COL1A1polymorphismandhighmyopia 窑Meta-Analysis窑 AssociationofCOL1A1polymorphismwithhigh myopia:aMeta-analysis

1 StateKeyLaboratoryofOphthalmology,Zhongshan ·KEYWORDS: collagentypeIalpha1;polymorphism;high OphthalmicCenter,SunYat-senUniversity,Guangzhou myopia;Meta-analysis 510060,GuangdongProvince,China DOI:10.18240/ijo.2016.04.22 2DepartmentofOphthalmology,theFifthAffiliatedHospital ofSunYat-senUniversity,Zhuhai519000,Guangdong JinGM,ZhaoXJ,ChenAM,ChenYX,LiQ.AssociationofCOL1A1 Province,China polymorphismwithhighmyopia:aMeta-analysis. 3 DepartmentofPharmacy,theFifthAffiliatedHospitalof 2016;9(4):604-609 Sun Yat-senUniversity,Zhuhai519000,Guangdong Province,China INTRODUCTION 4Departmentof Otorhinolaryngology,JiangmenCentral yopiaisacommoneyedisorderthathasbeenwidely Hospital,Jiangmen529030,GuangdongProvince,China M studiedinrecentyears [1-3],particularlybecauseofits Co-firstauthors: Guang-MingJin,Xiao-JingZhaoand increasing prevalenceacrosspopulationsworldwide [4]. Ai-MingChen Accordingtoepidemiologicalevidence,theincidenceof Correspondenceto: QinLi.DepartmentofOphthalmology, myopiaisincreasing,especiallyinEastAsia [3,5-6].Previous theFifthAffiliatedHospitalofSunYat-senUniversity,No. studieshaveindicatedthatthegreatestcontributortomyopic 52meihuaRoad,Zhuhai519000,GuangdongProvince, refractionistheaxiallength [7],andwhentheelongationof [email protected] theeyeballisexcessive,highmyopiaoccurs.Highmyopiais Received:2014-05-27Accepted:2015-07-20 anextremeformofmyopiathatischaracterizedbya sphericalequivalentoflessthan-6.00dioptersoranaxial lengthgreaterthan26mm.Manyseriouscomplicationsare Abstract associatedwithhighmyopia,includingretinaldetachment, ·AIM:Toinvestigatetheassociationbetweencollagen glaucoma,cataracts,maculardegeneration,and scleral typeIalpha1(COL1A1)geneandhighmyopia. staphyloma [8].Highmyopiaisoneofthemajorcausesof · METHODS:InthisMeta-analysis,weexamined5 blindnessinmanycountriesandisanimportantpublichealth [9] publishedcase-controlstudiesthatinvolved1942high problemworldwide . myopiacasesand2929healthycontrolstoassessthe Unfortunately,thepathogenesisofhighmyopiaremains association between the COL1A1 rs2075555 unclear.Althoughenvironmentalfactorssuchasnearwork, polymorphismandhighmyopiarisk.Wecalculatedthe highereducationallevels,andpooreconomicdevelopment pooled odds ratios (ORs) of COL1A1 rs2075555 levelshavebeenimplicatedintheoccurrenceofhigh polymorphisminhighmyopiacases healthycontrols myopia [10-11],theycannotexplainallcases.However,many toevaluatethestrengthoftheassociation. studiessuggestthatgeneticfactorsmayberesponsiblefor ·RESULTS:Overall,therewasnosignificantdifference highmyopia [12-14].Recentreportshavedemonstratedthat bothinthegenotypeandalleledistributionsofCOL1A1 interactionsofmultiplegeneticandenvironmentalfactors [15-16] rs2075555polymorphismbetweenhighmyopiacasesand maycontributetothedevelopmentofhighmyopia ,and healthycontrols:CC AAOR=1.10,95%confidence severalgeneshavebeenconfirmedtohaveanassociation interval(CI)=0.76-1.58;AC AAOR=0.98,95%CI0.80- withsusceptibilitytohighmyopia [12,14,17-19].However,other 1.20;CC/AC AA/OR=1.01,95%CI0.84-1.22;CC