ORIGINALARTICLE
Heritabilityofborderlinepersonalitydisorderfeaturesissimilaracrossthreecountries
M.A.Distel1*,T.J.Trull2,C.A.Derom3,E.W.Thiery4,M.A.Grimmer5,N.G.Martin5,G.Willemsen1andD.I.Boomsma112DepartmentofBiologicalPsychology,VUUniversity,Amsterdam,TheNetherlands
DepartmentofPsychologicalSciences,UniversityofMissouri-Columbia,Columbia,MO,USA3DepartmentofHumanGenetics,KatholiekeUniversiteitLeuven,Leuven,Belgium4AssociationforScientificResearchinMultipleBirths,Gent,Belgium5QueenslandInstituteofMedicalResearch,Brisbane,Australia
Background.Mostofourknowledgeaboutborderlinepersonalitydisorderfeatureshasbeenobtainedthroughthestudyofclinicalsamples.Althoughthesestudiesareimportantintheirownright,theyarelimitedintheirabilitytoaddresscertainimportantepidemiologicalandaetiologicalquestionssuchasthedegreetowhichthereisageneticinfluenceonthemanifestationofborderlinepersonalitydisorderfeatures.Thoughfamilyhistorystudiesofborderlinepersonalitydisorderindicategeneticinfluences,therehavebeenveryfewtwinstudiesandthedegreeofgeneticinfluenceonborderlinepersonalitydisorderremainsunclear.
Method.DataweredrawnfromtwinsamplesfromTheNetherlands(n=3918),Belgium(n=904)andAustralia(n=674).Intotal,datawereavailableon5496twinsbetweentheagesof18and86yearsfrom3644familieswhoparticipatedinthestudybycompletionofamailedself-reportquestionnaireonborderlinepersonalitydisorderfeatures.
Results.Inallcountries,femalesscoredhigherthanmalesandtherewasageneraltendencyforyoungeradultstoendorsemoreborderlinepersonalitydisorderfeaturesthanolderadults.Model-fittingresultsshowedthatadditivegeneticinfluencesexplain42%ofthevariationinborderlinepersonalitydisorderfeaturesinbothmenandwomenandthatthisheritabilityestimateissimilaracrossTheNetherlands,BelgiumandAustralia.Uniqueenvironmentalinfluencesexplaintheremaining58%ofthevariance.
Conclusion.GeneticfactorsplayaroleinindividualdifferencesinborderlinepersonalitydisorderfeaturesinWesternsociety.
Received17April2007;Revised21August2007;Accepted8September2007Keywords:Borderlinepersonalitydisorder,genetics,heritability,twinstudy.
Introduction
Borderlinepersonalitydisorder(BPD)isasevereper-sonalitydisorderwhosefeaturesincludeimpulsivity,affectiveinstability,relationshipproblemsandident-ityproblems(APA,2000).BPDisassociatedwithin-terpersonalandoccupationalimpairment,increasedriskforsuicideandhigherratesoftreatmentinbothmedicalandpsychiatricsettings(Skodoletal.2002).Inaddition,BPDisfrequentlyco-morbidwithAxisIdisorders,especiallysubstanceusedisordersinmales,eatingdisordersinfemales,anxietydisordersandmooddisorders(Zanarinietal.1998;Zimmerman&
*Addressforcorrespondence:M.A.Distel,M.Sc.,VUUniversityAmsterdam,DepartmentofBiologicalPsychology,vanderBoechorststraat1,1081BTAmsterdam,TheNetherlands.(Email:ma.distel@psy.vu.nl)
Mattia,1999),andthisco-morbiditypredictspoorershort-andlong-termoutcome(Skodoletal.2002).MostofourknowledgeaboutBPDhasbeenob-tainedthroughthestudyofclinicalsamples.Clinicalsamplesareimportantforcharacterizingthesyn-dromeasittypicallyispresentedfortreatment,assessingthelongitudinalcourseofthedisorder,andevaluatingthedisorder’sresponsetoformsoftreat-ment.However,clinicalsamplesarelimitedintheirabilitytoaddresscertainimportantepidemiologicalandaetiologicalquestionsastheyarelikelytocontainmoreseverecasesandmaythereforenotberepresen-tativeofthedisorderasitappearsinthegeneralpopulation.Also,theseclinicalcasesoftenexhibitmoreco-morbiditythancasesfromthecommunity(Skodoletal.2002),therebyfurthercloudingtheaetiologicalpicture.Inadditiontoclinicalstudies,itisthereforeinformativetoidentifyBPDfeaturesinthe
2M.A.Disteletal.
generalpopulationtogainafullunderstandingofthenatureofBPDandthedevelopmentalpathwaysleadingtoBPD.
Oneimportantaetiologicalissueforwhichcom-munitysamplesareessentialisthedegreetowhichthereisageneticinfluenceonthemanifestationofBPDsymptoms.IncreasedratesofBPDhavebeenfoundintherelativesofindividualswithBPD(e.g.Lorangeretal.1982;Baronetal.1985;Johnsonetal.1995;Zanarinietal.2004),andtheheritabilityoftraitsthatarehighlyassociatedwithBPD(e.g.neuroticism,negativeemotionality)iswelldocumented(Nigg&Goldsmith,1994).However,ourknowledgeofthegeneticinflu-enceonBPDsymptomsandfeaturesisratherlimited.OnlytwotwinstudiessofarhaveprovideddataonBPDdiagnosesandfeatures.Torgersen(1984)re-portedamonozygotic(MZ)concordancerateof0.0%andadizygotic(DZ)concordancerateof11.1%forBPD,suggestingthatsharedenvironmentalfactorsinfluencethevarianceinBPD.However,method-ologicalproblemsofthatstudylimitanyconclusions.Morerecently,Torgersenetal.(2000)reportedonthelargesttwinstudytodate(n=221twinpairs)thatexaminedBPD.ResultssuggestedageneticliabilityforBPDof69%,thoughthisheritabilityestimatemustbeconsideredapproximateduetothesmallnumberoftwins,theascertainmentmethod(samplingthosewhoweretreatedformentaldisorder)andthefactthatthezygosityanddiagnosticstatusofco-twinswasnothiddenfromtheinterviewers.
ToextendtheworkofTorgersenetal.(2000),weinitiatedatwinstudyofBPDfeaturesinthegeneralpopulation.Specifically,wesoughttoassessalargenumberofcommunity-basedadulttwinsfromawideagerangeandfrommultiplecountries.Inthisway,wewereabletoprovidepreciseestimatesofthegeneticinfluenceonBPDfeatures,totestforquantitativeandqualitativesexdifferencesandtodeterminewhetherourestimateswereconsistentacrossTheNetherlands,BelgiumandAustralia.MethodParticipants
DatawerecollectedaspartofaninternationalprojectonBPDfeaturesinDutch,BelgianandAustraliantwincohorts.Protocolsinallthreecountrieswereapprovedbytheirrespectiveethicscommittees.TheNetherlands
InTheNetherlands,thisstudyispartofanongoingstudyonhealthandlifestyleintwinfamiliesregis-teredwithTheNetherlandsTwinRegister(Boomsmaetal.2002b,2006;Vinketal.2004;Stubbeetal.2005).
