AnnalsoftheRheumaticDiseases199049412420REVIEWARTICLESubchondralava - PDF document

AnnalsoftheRheumaticDiseases1990;49:412-420REVIEWARTICLESubchondralavascularnecrosis:acommoncauseofarthritisPeterGBullough,EdwardFDiCarloArthritisisthetermweusetoindicatedysfunc-tionofajoint,clinicallycharacterisedbypain,instability,andlossofmotion.Normalfunctiondependsonintactjointanatomyandneuro-muscularcontrol.'Therefore,arthritismayresultfromanyconditionwhich(a)alterstheconfigurationofthecomponentsofajoint,inparticular,thecontourofthearticulatingsur-faces;(b)affectsthemechanicalpropertiesofthematerialsthatmakeupthesecomponents-forexample,bone,cartilage,andcollagenoustissue;(c)disturbstheneuromuscularcontrolofthejoint.Necrosisinthesubchondralregionofaboneisacommoncauseofarthritisinourexperience,thoughitisregardedasuncommonbyothers.2Withnecrosisinthesubchondralregion,arthritisresultsfromthechangeinthecontourofthearticularsurfacecausedbycollapseofthenecroticsubchondralbone.Afterashorthistoricaldiscussionoftermi-nologyandincidencethemorbidanatomyofthediseasewillbepresentedsothatthereadermaybebetterabletointerprettheradiologicalimagesandunderstandtheclinicalpresenta-tions,bothofwhichwillbepresentedlater.Abriefdiscussionoftreatmentwillthenbepresented.DepartmentofPathology,CornellUniversityMedicalCollege,andDepartmentofLaboratoryMedicine,HospitalforSpecialSurgery,NewYork,USAPGBullougliEFDiCarloCorrespondenceto:ProfessorPeterGBullough,HospitalforSpecialSurgery,535East70thStreet,NewYork,NewYork10021,USA.TERMINOLOGYItisgenerallyacceptedthatdeathoftissue-'necrosis-resultsfromoneoffourprimarytypesofinjury:(a)physical,(b)thermal,(c)toxic,or(d)circulatory.Whennecrosisoccurssecondarytocirculatorydisturbances,whetherbecauseofarterialdisease,embolism,orobstructionofthevenoussystem,theresultantregionofnecrosisisreferredtoasan'infarct'.Althoughcirculatorydisturbanceisassumedtobetheprincipalmechanismofnecrosisintheendsofbones,thecauseofthedisruptioncannotusuallybeshownbyanatomicaldis-section.Becauseofthisalargenumberofnameshasbeengiventonecrosisoccurringinbones.Someofthesenamesincludesegmentalsub-chondralinfarction,osteonecrosis,ischaemicnecrosis,idiopathicavascularnecrosis,asepticnecrosis,steroidnecrosis,andmanyeponymicdesignations(includingPerthes'disease,Freiberg'sdisease,Kienbock'sdisease,etc)dependingontheanatomicallocationsandjointsinvolved.Inthisreviewwewillusetheterm'sub-chondralavascularnecrosis'todescribealocal-isedareaofboneandbonemarrownecrosisoccurringimmediatelybeneaththearticularsurface,usuallyintheconvexcomponentofajoint.Wehavechosenthisterm,firstly,todistinguishthisconditionfromnecrosisoccur-ringinotherpartsoftheskeletonand,secondly,becausetheinvolvedregions,regardlessofthecause,arewithoutafunctionalbloodsupply.INCIDENCEANDDEMOGRAPHICSBeforethe1960ssubchondralavascularnecrosiswithoutahistoryoffracturewasconsideredunusual,3andby1962only27reportedcasesofsuchnecrosiscouldbefoundbyMankinandBrower.4Theoverallincidenceofsubchondralavascularnecrosisinanypopulationisdifficulttodeterminebecausethelesionmaybeclinicallysilentinasignificantproportionofcases.Certaingroupsofpeoplehaveapredilectionforthecondition,however.Someofthesegroupsatrisk,whoarelikelytobeencounteredinclinicalpractice,arealcoholics,8peopletakingcorti-costeroids,589blacksubjectswithsicklecelldisease,'0andJewswithGaucher'sdisease."Inourlaboratoryabout18%oftheapproxi-mately1000femoralheadsremovedeachyearintotalhipreplacementproceduresfornon-traumaticcausesshowevidenceofsubchondralavascularnecrosis.Approximately60%ofthecasesarebilateral;thereportedincidenceofbilateralityisestimatedtobe75%.12Theincidenceinourpopulationisslightlyhigheramongwomenthanamongmen,about1-2to1.Assummarisedinthetablethisdistributionisatvariancewiththemalepredominancereportedbefore1971,anditmaybetheresultoftheincreasingincidenceofsteroidrelatedcasesaswellastheparticularcaseloadatourinstitution.(TheHospitalforSpecialSurgeryisexclusivelydevotedtoorthopaedicsurgeryandrheuma-tologicaldiseases.)ThemeanageofpatientswithsubchondralavascularnecrosisofthehippresentinginourMale/femaledistributionofsubchondralavascularnecrosisofthefemoralheadAuthorsPatientsM:FReference(n)Voile1994:113Pattersonetal524:114Merled'Aubigneetal1254-5:115HastingsandMacNab144:316McCollumetal684-25:117Zinn483-8:118412 Subchondralavascularnecrosishospitalis55yearsascomparedwiththemeanageof67inpatientswithprimaryosteoarthri