AC/ studieshavenotbeenabletoreplicatetheoriginalfindings AAOR=1.06,95%CI=0.93-1.20;C AOR=1.06,95%CI forthesegenes [20-22].Inparticular,thecollagentypeIalpha1 0.91-1.23).Inaddition,inthestratifiedanalysesby (COL1A1)genehasbeenstudied. ethnicity,nosignificantassociationswerefoundinany TheCOL1A1geneislocatedonchromosome17 geneticmodelbothinEuropeanandAsiacohorts. (17q21.23),whichcontainsthemyopia5(MYP517q21-22) · CONCLUSION:OurresultsindicatethattheCOL1A1 locus [23].Thisgeneencodesthemajorcomponent(pro- 1 琢 rs2075555polymorphismmaynotaffectsusceptibilityto chains)oftypeIcollagen.Mutationsorsinglenucleotide highmyopia. polymorphisms(SNPs)oftheCOL1A1genemayaffectthe 604 陨灶贼允韵责澡贼澡葬造皂燥造熏灾燥造援 9熏晕燥援 4熏 Apr.18, 圆园16 www.ijo.cn 栽藻造押8629原愿圆圆源缘员苑圆 8629-82210956 耘皂葬蚤造押ijopress岳员远猿援糟燥皂 formationofCOL1A1productsbyalteringCOL1A1gene expression,whichasaresultcontributestothesusceptibility tocollagen-relateddiseasessuchasosteoporosis,osteogenesis imperfecta,Ehlers-Danlossyndrome,andMarfansyndrome, aswellasscleralthinning [24].TheCOL1A1genealso reportedlyplaysanimportantroleinthedevelopmentof animalmyopia [25-26].Recently,severalstudiesperformedin differentregionsinvestigatedageneticmutationinthe COL1A1sequence[rs2075555( ),adenineto cytosine,A>C],asacandidatebiomarkerassociatedwith highmyopia [19,21,27-29].However,theconclusionsofthese previousstudiesremaincontroversialandconflicting.To further investigatetheroleofCOL1A1 rs2075555 Figure1Flowchartfortheselectionofstudiesaccordingto polymorphisminhighmyopia,weperformedaMeta-analysis thecriteriaofthisMeta-analysis. involvingalltherelevantpublishedstudiesavailable. MATERIALSANDMETHODS StatisticalAnalysis TheallelicfrequencyofCOL1A1 Literature Search Strategy Potentialarticleswere rs2075555wascalculated,andHardy-Weinbergequilibrium identifiedbyasystematicsearchontheISIWebofScience, (HWE)wasassessedforthecontrolgroupineachincluded PubMed,EMBASE,WileyOnlineLibrary,andScience study.Pearson's 2 testwasused,andsignificant Directdatabasesupto December15,2013,usinga disequilibriumwasdefinedas <0.05.Thestrengthofthe combinationofsearchterms:"collagentypeIalpha1"OR associationbetweenCOL1A1rs2075555polymorphismand "COL1A1","polymorphism"OR"variation"OR"mutation" highmyopiasusceptibilitywasassessedbyORswith95% AND "myopia"without languageorpublicationdate confidenceintervals(CIs).Weused -testtojudgethe restrictions.Allrelevantpublicationsandtheirreference significanceofthepooledORs,andstatisticalsignificance weremanuallyscreenedtoidentifyeligiblestudies. wasconsideredwhen <0.05.ThepooledORswere SelectionCriteria Papersidentifiedduringtheliterature calculatedfor5geneticmodels:allelemodel(C A); searchhadtomeetthefollowinginclusioncriteriainorderto homozygotemodel(CC AA);heterozygotemodel(AC beincludedinourstudy:1)case-controlorcohortdesign AA);dominantgeneticmodel(AC/CC AA);and studiesofhumans;2)evaluationoftheassociationof recessivegeneticmodel(CC AC/AA).Stratifiedanalysis COL1A1rs2075555polymorphismwithhighmyopia;3) indifferentgeneticmodelswasperformedbyethnicity. sufficientpublisheddataavailableforourteamtoestimate