RESEARCHARTICLE Efficacy and Safety of Transcutaneous Electrical Acupoint Stimulation to Treat Muscle Spasticity following Brain Injury: A Double-Blinded, Multicenter, Randomized Controlled Trial WenliZhao1‡,ChaoWang2‡,ZhongzhengLi2,LeiChen3,JianboLi3,WeidongCui3, ShashaDing2,QiangXi2,FanWang2,FeiJia2,ShuhuaXiao2,YiGuo2,YeZhao4* 1 DepartmentofNeurology,TianjinNankaiHospital,Tianjin,300100,China,2 DepartmentofAcupuncture andMoxibustion,TianjinUniversityofTraditionalChineseMedicine,Tianjin,300193,China,3 Departmentof AcupunctureandMoxibustion,TianjinNingheHospital,Tianjin,301500,China,4 DepartmentofClinical Research,TianjinNankaiHospitalandTianjinAcademyofIntegrativeMedicine,Tianjin,300100,China ‡Theseauthorscontributedequallytothiswork. * [email protected] OPENACCESS Citation:ZhaoW,WangC,LiZ,ChenL,LiJ,CuiW, etal.(2015)EfficacyandSafetyofTranscutaneous Abstract ElectricalAcupointStimulationtoTreatMuscle SpasticityfollowingBrainInjury:ADouble-Blinded, Multicenter,RandomizedControlledTrial.PLoSONE Objective 10(2):e0116976.doi:10.1371/journal.pone.0116976 Thisstudywasaimedatevaluatingtheclinicalefficacyandsafetyoftranscutaneouselectri- AcademicEditor:MicheleSterling,Griffith calacupointstimulation(TEAS)totreatmusclespasticityafterbraininjury(ChineseClinical University,AUSTRALIA TrialRegistry:ChiCTR-TRC-11001310). Received:August13,2014 Methods Accepted:December7,2014 Published:February2,2015 Atotalof60patientswithmusclespasticityafterbraininjurywererandomizedtothefollow- ing3groups:100,2,and0Hz(sham)TEAS.TheacupointsHegu(LI4)—Yuji(LU10)and Copyright:©2015Zhaoetal.Thisisanopen accessarticledistributedunderthetermsofthe Zusanli(ST36)—Chengshan(BL57)ontheinjuredsidewerestimulatedat0,2,or100Hz, CreativeCommonsAttributionLicense,whichpermits 5timesperweekfor4weeks.Thepatientswerefollowedupfor1and2monthsafterthe unrestricteduse,distribution,andreproductioninany treatments.Theeffectsofthetreatmentsonmusclespasticityatthewrist,thumb,theother medium,providedtheoriginalauthorandsourceare 4fingers,elbow,shoulder,knee,andanklewereevaluatedbytheModifiedAshworth credited. Scale,andtheeffectsondisabilitywereassessedbytheDisabilityAssessmentScale.The DataAvailabilityStatement:Allrelevantdataare walkingcapabilitywasevaluatedbytheHoldenfunctionalambulationclassificationscore. withinthepaperanditsSupportingInformationfiles. TheoverallperformancewasassessedbytheGlobalAssessmentScalescoreandtheim- Funding:Author:ChaoWang.Thisworkwas provedBarthelIndex.Thesafetyofthetreatmentsadministeredwasalsomonitored. supportedbytheScienceFoundationonTraditional ChineseMedicine(TCM)/IntegrativeMedicineofthe TianjinAdministrationofTCM(grantno.11031).The Results fundershadnoroleinstudydesign,datacollection Thewristspasticitywassignificantlyreducedfrombaselineatweeks2,3,and4oftreat- andanalysis,decisiontopublish,orpreparationof themanuscript. mentandatthe1-and2-monthfollow-upvisitsinthe100Hzgroup(P<0.01).Compared with2HzorshamTEAS,100HzTEASdecreasedwristspasticityatweeks2,3,and4of CompetingInterests:Theauthorshavedeclared thatnocompetinginterestsexist. treatmentand1monthaftertreatment(P<0.001).Theotherendpointswerenotaffected PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 1/15 TEASTreatsMuscleSpasticity