tis.HISTORYInhislecturesonsurgicalpathologyin1860JamesPagetclearlydescribedthegrossappear-anceofbonenecrosis."In1888thefirstdescriptionofnecrosisinafemoralheadassociatedwithCaisson'sdiseasewaspublishedbyTwynham,20andin1911BornsteinandPlatemadeastudyof500constructionworkersontheElbetunnel.2'Inthatpapertheydescribedtheclinicalpictureofdecompressiondisease,andgavedetailedaccountsofthreepatientswithhippainwhoseradiographsshowedthetypicalpictureofsub-chondralavascularnecrosisofthefemoralhead.In1915Phemisterdescribedthemicroscopicfindingsinnecroticbone,comparingthechangesinbonedyingasaresultofinfection('septicnecrosis')withthoseresultingfromacircula-toryinterference('asepticnecrosis').22In1920herelatedthepathologicalandradiologicalchangesseenindeadbone,23andthen10yearslater,inapaperreadtotheAmericanOrtho-paedicAssociation,hediscussedthereparativeprocessesoccurringarounddeadbonefollowingfracture,grafting,andvascularocclusion.24InthisclassicpaperPhemistercoinedtheterm'creepingsubstitution'fortheprocesswherebythedeadboneisfinallyremovedafteralayeroflivingbonehasbeendepositedontothepre-existingdeadbone.In1922Axhausenreportedtheoccurrenceofasepticnecrosisinpatientswithalcoholism.7TherelationbetweensubchondralavascularnecrosisandalcoholismwaslatertobedevelopedbyJonesetalin1968.8TherelationofsubchondralavascularnecrosistosteroidswasfirstnotedbyPietrograndeandMastromarinoin1957.25ConfirmationofthiswasmadebyMandelandFreeman,whoin1964reportedacaseofasepticnecrosisofthefemoralheadinapatientwithCushing'sdisease.26Increasingrenaltransplantationandpostopera-tivesteroidtreatmentprovidedCreussetalwithfurtherclinicalevidenceoftheassociationbetweenthisdiseaseandsteroidtreatment.27In1965Cattopublishedtwoexcellentpapersgivingdetailsofthedestructiveandreparativechangesin'subchondralavascularnecrosisofthefemoralhead.2829Inthesepaperssheconsideredthechangesoccurringinthefemoralheadaftersubcapitalfracturesoftheneckofthefemur,anddealtwiththepathologicalprocessesindetail.MorbidanatomyFourstagesinthedevelopmentofsubchondralavascularnecrosishavebeendefinedmorpho-logicallyandthesemaybeshowntocorrelatewithobservedradiographicappearances.3133Thefirststageischaracterisedpredominantlybythepresenceofnecrosisofbothboneandbonemarrowwithoutevidenceofrepair.Inthesecondstagereparativeprocessesareevidentattheperipheryofthenecroticregion.Themajorfeatureofthethirdstageissegmentalcollapseofthearticularsurface.Inthefourthstagefeaturesofsecondaryosteoarthritishavedeveloped.Itisimportanttorecognisethatthemorphologicalfeaturesofsubchondralavascularnecrosisareacompositeofbothnecrotisingandreparativeprocessesandthatatleastsomeoftheapparentlydegenerativefeatures-forexample,segmentalcollapseinstageIII,whichgivesrisetoclinicalsymptoms,maybecontributedtobythereparativeprocesses.STAGEIInthisstagetheshapeofthejointisunalteredandexternalexaminationofthejointshowsnoabnormalities.Oncutsection,however,thenecroticzone,recognisedasasomewhatwedge-shapedregioninwhichthemarrowisdullyellow,chalky,andopaque,maybeseeninanimmediatelysubarticularlocation(fig1).Thisregionisusuallywelldemarcatedandseparatedfromthesurroundingbonemarrowbyathin,red,'hyperaemic'border.Themarrowbeyondthisbordershowsnospecificabnormalityrefer-abletothenecrotisingprocess,butmayhaveanabnormalappearancedependingonanyassoci-atedorpredisposingconditions-forexample,sicklecelldiseasewithitsdarkredbonemarrow,andGaucher'sdiseasewithitspaleandwaxymarrow.Atthisstagechangesinthetrabeculararchitecturearenotappreciableeitherbygrossinspectionoronspecimenradiographs,whichtypicallyshownoabnormality.Onmicroscopicexaminationtheoverlyingarticularcartilageappearsviabledowntothecalcifiedzone.Thebonyend-plate,however,isnecrotic.Thesubchondralbone,correspondingtotheopaqueyellowregionseengrossly,ischaracterisedbynecroticboneandbonemar-row.Themarrowelementsarereplacedbygranular,eosinophilicmateriallackingcellularelementsexceptfortheoccasionalghostsofdisruptedfatcells(fig2).Theremayalsobecystsoflipidmaterialwhich,withcalcification,mayappearassaponifiedfat('soap').Inthebonetheosteocyticlacunaemaybeempty,containcellulardebris,orhaveapale-stainingnucleus.AtthemarginoftheinfarctthereisFigure1:CutsurfaceofafemoralheadremovedaftersubcapitalfractureshowingthewelldefinednecroticregionaffectingmostoftheheadasitappearsinstageIsubchondralavascula

rnecrosis.413 Bullough,DiCarlo.4:...e..u.t,4.iFigure2:NecroticboneandmarappearanceofthemarrowandthThecysts(arrows)resultfromthe(Haematoxylinandeosin.)Figure3:AspecimenradiograpiakneejointwithstageIIsubchoshowingtheencroachingsclerosis'creepingsubstitution'occurringirelativelylucentnecroticzone.-E:F't,':;Figure4:Theadvancingfrontofvasculargranulationtissue(arrow)progressingintothenmcroticzoneatthetop.