Three-dimensional evaluation of skeletal deformities of the pelvis and lower limbs in ambulant children with cerebral palsy

Science Arts & Métiers (SAM)

is an open access repository that collects the work of Arts et Métiers Institute of

Technology researchers and makes it freely available over the web where possible.

This is an author-deposited version published in:

https://sam.ensam.eu

Handle ID: .

http://hdl.handle.net/10985/15878

To cite this version :

Abir MASSAAD, Ayman ASSI, Ziad BAKOUNY, Christophe SAURET, Nour KHALIL, Wafa

SKALLI, Ismat GHANEM - Three-dimensional evaluation of skeletal deformities of the pelvis and

lower limbs in ambulant children with cerebral palsy - Gait & Posture - Vol. 49, p.102-107 - 2016

Any correspondence concerning this service should be sent to the repository

Administrator :

archiveouverte@ensam.eu

Three-dimensional

evaluation

of

skeletal

deformities

of

the

pelvis

and

lower

limbs

in

ambulant

children

with

cerebral

palsy

Abir

Massaad

,

Ayman

Assi

*

,

Ziad

Bakouny

,

Christophe

Sauret

,

Nour

Khalil

,

Wafa

Skalli

,

Ismat

Ghanem

a

LaboratoryofBiomechanicsandMedicalImaging,FacultyofMedicine,UniversityofSaint-Joseph,Beirut,Lebanon

b

GaitAnalysisLaboratory,SESOBEL,ZoukMikael,Lebanon

c

InstitutdeBiomécaniqueHumaineGeorgesCharpak,ArtsetMétiersParisTech,Paris,France

d_{Htel-DieudeFranceHospital,UniversityofSaint-Joseph,Beirut,Lebanon}

Articlehistory:

Received29December2015 Receivedinrevisedform21June2016 Accepted22June2016 Keywords: Lowerlimbs Cerebralpalsy 3Dreconstructions Children Pelvis Posture ABSTRACT

Skeletalabnormalities, affecting posture andwalking pattern,increase withmotorimpairment in childrenwithcerebralpalsy(CP).However,itisnotknownwhethertheseskeletalmalalignmentsoccur inchildrenwithslightmotorimpairment.Ouraimwastoevaluateskeletalmalalignmentatthelevelof thepelvisandlowerlimbsinambulantchildrenwithCP,withslightmotorimpairment,usingalowdose biplanarX-raytechnique.

Twenty-sevenchildrenwithspasticCP(meanage:10.94years,7Hemiplegia,20Diplegia,GMFCS levelsI:17,II:10),withnoprevioustreatmentsatthehipsandknees,underwentEOS1biplanarX-rays.A controlgroup consisting of22 typicallydeveloping childrenwasalso included. Three-dimensional reconstructionsof thepelvis andlowerlimbswereperformedin ordertocalculate11radiological parametersrelatedtothepelvis,acetabulumandlowerlimbs.

PelvicincidenceandsacralslopeweresignificantlyincreasedinchildrenwithCPcomparedtoTD children(48_7_vs.43_8_,42_7_vs.38_5_{,respectively,p=0.003).Acetabularparametersdid} not significantly differ betweenthe two groups. Femoral anteversion and neck shaft angle were significantlyincreasedin childrenwith CP(2512 _vs.14_7_{,p<0.001; 134}_5 _vs.131_5_, p=0.005respectively).Nodifferencewasfoundfortibialtorsion.

ThisstudyshowedthatevenslightlyimpairedchildrenwithCPhaveanantevertedandabductedfemur andpresentpositionalandmorphologicalchangesofthepelvisinthesagittalplane.Theorientationof theacetabulumin3DseemstonotbeaffectedwhenchildrenwithCPpresentslightmotorimpairment. ã2016ElsevierB.V.Allrightsreserved.

1. Introduction

Cerebral Palsy(CP)is a neurologicaldisorderthat cancause

musclespasticity.TheincreasedmuscletoneofpatientswithCPis

knowntoaffecttheirpostureand walkingpattern[1].Skeletal

malalignments of the lower limbs are often encountered in

childrenwithCPinthethreeplanes[2–4]andaremainlycausedby

spasticity, abnormal posture, delays in gaining independent

walkingaswellasgaitalterations[5].

Previousstudieshavereportedthatskeletalmalalignmentsof

the lower limbs increase with motor impairment [2,6]. These

studieshavedescribedonlyafewfemoralparameters(neckshaft

angle and femoral parameters) in different groups of CP that

differed bytheir motor impairment but didnot compare their

resultstothoseofcontrolgroups[2,7].Moreover,itisnotknown

whethertheseskeletalmalalignmentsoccureveninchildrenwith

slightmotorimpairment.

Otherstudieshavereportedskeletalabnormalitiesofthehip

and acetabulum in children with CP, such as hip dysplasia or

acetabulardeficiencies[8,9].Itwas suggestedthatthese

abnor-malities,which had beenreportedin severelyaffectedchildren

* Correspondingauthorat:LaboratoryofBiomechanicsandMedicalImaging, CampusofInnovationandSport,UniversityofSaint-Joseph,DamascusStreet,P.O. Box1104-2020Beirut,Lebanon.

