprocedureson a 24/7 basis.Periprocedurecompli­ cationratesshouldbetrackedandreviewed,andanysignif­ icantdeviations should be addressed at regular qualityim­ provement/qualityassessment(QI/QA)meetings.

IntegrationofacutestrokecarebeginswiththeEMSsystem and extends to the ED. Many of these components were reviewedinthePSCpublication.3_{Writtencareprotocolsfor} acute stroke patients should be available to EMS and ED personnel.Suchprotocolsshouldbereviewedandrevisedat least annually. EMS areas that should be included in such protocolsinclude(1)rapid,efficient patientassessmentand triage; (2) prehospital EMS communication with hospital staff;and(3)medicalstabilizationenroute.Systemsshould beinplacetoallowforrapidcommunicationbetweenEMS andED personnel during the transportation ofacute stroke patients.208,209 _{The ED should have well-defined and} -documented procedures for calling the acute stroke team, including a call schedule. The ED should have a door-to­ needletimeof 60 minutesfortheadministration ofintra­ venous tissue plasminogen activator (tPA) to stroke pa­ tients.3,210 _{The EMS and ED staff should meetand review} patientcareissueswiththeCSCstaffatleasttwiceperyear. Atleast2specificassessmentcriteriaandbenchmarksrelated to acute stroke care should be defined, measured, and reviewedannually.ItisrecommendedthattheEMSandED physicians, nurses, and paramedics attend :2 in-service programs(orequivalent educationalprogram)annuallythat focusonacutestrokecare.

TheCSCshouldbeviewedasa communityandregional resourceinthemanagementofstrokepatients.EMSandED personnel, along with members of the stroke team, should playanactiveroleinthetriageofacutestrokepatients.This might include advice about diagnostic procedures, acute therapies,andreferraltoanappropriatefacilityiftransferis required or requested. For example, personnel at a CSC shouldbeavailabletohealthcarepersonnelatotherhospitals andprovidethemwithguidanceandrecommendationsabout thediagnosisandtreatmentofspecificpatients.Inanemer­ gencysetting,suchguidancemightalsoincludethereviewof radiologic studies via teleradiology techniques as well as adviceabouttheuseofacutetherapiessuchastPA.ACSC mightalsocoordinateacutecarewithinageographicregion to ensure that patients are transferred appropriately in a timely manner to the facility best suited to care for them. Several examples exist of successful regional acute triage systemsforstrokepatients.211–214_{Wheneversuchcommuni­} cations occur, CSC personnel should not be legally liable, assumingthattheyprovideprudentadvicethatisconsistent withcommunitymedicalstandards.

StrokeUnitandICU

The PSC recommendations include a discussion about the importanceofastrokeunitifpatientsaregoingtobeadmitted tothehospital(gradeIA).3_{Morerecentstudieshavegener­} ated additional data supporting the efficacy and cost-effectivenessofstrokeunits.215,216_{Arecentstudyfoundthat} stroke patients cared for in a stroke unit with continuous cardiac telemetry monitoring were more likely to have significant cardiac arrhythmias detected compared with strokeunitpatientswhodidnothavesuchmonitoring.217_This observationissharedbymanyvascularneurologists.

There-fore, we recommend that a stroke unit include continuous ated to prevent secondary strokes and other vascular cardiac andrespiratory monitoring whendeemed clinically events.230–233_{Frequentcommunicationsbetween andamong} appropriate. the care team, the patient, and their family will alleviate

A CSCmusthavea fullICUbecause somepatients ata anxietyandimproveplanningforpoststrokecare. CSC (ie, those with large ICHs or SAHs) will require the

servicesofa typicalICU(ie,intubation, ventilatorsupport, peripheralarteryandpulmonaryarterycatheters,ventriculos­ tomies,anduseofparenteralvasoactive medications;Table 3).218–221_{AdedicatedneurosciencesICU,althoughdesirable,} is not required. The stroke unit may be part of the ICU, althoughthismaynot be an efficient use ofICUbeds and nursingresourcesinsome hospitalsbecausemanyischemic strokepatientsdonot requirethecostly infrastructureofan ICU.