(Haematoxylinandeosin.)increasedosteoclasticactivityandaninter-trabecularinfiltrateofproliferatingfibroblastsandcapillaries.Thiszonecorrespondstothethinredrimseeninthegrossspecimen.Beyondthishypervascularzonetheboneandbonemarrowareunchangedbythenecrotisingprocessandreflectthestateofthemarrowbeforethenecrotisingevent.STAGEIIAsinstageI,theoverallshapeoftheboneispreservedandthearticularsurfaceisintact.Onsectioning,however,arimofbonysclerosiscanbeseenattheperipheryofthenecroticzoneattheboundarybetweenthenecroticzoneandtheunaffectedmarrow.Thisfeatureisbestseenonspecimenradiographs(fig3).ThecentralregionofnecrosisisunchangedfromstageI,butthehyperaemiczoneisgenerallythickerandmay�4-*nowcontainamixtureoftanandwhitetissue.Onmicroscopicexaminationanadvancingfrontofgranulationtissuecomposedoflipid-ladenmacrophages,proliferatingfibroblasts,rrowshowingthegranularandcapillariescanbeseenattheperipheryandeghostsofnecroticfatcells.extendsintothecentralregionofthenecroticedisintegrationoffatcells.zone(fig4).Followingcloselybehindthis'cleanup'frontisasecondfrontcomposedofosteo-blastsdepositingalayerofnewboneonthepre-existingdeadtrabecularbone.Followingthissecondfrontatavariabledistance,limitedamountsofboththeolddead,andnewlivingbonearebeingremovedbyosteoclasts.Theoveralleffectoftheseclimbingprocessesistoremovethenecroticmarrowandbonewhilemaintainingthestructuralintegrityofthebone.ThisseriesofprocessesiswhathasbeencalledcreepingsubstitutionbyPhemister.24Theincreasedosteoblasticactivityandthelayerofnewbonegiverisetotheclinicalradiographichofthefemoralcondyleofappearancesofbonysclerosisandtothein-ndralavascularnecrosiscreaseduptakeoftheradioactivetechnetiumswhichrepresentsthedpiattheperipheryofthediphosphonateisotopeonabonescan.STAGEIII"_fffAlterationoftheshapeoftheboneisfirstencounteredinthisstage.Thisistheresultof'V.''*4collapseinthenecroticregionandmaybe^*ittwapparentonexternalexaminationoftheboneas|,".9^.,abucklingorfragmentationofthearticularcartilage(fig5).Figure5:BucklinginthearticularcartilageseenattheperipheryofthecollapsedsegmentinahumeralheadwithstageIIIsubchondralavascularnecrosis.414 SubchondralavascularnecrosisOncutsectionitcanbeseenthatthereasonforcollapseisfractureoftrabecularbonewithassociatedfractureoftheend-plateitself(fig6).Fractureofthetrabeculaoccurseitherjustbelowthebonyend-plate,deepwithinthenecroticregion,oronthenecroticsideoftheadvancingsclerosisinthereparativefront.Thefractureofthetrabeculamayresultfromanyoneofthreecauses:(a)thecumulativeeffectofmicrofracturesinducedbyfatiguewithinthenecroticzone;(b)weaknessoftrabeculaeinthereparativefrontduetoosteo-clasticactivity;or(c)focalconcentrationsofstressatthejunctionsbetweenthethickenedsclerotictrabeculaofthereparativezoneandthenecrotictrabecula.ThislastcausemaybeIFigure6:Acutsectionthroughthehumeralheadshowninfig5demonstratingthefractureinthenecroticboneimmediatelybeneaththecollapsedarticularsurface.Thefracturelinecorrespondstotheradiolucentlineknownasthe'crescentsign'seenonclinicalradiographs.Figure7:CutsurfaceofafemoralheadwithstageIVsubchondralavascularnecrosisshowingthemarkeddestructionwhichoccursanddemonstratingadeepsaddleshapeddeformnitywhichresultsfromlossofthenecroticsegment.*~~~~1t-re+04.-+tS;A7*fs*_.+,_�tXA*-z9-v-,,,r.;_~~~~~~~wr-_.iF;igure8:SvnoviumshowingthebonvandcartilaginousdetrituswhichmavaccumulateinstageIVsubchondralavascularnecrosis.mostimportantindeepfracturesandresultfromthebioengineeringconceptof'stress'34risers.Thelinearfracturebeneaththeend-platecorrespondstotheradiolucentzone,referredtoasthe'crescentsign',seenonclinicalradio-graphs.34Microscopically,thefracturedtrabeculainthenecroticregionappearsasfragmentsofpulverisedbonyandcartilaginousdetritus.Theoverlyingcartilagemaystillappearviable.Fractureswithinthereparativezonehaveanabundanceoffibroustissue,cartilaginoustissue,andreactivewovenboneandhavetheusualappearanceofun-united,uns

tablefractureselsewhereintheskeleton.STAGEIVThemajorfeatureofthisstageistheappearanceofmorphologicalchangesusuallyassociatedwithosteoarthritis.Dependingonthedegreeoftheosteoarthriticchanges,itmaynolongerbepossibletorecognisetheinitialeventsasthoseofsubchondralavascularnecrosisontheclinicalradiographs.Inmostcases,however,thereissufficientevidenceongrossandmicroscopicexaminationtoallowproperdiagnosis.Ingeneral,asbestseeninthefemoralhead,formationofosteophytesisnotpronouncedandtheareasofcartilaginouserosionandbonyeburnationsurroundthecollapsedsegment.Becauseofthecollapseoftheinfarctedsegmentfragmentedcartilagemaypersistinthisregion.Whenthechangesofosteoarthritisaremarkedtheonlycluethattheinitialeventmighthavebeensubchondralavascularnecrosisisthatthefemoralheadhasadeep,saddleshapeddefor-mity(fig7).Thecutsurfaceinthisstageofsubchondralavascularnecrosismayshowresidualfragmentsofarticularcartilageanddensefibrousconnec-tivetissueintheinfarctedarea,surroundedatthesurfacebyamarginofdenselyscleroticbonerepresentingtheeburnatedarticularsurface.Inmarkedlydestructivecasesthe'articularsurface'maybecomposedofpulverisedbonydetritus.Microscopicexaminationoflessdistortedspecimensshowsfragmentsofviableornecroticcartilageonorwithinalayeroffibrousandcartilaginoustissue,whichmayincludegranu-lationtissueandreactivewovenbone.Thesurroundingeburnatedarticularsurfacesareseenmicroscopicallyassmoothsurfacesover-lyingdenselyscleroticlamellarandoccasionallywovenbone.Inmoreadvancedcasestwousefulcluestothediagnosisofsubchondralavascularnecrosismaybetheabsenceofclearlyeburnatedboneatthearticularsurfaceandthepresenceofbonyandcartilaginousdebrisintheaccompanyingsynovialandcapsulartissue(fig8).BIOCHEMICALASPECTSOFTHENECROTICZONEInanefforttocharacterisefurtherthenecroticregionofsubchondralavascularnecrosisthetissuelipidsin18femoralheadsresectedforthisnecrosishavebeenstudied.36Thetotallipidsinthenecroticregionwereincreasedincompari-415 