E-mailaddress:ayman.assi@gmail.com(A.Assi).

1

LaboratoryofBiomechanicsandMedicalImaging,CampusofInnovationand Sport,UniversityofSaint-Joseph,DamascusStreet,P.O.Box1104-2020Beirut, Lebanon.

2

InstitutdeBiomécaniqueHumaineGeorgesCharpak,ArtsetMétiersParisTech, 151Boulevarddel’Hôpital,75013Paris,France.

3

Department ofOrthopaedicSurgery,Hotel-DieudeFrance Hospital,Alfred NaccacheAvenue,Achrafiyeh,Beirut,Lebanon.

withCP,weremainlyduetoabnormalweight-bearingonthelower

limbs. Since slightly impaired children with CP also present

abnormal weight-bearing, it would be interesting to elucidate

whethertheiracetabularparametersareaffectedaswell.

The assessment of skeletal malalignments is important in

surgicaldecision-makingforchildrenwithCP[10],andisusually

based on physical examination [11], two-dimensional X-Ray

imaging [12] and three-dimensional (3D) gait analysis [13].

However, torsional abnormalities, such as abnormal femoral

anteversion(FA)ortibialtorsion(TT),requiremeasurementsin3D.

Three-dimensionalcomputedtomography(CT)isareliabletool

to measure FA and TT [10]. This technique has been used to

measuretheseparametersinchildrenwithCP[6].However,the

useofthistechniqueislimitedbythehighexposuretoradiation

[14]andthefactthatitsaccuracyisaffectedbypositionalvariables

[10],whicharedifficulttocontrolinchildrenwithCP.Forthese

reasons,3D CTscan is not routinely usedtoassess lowerlimb

skeletalabnormalitiesinchildrenwithCP.

LowdosebiplanarX-rayshavebeenpreviouslyfoundtobea

fastandreliabletoolforthemeasurementofskeletal

malalign-ments in the standing position [15–17]. This technique allows

quantificationofsubject-specificskeletalparametersbasedon3D reconstructionofthepelvisandlowerlimbs[16,18]andcouldbea viablealternativetoCTindailypractice.

Theaim of this study was toinvestigate whetherambulant

childrenwithCP,withslightmotorimpairment,presentskeletal

deformities at the pelvisand lower limbs, using thelow dose

biplanarX-raytechnique.

2. Methods 2.1.Population

ThisisanIRBapprovedcross-sectionalcase-controlreviewofa

consecutiveseriesofchildrenwithCPwhounderwentfullbody

lowdosebiplanarradiographsusingEOS1system(EOSImaging,

ParisFrance),inthesettingoftheirpre-treatmentevaluation.

Twenty-sevenchildrenwithspasticCP(17boys,10girls)were

included.MotorimpairmentwasdefinedaccordingtoGMFCSE&R

levels [19]. Children with a history of previous orthopedic

interventions(botulinumtoxin,casting,surgery),otherthanthose

directed to treat equinus (botulinum toxin in gastrocnemius,

lengthening)atleastoneyearpriortotheenrollmentinthestudy,

wereexcluded.

Twenty-two typically developing (TD) children(11 boys,11

girls),forwhombiplanarradiographsoftheirlowerextremities

had been prescribedfor atypical pain and had not shown any

abnormality,wereincludedinthestudy(Table1).Theparentsof thechildrenintheCPandTDgroupswereinformedofthepossible

useof theirchild’s radiographs forresearchpurposes and they

approvedandsignedawrittenconsentform.

2.2.Acquisitionand3Dreconstruction

EOS1biplanarX-rayacquisitionwasperformedinthestanding

position.Thelowerlimbswereslightlyshiftedinthesagittalplane,

in order to avoid knee overlap on the lateral view therefore

facilitating3Dreconstruction[15].Typicallydevelopingchildren

and childrenwithCP wereaskedtostand inthe freestanding

position, with hands on cheeks and flexed elbows [20]. This

positionisknowntoreproducethenaturalpostureofthesubject.

ChildrenwithCPwereallowed,whennecessary,toputtheirhands

ontheembeddedbarintheEOS1cabinforbetterstability,without itaffectingtheirnaturalposture.

