Exercise for depression (Review) Cooney GM, Dwan K, Greig CA, Lawlor DA, Rimer J, Waugh FR, McMurdo M, Mead GE ThisisareprintofaCochranereview,preparedandmaintainedbyTheCochraneCollaborationandpublishedinTheCochraneLibrary 2013,Issue9 http://www.thecochranelibrary.com Exercisefordepression(Review) Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. TABLE OF CONTENTS HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 PLAINLANGUAGESUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 SUMMARYOFFINDINGSFORTHEMAINCOMPARISON . . . . . . . . . . . . . . . . . . . 4 BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Figure1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 ADDITIONALSUMMARYOFFINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . 26 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 AUTHORS’CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 CHARACTERISTICSOFSTUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 DATAANDANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Analysis1.1.Comparison1Exerciseversus’control’,Outcome1Reductionindepressionsymptomspost-treatment. 117 Analysis1.2.Comparison1Exerciseversus’control’,Outcome2Reductionindepressionsymptomsfollow-up. . . 119 Analysis1.3.Comparison1Exerciseversus’control’,Outcome3Completedinterventionorcontrol. . . . . . . 120 Analysis1.4.Comparison1Exerciseversus’control’,Outcome4Qualityoflife. . . . . . . . . . . . . . 122 Analysis2.1.Comparison2Exerciseversuspsychologicaltherapies,Outcome1Reductionindepressionsymptomspost- treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Analysis2.2.Comparison2Exerciseversuspsychologicaltherapies,Outcome2Completedexerciseorpyschological therapies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Analysis2.3.Comparison2Exerciseversuspsychologicaltherapies,Outcome3Qualityoflife. . . . . . . . . 125 Analysis3.1.Comparison3Exerciseversusbrightlighttherapy,Outcome1Reductionindepressionsymptomspost- treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Analysis4.1.Comparison4Exerciseversuspharmacologicaltreatments,Outcome1Reductionindepressionsymptoms post-treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Analysis 4.2. Comparison 4 Exercise versus pharmacological treatments, Outcome 2 Completed exercise or antidepressants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Analysis4.3.Comparison4Exerciseversuspharmacologicaltreatments,Outcome3QualityofLife. . . . . . . 128 Analysis5.1.Comparison5Reductionindepressionsymptomspost-treatment:Subgroupanalyses,Outcome1Exercisevs controlsubgroupanalysis:typeofexercise. . . . . . . . . . . . . . . . . . . . . . . . . 129 Analysis5.2.Comparison5Reductionindepressionsymptomspost-treatment:Subgroupanalyses,Outcome2Exercisevs controlsubroupanalysis:intensity. . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Analysis5.3.Comparison5Reductionindepressionsymptomspost-treatment:Subgroupanalyses,Outcome3Exercisevs controlsubroupanalysis:numberofsessions. . . . . . . . . . . . . . . . . . . . . . . . 133 Analysis5.4.Comparison5Reductionindepressionsymptomspost-treatment:Subgroupanalyses,Outcome4Exercisevs controlsubroupanalysis:diagnosisofdepression. . . . . . . . . . . . . . . . . . . . . . 135 Analysis5.5.Comparison5Reductionindepressionsymptomspost-treatment:Subgroupanalyses,Outcome5Exercisevs controlsubgroupanalysis:typeofcontrol. . . . . . . . . . . . . . . . . . . . . . . . . 137 Analysis6.1.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome1Reductionindepressionsymptoms post-treatment:peer-reviewedjournalpublicationsanddoctoralthesesonly. . . . . . . . . . . . . 139 Analysis6.2.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome2Reductionindepressionsymptoms post-treatment:studiespublishedasabstractsorconferenceproceedingsonly. . . . . . . . . . . . . 141 Analysis6.3.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome3Reductionindepressionsymptoms post-treatment:studieswithadequateallocationconcealment. . . . . . . . . . . . . . . . . . 