bythetreatments.Notreatment-emergentadverseeventswerereportedduringtreatments andfollow-upvisits. Conclusions TEASappearstobeasafeandeffectivetherapytorelievemusclespasticityafterbraininju- ry,althoughlarge-scalestudiesarerequiredtofurtherverifythefindings. TrialRegistration ChineseClinicalTrialRegistryChiCTR-TRC-11001310http://www.chictr.org INTRODUCTION Spasticity—characterizedbyexaggeratedtendonjerksandclonus—isacommoncomplication followingbrainorspinalcorddamages,andadverselyaffectsworkandqualityoflifeofpa- tients[1–4].Theexactpathophysiologyofspasticityremainsunclear.Excessiveactivityofthe alphamotorneuron(-MN)poolcausedbytheimbalancebetweeninhibitoryandexcitatoryef- fectsisthoughttocontributetospasticityafterbraininjury[5].Brainandspinalcordinjuries frequentlydamagethedorsalreticulospinaltract,whichmediatesinhibitoryeffectsand,thus, consequentlyleadtotheexaggerationofexcitatoryinputsfromthemedialreticulospinaland vestibulospinaltractsand,ultimately,spasticity[6]. Conventionaltherapiesforspasticityincludingphysicaltherapy,surgery,andpharmaco- therapyoftenproduceunsatisfactoryoutcomes.Theoutcomesofsurgicalinterventionvary, dependinguponthelocationandtheseverityofspasticity[7],whereastheeffectsofpharmaco- logictreatmentsareoftenshorttermformanypatientsandseveresideeffectsoccursometimes [8–10].Transcutaneouselectricalnervestimulation(TENS)hasalsobeenshowntoimprove clinical,electrophysiological,andfunctionalvariablesinpatientswithspasticitywithasimilar efficacyasbaclofen[11].SimilartoTENS,transcutaneouselectricalacupointstimulation (TEAS)isalsoanelectricalstimulationmethodinwhichacupointsspecificallyassociatedwith medicalconditionsarestimulated. TEAShas,recently,beenemergingasapopulartherapeuticapproachandhasbeenwidely testedinmanymedicalconditions,suchaspain,gastrointestinaldisorder,inflammation,can- cer,andAlzheimer[12–17].Althoughcontroversialresultshavebeenreported,TEASappears tobeincreasinglyusedinclinicalpracticenow[18–20].ThemechanismofactionofTEAS mightbesimilartothatofelectricalacupuncture(EA),inwhichacupointsarestimulatedby electricalimpulsesgiventhroughneedles.StudiesonpainmanagementusingEAinanimal modelsandpatientssuggestthatEAblockspainbyactivatingabroadspectrumofbioactive chemicalssuchasopioids,serotonin,andnorepinephrinetodesensitizeperipheralnociceptors andreduceproinflammatorycytokinesperipherallyandinthespinalcord[20,21]. StudiesontherapeuticapplicationofTEASinthemanagementofspasticityaresparse,but theresultsfromthosestudiesconsistentlyshowthebeneficialeffectsofTEASonpatientswith spasticityfollowingspinalcordinjuryorstroke[22–26].Yanetal.conductedarandomized controlledtrialtocomparetheeffectsofTEAS,placebostimulation,andstandardrehabilita- tionaloneonmusclefunctionafteracutestrokeandfoundthatTEASsignificantlyincreased theproportionofpatientswithimprovedmusclefunction[22].Wangetal.evaluatedtheeffi- cacyof0,2,and100HzTEASonspasticityafterspinalcordinjuryandfound100HzTEASre- lievedspasticitymoreefficientlyandeffectivelythan2HzTEASinpatients[23].Althoughthe PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 2/15 TEASTreatsMuscleSpasticity resultsareconsistent,mostofthestudiesexceptthereportbyYanetal.arenotrandomized