Onetrainedoperatorperformedall3Dreconstructionsofboth

lower limbs using SterEOS1 (version 1.6.4.7977) based on a

previouslydescribed method[15]: afewgeometricalprimitives

and axes are _first digitalized, in order to obtain a simpli_fied

personalized parametrical model. Then, a 3D morpho-realistic

parametricmodel,basedonadatabaseofCT-Scanreconstructions,

is deformed. This deformation is based on statistical shape

modeling.Pelvicreconstructionisbasedonthesametechnique

used for the lower limbs and is performed using a dedicated

softwaredevelopedatArtsetMétiersParisTech[21](Fig.1). 2.3.Radiologicalparameters

Pelvic, acetabular, femoral and tibial parameters were

auto-maticallycalculatedfromthe3Dreconstructionsofthepelvisand

lowerlimbsofeachsubject.Elevenradiologicalparameterswere

chosenaccordingtotheirclinicalrelevanceinthediagnosisand

decision-makingforchildrenwithCP.Someoftheseparameters

areroutinelyused:pelvicincidence(PI),sacralslope(SS),pelvictilt

(PT),femoralanteversion(FA),neckshaftangle(NSA)andtibial

torsion(TT).Acetabularparametersarelessusedinthe

manage-mentofCP:1)acetabularanteversion(AAnt)reflectstherotation oftheacetabulumaroundtheverticalaxis[22].Inordertorender

this angle morphological, the vertical axis was defined as the

perpendiculartothesacralplate;2)acetabularinclination(AInc) reflectstherotationoftheacetabulumaroundthepostero-anterior axis,definedasaparalleltothesacralplate[22,23];3)acetabular

tilt reflects the rotation of the acetabulum around the

medio-lateralaxisofthepelvis[24].Thelistandthedescriptionofeachof theseparametersaredetailedinTable2anddisplayedinFig.2. Theseparameterswerecalculatedforbothlimbsofthechildrenin

the CP and TD groups. The unaffected limbs of childrenwith

hemiplegiawereexcludedfromfurtheranalysis.Themeanvalues

of the pelvic, acetabular and lower limb parameters werealso

separatelycalculatedforchildrenwithGMFCSlevelsI(N=17)and II(N=10).

2.4.Statistics

Comparisons between demographic and radiological

param-etersofCPandTDchildrenwereperformedusingStudent’st-test, Mann–Whitney’stestorWelch’sunequalvariancest-test.Signi fi-cancelevelwassetat0.05.

Means comparison betweenTD children and CP sub-groups

(GMFCSlevelsIandII)wereperformedusingone-wayanalysisof

variance(ANOVA)withHochberg_’sGT2post-hoctest.Variancesof

parameters between CP and TD groups were compared using

Levene’stest. Table1

Demographiccharacteristicsofthetypicallydevelopingchildren(TD)andchildrenwithcerebralpalsy(CP).DistributionofthechildrenwithCPaccordingtotheGrossMotor ClassificationSystem(GMFCS).

Groups Age(years) Weight(kg) Height(cm) N SpasticHemiplegia/Diplegia(N) GMFCSLevels:I/II(N)

MeanSD(min–max) MeanSD MeanSD

TDchildren 123(6.5–19) 44.413.8 149.915 22

A post-hocpoweranalysiswas performedbycalculatingthe effectsizeofeachsignificantresult.Cohen’sdwascalculatedwhen

the t-test is applied for comparisons between CP and TD

populations. Cohen’s f was applied when the ANOVA test was

performedbetweenTDand CPwithGMFCSlevelsIand II[25].

Matlab1 (The Mathworks, Inc, Natick, MA, USA) and Xlstat1

(Addinsoft,Paris,France)wereusedforstatisticalanalysis. 3. Results

3.1.ComparisonbetweenTDandCPchildren

Comparisonofage,weightandheight(Table1)betweenthe

twogroupsshowednosignificantdifferences(p=0.355,p=0.441

and p=0.129 respectively). Mean values of the radiological

parametersfortheTDandCPchildrenarereportedinTable3.

Themeanvalueofthepelvicincidence(PI)was significantly

increased in the CP population compared to TD (487 vs.

438 respectively, p=0.003, d=0.66). A significantly more verticalsacralplatewasfoundinchildrenwithCPcomparedtothe controlgroup(427vs.385respectively,p=0.003,d=0.65).

AmoreantevertedandabductedfemoralneckwasfoundintheCP

groupcomparedtocontrols(FA:2512_vs.14_7_{respectively,}

p<0.001, d=1.11 and NSA: 134_5 _{vs 131}_5 _{respectively,}

p=0.005,d=0.6).

Significantly higher variability was found for the following

parameters in children with CP compared to TD children: FA

(p=0.007),PT(p=0.002),andtheacetabulartiltangle(p=0.012). Table2

Thedefinitionoftheacetabularparameters.

