Rev. bras. ortop. vol.49 número6

w w w . r b o . o r g . b r

Original

Article

Evaluation

of

the

reproducibility

of

the

Tronzo

classification

for

intertrochanteric

fractures

of

the

femur

夽

,

夽夽

Fernando

Abdala

Silva

Oliveira,

Ricardo

Basile

_,

_Bruno

_Cézar

_Brabo

_Pereira,

Rafael

Levi

Louchard

Silva

da

Cunha

CarminoCaricchioMunicipalHospital,SãoPaulo,SP,Brazil

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received24September2013 Accepted21November2013 Availableonline7November2014

Keywords:

Femoralfractures/classification Femoralfractures/radiographs Reproducibilityoftests

a

b

s

t

r

a

c

t

Objective:To evaluate the inter-observerreproducibility of the Tronzo classificationfor transtrochantericfracturesusingthekappaconcordancecoefficient(␬).

Methods:Twentyradiographicimagesinanteroposteriorviewonhipswithtranstrochanteric fracturesofthefemurwereused.Thesewereclassifiedby12observersusingtheTronzo method.Theimageswerepresentedinsequenceandaquestionnairecontainingallthe options oftheTronzoclassificationwasfilled out,alongwithasimplifiedclassification usingTronzodivisionsintotwogroups(stableandunstable).Thedatawereanalyzedby meansofthekappaconcordancetest.

Results:Thefollowingkappaindiceswerefound:forimageswithstablefractures(Tronzo1 and2),0.11;forimageswithunstablefractures(Tronzo3,3variant,4and5),0.52;andforthe completeclassification,0.44(moderateconcordance).Inturn,thesimplifiedclassification didnotincreasetheconcordancerates.

Conclusion: TheTronzoclassification isunsuitableforclinicalpractice.We suggestthat anothersystemshouldbeusedorcreatedforthistypeoffracture.

©2014SociedadeBrasileiradeOrtopediaeTraumatologia.PublishedbyElsevierEditora Ltda.Allrightsreserved.

Avaliac¸ão

da

reprodutibilidade

da

classificac¸ão

de

Tronzo

para

fraturas

intertrocantéricas

do

fêmur

Palavras-chave:

Fraturasdofêmur/classificac¸ão Fraturasdofêmur/radiografias Reprodutibilidadedostestes

r

e

s

u

m

o

Objetivo:Avaliara reprodutibilidade, interobservadores, da classificac¸ãode Tronzopara fraturastranstrocantéricascomousodocoeficientedeconcordânciakappa(␬).

Métodos:Foram usadas 20 imagens de radiografias do quadril na incidência antero-posterior com fraturas transtrocantéricasdo fêmur, classificadas, segundo Tronzo,por

夽

Pleasecitethisarticleas:OliveiraFAS,BasileR,PereiraBCB,daCunhaRLLS.Avaliac¸ãodareprodutibilidadedaclassificac¸ãodeTronzo parafraturasintertrocantéricasdofêmur.RevBrasOrtop.2014;49:581–585.

夽夽

WorkdevelopedintheHipGroupoftheCarminoCaricchioMunicipalHospital,SãoPaulo,SP,Brazil. ∗ _{Correspondingauthor.}

E-mail:ricardo-basile@uol.com.br(R.Basile).

http://dx.doi.org/10.1016/j.rboe.2014.11.001

582

12observadores.Asimagensforamapresentadasemsequênciaefoipreenchidoum ques-tionáriocomtodasasopc¸õesdaclassificac¸ãodeTronzo,alémdaclassificac¸ãosimplificada, comadivisãodeTronzoemdoisgrupos(estáveiseinstáveis).Osdadosforamanalisados pormeiodotestedeconcordânciadekappa.

Resultados: Foramencontradososseguintesíndicesdekappa:paraimagenscomfraturas estáveis(Tronzo1e2),0,11;paraimagenscomfraturasinstáveis(Tronzo3,3variante,4 e5),0,52;eparaaclassificac¸ãocompleta,0,44(concordânciamoderada).Porsuavez,a classificac¸ãosimplificadanãoaumentouosíndicesdeconcordância.

Conclusão: Aclassificac¸ãodeTronzonãoéadequadaparaapráticaclínica.Sugerimosouso ouacriac¸ãodeoutrosistemaparaessetipodefratura.

©2014SociedadeBrasileiradeOrtopediaeTraumatologia.PublicadoporElsevier EditoraLtda.Todososdireitosreservados.