Bullough,DiCarlosonwithboththeinon-necroticregionsofthesamefemoralheadandotherfemoralheadswithoutevidenceofnecrosis.Thecholesterolcontentwashigherinthenecroticregionsthaninthenon-necroticregionsofthesamefemoralhead,butbothwerehigherthaninsevennormalcontrolspecimensandfourosteoarthriticspecimenswithoutevidenceofnecrosis.Interestingly,thegreatestincreasesinthecholesterolcontentwereencounteredinthosepatientswithhistoriesofcombineduseofsteroidsandalcohol.Thecholesterolcontentcorrelated(r=0-82)withtheproportionofnecrotictissueinthespecimenandmayeithercontributetocelldeathbyalteringmembranemetabolismormighthavebeenreleasedfromthecellasaconsequenceofcelldeath.ImagingmodalitiesTYPESThecurrentarmamentariumusedintheradio-graphicdiagnosisofsubchondralavascularnecrosisincludesfourmainimagingmodalities:plainradiographs,radionuclidescintigraphy,computedtomography,andmagneticresonanceimaging.Eachhascontributedtoourunder-standingofthecausesandprogressionofsubchondralavascularnecrosisanditsassociatedcomplications-mostnotably,arthritis.Asisthecasewithalladvancesinmedicaltechnology,thelatermodalitieshavebeenpromotedasreplacementsfortheirpredecessorsinthediag-nosisofmanyconditionsingeneralandsub-chondralavascularnecrosisinparticular.Ithasbeenfound,however,thatnoneofthenewermethodsisinfallible,andacombinationoftechniquesisstilloftennecessary.PLAINRADIOGRAPHSThenaturalcontrastaffordedbythemineralinbonematrixmakestheplainradiographreadilyapplicabletothediagnosisofbonedisease.Becauseofitsabilitytoshowstructuralaltera-tionstheplainradiographcanidentifyjointsaffectedbysubchondralavascularnecrosiswhenthereparativeprocessesarewelldeve-loped.Plainradiographsarelimited,however,byaninabilitytoshowtheearlychangesofthediseaseatatimewhentherapeuticinterventionhasthegreatestchancetopreventdevelopmentofarthritisandothercomplications-thatis,instageIwheretheplainradiographsarenormal.Theearliestplainradiographicfindinginanendofaboneaffectedbysubchondralavascularnecrosisisthepresenceofapoorlydefinedregionofsclerosis,whichdoesnotgenerallyreachtothesubchondralend-plate.Theappear-anceofthissclerosiscorrespondstothedeposi-tionofboneinthepresenceofcreepingsubstitutionduringstageII.Overtimethissclerosisbecomeswelldevelopedandreachestheend-plate.Becausetherepairprocesspro-ceedsfromtheperipheryincontactwithsurviv-ingbonemarrowinthreedimensionalspacethesclerosisseeno

ntheplain(twodimensional)radiographtendstooverestimatetheextentofrepair.InstageIII,wherecollapsehasoccurred,theplainradiographshowsthealterationintheshapeofthejoint.Thecharacteristicfindingofthecrescentsign,seenbestinthe'froglateral'viewofthefemoralhead,iseasilyrecognisableinplainradiographs(fig9).35Similarevidenceofcollapseisalsoapparentinotheraffectedjoints.AftercollapseinstageIV,changesofosteoarthritismaysuperveneandobscureorevenobliteratethefeaturescharacteristicofsubchondralavascularnecrosis(fig10).Insymptomaticsubchondralavascularnecrosis,ifthepainissufficienttocauseFigure9:Plainradiographofahipinthe'froglateral'viewshowingtheradiolucentline('crescentsign'-arrow)representingthesubchondralfracturewithdeformityofthesurfaceinstageIIIsubchondralavascularnecrosis.Figure10:PlainradiographofafemoralheadwithstageIVsubchondralavascularnecrosisshowingthedestructivearthritiswithosteophytesandsclerosisaffectingbothsidesofthejointmimickingosteoarthritis.416 Subchondralavascularnecrosisreduceduseol-thejointorextremity,thebonearoundthenecroticregionmayundergodisuseosteoporosis,causingthenecroticregiontoappeardense.RADIONUCLIDESCINTIGRAPHY(BONESCAN)Thebonescanisusedinthediagnosisofsubchondralavascularnecrosisbecauseofitsabilitytoidentifyregionsintheskeletonwherebonedepositionandmineralisationareoccurr-ing.EarlyinstageIIthepresenceofanincorporatedradioactiveisotopecanbedetectedwellbeforetheplainradiographsshowincreasedsclerosis(fig11).3Animportantfeatureofthebonescanisthatitiscapableofindicatingmultipleaffectedsiteswithoutincreasingtheexposureofthepatienttoionisingradiation.COMPUTEDTOMOGRAPHY(CT)OneoftheadvantagesoftheCTscanoverplainradiographsisitsabilitytoobtain'slices'throughthebone,therebyreducingoreliminat-ingoverlapandprovidingaclearerimageofthenecroticzonewithoutinterferencefromsur-roundingstructuressuchassofttissuesandregionalcorticalbone(fig12).ThechangesseenontheCTscanparallelthoseseenwithplainradiographsandincludeprogressiveperipheralsclerosisofstageIIand,whenitoccurs,thecrescentsignofstageIII.Theabilitytoobtainslicesprovidesabetterestimateoftheextentofrepairwithinthenecroticregion.MAGNETICRESONANCEIMAGING(MRI)Thethreeimagingmodalitiesjustdiscussedarehinderedbytheirinabilitytoidentifytheearlieststageofsubchondralavascularnecrosis.Theycanonlyidentifythereparativeprocessesaftertheyareestablished.Magneticresonanceimagingofferstheconsiderableadvantageofbeingabletoidentifychemicalchangesinnecroticbonemarrowwellbeforechangesinthebonecanbedetectedbytheothermethods.38.