Pelvicparameters Line1 Line2

PelvicIncidence(_{) PI Sacro-acetabularline(1) Normalsacralline(2)}

SacralSlope() SS Horizontalline Thelinetangenttothesacralplate

PelvicTilt() PT Verticalline Sacro-acetabularline

Acetabular parameters

Line1 Line2

Acetabular coverage(%)*

ACov Surfaceofthespheremodelingtheacetabulum Surfaceoftheprojectionoftheacetabulumonthespheredefinedin “Line1”

Sacro-acetabular angle(₎

SA Thelinedefinedbytheintersectionbetweentheacetabularedgeplane (3)andEOS_Splane(4)

ThetangenttothesacralplateprojectedinEOS_Splane Acetabulartilt(_{) AT Thelinedefinedbytheintersectionbetweentheacetabularedgeplane}

(3)andEOS_Splane(4)

Thepostero-anterioraxisoftheEOS1system,passingthroughthe centeroftheacetabulum

Acetabular inclination()

AInc Thelinenormaltotheacetabularedgeplane(3) andpassingthroughthecenteroftheacetabulum

ThenormaltotheSplane(5) Acetabular

anteversion(₎ AAnt ₍₃₎The_andline_Sdefined_plane(5)bytheintersectionbetweentheacetabularedgeplane _planeTheline₍₆₎defined_andSplanebythe₍₅₎intersectionbetweenthepelvicsagittal

Lowerlimbparameters Line1 Line2

FemoralAnteversion(_{) FA Femoralneckaxis Posteriorcondylarline}

NeckShaftAngle(_{) NSA Femoralneckaxis Proximaldiaphysisaxis}

TibialTorsion() TT Posteriortibialplatesline Bi-malleolaraxis

(1)Sacro-acetabularline:linejoiningthecenteroftheacetabulumandthecenterofthesacralplate. (2)Normalsacralline:theperpendiculartothesacralplate.

(3)Acetabularedgeplane:istheplanepassingthroughtheedgeoftheacetabulum. (4)EOS_Splane:EOSsagittalplanepassingthroughtheacetabularcenter.

(5)SPlane:istheplaneparalleltothesacralplatepassingthroughtheacetabularcenter.

(6)Pelvicsagittalplane:isaperpendicularplanetotheacetabularaxis,passingthroughthecentersofbothacetabulae.

*

ACovisaratiobetween“line1”and“line”.

Fig1.Exampleof(a)frontalandlateralEOS1radiographyofachildwithcerebralpalsy(F,8yearsold)with3Dreconstructionsofthelowerlimbs(b)andthepelvis(c)in coronal,lateralandtransverseviews.

3.2.ComparisonbetweenTDchildrenandCPchildrenwithGMFCS levelsIandII

Whilenodifferenceswerenotedbetweenthetwosub-groups

ofchildrenwithCPandTDchildrenfortheacetabularparameters,

femoralanteversionwassignificantlyincreasedinbothgroupsof

childrenwith GMFCS levels I and II compared to TD children

(p=0.001 and p<0.001 respectively, f=0.58; Table 3). Pelvic

incidence was significantly increased in children with GMFCS

levelsIand IIcomparedtoTDchildren(p=0.020andp=0.033

respectively,f=0.23).Thesacralslopewassignificantlyincreased

inchildrenwithGMFCSlevelIcomparedtotheTDgroup(p=0.001, f=0.26).

4. Discussion

Inthisstudy,acomparisonof3Dposturalandmorphological

skeletalparametersofbothpelvisandlowerlimbs,betweenTD

andambulantchildrenwithCPpresentingGMFCSlevelsIandII,

was performed using the EOS1 biplanar X-ray system. The

childrenwithCPincludedinthisstudyhadneithermedicalnor

surgical history. This is the first study that calculated the

acetabularparametersinCPandTDgroups.Whilenodifferences

werefoundfortheacetabularparametersandtibialtorsion,the

Fig.2.Illustrationofthelowerlimbsandacetabularparameters.

Pelvicparameters:a)pelvicincidence(PI),b)sacralslope(SS),c)pelvictilt(PT).Acetabularparameters:d)acetabularcoverage(ACov),e)sacro-acetabularangle(SA),f) acetabulartilt(AT),g)acetabularinclination(AInc),h)Acetabularanteversion(AAnt).Femoralandtibialparameters:i)femoralanteversion(FA),j)neckshaftangle(NSA),j) tibialtorsion(TT).

Table3

3Dsubject-specificradiologicalparametersrelatedtothepelvisandlowerlimbsfortheCPandTDchildren:Pelvicincidence(PI),sacralslope(SS),pelvictilt(PT),acetabular coverage(ACov),sacro-acetabularangle(SA),acetabulartilt(AT),acetabularinclination(AInc),acetabularanteversion(AAnt),femoralanteversion(FA),neckshaftangle (NSA),tibialtorsion(TT).

TDchildren CPchildren CPvsTD GMFCSI GMFCSII

Mean(SD) Mean(SD) pvalue Mean(SD)

pvalue(GMFCSIvsTD) Mean(SD) pvalue(GMFCSIIvsTD) Pelvicparameters PI(_{)* 43(8) 48(7) 0.003 47(6) 48(8)} 0.020 0.033 SS(_{)* 38(5) 42(7) 0.003 44(6) 39(8)} 0.001 0.983 PT() 6(5) 7(10) 0.502 3(8) 11(10) 0.299 0.070 Acetabularparameters ACov(%) 42(4) 41(4) 0.177 40(4) 42(4) 0.153 0.999 SA() 58(11) 62(12) 0.144 62(11) 61(14) 0.370 0.644 AT(_{) 20(9) 21(12) 0.700 17(11) 26(12)} 0.743 0.115 AInc(_{)* 72(6) 74(9) 0.062 75(5) 73(12)} 0.092 0.175 AAnt(_{) 31(4) 32(5) 0.205 33(4) 31(6)} 0.325 0.965

Femoralandtibialparameters

FA(_{)* 14(7) 25(12) p<0.001 24(13) 27(11)} 0.001 p<0.001 NSA(_{)* 131(5) 134(5) 0.005 134(6) 134(5)} 0.052 0.074 TT() 28(6) 28(8) 0.854 27(7) 29(8) 0.941 0.997

pelvic and femoral parameters were significantly different in

childrenwithCP.