142 Exercisefordepression(Review) i Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. Analysis6.4.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome4Reductionindepressionsymptoms post-treatment:studiesusingintention-to-treatanalysis. . . . . . . . . . . . . . . . . . . . 143 Analysis6.5.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome5Reductionindepressionsymptoms post-treatment:studieswithblindedoutcomeassessment. . . . . . . . . . . . . . . . . . . 144 Analysis6.6.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome6Reductionindepressionsymptoms post-treatment:allocationconcealment,intention-to-treat,blindedoutcome. . . . . . . . . . . . . 145 Analysis6.7.Comparison6Exerciseversuscontrol:sensitivityanalyses,Outcome7Reductionindepressionsymptoms post-treatment:Lowestdoseofexercise. . . . . . . . . . . . . . . . . . . . . . . . . . 146 ADDITIONALTABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 FEEDBACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 WHAT’SNEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 CONTRIBUTIONSOFAUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 DECLARATIONSOFINTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 SOURCESOFSUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 DIFFERENCESBETWEENPROTOCOLANDREVIEW . . . . . . . . . . . . . . . . . . . . . 156 INDEXTERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Exercisefordepression(Review) ii Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. [InterventionReview] Exercise for depression GaryMCooney1,KerryDwan2,CarolynAGreig3,DebbieALawlor4,JaneRimer5,FionaRWaugh6,MarionMcMurdo7,GillianE Mead8 1DivisionofPsychiatry,RoyalEdinburghHospital,NHSLothian,Edinburgh,UK.2InstituteofChildHealth,UniversityofLiverpool, Liverpool,UK.3UniversityofBirmingham,Birmingham,UK.4MRCCentreforCausalAnalysesinTranslationalEpidemiology,School ofSocialandCommunityMedicine,UniversityofBristol,Bristol,UK.5UniversityHospitalsDivision,NHSLothian,Edinburgh,UK. 6GeneralSurgery,NHSFife,VictoriaHostpitalKirkcaldy,Kirkcaldy,UK.7CentreforCardiovascularandLungBiology,Divisionof MedicalSciences,UniversityofDundee,Dundee,UK.8CentreforClinicalBrainSciences,UniversityofEdinburgh,Edinburgh,UK Contactaddress:GillianEMead,CentreforClinicalBrainSciences,UniversityofEdinburgh,RoomS1642,RoyalInfirmary,Little FranceCrescent,Edinburgh,EH164SA,[email protected]@staffmail.ed.ac.uk. Editorialgroup:CochraneDepression,AnxietyandNeurosisGroup. Publicationstatusanddate:Newsearchforstudiesandcontentupdated(nochangetoconclusions),publishedinIssue9,2013. Reviewcontentassessedasup-to-date: 13July2012. Citation: Cooney GM, Dwan K, Greig CA, Lawlor DA, Rimer J,Waugh FR, McMurdo M, Mead GE. Exercisefor depression. CochraneDatabaseofSystematicReviews2013,Issue9.Art.No.:CD004366.DOI:10.1002/14651858.CD004366.pub6. Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. ABSTRACT Background Depressionisacommonandimportantcauseofmorbidityandmortalityworldwide.Depressioniscommonlytreatedwithantide- pressants and/or psychological therapy,butsome peoplemaypreferalternativeapproachessuch asexercise.Thereare anumber of theoreticalreasonswhyexercisemayimprovedepression.Thisisanupdateofanearlierreviewfirstpublishedin2009. Objectives To determine the effectiveness of exercise in the treatment of depression in adults compared with no treatment or a comparator intervention. Searchmethods WesearchedtheCochraneDepression,AnxietyandNeurosisReviewGroup’sControlledTrialsRegister(CCDANCTR)to13July 2012.Thisregisterincludesrelevantrandomisedcontrolledtrialsfromthefollowingbibliographicdatabases:TheCochraneLibrary (allyears);MEDLINE (1950 todate);EMBASE(1974todate)andPsycINFO (1967 todate).Wealsosearchedwww.controlled- trials.com, ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform. No date or language restrictions were appliedtothesearch. WeconductedanadditionalsearchoftheCCDANCTRupto1stMarch2013andanypotentiallyeligibletrialsnotalreadyincluded arelistedas’awaitingclassification.’ Selectioncriteria Randomisedcontrolledtrialsinwhichexercise(definedaccordingtoAmericanCollegeofSportsMedicinecriteria)wascomparedto standardtreatment,notreatmentoraplacebotreatment,pharmacologicaltreatment,psychologicaltreatmentorotheractivetreatment inadults(aged18andover)withdepression,asdefinedbytrialauthors.Weincludedclustertrialsandthosethatrandomisedindividuals. Weexcludedtrialsofpostnataldepression. Exercisefordepression(Review) 1 Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. Datacollectionandanalysis Tworeviewauthorsextracteddataonprimaryandsecondaryoutcomesattheendofthetrialandendoffollow-up(ifavailable).We calculatedeffectsizesforeachtrialusingHedges’gmethodandastandardisedmeandifference(SMD)fortheoverallpooledeffect, usingarandom-effectsmodelriskratiofordichotomousdata.Wheretrialsusedanumberofdifferenttoolstoassessdepression,we includedthemainoutcomemeasureonlyinthemeta-analysis.Wheretrialsprovidedseveral’doses’ofexercise,weuseddatafromthe biggest ’dose’of exercise,andperformedsensitivity analysesusing thelower’dose’.Weperformedsubgroup analysestoexplore the influenceofmethodofdiagnosisofdepression(diagnosticintervieworcut-offpointonscale),intensityofexerciseandthenumberof sessionsofexerciseoneffectsizes.Twoauthorsperformedthe’Riskofbias’assessments.Oursensitivityanalysesexploredtheinfluence ofstudyqualityonoutcome. Mainresults Thirty-ninetrials(2326participants)fulfilledourinclusioncriteria,ofwhich37provideddataformeta-analyses.Thereweremultiple sourcesofbiasinmanyofthetrials;randomisationwasadequatelyconcealedin14studies,15usedintention-to-treatanalysesand12 usedblindedoutcomeassessors. Forthe35trials(1356participants)comparingexercisewithnotreatmentoracontrolintervention,thepooledSMDfortheprimary outcomeofdepressionattheendoftreatmentwas-0.62(95%confidenceinterval(CI)-0.81to-0.42),indicatingamoderateclinical effect.Therewasmoderateheterogeneity(I²=63%). Whenweincludedonlythesixtrials(464participants)withadequateallocationconcealment,intention-to-treatanalysisandblinded outcomeassessment,thepooledSMDforthisoutcomewasnotstatisticallysignificant(-0.18, 95%CI-0.47to0.11).Pooleddata fromtheeighttrials(377participants)providinglong-termfollow-updataonmoodfoundasmalleffectinfavourofexercise(SMD- 0.33,95%CI-0.63to-0.03). Twenty-ninetrialsreportedacceptabilityoftreatment,threetrialsreportedqualityoflife,nonereportedcost,andsixreportedadverse events. For acceptability oftreatment(assessedbynumber ofdrop-outs duringtheintervention), theriskratiowas1.00(95% CI0.97to 1.04). Seventrialscomparedexercisewithpsychologicaltherapy(189participants),andfoundnosignificantdifference(SMD-0.03,95% CI-0.32to0.26).Fourtrials(n=300)comparedexercisewithpharmacologicaltreatmentandfoundnosignificantdifference(SMD -0.11,-0.34,0.12).Onetrial(n=18)reportedthatexercisewasmoreeffectivethanbrightlighttherapy(MD-6.40,95%CI-10.20 to-2.60). Foreachtrialthatwasincluded,twoauthorsindependentlyassessedforsourcesofbiasinaccordancewiththeCochraneCollaboration ’Riskofbias’tool.Inexercisetrials,thereareinherentdifficultiesinblindingboththosereceivingtheinterventionandthosedelivering theintervention.Manytrialsusedparticipantself-reportratingscalesasamethodforpost-interventionanalysis,whichalsohasthe potentialtobiasfindings. Authors’conclusions Exerciseismoderatelymoreeffectivethanacontrolinterventionforreducingsymptomsofdepression,butanalysisofmethodologically robusttrialsonlyshowsasmallereffectinfavourofexercise.Whencomparedtopsychologicalorpharmacologicaltherapies,exercise appearstobenomoreeffective,thoughthisconclusionisbasedonafewsmalltrials. PLAIN LANGUAGE SUMMARY Exercisefordepression Whyisthisreviewimportant? Depressionisacommonanddisablingillness,affectingover100millionpeopleworldwide.Depressioncanhaveasignificantimpact onpeople’sphysicalhealth,aswellasreducingtheirqualityoflife.Researchhasshownthatbothpharmacologicalandpsychological therapiescan beeffectiveintreating depression. However,many peoplepreferto tryalternative treatments. SomeNHS guidelines