Challenges in assessing the sunscreen‐melanoma association Corina S. Rueegg.pdf

该【Challenges in assessing the sunscreen‐melanoma association Corina S. Rueegg 】是由【探春文档】上传分享，文档一共【39】页，该文档可以免费在线阅读，需要了解更多关于【Challenges in assessing the sunscreen‐melanoma association Corina S. Rueegg 】的内容，可以使用淘豆网的站内搜索功能，选择自己适合的文档，以下文字是截取该文章内的部分文字，如需要获得完整电子版，请下载此文档到您的设备，方便您编辑和打印。:..ChallengesinAssessingtheSunscreen-MelanomaAssociationCorinaSRueegg1,JoSStenehjem2,MatthiasEgger3,RezaGhiasvand1,EunyoungCho4,5,EilivLund7,ElisabeteWeiderpass2,7,8,9,AdeleCGreen10,11,MaritBVeier?d11OsloCentreforBiostatisticsandEpidemiology,DepartmentofBiostatistics,InstituteofBasicMedicalSciences,UniversityofOslo,0317Oslo,Norway2DepartmentofResearch,CancerRegistryofNorway,InstituteofPopulation-BasedCancerResearch,Oslo,0304Oslo,Norway3InstituteofSocialandPreventiveMedicine,UniversityofBern,3012Bern,Switzerland4DepartmentofDermatology,WarrenAlpertMedicalSchool,BrownUniversity,Providence,RhodeIsland02912,USA5DepartmentofEpidemiology,BrownSchoolofPublicHealthatBrownUniversity,RhodeIsland02903,USAworkMedicine,DepartmentofMedicine,BrighamandWomen?sHospitalArticleandHarvardMedicalSchool,Boston02115,Massachusetts,USAmunityMedicine,FacultyofHealthSciences,UniversityofTroms?,TheArcticUniversityofNorway,9019Troms?,Norway8DepartmentofMedicalEpidemiologyandBiostatistics,KarolinskaInstitutet,17177Stockholm,SwedenicEpidemiologyGroup,Folkh?lsanResearchCenter,andFacultyofMedicine,HelsinkiUniversity,00290Helsinki,Finland10QIMRBerghoferMedicalResearchInstitute,4006Brisbane,Australia11CRUKManchesterInstitute,UniversityofManchester,ManchesterM204GJ,UnitedKingdomCorrespondingauthor:,DepartmentofBiostatistics,InstituteofBasicMedicalSciences,UniversityofOslo,,0317Oslo,NorwayMail:.******@:+4722840231AcceptedFax:+4722851280eptedforpublicationandundergonefullpeerreviewbuthasnot?beenthroughthecopyediting,typesetting,paginationandproofreadingprocess,:.Allrightsreserved:..Shorttitle:Sunscreenandmelanoma:asystematicreview(6words)Keywords(3-5):Sunscreen,melanoma,skincancer,meta-analysis,sunprotectionArticlecategory::..BriefdescriptionEffectivenessofsunscreeninreducingUV--analysisofdatafromfourdifferentstudydesigns,thefirsttostratifyhospital-andpopulation-basedcase-controlstudies,andthemeta-analysistoincludethemostprospectivestudies(n=5).Wefoundheterogeneoussummaryestimatesforthesunscreen-,GRADEdevelopmentandevaluationHRhazardratiopp-valueNNordNOSNewcastle-:..Abstract(248/250words),--effectsmeta-bineestimatesoftheassociation,-analysisandmeta-?069melanomacasesfrom28studiespublished1979-2018:23case-control(11hospital-based,12population-based),1ecological,3cohortand1randomizedcontrolledtrial(RCT).Therewasmarkedheterogeneityacrossstudydesignsandamongcase-controlstudiesbutadjustmentforconfoundingbysunexposure,--useofsunscreenwasinverselyassociatedwithmelanomainhospital-basedcase-controlstudies(adjustedoddsratio(OR)=,95%confidenceinterval(CI)-,pheterogeneity<),theecologicalstudy(rateratio=,95%-),andtheRCT(hazardratio(HR)=,95%-).Itwasnotassociatedinpopulation-basedcase-controlstudies(OR=,95%-,pheterogeneity<)andwaspositivelyassociatedinthecohortstudies(HR=,95%-,pheterogeneity=).Theassociationdifferedbylatitude(pinteraction=),region(pinteraction=),adjustmentfornaevi/freckling(pinteraction=),andproportionofnever-sunscreen-users(pinteraction=0·012).Evidencefromobservationalstudiesonsunscreenuseandmelanomariskwasweakandheterogeneous,:..INTRODUCTIONCutaneousmelanomaistheleadingcauseofskincancerdeath,ountingfor1–2%,3In2015,urredin351?880peopleandresultedin59?