4.1.TDchildren

Thecalculationoflowerlimbandpelvicparametersin3Dfor

TDchildrenwasrecentlyreportedintheliterature[18,26,27];in

thisstudy, wehad additionallyreportedacetabularparameters.

ThepresentlyreportedacetabularparametersinTDchildrencould

bevaluableforbetterunderstandingofnormal3Dmorphologyof

theacetabulumandhipinthestandingposition.

Thecomparisonofpelvicparameterswiththeresultsfoundby

Szuperetal.[27]isdifficultsincetheauthorsonlyreportedthe

meanvaluesforamalepopulation.OurresultsforPT,SSandPI

werecomparabletothoseofRampalandGaumetou[18,26],who

performeda3DreconstructionofthepelvisinTDchildren,using

theEOS1system.Furthermore,ourresultswereconsistentwith

those of Jacquemier et al. [28], who had established a large

databaseofmaleandfemalechildren,onwhomFAandTTwere

measuredclinically. 4.2.CPchildren

Inthis study,3Dsubject-specificskeletalparametersof the

pelvis,acetabulumandlowerlimbswerecalculatedinambulant

childrenwithCP(GMFCSlevelsIandII),presentingnoorthopedic

history affecting the pelvis, hips or knees. The femoral

ante-version (FA) and neck shaft angle (NSA) had been previously

evaluatedinambulantchildrenwithCPusingphysical

examina-tionandfluoroscopy[2,7].ThemeanvaluesofFAandNSAinour

populationwere2512 _and134_5 _{respectively,compared}

to309forFAand1397forNSAinthegroupofambulant childrenofBobroff’sstudy[7].Sinceitisknownthatthevaluesof

FAandNSAdecreaseduringgrowth[26,29,30],thedifferencein

resultsforFAcouldberelatedtotheincreasedmeanageofour

population (10.94.0 years) compared toBobroff’spopulation

(8.22.3years).OurresultsforthemeanvaluesofFAinchildren withGMFCSlevelsI(24_13)andII(27_{11)wereinferiorto}

thosepreviouslyobtainedbyphysicalexamination[2],wherethe

authorshadreported30and36forchildrenwithGMFCSlevelsI

andIIrespectively.Thisfindingmeansthatphysicalexamination

couldoverestimatethemeasurementofFAinambulantchildren

withCP,incomparisonwiththeEOS1system.Whencomparing

ourresultstothoseobtainedby3DCTscan[6],wefoundthatthe valuesofFAandNSAinourpopulation(2512 _and134_5

respectively)wereinferiortothoseof Gose’spopulation(41.7  11.9 _and146.7_7_{),consistingof a groupof childrenwith}

GMFCSlevelsIIand III.Thisresultcouldberelatedtothehigh

prevalenceof ambulantchildrenwithGMFCSlevel IIIinGose’s

population(N=41/61),whichcontrastedwiththepredominance

ofchildrenwithGMFCSlevelIinourstudy(N=17/27).Inaddition

tothefactthatFAandNSAarecorrelatedtothedegreeofmotor

impairmentinchildrenwithCP [2,7],thisdifference inresults

could also be related to the difference in age between the 2

studies(Gose’spopulation:5.3years,currentstudy:10.9years)

[26].

The added value of our results is that the CP population

includedinthisstudy,agedfrom6to19years,hadnohistoryof eithersurgicalormedicalinterventionsontheirhipsorknees,as

is common in the literature. The evaluation of such patients

would be highly informative of the natural evolution of the

musculoskeletalsystemin ambulantchildrenwithCP whoare

slightlyaffected.Allprevious studiesonpatientswith CP [4,7]

included subjects who hadbeenpreviously treatedby various

techniques. Because of the current advances in the early

management of children with CP, it is not often thatwe can

studythenaturalhistoryoftheevolutionofskeletaldeformities

inchildrenwithspasticCP.Theabsenceofprevioustreatmentin

our population could be due to the fact that the patients

originatedfromremote areaswhereaccesstoproperdiagnosis

andtreatmentforCPisscarce,orthatorthopedictreatmenthad

not been previouslyrequired. However, this groupis not fully

representativeofthenaturalevolutionofskeletaldeformitiesin

childrenwith CP because of its cross-sectionalnature. Sincea

prospectivestudyisnotpossibleforethicalreasons,ourresults

could be valuable for understanding the progression of this

pathologyinambulatorychildren.