Introduction

Transtrochantericfracturesofthefemuroccurintheregion betweenthegreaterand lessertrochantersofthe proximal femurandareexclusivelyextracapsular.1_InBrazil,ina sur-veyconductedbytheMinistryofHealththroughtheNational HealthSystem(SUS),itwasfoundthat90%ofthefunding des-tinedfororthopedicdiseaseswasconsumedbyninediseases, whichincludedtranstrochantericfractures.2

Theirconsequencesforsocietyarealarming.Around one-thirdofthepatientsdieduringthefirstyearaftertheinjury, approximately50%becomeincapableofwalkingunaidedor goingupstairsand20%starttoneedcontinualhomecare.3

Several systems have been used to classify transtrochanteric fractures and thus guide their treat-ment.ThecommonestofthesearetheTronzo,4_Evans5,6_and AOclassifications.5–7

TheTronzoclassificationiswidelyusedinBrazilian ortho-pedicsandtraumatologyservices.

Exact documentation of the fracture depends on the capacity for radiographic evaluation and classification. Its reproducibilitydependsonthesurgeon’sskillininterpreting aclassificationsystem.Thepositionofthefracturedlimb,the radiographictechniqueandthesurgeons’levelsofexperience arefactorsthatcontributetowardthereproducibilityofa clas-sificationsystem.8

Ideally, a classificationsystem should be easy to apply, reliableandhelpfulinmakingtreatmentdecisions,and con-sequentlyshouldinfluencethefinalresult.Anidealsystem shouldnothaveinterobserverdiscrepancies.9

Therefore, the aim of this study was to assess the interobserverreproducibilityoftheTronzoclassificationfor transtrochantericfracturesofthefemurusingthekappa coef-ficientofconcordance(␬).

Materials

and

methods

A cross-sectional observational study was conducted, in which20radiographicimagesofhipswithtranstrochanteric fracturesofthefemur,inanteroposteriorview,wereused.All thepatientswereovertheageof65yearsandhadsuffered low-energy trauma. The following were used as exclusion criteriafortheimages:pathologicalfracturescausedbybone

tumors,previoussurgeryintheregionoftheproximalfemur andimagesoftranstrochantericfracturesfrompatientsunder theageof65years.

TheradiographswereclassifiedinaccordancewithTronzo, by12 observers:four specialistswho weremembers ofthe society fororthopedictrauma surgery, fourthird-year resi-dents and four first-year residents of an orthopedics and traumatologyservice.Alloftheseobserversweregivenprior explanationsregardingtheTronzoclassification,withgraphic imagesonspreadsheets.

The radiographs ofthe fractures were presentedin the formofslidesinthePowerPoint®_{software,insequence,} indi-viduallynumberedfromonetotwenty,andeachimagewas analyzedforamaximumof45s.Duringthepresentation,a questionnairewasfilledouttogatherdataonalltheoptions oftheTronzoclassification(Anexo1).

Theseriesofradiographswereanalyzedbymeansofthe completeTronzoclassification(sixtypes).Subsequently,this classificationwassubdividedforanalysisintotwosimple sub-types:stablefractures(TronzoI_andII_{)andunstablefractures} (TronzoIII_,III_variant,IV_andV_).

In1974,Tronzosubdividedthesefracturesintofivetypes. TypesIandIIwerestable;typeIwasdescribedasan incom-plete transtrochantericfracture,whiletypeII_{could present} fracturingofthelessertrochanter,butwithoutposteromedial comminution.TypesIII_andIV_{presentedposteromedial} com-minution;intypeIII,thediaphysiswasbroughttoamedial positionandproximalcalcarwasfittedtoit.When fractur-ingofthegreatertrochanterwasalsopresent,thesituation wasclassifiedasIII_{variant(Fig.1).IntypeIV,thediaphysis} wasbroughttoalateralposition,thefracturelinewasmore verticalandthecomminutionwasgenerallygreater.TypeV hadaninvertedline,fromlateraltomedialandfromdistal toproximal,whichmadethefractureunstable(Fig.2).In pre-sentinghisclassification,Tronzodescribedtheosteosynthesis techniquesproposedforthevarioustypesoffracture.1,4

The datagathered were analyzed statisticallyusing the kappaconcordancetest.ThesoftwareusedcomprisedSPSS V16,Minitab15andExcelOffice2007.

Results

Fig.1–TranstrochantericfractureoftypeIIIvariant.Notethefracturingofthegreatertrochanter.

forimageswithunstablefractures,0.52;andforthecomplete classification,0.44(Table1).

Whenwedividedtheobserversintogroupsconsistingof alltheresidentstogetherversustheorthopedictrauma spe-cialists,wefoundthefollowingresults:forresidents,0.49for thecompleteclassification,0.18forstablefracturesand0.55 forunstablefractures;fororthopedictraumaspecialists,0.39 forcompletefractures,0.20forstablefracturesand0.52for unstablefractures(Table2).