(hereaftertermedmelanoma)plexic,,6Melanomaismainlycausedbyultraviolet(UV)radiationexposureinsun-sensitivesubjectsanditisestimatedthatmorethan85%-,12ItfollowsthatmelanomaispreventablethroughreductionofUVexposure,makingprimarypreventionhighlycost-,13UseofsunscreenisgenerallyArticleregardedasamajorprimarypreventionmeasurealongsideseekingshade,wearingprotectiveclothes,andavoidingsunbeds,14--induceddamagetotheskinhasbeenprovenonlyinexperimentalstudies,(RCT)ofdailysunscreenapplicationtopreventskincancerhasbeenperformed,showingareducedriskofmelanoma(hazardratio=,p-value=),pliancetodailysunscreenapplicationwasapproximately75%;themajorityofparticipantsinthediscretionarysunscreengroupeitherdidnotapplysunscreen(38%)orappliedatmostonceortwiceaweek(35%).21Allotherstudiesofsunscreenandmelanomariskhavebeenobservational,mainlycase-control,-40AcceptedThemainproblemwithinvestigatingthisquestionwithobservationalstudiesisconfoundingbyindication,--:..etheseknownlimitationsbyperformingin-,weaimedto1)systematicallysummarizetheexistingliteratureonsunscreenuseandmelanomainhumans;2)investigatetheeffectofever--useonmelanomarisk;3)assesstheeffectofdifferentlevelsandpatternsofsunscreenuse;4)identifysourcesofbiasandbetween-studyheterogeneity;and5)(PROSPEROID:CRD4201706398049)ArticleordingtoPRISMA-P50,(includingMedline),EmbaseandCochraneDatabaseofSystematicReviewswithsearchtermsadaptedforeachofthem(SupplementalAppendixI).Inaddition,-reviewedjournalsarisingfromcase-controlstudies,ecologicalstudies(population-levelratherthanindividual-Acceptedlevelobservationalstudies),-:..Studiesonchildhoodmelanomawereincludedinthequalitativesynthesisbutexcludedfromthemeta-(ThomsonReuters,versionX8),de-duplicatedandthenimportedtoMicrosoftExcel(version2010)(CSRandJSS)byfirstscreeningtitlesandabstracts,(SupplementalAppendixII)-,,ifreported:a)ever--useofsunscreenfromminimallyadjustedmodel;b)ever--useofsunscreenfrommaximallyadjustedmodel;c)three-levelestimateofsunscreenusefrommaximallyadjustedmodelsforfrequencyofuse,sunprotectionfactor(SPF)usedanddurationofuse(SupplementalAcceptedTable1).Theminimallyandmaximallyadjustedmodelwasthemodelwithnooronlybasicadjustmentandthemodelwithmostvariablesincluded,respectively,--uselabelbecausemostunderlyingstudiesanalysedever--,:..extractedbibliographicanddemographicinformationofthestudies,assessmentofsunscreenuse,?stoolforassessingriskofbias54andtheNewcastle-OttawaScale(NOS).55Levelofbias(high,medium,low)wasratedbythedataextractor(CSR)(StataCorpLP,).Intheanalysisofever--,ifastudyreportedanestimatewiththreecategoriesofsunscreenuse:never,sometimes,andoften,weaggregated?sometimes?and?often?intoever-,whileothersreflectedever-,-level,differentpatternsandhighsunscreenuse,weextractedallestimateswithatleastthreecategoriesonfrequencyofsunscreenuse,SPFused,,thelowestandhighestcategorieswerecategorizedaslowestandhighestgroups,-effectsmeta-analysis58stratifiedbystudydesignfortheminimallyandmaximallyadjustedestimatesofever--useofsunscreen,andforeachthree---indexwasusedtoquantifytheextentofheterogeneity,withI2-values>50%,and>75%beingindicativeofmoderateandhighheterogeneity,:Thed