Surveysonhealthandlifestyleweresenttothetwinfamiliesevery2–3years.Forthisstudy,datafromtheseventhsurveywereusedwhichwassentin2004–2005.Atotalof12785twinsfrom6764familieswereapproached,ofwhomsomeindividualshadparticipatedbefore(n=7712)andsomehadneverparticipated(n=5073).Intotal,4017(31%)twinsre-turnedthesurvey.Toexaminereasonsfornotparti-cipating,weperformedanon-responsestudybycontactingbytelephonetwosubgroupsofnon-respondents;non-respondentswhohadparticipatedbeforeandnon-respondentswhohadneverpartici-pated.Addressesprovedincorrectin23.8%and42.0%ofthetwogroups,respectively;thusasubstantialgroupoftargetedparticipantsneverreceivedthequestionnaire.Aftersubtractingtheestimatednumberofincorrectaddressesfromthenumberofsentques-tionnaires,theestimated‘true’responseratesforthetwogroupswere52.2%and13.6%,respectively.Thepair-wiseresponserateofthetargetedtwinswhohadandhadnotparticipatedbeforewas33.6%and6.2%,respectively.Detailsonresponseratesanddemo-graphiccharacteristicsofthesamplecanbefoundelsewhere(Disteletal.2007).ForasubsampleoftheDutchparticipantsretestdatawereavailable.At6monthsafterthefirstquestionnairewassent,theretestsurveywassentto240twins,siblingsandparents(oneperfamily),ofwhom199(83%)completedthequestionnaireasecondtime.Belgium
Dutch-speakingtwinsinBelgiumwereaskedtotakepartintheDutchhealthandlifestylestudy.BelgianparticipantswererecruitedthroughtheEastFlandersProspectiveTwinSurvey(EFPTS),apopulation-basedregisterofmultiplebirthsintheBelgianprovinceofEastFlanderswhichwasstartedin1964.Multiplesareascertainedatbirth.Basicperinataldata,choriontypeandzygosityhavebeenestablished(Loosetal.1998;Derometal.2006).Youngadulttwinswerecontactedbymailandinvitedtocompleteasurveywhichwasenclosedwiththeletter.Atotalof3979twinswereapproached,ofwhom932(23%)twinsreturnedthesurvey.AsmosttargetedBelgianparticipantshadnotparticipatedinastudyoftheEFPTSbefore,itisunknowntowhatextentaddresseswerecorrect.Thepair-wiseresponseratewas15.7%.Australia
AustraliansubjectsweredrawnfromtheAustralianTwinRegister(ATR)foundedin1978(Jardineetal.1984),aswellasfromatwingrouppreviouslyre-cruitedbytheQueenslandInstituteofMedicalResearch(QIMR).TwinsapproachedbytheATRwere
askedtoparticipateinthePersonalityFeaturesinAdulthoodstudy;thiswasrenamedHealth,LifestyleandPersonalitystudyfortheQIMRapproach.TargetedparticipantsincludedAustraliantwinsbornbetween1972and1987andwereinvitedbymailtoparticipateinthestudy.Atotalof155completeATRtwinpairs’(310twinindividuals)contactdetailswereforwardedtoQIMRforapproachwithdetailsforcompletingthesurveyeitheronlineoronpaper;268ofthe310twins(86.4%)completedthesurvey.Ofthe808twinsapproacheddirectlybyQIMR,431(53.3%)completedthesurvey,resultinginatotalof699com-pletedsurveys(493online,206paper).Thepair-wiseresponseratewas50.6%.Demographics
ThemeanageoftheDutchtwinswas34.9years(S.D.=11.6,range19–86years),oftheBelgiantwins28.4years(S.D.=6.9,range18–67years)andoftheAustraliantwins23.1years(S.D.=3.74,range18–33years).Triplets(n=51),twinswithunknownzygosity(n=55)orage(n=9)andtwinswithoutavalidscoreonthePersonalityAssessmentInventory-BorderlineFeaturesscale(PAI-BOR)(n=37)wereexcluded.Thisresultedinatotalsampleforanalysisof5496partici-pantsfrom3644families.Zygosity
InTheNetherlands,thezygosityof3135same-sextwinswasdeterminedeitherfromDNApolymor-phism(n=1203)orfromself-reportanswerstosurveyquestionsonphysicaltwinresemblanceandconfusionofthetwinsbyfamilymembersandstrangers(n=1932).Basedontheanswerstotheseitemsfromallavailablesurveys,zygositywasassigned.Whentherewereinconsistenciesovertimeorpersonsreporting,themajorityofthezygosityjudgementsdeterminedthefinaloutcome.Atotalof783twinswereofop-positesexandthereforeclassifiedasDZ(seeWillemsenetal.2005).
InBelgium,twinzygositywasdeterminedthroughsequentialanalysisbasedonsex,fetalmembranes,umbilicalcordbloodgroupsandplacentalalkalinephosphataseuntil1985.Afterthattime,DNAfinger-printingwasused.IncaseofmissingorinsufficientDNAinformation,thezygosityofthesame-sexDZtwinswasbasedonsurveyitemsonphysicaltwinre-semblanceandconfusionofthetwins(seeDerom&Derom,2005).
InAustralia,thezygosityof674twinswasdeter-minedeitherfromself-reportanswerstostandardquestions(n=299),becausethetwinswereofoppositesex(n=91)orfromDNAtesting(n=284)(seeNyholt,2006).
Heritabilityofborderlinepersonalitydisorderfeatures
3
Measures
BorderlinepersonalityfeaturesweremeasuredbythePersonalityAssessmentInventory–BorderlineFeaturesscale(PAI-BOR;Morey,1991).PAI-BORitemstapfeaturesofseverepersonalitypathologythatareclinicallyassociatedwithBPD.BasedonareviewofthehistoricalconceptualizationsofBPD,aswellasonempiricalstudiesofborderlinepatients,potentialPAI-BORitemsweregeneratedtoreflectcorefactorsoftheconstruct(affectiveinstability,identityproblems,negativerelationshipsandself-harm/impulsivity)(Morey,1991).Finalselectionofitemswasguidedbyboththeconceptualnatureoftheitemsaswellastheitems’psychometricproper-ties.ThefinalversionofthePAI-BORconsistsof24itemsthatareratedonafour-pointscale(0to3:false;slightlytrue;mainlytrue;verytrue).PreliminarystudieshavesupportedthereliabilityandthevalidityoftotalPAI-BORscoresinindexingthedegreetowhichborderlinepersonalityfeaturesarepresent(Morey,1988,1991;Trull,1995,2001).Kurtz&Morey(2001)forexampleshowedthatPAI-BORscorescor-related0.78withastructuredinterview-basedas-sessmentofBPD,indicatinghighconvergentvalidity.Morey(1991)alsopresenteddatasupportingtheval-idityofthefourPAI-BORsubscales,andthePAI-BORhasbeenusedinanumberofstudiesofnon-clinicalparticipantsaswell(Trull,1995,2001).ThePAI-BORwasscoredaccordingtoMorey’stestmanual(Morey,1991),whichstatesthatatleast80%oftheitemsmustbeansweredtocalculateasumscoreandthatmissingandambiguousanswersshouldbesubstitutedbyazeroscore.InTheNetherlandsandBelgium,theDutchadaptationofthePAI-BORwasused.TheEnglishPAI-BORwastranslatedintoDutchandtranslatedbackintoEnglishbyanativeEnglish-speakingtranslator.TheDutchtranslationofthePAI-BORwasreviewedandapprovedbythetestauthorandpublishingcompany(PsychologicalAssessmentResources).Statisticalanalysis
Twinstudiesmakeuseofthegeneticrelatednessoftwinsandtheirfamilymembers.MZtwinsaregeneticallyidenticalwhileDZtwinsshareonaverage50%oftheirsegregatinggenes,likeothersiblings(Boomsmaetal.2002a).ComparingtheresemblanceinBPDfeatureswithinMZtwinpairswiththeresem-blanceinBPDfeatureswithinDZtwinpairsprovidesinformationofhowtoexplainindividualdifferencesinBPDfeatures.