Heterogeneityamongincludedstudieswasevaluatedbythe oddsratios(ORs)ofdifferentgenotypefrequencies;4) 2-based test [32] andthe 2 index [33]; <0.10or 2> publishedoriginalfull-textliterature.Weexcludedstudies 50%wereconsideredstatisticallysignificant.Therandom basedonthefollowingcriteria:1)insufficientreporteddata; effectsmodelwasusedtoestimatethepooledORswhen 2)abstracts,reviewpapers,andcase-onlystudies;3) obvious heterogeneitywaspresent [34],otherwisethe duplicationofpreviouslypublishedliterature. fixed-effectsmodelwasused [35].Thesourcesofthe DataExtraction Threeinvestigators(JinGM,ZhaoXJ,and heterogeneitywereidentifiedbytheGalbraithplot [36]. ChenYX)independentlyextractedthedata.Discrepancies Sensitivityanalysesthatdeleted1studyatatimetoreflect betweendifferentinvestigatorswereadjudicatedbyanother theeffectsoftheindividualstudytothepooledORswere 2investigators(ChenAMandLiQ),whoreached usedtoestimatethestabilityofourresults [37].Publication consensus.Thecollecteddataincluded:nameofthefirst biasofarticleswasassessedusingBegg'sfunnelplotsand author,publicationdate,geographicallocation,ethnicity, Egger'stest [38].Significantpublicationbiaswasconsidered sourceofcontrol,genotypingmethods,genotypefrequencies, when <0.05.Allstatisticalanalyseswereperformedwith matchingvariables,andthenumbersofcasesandcontrols. STATA12.0software(StataCorpLP,CollegeStation,TX, QualityAssessment WeemployedtheNewcastle-Ottawa USA). scale(NOS),whichhasbeendescribedindetailinprevious RESULTS study [30] toevaluatethequalityoftheincludedstudiesby2 CharacteristicsofEligibleStudies Overall,151published investigators(JinGMandZhaoXJ).Astudyscoringless articleswereretrievedbyourliteraturesearchstrategy.From than3starswascategorizedas"lowquality",while4to6 thisgroupweultimatelyanalyzed5case-controlstudiesthat starsand7to9starswerecategorizedas"moderatequality" includedfourAsianstudiesandoneEuropeanstudyandin and"highquality",respectively [31].Discrepanciesbetween totalinvolved1942casesand2929controls.Figure1shows investigatorswereresolvedbydiscussion. theliteratureselectionprocess.Genotypefrequenciesinthe 605 COL1A1polymorphismandhighmyopia Table 1 Main characteristics of eligible studies included in the Meta-analysis First author Year Region Ethnicity Genotyping method Source Matching Inamori Y[19] 2007 Japan Asia Probe method HB Age, sex and ethnicity Liang CL[29] 2007 Taiwan Asia TaqMan PB Age, sex and ethnicity Vatavuk Z[28] 2009 Croatia European BeadChip assay PB NA Nakanishi H[21] 2009 Japan Asia TaqMan PB NA Zhang D[27] 2011 China Asia SNaPshot method HB Age and sex HB: Hospital-based study; PB: Population-based study; NA: Not available. Table 2 Distribution of COL1A1 rs2075555 genotypes among high myopia cases and controls included in the Meta-analysis No. Case Control First author P for HWE Case Control AA AC CC AA AC CC Inamori Y[19] 330 330 34 166 128 52 176 98 0.07 Liang CL[29] 471 623 36 161 183 53 269 296 0.46 Vatavuk Z[28] 17 794 0 5 12 19 210 565 0.92 Nakanishi H[21] 427 420 65 194 167 72 189 158 0.23 Zhang D[27] 697 762 91 333 273 79 374 309 0.03 COL1A1: Collagen type I alpha 1; HWE: Hardy-Weinberg equilibrium.