The ICUin aCSC shouldbe staffedby physicianswith trainingincerebrovasculardiseaseandcriticalcare,although theymay be different persons. Training in cerebrovascular diseasehasbeendefinedpreviously.Thecriticalcarephysi­ ciansmusthavecompletedanaccreditedcriticalcarefellow­ shipprogram.TheICUdirector shouldalsohave:8hours peryearofCME training(orequivalent educationalactivi­ ties)relatedtocerebrovasculardisease.Formaltrainingasa neurointensivist, although preferred, is not a requirement. Coveragebyattendingphysiciansorresidentswithexpertise incerebrovasculardiseaseandcriticalcaremustbeavailable 24/7withawrittencallschedule.Suchcoveragemayconsist ofa teamapproach, with some membersfrom criticalcare medicineoranesthesiologyandothersfromneurosurgeryor neurology.

The nurse:patientratio in an ICUcaringforcritically ill strokepatientsshouldbe 1:1or1:2.Itispreferred(butnot required)that nursescaringforneurosciencepatients inthe ICU be board certified in neuroscience nursing. The ICU nursingstaff mustbe trainedto assess neurologic function andtodealwithseveralaspectsofneurocriticalcare,includ­ ing(1)functionofventriculostomyandexternalventricular drainage apparatus, (2) treatment of increased intracranial pressure,(3)careofpatientswithICHandSAH,(4)careof patients after thrombolytic therapy, (5) treatment of blood pressureabnormalitieswithparenteralvasoactiveagents,(6) managementofintubated/ventilatedpatients,and(7)detailed neurologicassessmentsandscales(ie,NIHStrokeScaleand GlascowComaScale).Suchtrainingcanbedocumentedby attendanceatin-servicesessions,participationinregionalor nationalcourses, andother modalities asestablished bythe CSCstaffandhospitaladministration.Itisrecommendedthat nursesintheICUcaringforstrokepatientsreceive:10hours per year ofCEU credit (or other educational programs) in areasrelatedtocerebrovasculardisease.

Allstrokepatientsshouldbefullyevaluatedtodetermine theetiologyoftheirstroke,whichisvitalforplanningtheir treatmentanddecidingonapproachesforsecondarypreven­ tion.222_{Whilethestrokepatientishospitalized,stepsshould} betakentoreducetheriskofperistrokecomplicationssuchas cerebraledema,aspirationpneumonia,infection,myocardial infarction, and deep venous thrombosis (DVT).21,223–225 _A multidisciplinarycarepathwayisveryusefultoensureproper diagnostic and preventive measures are taken in all pa­ tients.226 –229 _{Medicalandsurgical therapies shouldbe}

initi-RehabilitationandPoststrokeCare

Rehabilitation after a stroke is a keyelement for patients. Thereare 6 areas offocusfor poststrokerehabilitation:(1) trainingformaximumrecovery,(2)preventandtreatcomor­ bidconditions,(3)enhancepsychosocialcoping,(4)promote integrationintothecommunity,(5)preventrecurrentstrokes andother vascularevents,and(6)enhancequalityoflife.18 Rehabilitation of stroke survivors should begin as soon as possible.Publishedstudieshavedemonstratedthatorganized multidisciplinarystrokerehabilitationreducesdeath,deathor disability,anddeathorinstitutionalization(gradeIA).17,18,229,234 –239 Mobilizationofthe strokesurvivor andresumption of self-care activities should occur as soon as medically feasible. Rehabilitationmayincreasethestrokepatient’squalityoflife and reduce the financial and physical burden on society (grade IA).236,240,241 _{In addition to inpatient rehabilitation,} outpatientrehabilitationprogramscanimproveoutcomesand preventdeterioration(gradeIIB).242