-N*::~~~~~~~.....Itisalsousefulinshowingotherclinicallysilentfociofsubchondralavascularnecrosisinpatientswhoalreadyhaveaclinicallyevidentfocus.TheearliestfindingofsubchondralavascularnecrosisasshownbyMRIisadecreaseintheusualhighsignalobtainedbymarrowfatontheTiweightedimage.Theregionappearsdarkerontheimagewhencomparedwiththesur-roundingmarrow(fig13).Thisdecreasedsignalresultsfromthechemicalalterationinthefatthatresultsintheopaque,yellow,andsoapyappearanceseenongrossexamination.Magneticresonanceimagingshowsonlyvariationsinsignalintensityanditremainsincumbentupontheradiologisttokeepinmindthatanyconditionaffectingthemarrowmaychangethesignalintensitysimplybyreplacingordisplacingmarrowfat.Insomeconditions,suchassicklecelldiseaseandGaucher'sdisease,themarrowsignalmayalreadybereducedandthusmakeitdifficulttoevaluatetheimage.AetiologyandpathogenesisRegardlessofanypredisposingconditions,thefinalcauseofsubchondralavascularnecrosisisthelossofadequateperfusionofbloodinthearticularendsofbones.ANATOMICALFACTORSSeveralanatomicalfeaturesoftheendsofbonesL_.Figure12:Acomputedtomography(CT)scanshowingtheirregularcontouroftherightfemoralheadaftercollapseofthearticularsurfaceinstageIIIsubchondralavascularnecrosis.TheradiolucentlinebeneaththecollapsedcartilageistheCTrepresentationofthe'crescentsign'seenonplainradiographs.ABFigure11:BonescansshowingtheincreaseduptakeofradionuclideinthefemoralheadinstageIsubchondralavascularnecrosis(A).Theuptakeispresentonlyononesideoftherighthipjoint,whichdistinguishesthisconditionfromosteoarthritis.(B)Thecentralnecroticzonehasnouptakeinearly,largelesions.b=urinarybladder;a=anterioriliacwing;arrows-jointmarginoffemoralhead.Figure13:AmagneticresonanceimagingscanshowingthedecreasedsignalintheTIweightedimageinthenecroticregionoftheleftfemoralhead.41

7 Bullough,DiCarlorenderthemsusceptibletocompromiseintheirbloodsupply:(a)subchondralbonehasalimitedcollateralcirculation,particularlyontheconvexsidesofjoints3940;(b)theperfusionpressureandbloodflowofepiphyses,andthatoffattymarrow,incomparisonwithreddia-physialmarrow,islow,makingthesubchondralregionsusceptibletodecreasedperfusionpres-suresinmuchthesamewayasinothersites,particularlyintheendocardiumand'watershed'regionsofthegastrointestinaltract4";(c)finally,andpossiblymostimportantly,bone,unlikethesofttissueandorgansofthebody,isrigidandnon-distensible,andcanneitherexpandnorcollapseinordertoassistinthemaintenanceofadequateperfusionpressures.Asisthecasefornecrosiselsewhere,necrosisofsubchondralboneresultsfromeitheranintravascularoranextravasculardisturbance.Examplesofpossibleintravasculardistur-bancesarenarrowingofarteriesandarterioles,resultingfrommanyprimaryvasculardiseases,embolism,thrombosis,increasedviscosityoftheblood,andhaemoglobinopathies.Extra-vascularcausesofdisturbedperfusionconsistofthoseconditionswhichinfiltratethemarrowspaceandgenerallyincreaseinterstitialpres-sure,suchasGaucher'sdiseaseandotherlipidstoragediseases,infection,hypercortisonism,alcoholism,andbothbenignandmalignanttumours,allofwhichresultinincreasedvenouspressure.PHYSICALINJURYIntracapsularfractureofthefemoralneckwithitsconcomitantphysicaldisruptionofbloodvesselscausesclinicalsymptomsresultingfromsubchondralavascularnecrosisinupwardsof20%offracturecaseswithposteriordis-location.42Extracapsularfracturesoftheproxi-malfemuraremuchlesslikelytoresultinsubchondralavascularnecrosis.43Necrosisofbonemayalsoresultfromthedirecttoxiceffectsofhighdoseradiation,chemotherapy,thermalandelectricalinjury,andfreezing.LIPIDMETABOLISMANDFATEMBOLISMApproximately90%ofpatientswithnon-traumaticbonenecrosishavebeenshowntohavedisorderswithassociateddisturbancesinlipidmetabolismorfatembolism."45Thedeve-lopmentoffatembolihasbeenstronglyimpli-catedasthecausativefactor.447Joneshasproposedthreemechanismsthroughwhichfatembolimaydevelopandleadtonecrosisofbone:(a)fattyliver;(b)destabilisationorcoalescenceofplasmalipoproteins;and(c)disruptionofmarrowfat.4Thetwomostcommondiseasesassociatedwithnon-traumaticsubchondralavascularnecrosisarehypercortisolismandalcoholism,whichaccountforabouttwothirdsofthecases.45HypercortisolismTheevidencelinkinghypercortisolism,whetherasCushing'sdiseaseorasaresultofsteroidtreatment,tosubchondralavascularnecrosisispersuasive,thoughitislikelythatmorethanonemechanismisinvolved.Studiesinsteroidtreatedrabbitshaveshownthepresenceofhyperlipaemia,fattyliver,andfatemboliinthelungsandbones.48ThesefindingsindicateamajoralterationinsystemicfatmetabolismandsuggestthatallofthemechanismsproposedbyJonesinthedevelopmentoffatembolismmightplayapart.Inadditiontothesemechanisms,Wangetalhaveshownincreasedsizeofmarrowfatcellsinsteroidtreatedrabbits.49Thisinturnwasassociatedwithincreasedintraosseouspres-sureanddecreasedbloodflow.IthasbeensuggestedbySweetandMadewellthatnecrosismayextendorbeepisodicduringsteroidtreatment,andhistologicalevidenceofrecurr-ingnecrosis-thatis,evidenceofdeathinpreviouslyhealingareas,wasreportedin83%ofthefemoralheadsthattheyexamined.50Subchondralavascularnecrosiscomplicatessystemiclupuserythematosusinapproximately5%ofpatientsandisbelievedtobeassociatedmainlywithsteroidtreatment.5AlcoholssmInalcoholicpatientsthepresenceofafattylivermayactasacontinuoussourceofsubclinicalfatembolism.5Inaddition,fatmaybereleasedbyblunttraumatotherightupperquadrant,andfocallivernecrosismayalsocontributetoembolisation.INTRAOSSEOUSHYPERTENSIONTheimportanceofincreasedintraosseouspres-surehasbeenemphasisedbyZizicetal.52Inpatientsinwhomtheymeasuredintraosseouspressuresaraisedlevelwaspredictiveofosteo-necrosisin50%ofcases,whereasonlyrarelydidpatientswithnormalpressuredeveloposteo-necrosis.Asalreadymentioned,increasedintra-osseouspressureassociatedwithincreasedsizeofmarrowfatcellsmaybesignificantinexperimentalsubchondralavascularnecrosisinsteroidtreatedrabbits.Raisedintraosseouspressurehasbeenreportedin38hipsfrom26alcoholicpatients.53DysbarismIncontrastwithothercausesofsubchondralavascularnecrosis,thehumeralheadismorecommonlyaffectedwithnecrosisowingtocompressedair.54Subchondralavascularnecrosisindysbarismistheresultofthereleaseofnitrogendissolvedinthetissues,whichproducesb

ubblesintheextracellular,extra-vascularspace,leadingtocompressionofthecapillarynetwork.Gaucher'sdiseaseInGaucher'sdisease"andotherstoragediseaseswhichreplacethenormalmarrowconstituentsaccumulatedstoragecellspackthemarrowspaceandconstrictthevascularnetworkwithsubsequentcompromiseinvenousreturn.418 SubchondralavascularnecrosisSicklecelldiseaseInsicklecelldiseaseocclusionofthesmall'vesselsprobablyoccursbecauseoftheincreasedviscosityofthebloodcausedbyclumpingoftheaffectedredcellsfollowinghypoxia.'0ClinicalpresentationInanextensivestudyofpatientsseenatourinstitution55theprimarypresentingcomplaintinsubchondralavascularnecrosisofthehipwaspainexperiencedonweightbearingand,intwothirdsofthecases,alsowhileatrest.Inmostrespectsthesymptomsofsubchondralavascularnecrosisandosteoarthritisaresimilar,thoughinsubchondralavascularnecrosisthefunctionaldisability,asassessedintermsofdifficultyinwalkingandperformingactivitiesofdailyliv-ing,islikelytobesomewhatlessthanthatex-periencedwithprimaryosteoarthritis.Mostpatientswithsubchondralavascularnecrosishaveahardtimewalkingupstairsandputtingontheirshoes,however.Themodeofonsetofthesymptomsisratherdifferentfromthatseenwithosteoarthritis.Mostpatientswithsubchondralavascularnecrosisreportasuddenonsetofacutepain,whereasinosteoarthritistheonsetisusuallymoregradual.Theonsetisconsideredsuddenifthepatientcanrelateittoaspecificincidence-suchaswhilemowingthelawn,orarisingfromachair-orspecificallyremembersthatthesymptomsstartedduringaparticularweekormonth.Thestatisticsfordurationofsymptomsalsoshowamarkeddisparitybetweensubchondralavascularnecrosisandosteoarthritis.Beforesurgicaltreatmentpatientswithsubchondralavascularnecrosisofthehip,inourexperience,hadbeensymptomaticforanaverageof3-6yearscomparedwith5-8yearsforprimaryosteoarthritis,93yearsforrheumatoidarthritis,and13-9yearsforsecondaryosteoarthritis.Patientswithsubchondralavascularnecrosisgenerallyhaverestrictionofallhipmovementsbuttoamuchlesserextentthanisfoundinosteoarthritis.Thelimitingfactorinsub-chondralavascularnecrosisseemstobepainratherthandeformityofthejointsurfaces.InourexperiencejustoverhalfoftheaffectedsubjectshowapositiveTrendelenburg'stest-thatis,whenthepatientstandsononelegusingtheaffectedsidethepelvisontheoppositesidefallsratherthanrising.InZinn'sseriesof40patients29,or59%,hadapositiveTrendelen-burg'stest.Thereareasyetnodefiniteclinicallaboratoryfindingswhichindicatethatapatienthassubchondralavascularnecrosis.In1975Mielantsetal,reportingontherelationbetweensub-chondralavascularnecrosisandlipidmeta-bolism,showedthattherewasasignificantincreaseintotalserumlipidandtriglycerideconcentrationsinthesepatients.56Wehavenotbeenabletoconfirmthesefindingsinourlaboratory,however.TreatmentTreatmentofsubchondralavascularnecrosisofthefemoralheadbycorebiopsydecompressionhasbeenproposed.5758InstageIandstageIIsubchondralavascularnecrosis-thatis,beforecollapse,ithasbeenshownthatcorebiopsyaloneresultedingoodlongtermresultsin90%ofthepatients.53Anotherstudyreportedthathalfthepatientstreatedinthiswaydidnotrequireanyfurthertreatment.59Othershavereportedaseriesofpatientswithearlysub-chondralavascularnecrosisofthehipwhohavebeentreatedbyelectricalstimulationandgraft-ingalone.60Althoughnostatisticallysignificantdifferencewasnotedwiththeadditionoftheelectricalstimulation,theauthorsremainhope-fulthatitwillhelpandpersistwiththisformoftreatment.Insubchondralavascularnecrosisofthekneeitisgenerallythoughtthatpatientswithtypicalsymptoms,positivebonescansbutnolesionsevidentonradiologicalexamination(stagesI-II),arenotbenefitedbysurgery.6'Instead,con-servativetreatmentusingprotectedweightbearing,andanalgesicsisrecommended.Thesepatientsarelikelytodowell,buttheyshouldbefollowedupcarefullybyserialxrayexamina-tions.FordiseaseinstagesIIIandIVandforlargelesions,whethersymptomaticorasympto-matic,withanassociatedangulardeformityoftheknee,tibialosteotomyisadvocated.TotalkneereplacementisgenerallyonlyadvocatedforpatientswithsymptomaticstageIIIorIVdiseasewhoarelessactive.Certainly,totaljointreplacementshouldbeusedwithcare.Inastudyfromourinstitutiontherewasa37%overallfailurerateintotalhipreplacementsdoneforsubchondralavascularnecrosisascomparedwitha10%failurerateforthosereplacementsdoneforothercauses.62Inthatstudythehighestfailureratewasseenamongthosepatientswithsicklecelldisease,

allofwhomexperiencedfailureoftheirimplants.Thenexthighestfailureratewasobservedinalcoholics(67%)andtheninthosepatientswhohadsteroidtreatment(30%).Thecaseswhichhadnoknownpredisposingconditionhadamuchlowerfailurerate(11%).Summary(1)Subchondralavascularnecrosisisanimpor-tantcauseofjointpainanddisabilityandaccountsforupwardsof20%oftotalhipreplacementsdoneinourhospital.(2)Earlydiagnosismaybemadewiththeaidofmagneticresonanceimagingandradioactiveisotopestudies.(3)Althoughthesignsandsymptomsaresimilartothoseofosteoarthritis,therearesignificantdifferences-namely,(a)ahistoryofsuddenonsetofpain,presentinmorethanhalfthepatients;(b)ayoungeragegroup;(c)ashorterdurationofsymptomsattimeofsurgery;(d)clinicallythelimitingfactorispainratherthanactualjointdeformitytoaccountforrestrictionofmovement;(e)ahighincidenceofmultiplesitesofinvolvement.(4)Thediseaseiscommonlyassociatedwithsteroidtreatmentoralcoholabuse.Althoughmanyothercausesarerecognised,theyarerareinWesternurbanpractice.419 Bullough,DiCarlo(5)PatientswithstageI-IIsubchondralavas-cularnecrosis,especiallyoftheknee,arebettertreatedconservatively.(6)Surgicaltreatmentgiveslesssatisfactoryresultsthanthetreatmentofosteoarthritisbysimilarmodalities.BulloughPG.Thepathologyofdegenerativechangeinthehip:themicroscopicchanges.In:ReynoldsD,FreemanMeds.Osteoarthritisintheyounghip.NewYork:ChurchillLivingstone,1989.2CushnerFD,FriedmanRJ.Osteonecrosisofthefemoralhead.OrthopedicReview1988;17:29-34.3McCarthyE.Asepticnecrosisofbone.ClinOrthop1982;168:216-21.4MankinHL,BrowerTD.Bilateralidiopathicasepticnecrosisofthefemurinadults-Chandler'sdisease.BulletinoftheHospitalforJointDiseases1962;23:42-57.JonesJP.Alcoholism,hypercortisolism,fatembolismandosseousavascularnecrosis.In:ZinnWM,ed.Idiopathicischemicnecrosisofthefemoralheadinadults.Stuttgart:Thieme,1971:112-32.6HungerfordDS,ZizicTM.Alcoholismassociatedischemicnecrosisofthefemoralhead:earlydiagnosisandtreatment.ClinOrthop1978;130:144-53.7AxhausenG.DieNekrosedesproximalenBruchstuecksbeimSchenkelhalsbruchundihreBedeutungfuerdasHueft-gelenk.LangenbecksArchivfurKlinischeChirurgie1922;120:325-46.8JonesJP,JamesonRM,EnglemanEP.Alcoholism,fatembolismandavascularnecrosis.J7BoneJointSurg[Am]1968;50:1065.9SolomonL.Drug-inducedarthropathyandnecrosisofthefemoralhead.JBoneJointSurg[Br]1973;55:246-61.10SennaraH,GorryF.Orthopedicaspectsofsicklecellanemiaandalliedhemoglobinopathies.ClinOrthop1978;130:154-7.11HermanG,GoldblattJ,LevyRN,GoldsmithSJ,DesnickRJ,GrabowskiGA.Gaucher'sdiseasetype1:assessmentofboneinvolvementbyCTandscintigraphy.AmericanJ7ournalofRheumatology1986;147:943-8.12TurekSL.Orthopaedicprinciplesandtheirapplications.2nded.Philadelphia:Lippincott,1984.13VolleK.L'osteonecroseidiopathiquedelatitefemoralchezl'adulte.Lyons:Rey,1963.14PattersonRJ,BickelWH,DahlinDC.Idiopathicavascularnecrosisoftheheadofthefemur:astudyof52cases.JBonejointSurg[Am]1964;46:267-82.15Merled'AubigneR,PostelM,MazabrandA,MassiasP,GueguenJ.Idiopathicnecrosisofthefemoralheadinadults.JBonejointSurg[Br]1965;47:612-33.16HastingsDE,MacNabI.Spontaneousavascularnecrosisofthefemoralhead.Aclinicalandpathologicalreview.CanjSurg1965;8:68-83.17McCollumDE,MatthewsRS,O'NeillMT.Asepticnecrosisofthefemoralhead.Associateddiseasesandevaluationoftreatment.SouthMedJ1970;63:241-53.18ZinnWM.Clinicalpictureandlaboratoryfindings.In:ZinnWM,ed.Idiopathicischaemicnecrosisofthefemoralheadinadults.Stuttgart:Thieme,1971:12.19PagetJ.Mortification.LectureonSurgtcalPathology.1860;lectureXIX:301-2.20TwynhamGE.AcaseofCaissondisease.BrMedJ31888;i:190-1.21BornsteinA,PlateE.UeberchronischeGelenkveraenderun-gementstandendurchPresslufterkrankung.FortschritteaufdemGebietederRoentgenstrahlen1911;18:197-206.22PhemisterDB.Necroticboneandthesubsequentchangeswhichitundergoes.JAMA1915;64:211-6.23PhemisterDB.Recognitionofdeadbonebasedonpatho-logicalandX-raystudies.AnnSurg1920;72:466-85.(ReadtoAmericanSurgicalAssociation5May1920.)24PhemisterDB.Repairofboneinpresenceofasepticnecrosisarisingfromfractures,transplantationsandvascularobstruction.JBoneJ7ointSurg[Am]1930;12:769-87.25PietrograndeV,MastromarinoR.Osteopatdaprolungatatrattamentocortisonico.OrtopTraumApparMat1957;25:791-810.26MandelSH,FreemanLM.AvascularnecrosisofboneinCushing'sdisease.Radiology1964;83:1068-70.27CreussRL,BlennerhassettJ,MacDonaldFR,MacLeanLD,DossetorJ.Asep

