TheLaryngoscopeVC2013TheAmericanLaryngological,RhinologicalandOtologicalSociety,Inc.WeCanPredictPostpalatoplastyVelopharyngealInsufficiencyinCleftPalatePatientsJacquesE.Leclerc,MD;AudreyGodbout,MD;IsabelleArteau-Gauthier,MD;SophieLacour,MOA;KatiAbel,MOA;
Elisa-MaudeMcConnell,MScObjectives/Hypothesis: To find an
anatomicalmeasurement ofthe cleft palate (or a calculated
parameter) that predictsthe occurrence of velopharyngeal
insufficiency (VPI) after palatal cleft repair.Study Design:
Retrospective cohort study.Methods: Chartswerereviewedfromcleft
palatepatientswhounderwent
palatoplastybytheVonLangenbecktech-niqueforisolatedcleft
palateorBardachtwo-flappalatoplastyforcleft lip-palate.
Sevenanatomical cleft parameterswereprospectively measured during
the palatoplasty procedure.Three blinded speechlanguage
pathologists retrospectively scoredthe clinically assessed VPI at 4
years of age.The recommendation of pharyngoplasty was also used as
an indicator of VPI.Results: From1993to2008, 67patients were
enrolledinthe study. The best predicting parameter was the
ratioa/(30 2b1),
inwhichaisdefinedastheposteriorgapbetweenthesoftpalateandtheposteriorpharyngealwall
andb1isthewidthofthecleftatthehardpalatelevel. Ana/(30
2b1)ratio>0.7to0.8isassociatedwithahigherriskofdevelopingVPI
(relative risk 52.25.1, sensitivity 572%81%,P 20other parameters
usingtwoor moremeasurements werecreated and tested in our quest to
find the best
possiblepredictor.EvaluationofVelopharyngealFunctionThe SLPs were
blindedto the results of the
anatomicalmeasurements.Clinicalchartswereretrospectivelyreviewedforinformationregardingpatients
speechas relatedtovelophar-yngeal functionby a teamof three SLPs
specializing in theevaluation of CP patients. The speech of each
patient wasassessed as close as possible to the age of 4 years.
Somepatients hadremainingsmall hardpalatefistulas but
nosoftpalatefistulas.Thequalityofspeechandresonanceisacknowl-edgedasbeingthemainoutcomemeasureforevaluatingpost-palatoplastypatients.3Perceptual
evaluationremainsthegoldstandardforevaluatingspeech, aswell
asthemost
commonlyusedmethod.46Fig.1.Cleftlip/palatemeasurementsdiagram(J.E.L.).
[Color figure can beviewedintheonlineissue, whichisavailable at
wileyonlinelibrary.com.]TABLEI.DescriptionofthePalatal
andCleftMeasurements.Variable Descriptiona
Distancebetweentheposteriorendofthesoftpalateandtheposteriorpharyngeal
wall intheplaneofthehardpalateb
Largestwidthofthecleftatthesoftpalatelevelb1
Widthofthecleftattheposteriorendofthehardpalatec
Lengthofthesoftpalate,fromitsposteriorendtotheposteriorendofthehardpalated
Largestwidthofthedentalarchmeasuredatthetopofthecrestse
Lengthofthehardpalate,fromitsposteriorendtothetopoftheanterior
dental archwitharigidruler;thedepthofthehardpalateisnotevaluatedf
Total
lengthofthecleft,fromtheposteriorendofthesoftpalatetotheanteriorendofthecleftLaryngoscope124:February2014
Leclercetal.:PostpalatoplastyVelopharyngeal
Insufficiency562SpeechParametersandRatingScalesThedatacollectedunderwent
amappingprocess,
involv-ingtheconversionofourlocalteamsevaluationprotocoltothepredeterminedsetof
chosenparametersbasedonHenningssonetal.4andSelletal.5Forthepurposeofthisstudy,threediffer-entcriteriaorsetsofcriteriawereusedtoclassifyeachcaseastothepresenceofVPI:
Perceptual speechevaluation(PSE)
wasconsideredpositiveifatleastoneofthesecriteriawasfulfilled:severityscoreof3
onthe hypernasality scale or presence of audible
nasalemission(scoreof1onthescale)orpresenceofnasalturbu-lence(scoreof1onthescale;TableII).