Additivegeneticeffects(A)aresuggestedifthecorrelationinMZtwinsislargerthanthecorrelationinDZtwins.WhentheDZcorrelationismore
4M.A.Disteletal.
thanhalftheMZcorrelation,thereisevidenceforenvironmentaleffectssharedbytwinsfromthesamefamily(C)butwhentheDZcorrelationislessthanhalftheMZcorrelation,thereisevidencefornon-additivegeneticeffects(dominance;D).DifferencesinBPDfeaturescoreswithinMZtwinpairsareduetouniqueenvironmentalinfluences(E),whichalsoincludemeasurementerror.TheobservedvarianceinBPDfeaturescanthusbedecomposedinfourpossiblesourcesofvariance;A,D,CandE(Neale&Cardon,1992)buttheobservedvariancesandcovariancesonlyprovideenoughinformationtomodeleitheranACEmodeloranADEmodel.Basedonthepatternoftwincorrelations(seeResultssection),A,DandEweremodelledinthisstudy.
StatisticalanalyseswereperformedusingstructuralequationmodellingasimplementedinthesoftwarepackageMx(VirginiaCommonwealthUniversity,Richmond,VA;Nealeetal.2003).TherawdatafullinformationmaximumlikelihoodapproachinMxwasusedtofitdifferentmodelstothedata.Testingofsubmodelswasdonebymeansoflikelihood-ratiotests,bysubtractingthenegativelog-likelihood(x2LL)forthemorerestrictedmodelfromthex2LLforthemoregeneralmodel.Thisyieldsastatisticthatisdistributedasx2withdegreesoffreedom(df)equaltothedifferenceinthenumberofparametersinthetwomodels.Ifthex2testyieldsapvaluehigherthan0.01,theconstrainedmodelisdeemednotsignifi-cantlyworsethanthepreviousmodelandiskeptasthemostparsimoniousmodeltowhichthenextmodelwillbecompared.Inaddition,Akaike’sInformationCriterion(AIC;Akaike,1987)(x2x2df)wasevaluatedbecauseitreflectsboththegoodnessoffitandtheparsimonyofthemodel.ThelowertheAICvalue,thebetterthefitofthemodelrelativetothenumberofparametersestimated.
Wefirstfittedasaturatedmodelforeachcountryseparatelyinwhichvariances,covariancesandmeanswereestimated.Zygositygroupswereseparatedbysexandbothageandsexwereincludedinthemeansmodelasacovariate.WetestedforhomogeneityofmeansandvariancesforMZtwinsandDZtwinsandforfixedeffectsofageandsexonBPDfeatures.Finallywetestedforquantitativesexdifferencesbyconstrain-ingthecorrelationsbetweenmenandwomenwithinzygositytobeequal,andforqualitativesexdifferencesbyconstrainingtheDZsame-sexcorrelationtoequaltheDZopposite-sextwincorrelations,whichimpliesthatthegeneticcorrelationforbothDZsame-sexandDZopposite-sextwinsis0.5.Foreachcountrythemostparsimoniousmodelwasretainedforsimul-taneousanalysisofdatafromthethreecountries.Wetestedfordifferencesinmeans,standarddeviationsandcorrelationstructurebetweenthethreecountries.
Toobtaintheestimatedproportionofvarianceex-plainedbyA,DandE,simultaneousgeneticanalysesofthedatafromthethreecountrieswerecarriedout.ThefirstmodeldecomposedthevarianceofBPDfeaturesintoA,DandEwithdifferentparameterestimatesforeachcountry.Next,wetestedthesig-nificanceofAandDseparatelybyconstrainingtheseparameterstozeroineachcountry.Finally,wecon-strainedthestandardizedestimatestobeequalacrossthecountriestoobtainpooledestimatesofthevari-ancecomponentsexplainingindividualdifferencesinBPDfeatures.Results
The6-monthtest–retestcorrelationoftheDutchPAI-BORwas0.78andtheinternalconsistencies(Cronbach’sa)ofthePAI-BORitemsintheDutchandBelgiumsampleswereboth0.84,suggestingthattheDutchtranslationofthePAI-BORisareliablemeasure.TheinternalconsistencyofthePAI-BORitemsintheAustraliansamplewas0.87.
AccordingtoMorey,atotalPAI-BORscoreof38ormoreindicatesthepresenceofsignificantBPDfea-tures,whereasascoreof60ormoreindicatesalikelyDiagnosticandStatisticalManualofMentalDisorders(DSM)-IVBPDdiagnosis.Thesampleprevalenceofsignificantborderlinefeatureswas2.2%inTheNetherlands,4.0%inBelgiumand5.3%inAustralia,whileaBPDdiagnosiswassuggestedfor0.03%inTheNetherlands,0.1%inBelgiumand0.7%inAustralia.ThesomewhathigherprevalenceinAustraliacouldbeduetotheyoungeragerangeoftheAustraliansampleastheprevalenceratesofBPDareknowntobehighestamongyoungadults(Parisetal.1987;Stone,1990;Bernsteinetal.1996;APA,2000;Johnsonetal.2000;Samuelsetal.2002;Coidetal.2006).
Becausethedatashowedasomewhatskeweddis-tribution,asquarerootdatatransformationwasper-formed.Table1displaysthetwincorrelationsusingthetransformeddata.Themeanborderlinescoresformalesandfemales(correctedforage),thestandarddeviationsandageregressioneffectsinTable1werebasedontheraw,untransformeddata.