Figure2ForestplotoftheassociationbetweenCOL1A1rs2075555polymorphismandsusceptibilitytohighmyopiainthe heterozygotemodel(AC AA)indifferentethnicities. controlsofonestudyshowedsignificantdeviationfromthe 0.84-1.22);(CC AC/AA:OR=1.06,95%CI0.93-1.20); HWE,whichrequiredsensitivityanalysisandevaluationof (C A:OR=1.06,95%CI0.91-1.23).Inthestratified possibleselectionbias.AsevaluatedbytheNOS,two analysisbyethnicity,wefoundnosignificantassociations studieswerescoredas"moderatequality"andthreestudies betweenrs2075555polymorphismandhighmyopiaforany werescoredas''highquality'',indicatingthatthequalityof geneticmodels(Figure2). theincludedarticleswasacceptablefortheMeta-analysis. Test of Heterogeneity InourcurrentMeta-analysis, Individualcharacteristicsofthestudies,patients,andthe significantheterogeneitywasobservedin2geneticmodels controlgroupsareshowninTables1and2. (CC AAandC A)(Table3).Thesourcesofthe ResultsofMeta-analysis Theprimaryresultsofthe heterogeneitywereassessedbytheGalbraithplotofall Meta-analysisregardingrs2075555polymorphismandhigh includedstudies.Inamori 's [19] studywasidentifiedas myopiariskareshowninTable3.Overall,noobvious themainsourceofheterogeneityfortheassociationbetween associationsbetweenrs2075555polymorphismandhigh rs2075555polymorphismandhighmyopiasusceptibility myopiasusceptibilitywerefoundinanygeneticmodels(CC bothintheCC AAandtheC Amodels.Thepooled AA:OR=1.10;95%CI0.76-1.58);(AC AA:OR=0.98, ORsinthese2modelswerenotsignificantlyinfluenced 95%CI0.80-1.20);(CC/AC AA:OR=1.01,95%CI whenthisstudywasremoved. 606 陨灶贼允韵责澡贼澡葬造皂燥造熏灾燥造援 9熏晕燥援 4熏 Apr.18, 圆园16 www.ijo.cn 栽藻造押8629原愿圆圆源缘员苑圆 8629-82210956 耘皂葬蚤造押ijopress岳员远猿援糟燥皂 Table 3 Meta-analysis of COL1A1 rs2075555 polymorphism and susceptibility to high myopia Test of association Test of heterogeneity Comparisons Population N 2 OR (95%CI) Z P Model χ P I2 (%) CC vs AA Overall 5 1.10 (0.76-1.58) 0.49 0.62 R 10.10 0.04 60.4 (Homozygote) Asian 4 1.10 (0.75-1.62) 0.50 0.62 R 10.08 0.11 70.3 European 1 0.86 (0.05-15.1) 0.10 0.92 R - - - AC vs AA Overall 5 0.98 (0.80-1.20) 0.16 0.87 F 5.15 0.27 22.4 (Heterozygote) Asian 4 0.98 (0.80-1.20) 0.17 0.87 F 5.15 0.16 41.8 European 1 1.02 (0.05-19.1) 0.01 0.99 F - - - CC /AC vs AA Overall 5 1.01 (0.84-1.22) 0.12 0.90 F 7.76 0.10 48.4 (Dominant) Asian 4 1.01 (0.84-1.23) 0.13 0.90 F 7.75 0.05 61.3 European 1 0.88 (0.05-15.2) 0.09 0.93 F - - - CC vs AC/AA Overall 5 1.06 (0.93-1.20) 0.86 0.39 F 5.55 0.24 27.9 (Dominant) Asian 4 1.06 (0.93-1.20) 0.88 0.38 F 5.52 0.14 45.7 European 1 0.97 (0.34-2.80) 0.05 0.96 F - - - C vs A Overall 5 1.06 (0.91-1.23) 0.73 0.47 R 8.62 0.07 53.6 (Allele) Asian 4 1.06 (0.90-1.24) 0.69 0.49 R 8.61 0.04 65.2 European 1 1.07 (0.41-2.80) 0.14 0.89 R - - - COL1A1: Collagen type I alpha 1; OR: Odds ratio; R: Random-effect model; F: Fixed-effect model.