suggestthatexercisecouldbeusedasadifferenttreatmentchoice.However,itisnotclearifresearchactuallyshowsthatexerciseisan effectivetreatmentfordepression. Exercisefordepression(Review) 2 Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. Whomaybeinterestedinthisreview? Patientsandfamiliesaffectedbydepression. GeneralPractitioners. Mentalhealthpolicymakers. Professionalsworkinginmentalhealthservices. Whatquestionsdoesthisreviewaimtoanswer? ThisreviewisanupdateofapreviousCochranereviewfrom2010whichsuggestedthatexercisecanreducesymptomsofdepression, buttheeffectwassmallanddidnotseemtolastafterparticipantsstoppedexercising. Wewantedtofindoutifmoretrialsoftheeffectofexerciseasatreatmentfordepressionhavebeenconductedsinceourlastreview thatallowustoanswerthefollowingquestions: Isexercisemoreeffectivethannotherapyforreducingsymptomsofdepression? Isexercisemoreeffectivethanantidepressantmedicationforreducingsymptomsofdepression? Isexercisemoreeffectivethanpsychologicaltherapiesorothernon-medicaltreatmentsfordepression? Howacceptabletopatientsisexerciseasatreatmentfordepression? Whichstudieswereincludedinthereview? Weusedsearchdatabases tofindallhigh-quality randomisedcontrolledtrialsofhoweffectiveexerciseisfortreating depressionin adultsover18yearsofage.WesearchedforstudiespublishedupuntilMarch2013.WealsosearchedforongoingstudiestoMarch 2013.Allstudieshadtoincludeadultswithadiagnosisofdepression,andthephysicalactivitycarriedouthadtofitcriteriatoensure thatitmetwithadefinitionof‘exercise’. Weincluded39studieswithatotalof2326participantsinthereview.Thereviewersnotedthatthequalityofsomeofthestudieswas low,whichlimitsconfidenceinthefindings.Whenonlyhigh-qualitytrialswereincluded,exercisehadonlyasmalleffectonmood thatwasnotstatisticallysignificant. Whatdoestheevidencefromthereviewtellus? Exerciseismoderatelymoreeffectivethannotherapyforreducingsymptomsofdepression. Exerciseisnomoreeffectivethanantidepressantsforreducingsymptomsofdepression,althoughthisconclusionisbasedonasmall numberofstudies. Exerciseisnomoreeffectivethanpsychologicaltherapiesforreducingsymptomsofdepression,althoughthisconclusionisbasedon smallnumberofstudies. The reviewers also note that when only high-quality studies were included, the difference between exercise and no therapy is less conclusive. Attendanceratesforexercisetreatmentsrangedfrom50%to100%. Theevidenceaboutwhetherexercisefordepressionimprovesqualityoflifeisinconclusive. Whatshouldhappennext? Thereviewersrecommendthatfutureresearchshouldlookinmoredetailatwhattypesofexercisecouldmostbenefitpeoplewith depression,andthenumberanddurationofsessionswhichareofmostbenefit.Furtherlargertrialsareneededtofindoutwhether exerciseisaseffectiveasantidepressantsorpsychologicaltreatments. Exercisefordepression(Review) 3 Copyright©2013TheCochraneCollaboration.PublishedbyJohnWiley&Sons,Ltd. CopyrighxerciseE SUMMARY OF FINDINGS FOR THE MAIN COMPARISON [Explanation] tfo ©r 2013depre Exercisecomparedtonointerventionorplaceboforadultswithdepression Ts heCsion Patientorpopulation:adultswithdepression o(R Settings:anysetting chraevie Intervention:Exercise nw e) Comparison:nointerventionorplacebo C o lla bo Outcomes Illustrativecomparativerisks*(95%CI) NoofParticipants Qualityoftheevidence Comments r atio (studies) (GRADE) n . P u Assumedrisk Correspondingrisk b lis h ed Nointerventionorplacebo Exercise b y Jo hn Symptomsofdepression The mean symptoms of de- The mean symptoms of de- 1353 ⊕⊕⊕(cid:13) SMD-0.62(95%CI:-0.81to- W Differentscales pressioninthecontrolgroups pression in the intervention (35studies) moderate2,3,4 0.42). iley Follow-up:post-treatment was groupswas Theeffectsizewasinterpreted & S 0 0.62 standard deviations as ’moderate’ (using Cohen’s o n lower ruleofthumb) s , L (0.81to0.42lower)1 td . Symptoms of depression The mean symptoms of de- The mean symptoms of de- 377 ⊕⊕(cid:13)(cid:13) SMD-0.33(95%CI:-0.63to- (long-term) pression (long-term) in the pression(long-term)inthein- (8studies) low4,5 0.03). differentscales