controlledtrials,significantlycompromisingthequalityofthedata.Here,weconductedaran- domizedcontrolledtrialtoevaluatetheclinicalefficacyandsafetyofTEASandcomparetheef- ficacyofhighversuslowTEASfrequencyonpatientswithspasticityfollowingbraininjury. MATERIALSANDMETHOD Theprotocolforthistrial,supportingCONSORTchecklist,andtheoriginaldataoftheclinical trialareavailableassupportinginformation;seeS1Protocol,S1CONSORTChecklist,andS1 Table. Trialdesign Thisdouble-blinded,multicenter,randomizedcontrolledtrialwasconductedat6clinicalcen- tersinthecityofTianjinanditssurroundingcountiesbetweenMarchandSeptember2011. ThistrialwasconductedinaccordancewiththeprinciplesoftheDeclarationofHelsinki (VersionEdinburgh2000)andapprovedbytheChineseEthicsCommitteeofRegisteringClin- icalTrial.ThestudywasregisteredattheChineseClinicalTrialRegistry(http://www.chictr. org/cn,UniqueIdentifier:ChiCTR-TRC-11001310).Thestudyprotocolwasapprovedbythe InstitutionalEthicsCommitteeofTianjinUniversityofTraditionalChineseMedicine (TJUTCM-EC20110001).Writteninformedconsentwasobtainedfromallpatients.TheCON- SORTflowdiagramisdisplayedinFig.1.ThetrialdesignisillustratedinFig.2. Figure1.TheCONSORTflowdiagram. doi:10.1371/journal.pone.0116976.g001 PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 3/15 TEASTreatsMuscleSpasticity Figure2.Studydesign. doi:10.1371/journal.pone.0116976.g002 Patients Eligiblepatientswereadults18–85yearsofage(bothyearsinclusive)whodidnotreceivepre- viousTENSorTEAStreatmentsandhad:experiencedatleast1episodeofbraininjurysuchas cerebralhemorrhage,cerebralinfarction,orbraintrauma(confirmedbyexaminingpatients’ medicalrecordsandcomputedtomographyormagneticresonanceimagingofthebrain)at least3monthsbeforestudyenrollment;increasedmuscletoneoftheflexorcarpiontheaffect- edsidewithModifiedAshworthScale(MAS)score(cid:1)3;anddisabilityscore(cid:1)2accordingto theDisabilityAssessmentScale(DAS).Patientswiththefollowingconditionswereexcluded fromthestudy:severecontractureatthejointsofaffectedlimbs(inabilitytopassivelymove thejointby>10°),previousAchillestendonlengtheningortransplantationsurgery,previous nerveresectionsurgery,injectionofbotulinumtoxinintothetargetlimbwithin4monthsbe- foreenrollment,localneuromuscularblockingtherapyororalantispasmodicmedicationwith- in2weeksbeforeenrollment,severemuscleatrophyorinfectionintargetlimbs,pregnantor planningtobecomepregnantduringthecourseofthestudy,orsevereconditionsinotheror- ganssuchassevereliverandrenaldysfunctionorwearingacardiacpacemaker. Randomization,allocation,andblinding Acentralrandomizationschemewasusedtoassigneligiblepatientsto3groups—100,2,or0 HzTEAS(sham)—ina1:1:1ratioattheTianjinUniversityofTraditionalChineseMedicine. Apermuted-blockrandomizationprocedurewasusedtogenerateanallocationsequence.The blocksizeforrandomizationwas6.Sealedopaqueenvelopescontainingthepatientnumber andtheallocatedtreatmentwereprepared.Theenvelopesweresequentiallynumbered.Several levelsofblindingstrategywereimplementedtominimizepossiblebias.Thepersonnelwhore- cruitedpatients,professionalassessorswhoevaluatedpatientperformance,andthestatistician wereblindedtotreatmentallocation.Thepatientstoo,werecompletelyblindedtothetreat- mentadministered.Attherecruitmentinterview,theparticipantswereinformedthatthey weretoundergo1of3newtherapiesformusclespasticity.Thefollowingdetailedexplanations