Figure3SensitivityanalysisperformedtoevaluatetheinfluenceofasinglestudyonthepooledORsintheheterozygotemodel (AC AA).

SensitivityAnalysis Asinglestudyincludedinthecurrent publicationbiasinanygeneticmodel.Theseresultsindicate Meta-analysiswasexcludedatthetimeofsensitivity thatthereisnoevidenceofpublicationbiasinour analysis.Similarresultswererevealed,andpooledORswere Meta-analysis. notobviouslyalteredinanyofthegeneticmodels(Figure3) DISCUSSION showsthesensitivityanalysisresultsintheCA AA AlthoughtheassociationbetweenCOL1A1rs2075555 model),suggestingthereliabilityandstabilityofourresults. polymorphismandhighmyopiahasbeenreportedmany PublicationBias Publicationbiaswasevaluatedusing timesinseveralgeographiclocations,theresultsgenerally Begg'sfunnelplotandEgger'stest.AsFigure4shows,the wereconflictingratherthanconsistent [18,20,27-29].Different funnelplotshapefortheheterozygotemodel(CA AA) geneticbackgrounds,sourcesofcontrols,andstudydesigns seemsapproximatelysymmetrical,whichsuggeststhat mayberesponsiblefortheseinconsistenciesamongdifferent publicationbiasinthisstudycanbeneglected.Inaddition, studies.Anindividualstudywithasmallsamplesizemaybe Egger'stestprovidesfurtherstatisticalevidenceofthefunnel underpoweredtoevaluatetherelationshipbetweengene plots'symmetry(=0.712).Overall,neitherBegg'sfunnel mutationandhighmyopia,butpoolingalleligibledataby plotnorEgger'stestsuggestedanystatisticallysignificant Meta-analysiscanpossiblyprovidemorepowerfuland 607 COL1A1polymorphismandhighmyopia Meta-analysiswereconductedinJapanesepatients,andthe samplesizewastoosmalltoprovideapowerfulandprecise estimateforCOL1A1rs2075555polymorphismandhigh myopiarisk.Incontrast,ourcurrentMeta-analysiscombined alltheeligiblepublishedstudieswecouldidentifyandused enhancedstatistical methodstoprovidemore reliable evidencethatCOL1A1rs2075555polymorphismdoesnot contributetohighmyopiasusceptibility,evenwhendifferent ethnicitiesaretakenintoconsideration. SeveralpotentiallimitationsofourMeta-analysisshouldbe considered. Firstly,althoughastatisticallysignificant publicationbiaswasnotfoundinallthegeneticmodels, publicationbiascouldstillhaveoccurredamongthe Figure 4 Begg's funnel plot of COL1A1 rs2075555 publishedarticlesmeetingtheinclusioncriteria.Secondly,a polymorphism and high myopia susceptibility in the heterozygotemodel(AC AA). lackofsufficientrawdatasuchasagedistributionandthe sexofthepatientsincludedintheidentifiedarticleslimited credibleevidence.ThecurrentMeta-analysisevaluatingthe furtherexplorationofpotentialinteractions,andmoreprecise possibleeffectoftheCOL1A1rs2075555polymorphismon andconvincinganalysisshouldbeperformed.Thirdly,allof highmyopiasusceptibilitywasbasedonasubstantialamount thestudiesincludedinourstudywereconductedinAsians ofdatafrom5individualcase-controlstudiesthatincludeda andEuropeanscohorts,and,thus,conclusionsfromour Meta-analysismayberestrictedtothesetwopopulations. totalof1942highmyopiacasesand2929controls. However,someadvantagesofthecurrentMeta-analysiscan