Rehabilitationservicesshould be directedbya physician withboardcertificationinphysical medicineandrehabilita­ tion(ie,physiatrist)orbyother properlytrainedindividuals (ie,neurologistexperiencedinstrokerehabilitationor other physiciansorPhDswithfellowshiptraininginrehabilitation). Alltherapists,socialworkers,andnursecasemanagersmust meetrequirements forstatelicensure,andhave :1yearof experienceinthetreatmentofstrokesurvivors.Thephysical therapists, speechtherapists, andsocial workersmust com­ pleteamaster’sdegree,whereastheoccupationaltherapists andnursecasemanagersmustcompleteatleastabachelor’s degree.The nurse case managers andsocial workers must haveadequateknowledgeofinpatientrehabilitationfacilities andcommunityresourcesintheirgeographicregions.

A CSC should have physical, occupational, and speech therapists on staff or readily available by consultation for patientassessment andtherapyduringtheacutehospitaliza­ tion.Consultsforphysicalmedicineandrehabilitation,phys­ icaltherapy,occupationaltherapy,andspeechtherapyshould berequestedandcompleted(whendeemedmedicallyneces­ sary) usually within 24hours ofadmission sothat medical andtherapeuticevaluationsmaybeginassoonasthestroke survivor is medically stable. If the CSC does not have inpatient rehabilitation facilities on site, they should have documentedreferralprotocols andaworking knowledgeof nearbyfacilities.

Otherimportantmeasuresincludestepstopreventandtreat medicalcomplicationsofstrokesuchasaspirationpneumo­ nia and other infections, cerebral edema/herniation, DVT, pressure sores, and contractures.172,223,224 _{Poststroke care} shouldincludeassessmentandtreatment(whenpossible)of cognitive decline, depression, and social implications of stroke.225 _{These various therapies have been reviewed in} otherrecentpublicationsandarenotreiteratedhere.18

Education

Educationalprogramsare deemeda very importantcompo­ nentof a CSC. Suchprograms canbe divided into

profes-sional and public efforts. For professional programs, it is recommended that the CSC staff prepare and present :2 educationalcoursesperyearaimedathealthcareprofession­ alswithinoroutsideoftheCSC.

Public education is a vital component that can improve acutecarebydecreasingdelaytimesforpresentation(grade IA).243,244_{Publiceducationaboutstrokeriskfactorsmayalso} facilitate improved therapy (grade IIB).245–247 _{Other public} education programs related to vascular disease risk factors mayalsobeuseful,althoughresultsofrandomizedtrialshave beenmixed.248–251_{ItisrecommendedthataCSCsponsorat} least2 publiceducationalactivitieseach yearthatfocuson someaspectofstroke.Thesecouldincludelectures,screen­ ing forstroke risk factors, health fairs, andsimilar events. Suchevents could be advertisedand marketedto high-risk groupsinsuchareasasthosewithahighminoritypopulation, areaswithpooraccesstohealthcare,andotherunderserved populations.

Afellowshipincerebrovasculardiseaseaccomplishessev­ eralimportantgoals,suchasprovidingadditionaltrainingand experienceandpreparingphysiciansforacareerthatfocuses on cerebrovascular disease. The Accreditation Council for Graduate Medical Education (ACGME) has recently ap­ provedtheformation ofan officially recognizedfellowship programinvascularneurology.ACSCdoesnotneedtohave afellowshipprogram toprovideexcellentpatientcare.Itis recommendedthataCSCthatchoosestoofferafellowshipin cerebrovasculardiseasefollowtheguidelinesoftheACGME andworkcloselywiththeAmericanAcademyofNeurology indeveloping atrainingcurriculum.

Research

Researchprogramsareanimportantcomponentofacademic medical centers, and some community hospitals are also becominginvolvedinclinicalresearchtrials.252,253_Paststud­ ies have documented that patients involved in clinical re­ search have better outcomes than patients not involved in such research (grade IIIC).254–257 _{However, a hospital can} clearlyprovideexcellentcareasaCSCandnotbeinvolved inanyresearch.Therefore,researchisconsideredanoptional componentofaCSC.