Table1–Concordanceforthetotalsample.

kappa pvalue

Stable 0.11 <0.001

Unstable 0.52 <0.001 Complete 0.44 <0.001

Table2–Concordanceaccordingtotitle.

kappa pvalue

Residents

Stable 0.18 <0.001

Unstable 0.55 <0.001

Complete 0.49 <0.001

Orthopedictraumaspecialists

Stable 0.20 <0.001

Unstable 0.52 <0.001

Complete 0.39 <0.001

Lastly, when wetook into consideration the divisionof

theobserversintoallofthegroups,weobtainedthe follow-ing resultsforthegroupoffirst-year residents:0.50forthe completeclassification,0.24forstablefracturesand0.66for unstable fractures. For the third-yearresidents, the results were0.53forthecompleteclassification,0.37forstable frac-turesand0.51forunstablefractures.Meanwhile,theresultfor orthopedictraumaspecialistswas0.39forthecomplete classi-fication,0.20forstablefracturesand0.52forunstablefractures

(Table3).

Table3–Concordanceaccordingtogroup.

kappa pvalue

First-yearresidents

Stable 0.24 <0.001

Unstable 0.66 <0.001 Complete 0.50 <0.001

Third-yearresidents

Stable 0.37 <0.001

Unstable 0.51 <0.001 Complete 0.53 <0.001

Orthopedictraumaspecialists

Stable 0.20 <0.001

584

Fig.2–RadiographshowingTronzotypeV transtrochantericfracture.

Discussion

Thekappaconcordancecoefficient isoftenused whenthe intra- and interobserver reliability and reproducibility are evaluated.Thiscoefficientprovidespairedproportionsof con-cordancebetweenobservers,whichmaybychancebecorrect. Kappavaluescanrangefrom−1to+1;−1indicatestotal dis-cordance,zeroindicatesafortuitouslevelofconcordanceand +1indicatestotalconcordance.Ingeneralterms,kappavalues below0.5areconsideredunsatisfactory, valuesbetween0.5 and0.75areconsideredadequateandsatisfactoryandvalues above0.75areconsideredexcellent.8

Landisclassifiedkappavaluesasfollows:poor(below0), slight(0–0.2),weak(0.2–0.41),moderate(0.41–0.6),substantial (0.61–0.8)andalmostperfect(0.81–1).10

Theclassificationofafractureisthebasisforthechoice of treatment. Thus, it is important to assess the validity of a classification system. Gusmão et al.11 _{evaluated the} Gardenclassificationsystemforfemoralneckfracturesand foundpoorlyreproducibleconcordance(0.32).Schwartsmann etal.7 _{didthesame inrelationtothe AOclassificationfor} transtrochantericfracturesandfoundweakconcordancefor thecomplete AOclassification(0.34).Pervez etal.5 _founda weakconcordanceindex(0.34)fortheEvans/Jensen classifi-cationfortranstrochantericfractures.

InoursearchesintheLilacsandMedlinedatabases,wedid notfindanystudiesthatevaluatedtheconcordanceindexof theTronzoclassification.

ThegreatestdifficultyregardingtheTronzoclassification4 isininterpretingthestabilityofthefracture,sincetheauthor’s description states that the stability isin the comminution of the posteromedial wall, and also that in type II frac-tures(whichareconsideredstable),thelessertrochantermay befractured.Thisprobablyexplainstheinsignificant repro-ducibilityforstablefractures(0.11)thatwefoundinourstudy, whichgaverisetolargebiasintheobservers’classifications oftypeIIandIIIfractures.

Theunstablefracturespresentedahigherdegreeof con-cordancethanshownbythestablefractures(moderate;0.52), betweentheobserversingeneral.Thiscanbeexplainedby Tronzo’sdescriptionsofthepatternsoffracturetypesIVand V,giventhattheyarespecificanddonotleaveroomfordoubt regardingtheirpatternofinstability.

Schipperetal.6_{studied20fracturesusing15observersto} makethecompleteAOclassification.Theinterobserverkappa valuewas 0.33.When theyusedthesimplified AO classifi-cation, thekappa valuewas0.67.They concludedthat the simplifiedAOclassificationhadgoodreproducibilitybutthat this wasnotrepeated inthe completeclassification.Inour study,simplificationofthe Tronzo classificationintostable andunstablefracturesdidnotincreasetheconcordanceindex inrelationtothecompleteclassification.

Fungetal.9_{reportedthat,unexpectedly,inastudy} eval-uatingthereproducibilityoftheEvans/Jensenclassification, increasingexperienceamongtheobserverscorrelatedwitha decliningconcordanceindexandtheyinferredthatthis clas-sification system was unclear or that another system was necessary.