Unfortunately,novalidassociationbetweenCOL1A1 behighlighted.Firstly,basedonthestrictcriteriaof rs2075555polymorphismandhighmyopiawasdetected inclusionandexclusion,alltheincludedstudieswereof accordingtothepooledORsinthedifferentgeneticcontrast acceptablequality,whichsignificantlyenhancesthe models.Populationstratificationisafactorthatcanleadto statisticalpowerofour Meta-analysis.Secondly,we falseevidenceabouttheassociationbetweengenemarkers attemptedto obtainadjustedestimatesindifferent andhighmyopia.Previousstudieshavedemonstratedthat populationsbyperformingasubgroupanalysisbetween theprevalenceofmyopiavariesaccordingtoethnicity[39] and differentethnicities.Thirdly,no evidenceofobvious ismuchhigherintheAsianpopulation [40-41],especiallyin publicationbiaswasdetectedforanyofthegeneticmodels, China[42].Becauseofthis,stratifiedanalysisbyethnicitywas whichsuggeststhatourresultsareunbiasedandstatistically performedbutdidnotrevealasignificantrelationbetween reliable. theCOL1A1rs2075555polymorphismandhighmyopiain Insummary,ourMeta-analysissuggeststhattheCOL1A1 eithertheAsianorEuropeanpopulations.Theresultsofthis rs2075555polymorphismisnotariskfactorfor studysuggestthatethnicityisnotafactorthatinfluencesthe susceptibilitytohighmyopia.Nevertheless,highmyopiaisa relationshipbetweenCOL1A1rs2075555polymorphismand complexmultifactorialeyediseasethatisinfluencedbyboth highmyopiasusceptibility. geneticand environmentalfactorsaswellastheir [21] ApreviousMeta-analysisperformedbyNakanishi interactions.Unfortunately,duetoalackofsufficientraw showedalimitedsignificancebetweenCOL1A1rs2075555 data fromtheincluded studies,interactionssuchas polymorphismandhighmyopiasusceptibility(OR=1.19; gene-gene,SNP-SNP,andgene-environmentcouldnotbe 95%CI1.03-1.38, <0.05)whichisaresultthatis evaluated.Furtherrelatedresearchwithlargersamplesizes inconsistent with ourfindings.However,intheir andenhanceddesignsareneeded. [21] Meta-analysis,Nakanishi combineddataoftheir ACKNOWLEDGEMENTS ownwithdatafromapreviouslypublishedJapanesestudy ConflictsofInterest:JinGM, None; ZhaoXJ, None; thatwasthefirstreportedpositiveassociation.Asweknow, ChenAM, None; ChenYX, None; LiQ, None. thereportedORsinthefirstpositivestudiesareusually REFERENCES higherthanORsreportedinsubsequentreplicationstudies[43], 1HeM,XuJ,YinQ,EllweinLB.Needandchallengesofrefractive sotherewasapossiblecorrelationbetweenCOL1A1 correctioninurbanChineseschoolchildren. 2005;82(4): rs2075555polymorphismandhighmyopiasusceptibility [19]. 229-234. Inaddition,itispossiblethatpublicationbiasaffectedthe 2LamCS,LamCH,ChengSC,ChanLY.Prevalenceofmyopiaamong HongKongChineseschoolchildren:changesovertwodecades. resultsinthefirstpositivestudy,andtheactualORofthe 2012;32(1):17-24. SNPwasoverestimatedintheiranalysis,whichincluded 3PanCW,RamamurthyD,SawSM.Worldwideprevalenceandriskfactors only2studies.Lastly,bothofthestudiesincludedintheir formyopia. 2012;32(1):3-16. 608 陨灶贼允韵责澡贼澡葬造皂燥造熏灾燥造援 9熏晕燥援 4熏 Apr.18, 圆园16 www.ijo.cn 栽藻造押8629原愿圆圆源缘员苑圆 8629-82210956 耘皂葬蚤造押ijopress岳员远猿援糟燥皂