ResultsofthetestsperformedinthesaturatedmodelsforeachcountryareshowninTable2.Ineachcountrymeanborderlinescoresdidnotdiffersignifi-cantlybetweenMZandDZtwins.Standarddevi-ationswereequalbetweenmalesandfemalesandDZandMZtwins.Sexeffectsonthemeansweresignifi-cantintheDutchsample,wherewomenscoredonaverage1.94pointshigherthanmen.Thesamedirec-tionofeffectwasobservedintheothertwosamples,butduetothesmallersamplesizetheseeffectswerenotsignificant.Theageregressioncoefficientsonthe
Heritabilityofborderlinepersonalitydisorderfeatures
Table1.Numberofparticipantsfromcompleteandincompletetwinpairsineachzygositygroup,themaximumlikelihoodestimatesoftwincorrelationsa(95%CIs)andestimatesformeanborderlinescoresformalesandfemales,standarddeviationsandageregression
Dutchtwins
Monozygoticmales
n,Fromcomplete(incomplete)twinpairs
Maximumlikelihoodestimate(95%CI)Dizygoticmales
n,Fromcomplete(incomplete)twinpairs
Maximumlikelihoodestimate(95%CI)Monozygoticfemales
n,Fromcomplete(incomplete)twinpairs
Maximumlikelihoodestimate(95%CI)Dizygoticfemales
n,Fromcomplete(incomplete)twinpairs
Maximumlikelihoodestimate(95%CI)Dizygoticoppositesex
n,Fromcomplete(incomplete)twinpairs
Maximumlikelihoodestimate(95%CI)Allmonozygotictwins
Maximumlikelihoodestimate(95%CI)bAlldizygotictwins
Maximumlikelihoodestimate(95%CI)bMeanscoremales,untransformed(transformed)cMeanscorefemales,untransformed(transformed)cStandarddeviation,untransformed(transformed)cRegressionofageperyear,untransformed(transformed)cBelgiantwins
Australiantwins
5
374(189)
0.46(0.34to0.56)
118(45)
0.48(0.23to0.64)
100(36)
0.28(x0.02to0.50)
154(151)
0.27(0.05to0.45)
32(33)
0.19(x0.25to0.53)
58(18)
0.12(x0.36to0.50)
1120(396)
0.42(0.35to0.48)
242(73)
0.43(0.28to0.56)
170(23)
0.49(0.32to0.62)
476(275)
0.11(x0.03to0.24)
86(59)
0.12(x0.21to0.40)
96(12)
0.32(x0.03to0.55)
410(373)
0.24(0.12to0.35)
142(74)
0.12(x0.11to0.33)
126(35)
0.16(x0.07to0.36)
0.43(0.37to0.48)0.45(0.32to0.55)0.43(0.28to0.55)
0.19(0.11to0.27)16.04(3.83)17.98(4.09)8.41(1.07)x0.07(x0.01)
0.13(x0.05to0.29)21.01(4.44)22.30(4.58)8.94(1.07)x0.25(x0.03)
0.22(0.05to0.32)21.43(4.52)22.94(4.63)9.86(1.10)x0.31(x0.03)
CI,Confidenceinterval.aCorrelationswereestimatedfromthesquareroot-transformeddata.bAfterconstrainingthesecorrelationstobeequal.cEstimatesaregivenfortheuntransformeddataandusingsquareroot-transformeddata.
meanswerenegativeinallsamples,indicatingthatBPDfeaturesdecreasewithage.TheageeffectwassignificantinTheNetherlandsandinBelgiumbutnotinAustralia.Forallthreecountries,thetwincorre-lationsforMZmalesandMZfemaleswereequalaswerethetwincorrelationsforDZmales,DZfemalesandDZopposite-sextwins.ThisindicatesthatthereisnosexdifferenceintheheritabilityofBPDfeaturesandthatthesamegenesinfluenceBPDfeaturesinmalesandfemales.
Table3showstheresultsofthesimultaneousmodellingofthedatafromthethreecountries.Mean
6M.A.Disteletal.
Table2.Saturatedmodel-fittingresultsforborderlinepersonalitydisorderfeaturesintheDutch,BelgianandAustraliantwindata
Test
TheNetherlands0.Saturatedmodel
1.MeanMZ=meanDZ
2.S.D.males=femalesandS.D.MZ=DZ3.Sexeffectonmean4.Ageeffectonmean
5.CorrelationMZM=MZF,DZM=DZF6.CorrelationDZM=DZF=DOSBelgium
0.Saturatedmodel
1.MeanMZ=meanDZ
2.S.D.males=femalesandS.D.MZ=DZ3.Sexeffectonmean4.Ageeffectonmean
5.CorrelationMZM=MZF,DZM=DZF6.CorrelationDZM=DZF=DOSAustralia
0.Saturatedmodel
1.MeanMZ=meanDZ
2.S.D.males=femalesandS.D.MZ=DZ3.Sexeffectonmean4.Ageeffectonmean
5.CorrelationMZM=MZF,DZM=DZF6.CorrelationDZM=DZF=DOS
x2LL
df
Dx2Ddf
p
AIC
1v.02v.13v.24v.25v.26v.5
11467.7511469.7811474.4911519.2211509.7811476.6611477.792627.952632.982642.712645.702672.312645.932645.971993.631996.172007.142008.642015.222017.812018.26
3899390039093910391039113912885886895896897898899655656665666667669670
2.024.7244.7335.292.171.131911210.160.860.000.000.340.290.02x13.2842.7333.29x1.83x0.87
1v.02v.13v.24v.35v.36v.55.039.742.9926.610.230.041911210.030.370.080.000.890.843.03x8.260.9924.61x3.77x1.96
1v.02v.13v.24v.35v.46v.52.5310.971.506.582.590.451911210.110.280.220.010.270.050.53x7.03x0.504.58x1.41x1.55
x2LL,x2log-likelihood;df,degreesoffreedom;AIC,Akaike’sInformationCriterion;S.D.,standarddeviation;MZ,monozygotic;DZ,dizygotic;MZM,monozygoticmale;MZF,monozygoticfemale;DZM,dizygoticmale;DZF,dizygoticfemale;DOS,dizygoticoppositesex.
Table3.Saturatedmodel-fittingresultsincludingthedatafromthreecountriesaTest
0.Saturatedmodel
1.MeansNL=BE=AU2.S.D.NL=BE=AU
3.TwincorrelationsNL=BE=AU
x2LL16130.9016207.0316131.8116132.52
df5478548054805484
Dx2Ddf
p
AIC
1v.02v.03v.276.120.910.712240.000.640.9572.12x3.09x7.29
x2LL,x2log-likelihood;df,degreesoffreedom;AIC,Akaike’sInformationCriterion;NL,TheNetherlands;BE,Belgium;AU,Australia;S.D.,standarddeviation.aEffectsofsexandagearemodelledforeachcountryseparately.
scoresdifferedbetweenthethreecountries,butstandarddeviationscouldbeequated.Thelowestmeanscore(correctedforage)wasfoundinTheNetherlands(16.04formalesand17.98forfemales).ThecorrelationsforMZtwinswereequalforTheNetherlands,BelgiumandAustraliaandthesamewastrueforDZtwins.