ofthenewtherapieswerealsoprovidedtotheparticipants:“Duringtreatment,acupuncture pointswillbestimulatedbymicro-currentsgeneratedbyaspecialelectricaldevice.Themagni- tudeofthemicro-currentsissolowthatapersonmightexperienceonlymildelectriccurrent sensationorfeelnothingatall.”Topreventanypossiblepsychologicalinterferenceonthere- sults,thepatientswereallowedtowatchtheoperationoftheelectricalstimulatorandbelieved PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 4/15 TEASTreatsMuscleSpasticity thattheyreceivedrealtreatments.IntheshamTEASgroup,theelectrodeswereplacedonthe sameacupointsandtreatedforthesamedurationastheother2groups.Tofurtherensurethe participantswereblindedfortreatmentsadministeredandtopreventanypossiblepsychologi- calinterference,ateachtreatmentsession,atrainednursewhowasblindedtotreatmentalloca- tionwaspresentandengagedaconversationwiththeparticipantstodistractandhelp themrelax. Astheacupuncturistshadtoadministertreatment,itwasimpossibletocompletelyblind themtopatients’informationsuchassymptoms.However,theacupuncturistswereforbidden tospeaktopatientsduringtreatments.Inthecasethatapatientexpressedconcernsorcom- plainsregardingtreatments,thepatientspoketothenurseandthenurseforwardedtheinfor- mationtotheacupuncturists,whowouldadjustthetreatments.AftersettinguptheTEAS instrumentandensuringtheinstrumentworkedproperly,theacupuncturistsleftthetreatment room.Theunblindingwasconductedin2steps.Thefirstdataunblindingwasperformedafter thetrialwascompletedandthedatabasewaslocked.Thestatisticianwasexposedtothedata organizedaccordingtotreatmentgroups,representedasA,B,andC.Theannotationofeach letterwasnotrevealedtothestatistician.Theseconddataunblindingwasperformedafterdata analyseswerecompleted.Alltheinformationonpatientsandtreatmentallocationwerethen disclosedtoinvestigatorsparticipatinginthetrial. Interventions Allpatientswereallowedtocontinuetheirroutinerehabilitationtraininganddietaryregimen duringthestudy.Alleligiblepatientswerealsoallowedtomaintaintheirexistingmedication regimenaslongasitdidnotviolatetheinclusioncriteria.However,theywererequiredtore- corddetailsoftheirmedicationregimenthroughoutthestudy.Atotalof2electricalcircuits werestimulatedamongacupointsusingaTEASinstrument—theacupointnerveelectrical stimulator(HANS-100A,NanjingGensunmedicaltechnologycompany,Nanjing,China).The locationoftheacupointsisillustratedinFig.3.Onepairofroundpercutaneouselectrodes (diameter,24mm)wasplacedontheacupointsHegu(LI4)–Yuji(LU10)ontheaffectedsideto induceanelectriccircuit.Theotherelectriccircuitwasstimulatedbyplacinganotherpairof electrodesontheacupointsZusanli(ST36)–Chengshan(BL57)onthesameside.These2pairs ofelectrodeswereconnectedtotheacupointnervestimulator.Theacupointswerestimulated at100or2Hzfor0.2ms.Frequencieswerechosenbasedonthereportsdescribingthethera- peuticapplicationandmolecularmechanismofTEASinspasticity[23,27–29].Totreatpa- tientsreceiving100or2HzTEAS,theacupuncturistsgraduallyincreasedtheintensityofthe Figure3.LocationoftheacupointsHegu(L14)–Yuji(LU10)andZusanli(ST36)–Chengshan(BL57). doi:10.1371/journal.pone.0116976.g003 PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 5/15 TEASTreatsMuscleSpasticity