Speechacceptabilitywasassessedindependently.Acceptabil-ityisdefinedasthedegreetowhichspeechcallsattentiontoitself
apartfromthecontentof thespokenlanguage,
anditiscloselyrelatedtopatients
qualityoflife.Scoresof2and3wereusedasanindicationofVPI(TableII).
Pharyngoplastyrecommendation(PR)wasusedasthethirdindicatorof
significantpersistingVPI. Asurgical treatmentwas
offeredtothepatient when1)
speechtherapydidnotimproveVPIaftercompleteSLPfollow-upand2)anobviousanatomical
defect couldexplainthe problem(Table II). Nopharyngoplasty was
recommended or performed in theabsenceofVPI.StatisticalAnalysisIn
regard to each outcome, the discriminating capacity(dc)
wasevaluatedforeachvariablebythec-indexcorrespond-ing to the area
under the receiver operating characteristic(ROC) curve: 0dc 1 using
the SAS LOGISTICprocedure(version9.2.3; SASInstitute, Cary, NC).
Onlyvariableshavingdc 0.6wereretainedforfurtheranalysis.
UsingtheOUTROCoptionof theLOGISTICprocedure,
eachselectedvariablewasdichotomizedatacorrespondinglevelthatmaximizedthesensi-tivityandthe
specificity of the measurement. Arelative
riskratiowascalculatedforeachvariable. All
Pvaluesweredeter-minedusingthe
v2testortheFisherexacttest.RESULTSThechartsof272CLPpatientswhounderwentpal-atoplastybetween1993and2008were
reviewed. Only67patients met all the inclusioncriteria(Fig. 2).
TheclinicalfeaturesofourseriesareshowninTableIII.TheCPmeasurementdataareshowninFigure3.
Theinde-pendent variables usedinour analysis aredepictedinTableIV.
OurstudyhasshownarespectiveincidenceofTABLEII.ParametersUsedbytheSpeechLanguagePathologists.PerceptualSpeechEvaluationHypernasality(sentences)0
5withinnormallimits1
5mild,indicateshypernasalityresonanceperceivedonvowels2
5moderate,hypernasalityresonanceperceivedonvowels,approximants/semivowels,andliquids,andpresenceofweakenedconsonants3
5severe,indicatesall
oftheaboveandthereplacementofvoicedconsonantsbytheirnasal
equivalents*Audiblenasal emission(sentences)0 5normal1
5present*Audiblenasal turbulence(sentences)0 5normal1
5present*SpeechAcceptability0 5normal1 5normal tomild2
5moderate(speechdeviatesfromnormal toamoderatedegree)*3
5severe(speechdeviatesfromnormal
toaseveredegree)*PharyngoplastyRecommendationFailedadequatespeechtherapyandanatomicallimitation*Positiveforvelopharyngeal
insufficiency.Fig.2.Flowchartforpatientrecruitment.Laryngoscope124:February2014
Leclercetal.:PostpalatoplastyVelopharyngeal Insufficiency56320.9%,
30.8%, and50%forPR, speechacceptability, andPSE.
Addedtoouranatomical measurements, >20com-binationsof
theanatomical measurementswerecreatedas other parameters. For
eachdependent variable(PR,PSE, acceptability), these combinations
of anatomicalmeasurements were tested. The 10 that best
discrimi-natedaccordingto the dc value of the ROCcurve
arepresentedindecreasingorderinTablesVtoVI.Tables VandVI,
respectively, present the dataofonly 60 and 65 patients of 67. In
their retrospectivechart evaluation, the SLPs felt that they could
notadequatelyclassifythemissingothers
accordingtotheavailableinformation. Wedidnot get b1values for
allthe patients in each table. The missing values corre-spondto the
number of patients withclefts limitedtothe soft palate. We elected
to assign no value for b1insteadofzero.