InthefullgeneticmodelwithoutsexdifferencesthevariancecomponentsA,DandEwereestimatedforthethreecountries,explaining34.3,8.4and57.3%ofthevarianceinBPDfeaturesintheDutchsample,6.7,37.8and55.5%intheBelgiumsampleand33.6,5.3and58.1%intheAustraliansample,respectively.Model-fittingresultsareshowninTable4.RemovingDfromthefullmodeldidnotgiveasignificantworseningofthegoodnessoffit(p=0.71)ofthemodelbutremovingAdid(p=0.00),resultinginmodel3asthebest-fittingmodel.Inaddition,theAICofmodel3waslowerthantheAICofmodels0,1and2,indi-catingthatmodel3wasthemostparsimoniousmodel.
Heritabilityofborderlinepersonalitydisorderfeatures
Table4.Geneticmodel-fittingresultsincludingthedatafromthreecountriesaTest
0.ADE
1.AEforeachcountry2.Eforeachcountry
3.StandardizedestimatesAandEequal,NL=BE=AUbx2LL16130.9016132.2916356.1616132.30
df5478548154845485
Dx2Ddf
p
AIC
7
1v.02v.13v.11.38223.870.013340.710.001x4.62217.87x7.99
x2LL,x2log-likelihood;df,degreesoffreedom;AIC,Akaike’sInformationCriterion;A,additivegeneticfactors;D,non-additivegeneticfactors(dominance);E,uniqueenvironmentalfactors;NL,TheNetherlands;BE,Belgium;AU,Australia.aEffectsofsexandagearemodelledforeachcountryseparately.bBest-fittingmodel.
Table5.Maximumlikelihoodestimatesofproportionsof
varianceexplainedbyadditivegeneticanduniqueenvironmentaleffects
A
TheNetherlands(%)Belgium(%)Australia(%)
Estimatesconstrainedtobeequal(%)
42.342.541.642.2
E57.757.558.457.8
A,Additivegeneticeffects;E,uniqueenvironmentaleffects.
Table5showstheestimatesoftheproportionofvari-anceexplainedbyAandEforeachcountryandthethreecountriespooled.Discussion
Thepresentstudyisalarge-scalemultinationaltwinstudyspecificallyfocusingonBPDsymptomsandfeaturesincommunitysamples.TheaimofthisstudywastoexaminethegeneticliabilitytoBPDfeaturesinalargesampledrawnfromgeneralpopulations,totestquantitativeandqualitativesexdifferences,andfordifferencesbetweenTheNetherlands,BelgiumandAustralia.
WefoundthatBPDfeaturesaregeneticallyinflu-enced(42%)andthatthisgeneticinfluence,similaracrossthethreecountries,doesnotdifferbetweenmenandwomenandactsinanadditivemanner.Environmentalfactorsuniquetoanindividualac-countedfortheremaining58%ofthevarianceinBPDfeatures.Torgersenetal.(2000)reportedahigherheritabilityestimate(69%),thoughthisestimateisprobablytoohighduetomethodologicallimitations.AlthoughBPDismoreoftendiagnosedinwomenthaninmen(Gunderson&Zanarini,1987;Widiger&
Weissman,1991;APA,2000),researchfindingsaboutthesexdifferenceintheprevalenceofBPDareincon-clusive.SeveralclinicalstudieshavetestedforsexdifferencesinDSMpersonalitydisorders(Jacksonetal.1991;Golombetal.1995;Griloetal.1996;Carteretal.1999;Grilo,2002)butonlyone(Carteretal.1999)foundasexdifference,theprevalencebeingun-expectedlyhigherinmen.Resultsfromnon-clinicalstudiesarealsoinconsistent;somereportedhigherprevalenceratesinwomen(Zimmerman&Coryell,1989),othersinmen(Samuelsetal.2002;Coidetal.2006),whiletheonlylargerepresentativepopulation-basedstudy(Torgersenetal.2001)didnotfindsexdifferences.Inourstudy,meanscoresonthePAI-BORdidnotsignificantlydifferbetweenmenandwomeninBelgiumandAustraliawhileinTheNetherlandswomenscoredhigherthanmen.However,thissexdifferencewasrelativelysmallwithameandifferenceof1.97points(onascalerangingfrom0to72).
GenerallyBPDsymptomsappearbyearlyadult-hood,andthedisorderoccurslessfrequentlywithin-creasingage(Parisetal.1987;Stone,1990;Bernsteinetal.1996;APA,2000;Johnsonetal.2000;Samuelsetal.2002;Coidetal.2006).Inthepresentstudy,allageregressioncoefficientsonthemeanborderlinefeaturesscorewerenegative,indicatingthatBPDfeaturesde-creasewithincreasingage,althoughtheeffectsweresmall.IntheAustraliancohortthisageeffectwasnotsignificant,probablyduetothenarrowagerangeintheAustraliansample(18to33years)andthesmallersamplesize.TheyoungageoftheAustraliancohortmayalsoexplainwhythenumberofsubjectsscoring>60ishigherintheAustraliansamplethanintheDutchandBelgiansample.
Recently,thenatureofpersonalitydisordersanditsrelationshiptonormalpersonalityhasreceivedex-tensiveattention(Widiger&Trull,2007).TheDSM-IV-Rdefinespersonalitydisorderswithinacategoricalsystem,buttheinclusionofadimensionalmodelof
8M.A.Disteletal.
personalityisincreasinglyrecommended(Trulletal.1990,2007;Livesley,2007;Widiger&Trull,2007).ThreeproposeddimensionalmodelsofpersonalityareLivesley’s18-factormodelofpersonalitypathology(Livesley,1986,1987),whichdistinguishesfourhigh-er-orderfactors(emotionaldysregulation,dissocialbehaviour,inhibitedness,compulsivity),Cloninger’spsychobiologicalmodel(Cloningeretal.1993),whichdistinguishesfourdimensionsoftemperament(noveltyseeking,harmavoidance,rewarddepen-denceandpersistence)andthreedimensionsofcharacter(self-directedness,cooperativenessandself-transcendence),andtheFiveFactorModel(FFM)ofpersonality(Costa&McCrae,1992),whichdistin-guishesfivepersonalitytraits(neuroticism,extraver-sion,opennesstoexperience,agreeablenessandconscientiousness).
Livesley’straitmodelofpersonalitypathologyisop-erationalizedthroughaself-reportquestionnaire,theDimensionalAssessmentofPersonalityPathology–BasicQuestionnaire(DAPP-BQ;Livesley,2006).Aseriesofsmall-sampletwinstudies(Livesleyetal.1993,1998;Jangetal.1996b,c),providedsupportfortheheritabilityofmostofthe18lower-orderDAPP-BQtraitsandofallofthefourhigher-orderfactors.AccordingtoLivesley,theemotionaldysregulationfactoranditsfirst-ordertraitsresemble,butarebroaderthan,thediagnosticconstructofBPD.Forex-ample,thecorrelationbetweenDAPP-BQemotionaldysregulationscoresandthenumberofBPDsymp-tomshasbeenestimatedtobe0.47inclinical(Pukropetal.2001)and0.62innon-clinical(Bagge&Trull,2003)samples.Theheritabilityforemotionaldysre-gulationhasbeenreportedat53%andfortheprimarytraitsmakingupemotionaldysregulationat44%to53%(Jangetal.1996c;Livesleyetal.1998).