electricalstimulationtothelevel(usuallybetween20and40mA)thatrhythmiccontractionof musclesonhandsorlegswasvisible.Patientsusuallydidnotexperienceanypainordiscomfort atsuchalevelofintensity.Theaverageintensityoftheelectricalstimulationusedinthisstudy wasapproximately30mA.Eachtreatmentsessionlasted30minutes,andthepatientsreceived 5treatmentsessions(1treatmentsessionperday)everyweekfor4weeks.IntheshamTEAS group,theelectrodeswereplacedonthesameacupointsandtreatedforthesamedurationas theother2groups;however,theintensityofelectricstimulationwas0mA.Allpatientswereal- lowedtowatchtheoperationoftheelectricalstimulatorandtoviewtheentireproceduredur- ingtreatmentsessions,andbelievedtheyreceivedrealtreatments. Assessments Patientswererequiredtocompletethequestionnairesweeklyafterthefifthtreatmenteveryweek andatthefollow-upvisits1and2monthsafterthetreatment.Theseverityofmusclespasticity anddisability,walkingcapability,andoverallperformanceofdailyactivityofpatientswereas- sessedbyaprofessionallytrainedassessoratenrollment,afterthefifthtreatmenteveryweek,and ateachfollow-upvisit.MASwasusedtoassessmuscletone(0representsnoincreaseinmuscle tone;4representsthehighestincreaseinmuscletonewithrigidityinflexionorextension)[30]. DisabilitywasassessedbyDASonascaleof0–3,with0indicatingnormaland3indicatingse- veredisability[31].WalkingcapabilitywasevaluatedbytheHoldenfunctionalambulationclas- sificationscoreonascaleof1–5,withhigherscoresindicatingbetterwalkingcapability[32], whereasoverallperformancewasassessedbytheGlobalAssessmentScale(GAS)scoreatarange of1–100withhigherscoressuggestingabetteroverallleveloffunctioningandcarryingouractiv- itiesofdailyliving[33]andtheimprovedBarthelIndex(BI)scoreintherangeof1–100with higherscoresindicatingabetterperformanceinactivitiesofdailyliving[34,35].Wealsocom- paredthepercentageofpatientsversusthepercentageofcaregiversexpectingpost-treatmentim- provementindailyactivitiesthatwereadverselyinfluencedbydisability. Endpoints TheprimaryendpointsincludedtheMASscoreofthewrist,thumb,andtheother4fingers. ThesecondaryendpointsweretheMASscoreoftheelbow,shoulder,knee,andankle,theDAS scoreofdisability,Holdenfunctionalambulationclassificationscore,GASscore,andimproved BIscores.Adverseeventwasdefinedasanoccurrenceofanyofthefollowingevents:bleeding, hematoma,syncope,severepain,andlocalinfection.Adverseeventsweremonitoredduring thestudyanddocumentedduringtreatmentandatfollow-upvisits. Samplesizecalculationandstatisticalanalysis Basedonourpreviousstudy,todetectasignificantdifferenceatthelevelof0.05betweenthe groupswith90%powerattheassumptionofadrop-outrateof15%anda1-tailedα=0.05,20 patients(total60patients)wererequiredforeachtreatmentgroup[23].Datawerepresentedas mean±standarddeviation(SD).Countvariableswithnormaldistributionwereanalyzedby one-wayorrepeated-measuresANOVAtoexaminedifferencesandbyPearson’schi-squaretest todeterminecorrelation.Countvariableswithpartialdistributionwereanalyzedusingarank- sumorSpearmanrankcorrelationtest.Forbaselinegeneralclinicalcharacteristics,one-way ANOVAwasperformedoncontinuousvariablesincludingage,spasticityduration,height, weight,andbloodpressure,whereasPearson’schi-squaretestwasusedoncountvariablesin- cludinggender,braininjurytype,andcomorbiddisease.One-wayANOVAwasalsoperformed