InanotherstudythatassessedtheEvans/Jensen classifica-tionusingfourresidents(twoseniorsandtwojuniors)and52 radiographstoevaluatetheconcordance,Gehrchenetal.12_did notfindanydifferencesandtookincreasingexperienceinto consideration.

Schipper et al.6 _{used the AO classification and did not} observeanyappreciabledifferenceintheconcordanceamong differentgroupsofprofessionals(residents,orthopedic sur-geonsandradiologists).

We observeda declineinthe kappa concordanceindex when the increase inthe observers’experience was taken into consideration:0.50forthe first-yearresidents, 0.53for thethird-yearresidentsand0.39fortheorthopedictrauma surgeons.Thisdifferencewasstatisticallyinsignificantwhen the groupsofresidentsweretakeninto consideration.This resultwasunexpectedbecausegreater trainingand greater experiencegenerallyoughttoimplygreaterreliability.

Although the Tronzo classification is used in many Brazilianorthopedicservices,itsvalueinpracticeisvery ques-tionable, given that we found concordance of 0.44 for the completeclassification,whichisconsideredmoderate accord-ingtoLandis.Itssubdivisionintostableandunstablefractures presentedmajorbias,inthatthestablefracturespresented slighttoweakconcordance(0.11)andtheunstablefractures, moderatetosubstantial(0.52).

Conclusion

classification into stable and unstable fractures did not increasetheconcordanceindices.

Therefore,theTronzoclassificationdoesnotfitwithinthe criteria for an adequate classification for clinical practice. Thus, we suggest that another system should be used or anotheroneshouldbecreatedforthistypeoffracture.

Conflicts

of

interest

Theauthorsdeclarenoconflictsofinterest.

Anexo

1.

Data-gathering

table

Observer:

Specialty:

IMAGE TRONZOCLASSIFICATION

– I II III III_VARIANT IV V

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

r

e

f

e

r

e

n

c

e

s

1.KöberleG.Fraturastranstrocanterianas.RevBrasOrtop.

2001;36(9):325–9.

2.DaniWS,AzevedoE.Fraturastranstrocanterianas.Periódico nainternet,2008.Consultadoem22dejulhode2011. Disponívelem:http://www.moreirajr.com.br/

revistas.asp?fase=r003&idmateria=3245

3.LimaALP,AzevedoAJ,AmaralNP,FranklinCE,GiordanoV.

Tratamentodasfraturasintertrocanterianascomplacae

parafusodeslizante.RevBrasOrtop.2003;38(5):271–9.

4.TronzoRG.Symposiumonfracturesofthehip.Special

considerationsinmanagement.OrthopClinNorthAm.

1974;5(3):571–83.

5.PervezH,ParkerMJ,PryorGA,LutchmanL,ChirodianN.

Classificationoftrochantericfractureoftheproximalfemur:

astudyofthereliabilityofcurrentsystems.Injury.

2002;33(8):713–5.

6.SchipperIB,SteyerbergEW,CasteleinRM,vanVugtB.

RealiabilityoftheAO/ASIFclassificationforpertrochanteric

femoralfractures.ActaOrthopScand.2001;72(1):36–41.

7.SchwartsmannCR,BoschinLC,MoschenGM,Gonc¸alvesRZ,

RamosASN,GusmãoPDF,etal.Classificac¸ãodasfraturas

trocantéricas:avaliac¸ãodareprodutibilidadedaclassificac¸ão

AO*.RevBrasOrtop.2006;41(7):264–7.

8.DirchlDR,CannadaLK.Classificationoffractures.In:Bucholz

RW,HeckmanJD,Court-BrownCM,TornettaP3rd,McQueen

MM,RicciWM,editors.Rockwood&Grennfracturesin

adults.7thed.Philadelphia:LippincottWilliams&Wilkins;

2009.p.45–51.

9.FungW,JonssonA,BuhrenV,BhandariM.Classifying

intertrochantericfracturesoftheproximalfemur:does

experiencematter?MedPrincPract.2007;16(3):198–202.

10.LandisJR,KochGG.Themeasurementofobserveragreement

forcategoricaldata.Biometrics.1977;33(1):159–74.

11.GusmãoPDF,MothuesFC,RubinLA,Gonc¸alvesRZ,Telöken

MA,SchwartsmannCR.Avaliac¸ãodareprodutibilidadeda

classificac¸ãodeGardenparafraturasdocolofemoral.Rev

BrasOrtop.2002;37(9):381–6.

12.GerchenPM,NeilsenJ,OlensenB.Poorreproducibilityofthe

Evan’sclassificationofthetrochantericfracture:assessment

of4observersin52cases.ActaOrthopScand.1993;64(1):