4VitaleS,SperdutoRD,FerrisFL3rd.Increasedprevalenceofmyopiain 2008;49(1):49-54. theUnitedStatesbetween1971-1972and1999-2004. 23PaluruP,RonanSM,HeonE,DevotoM,WildenbergSC,ScavelloG, 2009;127(12):1632-1639. HolleschauA,MakitieO,ColeWG,KingRA,YoungTL.Newlocusfor 5SawSM,TongL,ChuaWH,ChiaKS,KohD,TanDT,KatzJ.Incidence autosomaldominanthighmyopiamapstothelongarmofchromosome17. andprogressionofmyopiainSingaporeanschoolchildren. 2003;44(5):1830-1836. 2005;46(1):51-57. 24DalgleishR.ThehumantypeIcollagenmutationdatabase. 6HeM,ZhengY,XiangF.Prevalenceofmyopiainurbanandrural 1997;25(1):181-187. childreninmainlandChina. 2009;86(1):40-44. 25WangQ,ZhaoG,XingS,ZhangL,YangX.Roleofbonemorphogenetic 7WickremasingheS,FosterPJ,UranchimegD,LeePS,DevereuxJG, proteinsinform-deprivationmyopiasclera. 2011;17:647-657. AlsbirkPH,MachinD,JohnsonGJ,BaasanhuJ.Ocularbiometryand 26LinZ,ChenX,GeJ,CuiD,WuJ,TangF,TanJ,ZhongX,GaoQ. refractioninMongolianadults. 2004;45(3): Effectsofdirectintravitrealdopamineinjectiononscleraandretinain 776-783. form-deprivedmyopicrabbits. 2008;24(6):543-550. 8SawSM,GazzardG,Shih-YenEC,ChuaWH.Myopiaandassociated 27ZhangD,ShiY,GongB, AnassociationstudyoftheCOL1A1 pathologicalcomplications. 2005;25(5):381-391. geneandhighmyopiainaHanChinesepopulation. 2011;17: 9BuchH,VindingT,LaCourM,AppleyardM,JensenGB,NielsenNV. 3379-3383. Prevalenceandcausesofvisualimpairmentandblindnessamong9980 28VatavukZ,SkuncaHermanJ,BencicG, Commonvariantin Scandinavianadults:theCopenhagenCityEyeStudy. myocilingeneisassociatedwithhighmyopiainisolatedpopulationof 2004;111(1):53-61. KorculaIsland,Croatia. 2009;50(1):17-22. 10FanQ,WojciechowskiR,KamranIkramM,ChengCY,ChenP,Zhou 29LiangCL,HungKS,TsaiYY,ChangWS,WangHS,JuoSHH. X,PanCW,KhorCC,TaiES,AungT,WongTY,TeoYY,SawSM. SystematicassessmentofthetaggingpolymorphismsoftheCOL1A1gene Educationinfluencestheassociationbetweengeneticvariantsand forhighmyopia. 2007;52(4):374-377. refractiveerror:ameta-analysisoffiveSingaporestudies. 30DouH,MaE,YinL,JinY,WangH.Theassociationbetweengene 2014;23(2):546-554. polymorphismofTCF7L2andtype2diabetesinChineseHanpopulation:a 11SiegwartJT,Jr.,NortonTT.Perspective:howmightemmetropization meta-analysis. 2013;8(3):e59495. andgeneticfactorsproducemyopiainnormaleyes? 2011; 31deCarvalhoTB,SumanM,MolinaFD,PiattoVB,ManigliaJV. 88(3):E365-372. Relationshipofobstructivesleepapneasyndromewiththe5-HT2A 12KhorCC,FanQ,GohL,TanD,YoungTL,LiYJ,SeielstadM,GohDL, receptorgeneinBrazilianpatients. 2013;17(1):57-62. SawSM.SupportforTGFB1asasusceptibilitygeneforhighmyopiain 32HigginsJP,ThompsonSG.Quantifyingheterogeneityina individualsofChinesedescent. 2010;128(8):1081-1084. meta-analysis. 2002;21(11):1539-1558. 