onthebaselineprimary(MASscoreofthewrist,thumb,andtheother4fingers)andsecondary endpoints(DASscore,Holdenfunctionalambulationclassificationscore,GASscore,and PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 6/15 TEASTreatsMuscleSpasticity improvedBIscores).ThecorrelationsamongbaselineendpointswereanalyzedbyPearson’schi- squaretest.Toassesstreatmentefficacy,dataoftheprimaryandsecondaryendpointswereex- aminedbyMauchly’ssphericitytestfirst,andthenrepeated-measuresANOVAwasperformed toanalyzethedata.Allstatisticalanalyseswereconductedontheintention-to-treat(ITT)popu- lation.Theprimaryendpointsfromper-protocol(PP)datasetswerealsoanalyzedusingthe samestatisticalanalyses.TheITTsetincludedalltheenrolledpatients(n=60);thePPsetinclud- edpatientsintheITTsetwhodidnotdeviatefromtheprotocol.Intotal,9patients(15%) droppedoutofthestudy,including8notreceivingtreatmentand1diedfromaninfectiouscom- plicationunrelatedtotreatments.Thus,thePPsetincluded51patients.P<0.05(2-sided)was consideredstatisticallysignificant.AllanalyseswereperformedusingtheSPSS18.0software. RESULTS Baselineclinicaldataaresimilaramongthetreatmentgroups Ofthe123patientsscreened,60wererandomized,and51completedthestudy.Thepatient flowchartisshowninFig.1.Themajorityofpatientsweremale;themeanageofthepatients includedinthestudywas>60yearsinall3groups.Thebaselinecharacteristicsofthepatients werecomparableacrossall3treatmentgroups(P>0.05;Table1).Additionally,thebaseline musclespasticityatthewrist,thumb,theother4fingers,elbow,shoulder,knee,andankleeval- uatedbytheMASscore,theseverityofdisabilityreflectedbyDASscore,walkingcapability evaluatedbyHolderfunctionalambulationclassification,andoverallperformancereflectedby theGASscoreandimprovedBIwerenotsignificantlydifferentamongthe3groups(P>0.05, Table2). Table1.Baselinecharacteristics. Characteristic 100Hzgroup 2Hzgroup Shamgroup Age(years) 62.00±9.20 63.50±9.29 62.45±8.44 Gender,n Male 15 16 15 Female 5 4 5 Braininjurytype,n Cerebralinfarction 15 15 17 Trauma 0 1 0 Hemorrhage 5 4 3 Comorbiddisease,n Yes 14 9 11 No 6 11 9 Spasticityduration(years) 5.758±5.174 4.717±5.091 3.495±3.527 Height(cm) 171.70±7.47 170.85±6.35 169.75±7.33 Weight(kg) 69.33±8.35 66.40±8.46 66.15±8.94 Systolicpressure(mmHg) 136.35±12.487 138.50±8.599 135.25±12.298 Diastolicpressure(mmHg) 90.20±8.383 88.50±6.091 86.25±9.301 Dataarepresentedasmean±standarddeviationunlessotherwiseindicated.One-wayANOVAwas performedoncontinuousvariablesincludingage,spasticityduration,height,weight,andbloodpressure; theChi-squaretestwasusedoncountvariablesincludinggender,braininjurytype,andcomorbiddisease. Allvariableswerenotsignificantlydifferentamongthe3groups(P>0.05). doi:10.1371/journal.pone.0116976.t001 PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 7/15 TEASTreatsMuscleSpasticity Table2.BaselineMASscoreofthewrist,thumb,theother4fingers,elbow,shoulder,knee,and ankle,DASscore,walkingcapability,andoverallperformance. 