4.3.ComparisonbetweenCPandTDchildren

Thecurrentstudyshowedthatpelvicincidence(PI)andsacral

slope (SS) were significantly increased in children with CP

comparedtoTDchildren.Thedifferencebetweenthetwogroups

is greater than the previously reported uncertainties for these

parametersusingtheEOS1technique[18].PelvictiltandSSare

positional parameters of the pelvis and thus reflect dynamic

sagittalbalance.Therefore,thesignificantdifferenceinSS(Table3) reflectsachangeinthepostureofCPchildren,whichshouldaffect

theirsagittalbalancegiventheimportantcorrelationbetweenSS

andlumbarlordosis[31].However,PIisananatomicalparameter, whichisonlyaffectedbyamorphologicalchangeinthepelvis.The significantdifferencebetweenthepathologicalandcontrolgroups forPIsuggestsaskeletaldeformityofthepelvisintheCPgroupand

subgroups(GMFCSlevelsIandII).Futurestudiesshouldconsider

thefunctionalorclinicalimplicationsoftheseabnormalitiesand

whethertheymustbetreated.

Furthermore, femoral anteversion (FA) was significantly

in-creasedinambulantchildrenwithCPaswellasinchildrenwith

GMFCSlevelsIandII,comparedtotheTDchildren,whichisin

accordancewiththeresultsofpreviousstudies,wheredifferent

techniqueswereusedtoassessFA[2,7].Thedifferencefoundin

thisstudybetweenthetwogroupsisgreaterthanthepreviously

reported uncertainties for these parameters using the EOS1

technique [16,17]. Typically developing childrenare born with

elevatedFA,which laterdecreasesduringgrowthandstabilizes

aroundskeletalmaturity[26].TheelevatedvaluesofFAinchildren

withCP,comparedtothecontrolgroup,suggestthatthisnormal

reductionofFAdoesnotoccuratthesamerateasforTDchildren, eveninslightlyaffectedchildrenwithCP.Itwasalsonotablethat

FA was significantly increased in both groups of childrenwith

GMFCSlevelsIandIIwhencomparedtotheTDgroup.

The elevated FA in the CP group is known tobe related to

spasticity,abnormalmotionstrategiesduringgrowthanddelaysin

achieving independent ambulation, which can lead to skeletal

malalignmentandleverarmdysfunction[1,5,32].Moreover,the

neck shaft angle (NSA) was significantly increased in the CP

population. This could be related to the abnormal motion

strategiesandweight-bearingonthelowerlimbs.

TherewerenosignificantdifferencesbetweentheCPandTD

groupsforany oftheacetabularparameters.Thisindicatesthat

majoracetabulardeformitiesarenotencounteredinambulatory

childrenwith CP (GMFCS I and II), at least in our population;

however, futurestudies shouldassessacetabular parametersin

ambulant childrenwith GMFCSlevel III, which usuallypresent

moresevereskeletaldeformities.

Moreover, no significant differences were found for tibial

parameters between the TD and CP groups. This result could

indicatethatproximalskeletaldeformitiesofthepelvisandfemur

are more frequently encountered than tibial abnormalities in

childrenwithGMFCS levelsIand II.However,it isknownthat

abnormal tibial torsion, in the majorityof childrenwith CP, is

Therefore,aprospectivestudywouldbenecessarytounderstand thenaturalhistoryoftibialdeformitiesinambulantchildrenwith CP.

AlargersamplesizeofTDchildrenandambulantchildrenwith

CPisrequiredinordertobuildadatabaseforthecalculationof3D

subject-specific musculoskeletal parameters for the pelvis and

lowerlimbsinbothpopulations.Ourpopulationisnot

represen-tativeofalltypesofambulantchildrenwithCP;moreprecisely,it

does not include children with GMFCS level III. This group of

childrenrequiresassistive devicestowalkandusually presents

more severe skeletal abnormalities compared to the group of

childrenwithGMFCSlevelsIandII[2].Theabsenceofthisgroupof

children from our study is related to our restricted inclusion

criteria,requiringtheabsenceoforthopedictreatment, evenon

softtissue,atthelevelsofthehipsandknees. 5. Conclusion

Inconclusion, thisstudyshowed thatevenslightly impaired

ambulantchildrenwithCP presentanantevertedandabducted

femur.Theorientation oftheacetabulum andthetibial torsion

were not found to be affected; however morphological and

posturalchangesatthepelvisinthesagittalplanewerenoted.The lowdosebiplanarX-raytechniquecouldberoutinelyusedindaily practiceinordertoassessin3Dpelvicandlowerlimbdeformities

inchildren withCP. Thequantificationof thesesubject-specific

skeletalmalalignmentsrelatedtobothpostureandmorphology,

when combined with gait analysis, would be of paramount

importanceinordertounderstandtowhichextenttheyalterthe

walkingpatternofeachpatient.