ticnecrosisfollowingrenaltrans-plantation.JBoneJointSurg[Am]1968;50:1577-89.28CattoMA.Histologicalstudyofavaseularnecrosisofthefemoralheadaftertranscervicalfracture.JBoneJ3ointSurg[Br]1965;47:749-76.29CattoM.Thehistologicalappearanceoflatesegmentalcollapseofthefemoralheadaftertranscervicalfracture.J3BoneJointSurg[Br]1965;47:777-91.30FicatRP,ArletJ.Ischemitaandnecrosisofbone.Baltimore:WilliamsandWilkins,1980.31GlimcherMJ,KenzoraJE.Thebiologyofosteonecrosisofthehumanfemoralheadanditsclinicalimplications:PartIClinOrthop1979;138:284-309.32GlimcherMJ,KenzoraJE.Thebiologyofosteonecrosisofthehumanfemoralheadanditsclincalimplications:PartII.ClinOrthop1979;139:283-312.33GlimcherMJ,KenzoraJE.Thebiologyofosteonecrosisofthehumanfemoralheadanditsclinicalimplications:PartIII.ClinOrthop1979,140:273-312.34KenzoraJE,GlimcherMJ.Pathogenesisofidiopathicosteonecrosis:theubiquitouscrescentsign.OrthopClinNorthAm1985;16:681-96.35NormanA,BulloughPG.Theradiolucentcrescentline:anearlydiagnosticsignofavascularnecrosisofthefemoralhead.BullHospJrtDisOrthopInst1963;24:99-104.36BoskeyAL,RaggioCL,BulloughP,KinnettJG.Changesinthebonetissuelipidsinpersonswithsteroidandalcoholinducedosteonecrosis.ClinOrthop1983;172:289-95.37BonnarensF,HernandezA,D'AmbrosiaR.Bonescinti-graphicchangesinosteonecrosisofthefemoralhead.OrthopClinNorthAm1985;16:697-703.38JergesenHE,HellerM,GenantHK.Magneticresonanceimaginginosteonecrosisofthefemoralhead.OrthopClinNorthAm1985;16:705-16.39SevittS,ThompsonRG.Thedistributionandanastomosesofarteriessupplyingtheheadandneckofthefemur.JBoneJointSurg[Br]1965;47:560-73.40BrookesM.Thebloodsupplyofbone.NewYork:Appleton-Century-Crofts;1971:193.41BrookesM.Thebloodsupplyofbone.NewYork:Appleton-Century-Crofts;1971:81.42RoederLF,DeLeeJC.Femoralheadfracturesassociatedwithposteriorhipdislocations.ClinOrthop1980;147:121-30.43MannRL.Avascularnecrosisofthefemoralheadfollowingintertrochantericfracture.ClinOrthop1973;92:108-15.44JacobsB.Epidemiologyoftraumaticandnon-traumaticosteonecrosis.ClinOrthop1978;130:51-67.45JonesJP.Osteonecrosis.In:McCartyDJ,ed.Arthritisandalliedconditions.11thed.Philadelphia:LeaandFebiger,1989:1546.46JonesJP,EnglemanEP,SteinbachHL,MurrayWR,Rambo0N.Fatembolizationasapossiblemechanismproducingavascularnecrosis.ArthritisRheum1965;8:449.47JonesJP,SakovichL,AndersonCE.Experimentallyproducedosteonecrosisasaresultoffatembolism.In:BeckmanEL,ElliotDH,SmithEM,eds.Dysbarism-relatedosteonecrosis.Washington:DeptHEW,1974:117-32.48JonesJP.Fatembolismandosteonecrosis.OrthopClinNorthAm1985;16:595-633.49WangGJ,SweetDE,RegerSI,ThompsonRC.Fatcellchangesasamechanismofvascularnecrosisofthefemoralheadincortisone-treatedrabbits.JBonejointSurg[Am]1977;59:729-35.50SweetDE,MadewellJE.Pathogenesisofosteonecrosis.In:ResnickD,NiwayamaG,eds.Diagnosisofboneandjointdisorders.Vol3.Philadelphia:Saunders,1981:2781-831.51DuBoisEL.Lupuserythematosus.2nded.LosAngeles:USCPress,1974:332-42.52ZizicTM,MarcouxC,HungerfordDS,StevensMB.Predictivevalueofhemodynamicstudiesinischemicnecrosisofbone[Abstract].ArthritisRheum1983;26:S37.53HungerfordDS,ZizicTM.Alcoholismassociatedischaemicnecrosisofthefemoralhead.ClinOrthop1978;130:144-53.54DecompressionSicknessPanel,MedicalResearchCouncil.Asepticbonenecrosisincommercialdivers.Lancet1981;ii:384-8.55FitzpatrickDJ.Avascularnecrosisofthefemoralhead:itsnaturalhistoryandpathology.Dublin:TrinityCollege,1977.(Thesis.)56MielantsH,VeyesEM,DeBussereA,VanderJeughtJ.Avascularnecrosisanditsrelationtolipidandpurinemetabolism.JRheumatol1975;2:430-6.57FicatRP.Treatmentofavascularnecrosisofthefemoralhead.In:ProceedingsoftheeleventhopenscientificmeetingoftheHipSociety.StLouis:Mosby,1983:279-95.58HungerfordDS.Bonemarrowpressure,venographyandcoredecompressioninischaemicnecrosisofthefemoralhead.In:ProceedingsoftheseventhopenscientificmeetingoftheHipSociety.StLouis:Mosby,1979:218-37.59KenzoraJE.Treatmentofidiopathicosteonecrosis:thecurrentphilosophyandrationale.OrthopClinNorthAm1985;16:717-25.60SteinbergME,BrightonCT,HaykenGB,ToozeSE,SteinbergDR.Electricalstimulationinthetreatmentofosteonecrosisofthefemoralhead.OrthopClinNorthAm1985;16:747-56.61LotkePA,EckerML.Osteonecrosisoftheknee.OrthopClinNorthAm1985;16:797-808.62CornellCN,SalvatiE,PellicciPM.Long-termfollow-upoftotalhipreplacementinpatientswithosteonecrosis.OrthopClinNorthAm1985;16:757-