13ZhuG,HewittAW,RuddleJB,KearnsLS,BrownSA,MackinnonJR, 33HigginsJP,ThompsonSG,DeeksJJ,AltmanDG.Measuring ChenCY,HammondCJ,CraigJE,MontgomeryGW,MartinNG,Mackey inconsistencyinmeta-analyses. 2003;327(7414):557-560. DA.Geneticdissectionofmyopia:evidenceforlinkageofocularaxial 34DerSimonianR,LairdN.Meta-analysisinclinicaltrials. lengthtochromosome5q. 2008;115(6):1053-1057. 1986;7(3):177-188. 14HammondCJ,AndrewT,MakYT,SpectorTD.Asusceptibilitylocus 35MantelN,HaenszelW.Statisticalaspectsoftheanalysisofdatafrom formyopiainthenormalpopulationislinkedtothePAX6generegionon retrospectivestudiesofdisease. 1959;22(4):719-748. chromosome11:agenomewidescanofdizygotictwins. 36GalbraithRF.Anoteongraphicalpresentationofestimatedoddsratios 2004;75(2):294-304. fromseveralclinicaltrials. 1988;7(8):889-894. 15KleinAP,DuggalP,LeeKE,KleinR,Bailey-WilsonJE,KleinBE. 37PatsopoulosNA,EvangelouE,IoannidisJP.Sensitivityof Supportforpolygenicinfluencesonocularrefractiveerror. between-studyheterogeneityinmeta-analysis:proposedmetricsand 2005;46(2):442-446. empiricalevaluation. 2008;37(5):1148-1157. 16MorganIG.Thebiologicalbasisofmyopicrefractiveerror. 38BeggCB,MazumdarM.Operatingcharacteristicsofarankcorrelation 2003;86(5):276-288. 17LiaoX,YangXB,LiaoM,LanCJ,LiuLQ.Associationbetweenlumican testforpublicationbias. 1994;50(4):1088-1101. gene-1554T/CpolymorphismandhighmyopiainAsianpopulation:a 39KempenJH,MitchellP,LeeKE,TielschJM,BromanAT,TaylorHR, meta-analysis. 2013;6(5):696-701. IkramMK,CongdonNG,O'ColmainBJ.Theprevalenceofrefractiveerrors 18AhmedI,RasoolS,JanT,QureshiT,NaykooNA,AndrabiKI.TGIF1is amongadultsintheUnitedStates,WesternEurope,andAustralia. apotentialcandidategeneforhighmyopiainethnicKashmiripopulation. 2004;122(4):495-505. 2013;39(3):282-290.. 40SawadaA,TomidokoroA,AraieM,IwaseA,YamamotoT.Refractive 19InamoriY,OtaM,InokoH,OkadaE,NishizakiR,ShiotaT,MokJ,Oka errorsinanelderlyJapanesepopulation:theTajimistudy. A,OhnoS,MizukiN.TheCOL1A1geneandhighmyopiasusceptibilityin 2008;115(2):363-370. Japanese. 2007;122(2):151-157. 41PanCW,WongTY,LavanyaR,WuRY,ZhengYF,LinXY,MitchellP, 20ParkSH,MokJ,JooCK.Absenceofanassociationbetweenlumican AungT,SawSM.Prevalenceandriskfactorsforrefractiveerrorsin promotervariantsandhighmyopiaintheKoreanpopulation. Indians:theSingaporeIndianEyeStudy(SINDI). 2013;34(1-2):43-47. 2011;52(6):3166-3173. 21NakanishiH,YamadaR,GotohN,HayashiH,OtaniA,TsujikawaA, 42LiuHH,XuL,WangYX,WangS,YouQS,JonasJB.Prevalenceand YamashiroK,ShimadaN,Ohno-MatsuiK,MochizukiM,SaitoM,SaitoK, progressionofmyopicretinopathyinChineseadults:theBeijingEyeStudy. IidaT,MatsudaF,YoshimuraN.AbsenceofassociationbetweenCOL1A1 2010;117(9):1763-1768. polymorphismsandhighmyopiaintheJapanesepopulation. 43LohmuellerKE,PearceCL,PikeM,LanderES,HirschhornJN. 2009;50(2):544-550. Meta-analysisofgeneticassociationstudiessupportsacontributionof 22PertileKK,SchacheM,IslamFM,ChenCY,DiraniM,MitchellP, commonvariantstosusceptibilitytocommondisease. 2003;33 BairdPN.AssessmentofTGIFasacandidategeneformyopia. (2):177-182. 609