Acknowledgments

ThisstudywasfundedbytheresearchcounciloftheUniversity

of Saint-Joseph (grant FM244) and the CEDRE Project (grant

11SCIF44/L36).Bothinstitutionsdidnotinterfereinanypartofthis study.

Conflictofinterest None.

References

[1]J.R.Gage,T.F.Novacheck,Anupdateonthetreatmentofgaitproblemsin cerebralpalsy,J.Pediatr.Orthop.PartB10(2001)265–274,doi:http://dx.doi. org/10.1097/01202412-200110040-00001.

[2]J.Robin,H.K.Graham,P.Selber,F.Dobson,K.Smith,R.Baker,Proximalfemoral geometryincerebralpalsy:apopulation-basedcross-sectionalstudy,J.Bone JointSurg.Br.90(2008)1372–1379, doi:http://dx.doi.org/10.1302/0301-620X.90B10.20733.

[3]S.Aktas,M.D.Aiona,M.Orendurff,Evaluationofrotationalgaitabnormalityin thepatientscerebralpalsy,J.Pediatr.Orthop.20(2000)217–220,doi:http:// dx.doi.org/10.1097/01241398-200003000-00016.

[4]F.Miller,Y.Liang,M.Merlo,H.T.Harcke,Measuringanteversionandfemoral neck-shaftangleincerebralpalsy,Dev.Med.ChildNeurol.39(1997)113–118. [5]G.Chan,F.Miller,Assessmentandtreatmentofchildrenwithcerebralpalsy, Orthop.Clin.NorthAm.45(2014)313–325,doi:http://dx.doi.org/10.1016/j. ocl.2014.03.003.

[6]S.Gose,T.Sakai,T.Shibata,T.Murase,H.Yoshikawa,K.Sugamoto, Morphometricanalysisofthefemurincerebralpalsy:3-dimensionalCTstudy, J.Pediatr.Orthop.30(2010)568–574,doi:http://dx.doi.org/10.1097/ BPO.0b013e3181e4f38d.

[7]E.D.Bobroff,H.G.Chambers,D.J.Sartoris,M.P.Wyatt,D.H.Sutherland,Femoral anteversionandneckshaftangleinchildrenwithcerebralpalsy,Clin.Orthop. Relat.Res.(1999)194–204.

[8]T.Terjesen,Thenaturalhistoryofhipdevelopmentincerebralpalsy,Dev.Med. ChildNeurol.54(2012)951–957, doi:http://dx.doi.org/10.1111/j.1469-8749.2012.04385.x.

[9]A.Connelly,P.Flett,H.K.Graham,J.Oates,HipsurveillanceinTasmanian childrenwithcerebralpalsy,J.Paediatr.ChildHealth45(2009)437–443,doi: http://dx.doi.org/10.1111/j.1440-1754.2009.01534.x.

[10]J.R.Davids,A.D.Marshall,E.R.Blocker,S.L.Frick,D.W.Blackhurst,E.Skewes, Femoralanteversioninchildrenwithcerebralpalsy,J.BoneJointSurg.85 (2003)481–488.

[11]E.Viehweger,C.Bérard,A.Berruyer,M.C.Simeoni,Bilanarticulairedes membresinférieursd’unenfantatteintd’uneinfirmitémotricecérébrale, AnnalesdeReadaptationetdeMedecinePhysique50(2007)258–265,doi: http://dx.doi.org/10.1016/j.annrmp.2007.02.004.

[12]M.Wynter,N.Gibson,M.Kentish,S.Love,P.K.G.H.Thomason,TheConsensus StatementonHipSurveillanceforChildrenwithCerebralPalsy:Australian StandardsofCare,J.Pediatr.Rehabil.Med.4(2011)183–195.

[13]J.R.Davids,Identificationofcommongaitdisruptionpatternsinchildrenwith cerebralpalsy,J.Am.Acad.Orthop.Surg.22(2014)782–790.

[14]C.Delin,S.Silvera,C.Bassinet,P.Thelen,J.L.Rehel,P.Legmann,etal.,Ionizing radiationdosesduringlowerlimbtorsionandanteversionmeasurementsby EOSstereoradiographyandcomputedtomography,Eur.J.Radiol.83(2014) 371–377,doi:http://dx.doi.org/10.1016/j.ejrad.2013.10.026.

[15]Y.Chaibi,T.Cresson,B.Aubert,J.Hausselle,P.Neyret,O.Hauger,etal.,Fast3D reconstructionofthelowerlimbusingaparametricmodelandstatistical inferencesandclinicalmeasurementscalculationfrombiplanarX-rays, Comput.MethodsBiomech.Biomed.Eng.15(2012)457–466,doi:http://dx. doi.org/10.1080/10255842.2010.540758.