Other

AstrokeregistryisanimportantelementofaCSC.Astroke registry is a systematic collection of data that deals with strokecare,riskfactors, outcomes,andrelatedissues.258–262 Such a registry is important for tracking outcomes and definingareasinneedofimprovementandisincludedinthe recommendationsfora PSC.Itisrecommendedthat aCSC haveastrokeregistryoranothersimilardatacollectiontool (gradeIIIC).Apilotprojectfordeveloping anationalacute strokeregistryiscurrentlyunderway.263

Astrokecliniccouldhaveamultidisciplinaryfacultythat wouldaddressseveral componentsofstrokecare,including prevention,rehabilitation,medicalcomplications,andsocial issues.ItisrecognizedthatsomeCSCfacilitiesmaynothave the space fora multidisciplinary stroke clinic orthat other logisticalfactorscouldlimittheformationofsuchaclinic.In addition,suchcarecouldbeprovidedinanotherclinicsetting

such asa multidisciplinary vascular clinic. For thisreason, thestrokeclinicisanoptionalcomponentofa CSC.

Anairambulanceisanimportanttransportationoptionfor somehospitalsbecauseofspecificgeographicconsiderations. Inlargecities,suchaservicemayreducetransportationtimes because of high traffic volume on congested highways. In rural locations, an air ambulance may be a vital service offered by a CSC to transport acute stroke patients from outlyingfacilities.264_{Studieshaveshownthatsuchaservice} canreducetreatmenttimesforthrombolytictherapyinacute stroke (grade IIB).212,265 _{However, a CSC in some urban} settingscan certainlyfunction well withoutsuch a service. Therefore, an air ambulance is an important but optional featureofaCSC, dependinglargelyon localtransportation, referral,andgeographicfactors.

Quality assurance and improvement are important pro­ cessesfor maintaining and enhancingthe qualityof health careatanytypeofstrokecenter.3_{Thisisparticularlytruein} aCSC,inwhichveryillpatientswithcomplexdiseasesare sometimes treated with potentially dangerous medications and procedures. Because of the serious consequences of strokeanditsrelatedcomplications,aswellasthecomplexity ofsurgicalandendovascularproceduresperformedataCSC, itisessentialthatoutcomesbemonitored.Amultidisciplinary institutional qualityassurance committee should meeton a regular basis to monitor quality benchmarks and review complications.Thiscommitteeshouldincludememberswho participateinthecareofstrokepatientsaswellasotherswho arenotparticipatingdirectlyinsuchcarebutcouldprovide insight into overall patient care. The goal of this quality assurance program is quality improvement, correction of errors,andsystemsimprovement.

In addition to monitoring the outcomes of procedures performedataCSC,thequalityassurancecommitteeshould also monitor the overall care of patients. A database or registry should be established that allows for tracking of parameters such as length of stay, treatments received, dischargedestinationandstatus,incidenceofcomplications (such as aspiration pneumonia, urinary tract infection, and DVT),anddischargemedications. Specificbenchmarksand indicatorsshouldbesetandreviewedona regularbasis(at least annually). There are several multi-institutional or na­ tional databases that can be used to establish appropriate benchmarks.

Discussion

Thediagnosisandtreatmentofsomepatientswith complex typesofstrokeorwith severedeficitsandmultiorganprob­ lemsmay require more resourcesand a higher intensity of carethanisofferedinmanyhospitalsandinmostPSCs.Such patientsoften requireandmaybenefitfromadvanceddiag­ nosticandtreatmentproceduresdirectedbyspecialty-trained physiciansandotherhealthcareprofessionals.ACSCwould havethe staffing, expertise,infrastructure, andprograms to meettheneedsofthesepatientsandtoserveasaresourceto thePSCsintheregion.

Manyoftherecommendationsoutlinedabovearemeantto benefitthetypesofpatientslikelytobecaredforinaCSC, namely those with hemorrhagic strokes, large ischemic