Thefparameterwasourmarkerfortheanteroposteriorextentofthecleft.Therewasanobviousclerical
mistakeforonecvaluethat wedidnot
useforanyofourcalculations.Forallthreedependentvariables,
thedatainthesetables showthat the ratio a/(302b1), in which a
isdefinedastheposteriorgapbetweenthesoftpalateandtheposteriorpharyngealwallandb1isthewidthofthecleftatthehardpalatelevel,
hadagooddiscriminatingvalueandagloballybettersensitivitythatreachedsta-tistical
significance. The cutting ratio values for PSE,acceptability,
andPRwere respectively0.73, 0.83, and0.79. Table VIII shows the
calculated relative risk ofdevelopingVPI for these values.
Theparameter
a1b1wasaclosecontender(TablesV).b1andapresentedthehighest
discriminatingval-uesamongthesingleanatomicalmeasurements,
butdidnot get goodrankings inall threedependent
variables(TableVIII).Themeasurementb1performedwellintwoof
thethreeVPIevaluationsandthedepthof thenaso-pharynx(a) inone(PSE).
Theywerealsofoundwithinmanyofthetop10parameters(TablesV).Therewasnoassociation
between age, sex and PSE, and PR
andacceptability.DISCUSSIONVPIRates:WhyThreeEvaluationModalities?WeselectedtousethreeVPIevaluationmodalities.Withinthem,
we definedVPI differently to be able topickupthesubtleaswell
asthemoreseverecasesandtherefore cover the widest possible range of
VPI. Ourideal anatomical parameter or
combinationhadtoper-formwellinallthreemodalities.PRwastheindicatorofanobvious
problemcausedbyasignificant anatomicaldeficit that neededto
becorrected. The
procedurewasfoundtoberequiredin20.9%ofsubjectsinourseriesatapproximately
4 years old. For the second
modality,speechacceptability,weelectedtousethescores2(mod-erate)
and3(severe), whichincludedalargergroupofchildren. We found that
30.5%of the 4-year-olds pre-sentedasignificantdifficulty. Of
thisgroup,
12%hadamoderatedegreeand18.5%aseveredegreeofdifficulty.Withthethirdmodality,
perceptual evaluation, weoptedtoincludeall
patientswithevensubclinical audi-blenasal emissionorturbulence.
Wefoundthat50%ofthepatientswerecompletelynormal. Theresidual
50%includedawiderangeof VPI, fromclinicallynormal
tosevere.Thesefiguresmayseemelevated.However,12of67patients(18%)
scoredpositiveinonlyoneof
thetwoTABLEIII.CleftLip/PalateSeries.Characteristic ValueTotal
number 67Gender,male%/female% 58.2/41.7Weightatpalatoplasty,kg
9.1(61.4)Agepalatoplasty,mo 10.6(60.9)AgeatevaluationbySPL,mo
49.6(616.0)AgeatPR,mo 73.9(638.4)Isolatedcleftpalate
38(57%)Unilateral cleftlipandpalate 23(34%)Bilateral
cleftlipandpalate 6(9%)PR5pharyngoplasty recommendation; SPL
5speechlanguagepathologist.Fig.3.Cleftpalateanatomical
measurementsinmillimeters.TABLEIV.Incidence(%)ofVelopharyngeal
InsufficiencyasDeterminedbyPerceptualSpeechEvaluation,SpeechAcceptability,andPharyn-goplastyRecommendation.Velopharyngeal
InsufficiencyEvaluation %Perceptual speechevaluation
50Hypernasality0 31.81 16.72 33.33 18.2*Audiblenasal emissionscore1
24.1*Audiblenasal turbulencescore1 30.9*Speechacceptability
30.8Pharyngoplastyrecommendation 20.9*Respectivepercentagesfor
thecriteriausedfor perceptual speechevaluationinthestudy. Thesumof
thepercentagesof theindividual PSEcriteriadoesnot matchthe50%total
becausesomepatientsscoredaspositive>1criterion.Laryngoscope124:February2014
Leclercetal.:PostpalatoplastyVelopharyngeal
Insufficiency564parameters (constant or inconstant nasal
turbulenceoremission). They were classified as minimal VPI,
butexperiencednofunctional or social impact oneverydaylife. Of
theremaining32%, aspreviouslymentioned,2=3of the cases were offered
a pharyngoplasty (14 of 67patients). We found a close correlation
between
the20.9%pharyngoplastyrateandthescore3hypernasalityrateof18.2%(TableIV;P