[16]A.Assi,Y.Chaibi,A.Presedo,J.Dubousset,I.Ghanem,W.Skalli, Three-dimensionalreconstructionsforasymptomaticandcerebralpalsychildren’s lowerlimbsusingabiplanarX-raysystem:afeasibilitystudy,Eur.J.Radiol.82 (2013)2359–2364,doi:http://dx.doi.org/10.1016/j.ejrad.2013.07.006. [17]M.Thépaut,S.Brochard,J.Leboucher,M.Lempereur,E.Stindel,V.Tissot,etal.,

Measuringphysiologicalandpathologicalfemoralanteversionusinga biplanarlow-doseX-raysystem:validity,reliability,anddiscriminativeability incerebralpalsy,SkeletalRadiol.45(2016)243–250,doi:http://dx.doi.org/ 10.1007/s00256-015-2298-y.

[18]V.Rampal,J.Hausselle,P.Thoreux,P.Wicart,W.Skalli,Three-dimensional morphologicstudyofthechild’ship:whichparametersarereproducible?Clin. Orthop.Relat.Res.471(2013)1343–1348,doi:http://dx.doi.org/10.1007/ s11999-012-2660-5.

[19]R.J.Palisano,P.Rosenbaum,D.Bartlett,M.H.Livingston,Contentvalidityofthe expandedandrevisedGrossMotorFunctionClassificationSystem,Dev.Med. ChildNeurol.50(2008)744–750, doi:http://dx.doi.org/10.1111/j.1469-8749.2008.03089.x.

[20]J.S.Steffen,I.Obeid,N.Aurouer,O.Hauger,J.M.Vital,J.Dubousset,etal.,3D posturalbalancewithregardtogravityline:anevaluationinthetransversal planeon93patientsand23asymptomaticvolunteers,Eur.SpineJ.19(2010) 760–767,doi:http://dx.doi.org/10.1007/s00586-009-1249-5.

[21]D.Mitton,S.Deschenes,S.Laporte,B.Godbout,S.Bertrand,J.A.DeGuise,W. Skalli,3Dreconstructionofthepelvisfrombi-planarradiography,Comput. MethodsBiomech.Biomed.Eng.9(2006)1–5.

[22]D.W.Murray,Thedefinitionandmeasurementofacetabularorientation,J. BoneJointSurg.Br.75(1993)228–232.

[23]E.S.Stem,M.I.O’Connor,M.J.Kransdorf,J.Crook,Computedtomography analysisofacetabularanteversionandabduction,SkeletalRadiol.35(2006) 385–389.

[24]J.Y.Lazennec,N.Charlot,M.Gorin,B.Roger,N.Arafati,A.Bissery,etal., Hip-spinerelationship:aradio-anatomicalstudyforoptimizationinacetabular cuppositioning,Surg.Radiol.Anat.26(2004)136–144,doi:http://dx.doi.org/ 10.1007/s00276-003-0195-x.

[25]JacobCohen,StatisticalPowerAnalysisfortheBehavioralSciences,2nded., LawrenceErlbaumAssociates,UnitedStatesofAmerica,1988.

[26]E.Gaumétou,S.Quijano,B.Ilharreborde,A.Presedo,P.Thoreux,K.Mazda, etal.,EOSanalysisoflowerextremitysegmentaltorsioninchildrenandyoung adults,Orthop.Traumatol.100(2014)147–151,doi:http://dx.doi.org/10.1016/j. otsr.2013.09.010.

[27]K.Szuper,A.T.Schégl,E.Leidecker,C.Vermes,S.Somoskeöy,P.Than, Three-dimensionalquantitativeanalysisoftheproximalfemurandthepelvisin childrenandadolescentsusinganuprightbiplanarslot-scanningX-ray system,Pediatr.Radiol.45(2015)411–421,doi:http://dx.doi.org/10.1007/ s00247-014-3146-2.

[28]M.Jacquemier,Y.Glard,V.Pomero,E.Viehweger,J.-L.Jouve,G.Bollini, Rotationalprofileofthelowerlimbin1319healthychildren,GaitPosture28 (2008)187–193,doi:http://dx.doi.org/10.1016/j.gaitpost.2007.11.011. [29]S.S.Upadhyay,R.G.Burwell,A.Moulton,P.G.Small,W.A.Wallace,Femoral

anteversioninhealthychildren.Applicationofanewmethodusing ultrasound,J.Anat.169(1990)49–61.

[30]A.R.ShandsJr,M.K.Steele,Torsionofthefemur;afollow-upreportontheuse oftheDunlapmethodforitsdetermination,J.BoneJointSurg.Am.40-A(1958) 803–816.

[31]R.Vialle,Radiographicanalysisofthesagittalalignmentandbalanceofthe spineinasymptomaticsubjects,J.BoneJointSurg.(American)87(2005)260, doi:http://dx.doi.org/10.2106/JBJS.D.02043.

[32]T.F.Novacheck,J.R.Gage,Orthopedicmanagementofspasticityincerebral palsy,Child’sNerv.Syst.23(2007)1015–1031,doi:http://dx.doi.org/10.1007/ s00381-007-0378-6.