controlgroupswas terventiongroupswas Theeffectsizewasinterpreted 0 0.33 standard deviations as’small’(usingCohen’sruleof lower thumb) (0.63to0.03lower) Adverseevents Seecomment Seecomment 0 ⊕⊕⊕(cid:13) Seven trials reported no dif- (6studies) moderate ferenceinadverseeventsbe- tweenexerciseandusualcare groups. Dunn 2005 reported increased severity of depres- sivesymptoms(n=1),chest pain (n = 1) and joint pain/ 4 CE opyrighxercise swelling (n = 1); all these tfo ©r participantsdiscontinuedexer- 2013depre cise.Singh1997reportedthat Ts 1exerciserwasreferredtoher heCsion psychologist at 6 weeks due ochra(Revie tmouisnccurleoasskienlgetaslusicyimdaplittoym; sanidn nw e) 2participantsrequiredadjust- C o lla mentoftrainingregime.Singh bo 2005 r atio reportedadverseeventsinde- n. tail (visits to a health profes- P u sional, minor illness, muscu- b lish lar pain, chest pain, injuries e d requiring training adjustment, b yJo falls,deathsandhospitaldays) hn and found no difference be- W tween the groups. Knubben ile y 2007 reported ‘‘no negative & S effects of exercise (muscle o ns pain,tightnessorfatigue)’’;af- , Ltd terthetraininghadfinished,1 . person in the placebo group required gastric lavage and 1 person in the exercise group inflictedasuperficialcutonher arm.Sims2009 reported no adverse events or falls in either the exercise or control group. Blumenthal 2007 reported more side ef- fects in the sertraline group (see comparison below) but there was no difference be- 5 CE opyrighxercise tweentheexerciseandcontrol tfo ©r group. Blumenthal 2012a re- 2013depre ported more fatigue and sex- Ts ualdysfunctioninthesertraline heCsion groupthantheexercisegroup o(R chraneview Acceptabilityoftreatment Studypopulation 1363 ⊕⊕⊕(cid:13) RR1 eC) (29studies) moderate2 (95%CI:0.97to1.04) olla 865per1000 865per1000 bo (839to900) r a tio n. Qualityoflife Themeanqualityoflifeinthe 0 Seecomment Therewasnostatisticallysig- P u interventiongroupswas (4studies) nificant differences for the b lish 0higher mental (SMD-0.24;95%CI- e d (0to0higher) 0.76 to 0.29). psychological b yJo (SMD0.28;95%CI-0.29to0. hn 86)andsocialdomains(SMD W 0.19; 95% CI -0.35 to 0.74) ile y . Two studies reported a sta- & S tistically significant difference o ns for the environment domain , Ltd favouringexercise(SMD0.62; . 95% CI 0.06 to 1.18) and 4 studiesreportedastatistically significant difference for the physicaldomainfavouringex- ercise (SMD 0.45; 95% CI 0. 06to0.83) *Thebasisfor the assumedrisk(e.g.themediancontrol groupriskacross studies) isprovidedinfootnotes. Thecorresponding risk(andits95%confidence interval) isbasedonthe assumedriskinthecomparisongroupandtherelativeeffectoftheintervention(andits95%CI). CI:Confidenceinterval;RR:Riskratio; GRADEWorkingGroupgradesofevidence Highquality:Furtherresearchisveryunlikelytochangeourconfidenceintheestimateofeffect. Moderatequality:Furtherresearchislikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandmaychangetheestimate. Lowquality:Furtherresearchisverylikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandislikelytochangetheestimate. Verylowquality:Weareveryuncertainabouttheestimate. 6 CE opyrightxercisefo 1EBffleucmteenstthimala2te00c7alc(usltautdeydcbhyorsee-nexfporresbseiinnggtmheosStMrDeporensethnetaHtivaem).iltTohneDSeDprewsassiomnuRlatitpinligedSbcyalethuesipnogoltehdecSoMnDtrotlogprorouvpidSeDth(e7)effrfoemct ©r 2013depre 2LeasctkimoafteblionnditnhgeoHfDoRuStc.omeassessorsprobablyincreasedeffectsizesanddrop-outrateswerehigh.Alsosequencegenerationwas Thessio consideredunclearin23studies. Cn 3I²=63%andP<0.00001,indicatedmoderatelevelsofheterogeneity o(R chraneview 45PLaocpkuloaftiobnlinsdizinegisolfaorguetc,oemffeecatsssizeessisorasbopvroeb0a.b2lySDin,carenadstehdee9f5fe%ctCsIizdeosesanndotdcrorops-osuttheraltiensewofenreoheifgfehc.tA.lsosequencegenerationwas e) C consideredunclearin4studies. o lla b xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx o r a tio n . P u b lis h e d b y Jo h n W ile y & S o n s , L td . 7