100Hzgroup 2Hzgroup Shamgroup MASscore Wrist 3.150±0.366 3.200±0.410 3.100±0.308 Thumb 2.400±0.868 2.550±0.742 2.450±0.686 Other4fingers 2.625±0.944 2.525±0.638 2.525±0.896 Elbow 2.900±0.837 2.575±0.950 2.450±0.985 Shoulder 1.725±1.019 1.525±0.819 1.725±0.678 Knee 2.200±0.938 2.175±0.766 1.800±0.571 Ankle 2.675±0.694 2.775±0.734 2.900±0.718 DASscore Personalhygiene 2.800±0.410 2.850±0.366 2.750±0.550 Dressing 2.750±0.444 2.850±0.366 2.750±0.444 Posture 2.750±0.550 2.750±0.444 2.650±0.587 Painordiscomfort 1.350±1.040 1.300±0.657 1.450±0.759 Holderfunctionalambulationclassification 3.050±1.432 2.900±1.917 2.800±1.542 GASscore 56.65±15.83 55.00±16.16 53.40±11.27 ImprovedBarthelIndex 65.15±23.94 64.85±26.78 68.60±22.32 Dataarepresentedasmean±standarddeviation.MAS=ModifiedAshworthScale.DAS=Disability AssessmentScale.GAS=GlobalAssessmentScale.One-wayANOVAwasperformedonthedata.Allthe variableswerenotsignificantlydifferentamongthe3groups(P>0.05). doi:10.1371/journal.pone.0116976.t002 Table3.Correlationofwristandhandfunctionwithactivitycapabilitybeforetreatment. Analyzedparameters Pearsoncoefficient Pvalue(2sides) Wrist(MASscore) Thumb(MASscore) 0.360 0.005** Wrist(MASscore) Other4fingers(MASscore) 0.227 0.081 Thumb(MASscore) Other4fingers(MASscore) 0.403 0.001** Thumb(MASscore) Personalhygiene(DASscore) 0.333 0.009** Other4fingers(MASscore) Personalhygiene(DASscore) 0.357 0.005** Other4fingers(MASscore) Posture(DASscore) 0.432 0.001** Thumb(MASscore) Dressing(DASscore) 0.165 0.208 Other4fingers(MASscore) Dressing(DASscore) 0.211 0.106 Other4fingers(MASscore) GAS -0.240 0.065 Wrist(MASscore) ImprovedBarthelIndex -0.226 0.083 Other4fingers(MASscore) ImprovedBarthelIndex -0.312 0.015* GAS ImprovedBarthelIndex 0.518 <0.0001** Knee(MASscore) Holdenfunctionalambulationclassification -0.170 0.193 Ankle(MASscore) Holdenfunctionalambulationclassification -0.119 0.364 Knee(MASscore) ImprovedBarthelIndex -0.201 0.124 Ankle(MASscore) ImprovedBarthelIndex -0.055 0.676 DatawereanalyzedbyPearson’schi-squaretest. *P0.05. **P0.01. MAS=ModifiedAshworthScale.DAS=DisabilityAssessmentScale.GAS=GlobalAssessmentScale. doi:10.1371/journal.pone.0116976.t003 PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 8/15 TEASTreatsMuscleSpasticity Wethenanalyzedthepossibleassociationofmusclespasticityatthewristandfingerswith thecapabilityofdailyactivitiesandthecorrelationofmusclespasticityatthekneeandankle withwalkingcapabilityinourpatients.Theresultsshowsignificantcorrelationofmusclespas- ticityatthewrist,thumb,andtheother4fingerswitheachotherandwiththecapabilitiesfor personalhygieneandposture(P<0.01,Table3).Incontrast,theMASscoreofthekneeand ankledidnotseemtobeassociatedwiththewalkingcapability.AsbothGASandimprovedBI indicateoverallperformance,itisnotsurprisingthatthese2valueswerecloselyassociated witheachother.Becauseimprovementinactivityforpersonalhygienewasthekeyexpectation ofbothpatientsandcaregivers(S1Fig.)andwristandhandfunctionstronglycorrelatedwith thisactivity,wefocusedontheimprovementofwristandhandspasticityinthisstudy. Efficacy TEASwith100Hzstimulationsignificantlyreduceswristspasticity.Comparedwithsham treatment,100HzTEASsignificantlyreducedwristMASscoreatweeks2,3,and4ofthetreat- ment,and1monthafterthetreatment,whereas2HzstimulationonlydecreasedwristMAS Figure4.Efficacyof100Hz,2Hz,orshamTEAS.A.WristMASscorewassignificantlyreducedby100HzTEAScomparedwith2HzorshamTEAS. Datawereanalyzedbyrepeated-measuresANOVA.The*representssignificantdifferencebetween100Hzandsham,P<0.05;#representssignificant