ConcerningtraitsfromtheFFMandCloninger’spsychobiologicalmodel,heritabilityestimatesbe-tween41%and55%havebeenreportedforthebigfivefactorneuroticism(Jangetal.1996a;Johnsonetal.2004),andforCloninger’snovelty-seekingscale(Kelleretal.2005),bothhigher-orderpersonalitytraitsbelievedtobeassociatedwithBPD(Morey,1991;Saulsman&Page,2004;Korneretal.2007).ThesefindingssupportthepresentfindingofmoderategeneticeffectsonthemanifestationoftraitsrelatedtoBPDfeatures.
Inthepresentstudy,thePAI-BORquestionnairewasusedtomeasureBPDfeatures.ThePAI-BORdoesnotdiagnoseBPDperse,butassessesfeaturesrelatedtotheBPDsyndromewhicharealsocommontootherpersonalitydisorders(Morey,1991).Inaddition,ahighscoreonthePAI-BORisassociatedwithhigherprevalenceratesforseveralAxisIdisorders(Trull,1995).Theco-morbiditybetweenBPDandother
personalitydisordersaswellasAxisIdisordersisalsowelldocumentedbystudiesusingclinicalsamples(Zanarinietal.1998;Zimmerman&Mattia,1999;Beckeretal.2000;McGlashanetal.2000;Griloetal.2002).SeveralpriorstudieshaveshownthePAI-BORtobeareliableandvalidmeasureofBPDfeatures,andsupporttheusefulnessofthePAI-BORinasses-singBPDfeaturesinthegeneralpopulationaswellasBPDintheclinicalsetting(Kurtzetal.1993;Trull,1995).BellPringleetal.(1997)andSteinetal.(2007),forexample,showedthatthePAI-BORdiffer-entiatesbetweenpatientsdiagnosedwithBPDandpatientswithoutborderlinepersonalitypathologyorunscreenedcontrolswith75%to80%accuracy.Inaddition,Jacoboetal.(2007)administeredthePAI-BORtopatientsdiagnosedwithBPDandfoundasignificantcorrelationof0.58betweenthetotalnumberofBPDSCID-IIcriteriaandthePAI-BORscale.
Severalissuesshouldbekeptinmindwheninter-pretingtheresultsofthisstudy.First,whennon-responseinfluencesthedatacollectedinsurveyresearch,thismayseriouslylimitthevalidityofthefindings.Whileclinicalstudiestendtosamplethemostseverecases,non-responsebiasmightcauseaffectedindividualstobeunder-representedinpopu-lationstudies.BecauseBPDhasafamilialcomponent,twin-familystudiescanstudythispossiblenon-responsebiasbyusingdatafromrespondentsasaproxyforthedataoftheirnon-respondingfamilymembers.Disteletal.(2007)comparedborderlinepersonalityscoresfromhighlycooperativefamilies(i.e.manyofinvitedfamilymembersparticipate)withdataprovidedbytheparticipatingmembersoflesscooperativefamilies(i.e.fewinvitedfamilymembersparticipate).Asexpected,theparticipatingmembersoflesscooperativefamiliesshowedsomewhathigherscoresonthePAI-BORscale,suggestingnon-responsewillbehigheramongsubjectswithmoreBPDfeatures.However,thedifferencebetweenparticipantsfromlesscooperativeandhighlycooperativefamilieswasrelativelysmall,withameandifferenceoflessthan1pointonascalerangingfrom0to72.Thissuggeststhatalthoughthereisadifference,questionnairedataonBPDfeaturesarerelativelyunbiased,atleastintheDutchsample,whichconstitutedthelargestsampleinthepresentstudy.Second,wedidnotfindevidencefornon-additivegeneticeffectsthoughthetwincor-relationssuggestedacontributionofnon-additivegeneticinfluence.Theheritabilityestimateof42%mayincludesomenon-additiveeffects,buttheseareun-likelytobelarge.Inthefuturewewillcollectandincludedataofsiblingsandparentsoftwinsinthemodeltoincreasestatisticalpower,neededtoaddressthisissuemorethoroughly.
Inaddition,severalotherlinesoffutureresearchonBPDaresuggested.First,althoughourfindingswereconsistentacrossthreesamples,suggestingnosignifi-cantculturalroleinBPDfeatures,itwillbeimportanttotrytoreplicatethesefindingsinothersamplesandwithothermeasuresofBPD.Second,furtherpheno-typicandgeneticanalysesofPAI-BORitemsmaybeinformativeastheseanalysesmaypointtocohesive,geneticallyinfluenced,factorsthatcouldbeusedinfutureaetiologicalstudies.Finally,ourresultsandfuturestudiesusingthePAI-BORmayaidintheevaluationofendophenotypesthathavebeenpro-posedforthisdisorder,includinglaboratorytasks,neuroimagingfindingsandpsychophysiologicalin-dicators.
Acknowledgements
Thepresentstudywassupportedby:BorderlinePer-sonalityDisorderResearchFoundation;Spino-zapremie(NWO/SPI56-464-14192);CenterforNeurogenomicsandCognitiveResearch;CenterforMedicalSystemsBiology(NWOGenomics);Twin-familydatabaseforbehaviorgeneticsandgenomicsstudies(NWO480-04-004);Genome-wideanalysesofEuropeantwinandpopulationcohortstoidentifygenespredisposingtocommondiseases(EU/QLRT-2001-01254).
DeclarationofInterestNone.References
AkaikeH(1987).FactoranalysisandAIC.Psychometrika52,317–332.
APA(2000).DiagnosticandStatisticalManualofMentalDisorders,4thedn,revised.AmericanPsychiatricPress:Arlington,VA.
BaggeCL,TrullTJ(2003).DAPP-BQ:factorstructureandrelationstopersonalitydisordersymptomsinanon-clinicalsample.JournalofPersonalityDisorders17,19–32.
BaronM,GruenR,AsnisL,LordS(1985).Familial
transmissionofschizotypalandborderlinepersonalitydisorders.AmericanJournalofPsychiatry142,927–934.BeckerDF,GriloCM,EdellWS,McGlashanTH(2000).Comorbidityofborderlinepersonalitydisorderwithotherpersonalitydisordersinhospitalizedadolescentsandadults.AmericanJournalofPsychiatry157,2011–2016.BellPringleVJ,PateJL,BrownRC(1997).AssessmentofborderlinepersonalitydisorderusingtheMMPI-2andthepersonalityassessmentinventory.Assessment4,131–139.BernsteinDP,CohenP,SkodolA,BezirganianS,BrookJS(1996).Childhoodantecedentsofadolescentpersonalitydisorders.AmericanJournalofPsychiatry153,907–913.
Heritabilityofborderlinepersonalitydisorderfeatures
9
BoomsmaDI,BusjahnA,PeltonenL(2002a).Classicaltwinstudiesandbeyond.NatureReviewsGenetics3,872–882.BoomsmaDI,deGeusEJC,VinkJM,StubbeJH,DistelMA,HottengaJJ,PosthumaD,vanBeijsterveldCEM,HudziakJJ,BartelsM,WillemsenG(2006).NetherlandsTwinRegister:fromtwinstotwinfamilies.TwinResearchandHumanGenetics9,849–857.
BoomsmaDI,VinkJM,vanBeijsterveldtTC,deGeusEJC,BeemAL,MulderEJ,DerksEM,RieseH,WillemsenGA,BartelsM,vandenBergM,
KupperNH,PoldermanTJ,PosthumaD,RietveldMJ,StubbeJH,KnolLI,StroetT,vanBaalGC(2002b).NetherlandsTwinRegister:afocusonlongitudinalresearch.TwinResearch5,401–406.
CarterJD,JoycePR,MulderRT,SullivanPF,LutySE
(1999).Genderdifferencesinthefrequencyofpersonalitydisordersindepressedoutpatients.JournalofPersonalityDisorders13,67–74.
CloningerCR,SvrakicDM,PrzybeckTR(1993).A
psychobiologicalmodeloftemperamentandcharacter.ArchivesofGeneralPsychiatry50,975–990.
CoidJ,YangM,TyrerP,RobertsA,UllrichS(2006).
PrevalenceandcorrelatesofpersonalitydisorderinGreatBritain.BritishJournalofPsychiatry188,423–431.
CostaPT,McCraeRR(1992).ProfessionalManual:RevisedNEOPersonalityInventory(NEO-PI-R)andNEOFive-FactorInventory(NEO-FFI).PsychologicalAssessmentResources:Odessa,FL.
DeromC,DeromR(2005).TheEastFlandersprospectivetwinsurvey.InMultiplePregnancy:Epidemiology,GestationandPerinatalOutcome,2ndedn(ed.I.K.L.G.Blickstein),pp.39–47.TaylorandFrancis:Oxford.
DeromCA,VlietinckRF,ThieryEW,LeroyFOG,FrynsJP,DeromRM(2006).TheEastFlandersProspectiveTwinSurvey(EFPTS).TwinResearchandHumanGenetics9,733–738.
DistelMA,LigthartL,WillemsenG,NyholtDR,TrullTJ,BoomsmaDI(2007).Personality,healthandlifestyleinaquestionnairefamilystudy:acomparisonbetweenhighlycooperativeandlesscooperativefamilies.TwinResearchandHumanGenetics10,348–353.
GolombM,FavaM,AbrahamM,RosenbaumJF(1995).Genderdifferencesinpersonalitydisorders.AmericanJournalofPsychiatry152,579–582.
GriloCM(2002).AretheregenderdifferencesinDSM-IVpersonalitydisorders?ComprehensivePsychiatry43,427–430.
GriloCM,AnezLM,McGlashanTH(2002).DSM-IVaxisIIcomorbiditywithborderlinepersonalitydisorderinmonolingualHispanicpsychiatricoutpatients.JournalofNervousandMentalDisease190,324–330.
GriloCM,BeckerDF,WalkerML,EdellWS,McGlashanTH(1996).Genderdifferencesinpersonalitydisordersinpsychiatricallyhospitalizedyoungadults.JournalofNervousandMentalDisease184,754–757.
GundersonJG,ZanariniMC(1987).Currentoverviewoftheborderlinediagnosis.JournalofClinicalPsychiatry48,5–11.
JacksonHJ,WhitesideHL,BatesGW,BellR,RuddRP,EdwardsJ(1991).Diagnosingpersonalitydisordersin
10M.A.Disteletal.
psychiatricinpatients.ActaPsychiatricaScandinavica83,206–213.
JacoboMC,BlaisMA,BaityMR,HarleyR(2007).
Concurrentvalidityofpersonalityassessmentinventoryscalesinpatientsseekingdialecticalbehaviourtherapy.JournalofPersonalityAssessment88,74–80.
JangKL,LivesleyWJ,VernonPA(1996a).Heritabilityofthebigfivepersonalitydimensionsandtheirfacets:atwinstudy.JournalofPersonality64,577–591.
JangKL,LivesleyWJ,VernonPA(1996b).Thegeneticbasisofpersonalityatdifferentages:across-sectionaltwinstudy.PersonalityandIndividualDifferences21,299–301.JangKL,LivesleyWJ,VernonPA,JacksonDN(1996c).Heritabilityofpersonalitydisordertraits:atwinstudy.ActaPsychiatricaScandinavica94,438–444.
JardineR,MartinNG,HendersonAS(1984).Geneticcovariationbetweenneuroticismandthesymptomsofanxietyanddepression.GeneticEpidemiology1,89–107.JohnsonAM,VernonPA,HarrisJA,JangKL(2004).A
behaviorgeneticinvestigationoftherelationshipbetweenleadershipandpersonality.TwinResearch7,27–32.JohnsonBA,BrentDA,ConnollyJ,BridgeJ,MattaJ,ConstantineD,RatherC,WhiteT(1995).Familial
aggregationofadolescentpersonalitydisorders.JournaloftheAmericanAcademyofChildandAdolescentPsychiatry34,798–804.
JohnsonJG,CohenP,KasenS,SkodolAE,HamagamiF,BrookJS(2000).Age-relatedchangeinpersonalitydisordertraitlevelsbetweenearlyadolescenceand
adulthood:acommunity-basedlongitudinalinvestigation.ActaPsychiatricaScandinavica102,265–275.
KellerMC,CoventryWL,HeathAC,MartinNG(2005).Widespreadevidencefornon-additivegeneticvariationinCloninger’sandEysenck’spersonalitydimensionsusingatwinplussiblingdesign.BehaviorGenetics35,707–721.KornerA,GerullF,StevensonJ,MearesR(2007).Harmavoidance,self-harm,psychicpain,andtheborderlinepersonality:lifeina‘hauntedhouse’.ComprehensivePsychiatry48,303–308.
KurtzJE,MoreyLC(2001).Useofstructuredself-reportassessmenttodiagnoseborderlinepersonalitydisorderduringmajordepressiveepisodes.Assessment8,291–300.KurtzJE,MoreyLC,TomarkenAJ(1993).Theconcurrentvalidityofthreeself-reportmeasuresofborderlinepersonality.JournalofPsychopathologyandBehavioralAssessment15,255–266.
LivesleyWJ(1986).Traitandbehavioralprototypesofpersonalitydisorder.AmericanJournalofPsychiatry143,728–732.
LivesleyWJ(1987).Asystematicapproachtothedelineationofpersonalitydisorders.AmericanJournalofPsychiatry144,772–777.
LivesleyWJ(2006).Thedimensionalassessmentof
personalitypathology(DAPP)approachtopersonalitydisorder.InDifferentiatingNormalandAbnormalPersonality,2ndedn(ed.S.Strack),pp.401–425.SpringerPublishingCompany:NewYork.
LivesleyWJ(2007).Aframeworkforintegratingdimensionalandcategoricalclassificationsofpersonalitydisorder.JournalofPersonalityDisorders21,199–224.
LivesleyWJ,JangKL,JacksonDN,VernonPA(1993).Geneticandenvironmentalcontributionstodimensionsofpersonalitydisorder.AmericanJournalofPsychiatry150,1826–1831.