differencebetween100Hzand2Hz,P<0.05;&representssignificantdifferencebetween2Hzandsham.B.WristMASscoreinpatientstreatedwith100 HzTEASwassignificantlydecreasedfrombaselineduringtreatmentandatfollow-upvisits.Comparisonwithbaselinevalueswasanalyzedbyrepeated- measuresANOVA.ThewristMASscoreswereusedforthestatisticalanalysis.Datawerepresentedasmean±standarderror.The*representssignificant differencesbetweenpost-treatmentandthebaselinevalue,P<0.05.C.ThumbMASscorewasnotaffectedbytreatments.D.TheMASscoreoftheother4 fingerswasnotaffectedbytreatments.ResultsarefromtheanalysisonITTdataset.ANOVA=analysisofvariance.MAS=ModifiedAshworthScale.TEAS =transcutaneouselectricalacupointstimulation. doi:10.1371/journal.pone.0116976.g004 PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 9/15 TEASTreatsMuscleSpasticity Table4.MASscoreoftheelbow,shoulder,knee,andankle. MASElbow MASShoulder MASKnee MASAnkle Baseline 100Hz 2.3±1.01 1.5±1.03 1.88±0.67 2.68±0.95 2Hz 2.6±0.84 1.8±0.62 2.0±0.72 2.8±0.52 Sham 3±0.84 1.725±0.85 2.35±0.81 2.875±0.6 Week1 100Hz 1.83±0.0.77 1.53±1.04 1.83±0.88 2.7±0.91 2Hz 2.53±0.97 1.75±0.53 2±0.69 2.8±0.68 Sham 2.75±0.7 1.825±0.73 2.35±0.86 2.9±0.55 Week2 100Hz 1.75±0.72 1.25±0.88 1.575±0.8 2.23±0.94 2Hz 2.55±0.87 1.68±0.65 1.88±0.70 2.73±0.6 Sham 2.6±0.74 1.78±0.83 2.15±0.86 2.88±0.6 Week3 100Hz 1.9±0.64 1.38±0.9 1.78±0.82 2.38±0.96 2Hz 2.35±0.84 1.55±0.63 1.85±0.71 2.65±0.65 Sham 2.33±0.8 1.6±0.85 2.08±0.78 2.85±0.59 Week4 100Hz 2±0.73 1.33±0.91 1.65±0.83 2.38±0.81 2Hz 2.25±0.7 1.7±0.7 1.73±0.5 2.53±0.7 Sham 2.4±0.79 1.7±0.85 2.05±0.65 2.7±0.57 Month1 100Hz 2±0.76 1.25±0.87 1.75±0.94 2.45±0.79 2Hz 2.35±0.78 1.7±0.5 1.8±0.66 2.7±0.7 Sham 2.48±0.77 1.68±0.73 2.13±0.69 2.78±0.57 Month2 100Hz 2.2±0.68 1.43±0.98 1.78±0.88 2.55±0.74 2Hz 2.35±0.83 1.53±0.55 1.73±0.47 2.7±0.55 Sham 2.6±0.66 1.83±0.75 2±0.71 2.78±0.57 Datawereanalyzedbyrepeated-measuresANOVA.MAS=ModifiedAshworthScale. doi:10.1371/journal.pone.0116976.t004 scoresignificantlyatWeek4ofthetreatment(P<0.05;Fig.4A).Inaddition,TEAS100Hz significantlyreducedwristMASscorefrombaselineatweeks2,3,and4oftreatmentandat the1-and2-monthfollow-upvisits(P<0.05;Fig.4B).ThemaximalreductioninMASscore atthewristfrombaseline(28.3%±4.54%)wasobservedatWeek3ofthe100Hztreatment (Fig.4B).Theseresultssuggestthatwristspasticityfollowingbraininjurycanbesubstantially attenuatedby100HzTEAS.Neither100Hznor2HztreatmentaffectedtheMASscoreofthe thumb(Fig.4C)andtheother4fingers(Fig.4D).Theresultsfromrepeated-measures ANOVAalsoshowthat,inadditiontotheintervention(P=0.012)andtime(P<0.001),the interactionbetweentimeandinterventionalsosignificantlyaffectedwristMASscore (P=0.003),indicatingthattherateofchangeintheMASscoreovertimeisdifferentforeach treatmentgroup.TheMASscoreattheother4fingerswasonlysignificantlyaffectedbytime (P<0.001).Time,intervention,orinteractionbetweentimeandinterventiondidnotaffect thumbMASscoresignificantly. Secondaryendpoints Noneofthesecondaryendpoints,includingtheMASscorefortheelbow,shoulder,knee,and ankle,DASscore,Holdenfunctionalambulationclassificationscore,GASscore,andtheim- provedBIscorewerechangedwith100Hzor2Hztreatments(Tables4,5,6).Timeortheinter- actionoftimeandinterventiondidnotaffectthesecondaryendpointssignificantly. PLOSONE|DOI:10.1371/journal.pone.0116976 February2,2015 10/15