LivesleyWJ,JangKL,VernonPA(1998).Phenotypicandgeneticstructureoftraitsdelineatingpersonalitydisorder.ArchivesofGeneralPsychiatry55,941–948.
LoosR,DeromC,VlietinckR,DeromR(1998).TheEastFlandersprospectivetwinsurvey(Belgium):apopulation-basedregister.TwinResearch1,167–175.
LorangerAW,OldhamJM,TulisEH(1982).Familial
transmissionofDSM-IIIborderlinepersonalitydisorder.ArchivesofGeneralPsychiatry39,795–799.
McGlashanTH,GriloCM,SkodolAE,GundersonJG,SheaMT,MoreyLC,ZanariniMC,StoutRL(2000).Thecollaborativelongitudinalpersonalitydisordersstudy:baselineaxisI/IIandII/IIdiagnosticco-occurrence.ActaPsychiatricaScandinavica102,256–264.
MoreyLC(1988).Thecategoricalrepresentationofpersonalitydisorder:aclusteranalysisofDSM-III-Rpersonalityfeatures.JournalofAbnormalPsychology97,314–321.
MoreyLC(1991).ThePersonalityAssessmentInventory:
ProfessionalManual.PsychologicalAssessmentResources:Odessa,FL.
NealeMC,BokerSM,XieG,MaesHH(2003).Mx:StatisticalModeling,6thedn.VirginiaCommonwealthUniversity:Richmond,VA.
NealeMC,CardonL(1992).MethodologyforGeneticStudiesofTwinsandFamilies.KluwerAcademicPublishers:Dordrecht.
NiggJT,GoldsmithHH(1994).Geneticsofpersonalitydisorders:perspectivesfrompersonalityandpsychopathologyresearch.PsychologicalBulletin115,346–380.
NyholtDR(2006).Ontheprobabilityofdizygotictwinsbeingconcordantfortwoallelesatmultiplepolymorphicloci.TwinResearchandHumanGenetics9,194–197.
ParisJ,BrownR,NowlisD(1987).Long-termfollow-upofborderlinepatientsinageneralhospital.ComprehensivePsychiatry28,530–535.
PukropR,GentilI,SteinbringI,SteinmeyerE(2001).FactorialstructureoftheGermanversionofthe
dimensionalassessmentofpersonalitypathology–basicquestionnaireinclinicalandnonclinicalsamples.JournalofPersonalityDisorders15,450–456.
SamuelsJ,EatonWW,BienvenuOJ,BrownCH,CostaPT,NestadtG(2002).Prevalenceandcorrelatesofpersonalitydisordersinacommunitysample.BritishJournalofPsychiatry180,536–542.
SaulsmanLM,PageAC(2004).Thefive-factormodelandpersonalitydisorderempiricalliterature:ameta-analyticreview.ClinicalPsychologyReview23,1055–1085.
SkodolAE,GundersonJG,PfohlB,WidigerTA,
LivesleyWJ,SieverLJ(2002).TheborderlinediagnosisI:Psychopathology,comorbidity,andpersonalitystructure.BiologicalPsychiatry51,936–950.
SteinMB,Pinkster-AspenJH,HilsenrothMJ(2007).Borderlinepathologyandthepersonalityassessment
inventory(PAI):anevaluationofcriterionand
concurrentvalidity.JournalofPersonalityAssessment88,81–89.
StoneMH(1990).TheFateofBorderlinePatients:SuccessfulOutcomeandPsychiatricPractice.GuilfordPress:NewYork,NY.
StubbeJH,PosthumaD,BoomsmaDI,deGeusEJC(2005).Heritabilityoflifesatisfactioninadults:atwin-familystudy.PsychologicalMedicine35,1581–1588.
TorgersenS(1984).Geneticandnosologicalaspectsof
schizotypalandborderlinepersonalitydisorders.Atwinstudy.ArchivesofGeneralPsychiatry41,546–554.
TorgersenS,KringlenE,CramerV(2001).Theprevalenceofpersonalitydisordersinacommunitysample.ArchivesofGeneralPsychiatry58,590–596.
TorgersenS,LygrenS,OienPA,SkreI,OnstadS,EdvardsenJ,TambsK,KringlenE(2000).Atwinstudyofpersonalitydisorders.ComprehensivePsychiatry41,416–425.
TrullTJ(1995).Borderlinepersonalitydisorderfeaturesinnonclinicalyoungadults:1.Identificationandvalidation.PsychologicalAssessment7,33–41.
TrullTJ(2001).Structuralrelationsbetween
borderlinepersonalitydisorderfeaturesandputativeetiologicalcorrelates.JournalofAbnormalPsychology110,471–481.
TrullTJ,TragesserSL,SolhanM,Schwartz-MetteR(2007).Dimensionalmodelsofpersonalitydisorder:DiagnosticandStatisticalManualofMentalDisordersFifthEditionandbeyond.CurrentOpinioninPsychiatry20,52–56.
TrullTJ,WidigerTA,GuthrieP(1990).Categoricalversusdimensionalstatusofborderlinepersonalitydisorder.JournalofAbnormalPsychology99,40–48.
Heritabilityofborderlinepersonalitydisorderfeatures
11
VinkJM,WillemsenG,StubbeJH,MiddeldorpCM,LigthartRS,BaasKD,DirkzwagerHJ,deGeusEJ,BoomsmaDI(2004).Estimatingnon-responsebiasinfamilystudies:applicationtomental
healthandlifestyle.EuropeanJournalofEpidemiology19,623–630.
WidigerTA,TrullTJ(2007).Platetectonicsintheclassificationofpersonalitydisorder:shiftingtoadimensionalmodel.AmericanPsychologist62,71–83.WidigerTA,WeissmanMM(1991).Epidemiologyofborderlinepersonalitydisorder.HospitalandCommunityPsychiatry(Washington,DC)42,1015–1021.
WillemsenG,PosthumaD,BoomsmaDI(2005).EnvironmentalfactorsdeterminewheretheDutchlive:resultsfromtheNetherlandsTwinRegister.TwinResearchandHumanGenetics8,312–317.ZanariniMC,FrankenburgFR,DuboED,
SickelAE,TrikhaA,LevinA,ReynoldsV(1998).AxisIcomorbidityofborderlinepersonality
disorder.AmericanJournalofPsychiatry155,1733–1739.ZanariniMC,FrankenburgFR,YongL,RaviolaG,ReichDB,HennenJ,HudsonJI,GundersonJG(2004).Borderlinepsychopathologyinthefirst-degreerelativesofborderlineandaxisII
comparisonprobands.JournalofPersonalityDisorders18,439–447.
ZimmermanM,CoryellW(1989).DSM-IIIpersonality
disorderdiagnosesinanonpatientsample.Demographiccorrelatesandcomorbidity.ArchivesofGeneralPsychiatry46,682–689.
ZimmermanM,MattiaJI(1999).AxisIdiagnosticcomorbidityandborderlinepersonalitydisorder.ComprehensivePsychiatry40,245–252.
因篇幅问题不能全部显示，请点此查看更多更全内容