Identification of large-for-gestational age fetuses using antenatal customized fetal growth charts: Can we improve the prediction of abnormal labor course?

Full

length

article

Identi

fication

of

large-for-gestational

age

fetuses

using

antenatal

customized

fetal

growth

charts:

Can

we

improve

the

prediction

of

abnormal

labor

course?

Andrea

Dall'Asta

_,

_Giuseppe

_Rizzo

_,

_Ariane

_Kiener

_,

_Nicola

_Volpe

_,

_Elvira

_Di

_Pasquo

_,

Enrica

Roletti

,

Ilenia

Mappa

,

Alexander

Makatsariya

,

Giuseppe

Maria

Maruotti

,

Gabriele

Saccone

,

Laura

Sarno

,

Marta

Papaccio

,

Anna

Fichera

,

Federico

Prefumo

,

Chiara

Ottaviani

,

Tamara

Stampalija

,

Tiziana

Frusca

,

Tullio

Ghi

*

a

DepartmentofMedicineandSurgery,ObstetricsandGynaecologyUnit,UniversityofParma,Parma,Italy

b

DepartmentofMetabolism,DigestionandReproduction,InstituteofReproductiveandDevelopmentalBiology,ImperialCollegeLondon,UnitedKingdom

c

DivisionofMaternalandFetalMedicine,OspedaleCristoRe,UniversityofRomeTorVergata,Rome,Italy

d

DepartmentofObstetricsandGynecology,TheFirstI.M.SechenovMoscowStateMedicalUniversity,Moscow,RussianFederation

e

DepartmentofNeuroscience,ReproductiveSciencesandDentistry,SchoolofMedicine,UniversityofNaplesFedericoII,Naples,Italy

f

DepartmentofObstetricsandGynaecology,UniversityofBrescia,SpedaliCiviliDiBrescia,Brescia,Italy

g_{UnitofFetalMedicineandPrenatalDiagnosis,InstituteforMaternalandChildHealth,IRCCSBurloGarofolo,Trieste,Italy} h

DepartmentofMedical,SurgicalandHealthScience,UniversityofTrieste,Trieste,Italy

ARTICLE INFO Articlehistory:

Received12November2019

Receivedinrevisedform5March2020 Accepted9March2020 Availableonlinexxx Keywords: Fetalgrowth Macrosomia Birthcanal Caesareansection Prolongedlabor ABSTRACT

Introduction:Fetalovergrowthisanacknowledgedriskfactorforabnormallaborcourseandmaternal

andperinatalcomplications.Theobjectiveofthisstudywastoevaluatewhethertheuseofantenatal

ultrasound-basedcustomizedfetalgrowthchartsinfetusesatriskforlarge-for-gestationalage(LGA)

allowsabetteridentificationofcasesundergoingcaesareansectionduetointrapartumdystocia.

Materialandmethods:AnobservationalstudyinvolvingfourItaliantertiarycenterswascarriedout.

Womenreferredtoadedicatedantenatalclinicbetween35and38weeksduetoanincreasedriskof

havinganLGAfetusatbirthwereprospectivelyselectedforthestudypurpose.Thefetalmeasurements

obtainedandusedfortheestimationofthefetalsizewerebiparietaldiameter,headcircumference,

abdominalcircumferenceandfemurlength,wereprospectivelycollected.LGAfetusesweredefinedby

estimatedfetalweight(EFW)>95thcentileeitherusingthestandardchartsimplementedbytheWorld

HealthOrganization(WHO)orthecustomizedfetalgrowthchartspreviouslypublishedbyourgroup.

Patients scheduled forelectivecaesarean section(CS)or forelectiveinductionforsuspected fetal

macrosomiaorsubmittedtoCSorvacuumextraction(VE)purelyduetosuspectedintrapartumdistress

wereexcluded.TheincidenceofCSduetolabordystociawascomparedbetweenfetuseswithEFW>95th

centileaccordingWHOorcustomizedantenatalgrowthcharts.

Results:Overall,814womenwereeligible,however562wereconsideredforthedataanalysisfollowing

theevaluationoftheexclusioncriteria.Vaginaldeliveryoccurredin466(82.9%)women(435(77.4%)

spontaneousvaginaldeliveryand31(5.5%)VE)while96hadCS.TheEFWwas>95thcentilein194(34.5

%)fetusesaccordingtoWHOgrowthchartsandin190(33.8%)bycustomizedgrowthcharts,respectively.

CSduetodystociaoccurredin43(22.2%)womenwithLGAfetusesdefinedbyWHOcurvesandin39

(20.5%)womenwithLGAdefinedbycustomizedgrowthcharts(p0.70).WHOcurvesshowed57%

sensitivity,72%specificity,24%PPVand91%NPV,whilecustomizedcurvesshowed52%sensitivity,73%

specificity,23%PPVand91%NPVforCSduetolabordystocia.

Conclusions:Theuseofantenatalultrasound-basedcustomizedgrowthchartsdoesnotallowabetter

identificationoffetusesatriskofCSduetointrapartumdystocia.

©2020ElsevierB.V.Allrightsreserved.

Abbreviations:US,ultrasound;2D,two-dimensional;3D,three-dimensional;LGA,large-for-gestationalage;CS,caesareansection;EFW,estimatedfetalweight;WHO, WorldHealthOrganization;SVD,spontaneousvaginaldelivery;VE,vacuumextractor;PPV,positivepredictivevalue;NPV,negativepredictivevalue.

* Correspondingauthorat:DepartmentofMedicineandSurgery,ObstetricsandGynaecologyUnit,UniversityofParma,ViaGramsci14,43126,Parma,Italy. E-mailaddress:tullioghi@yahoo.com(T.Ghi).

https://doi.org/10.1016/j.ejogrb.2020.03.024

0301-2115/©2020ElsevierB.V.Allrightsreserved.

ContentslistsavailableatScienceDirect

European

Journal

of

Obstetrics

&

Gynecology

and

Reproductive

Biology

Introduction

Fetalmacrosomiaisamongtheacknowledgedriskfactorsfor intrapartumdystociaandothermajorcomplicationsoflaborfor the mother and the fetus which include emergency caesarean section,postpartumhemorrhage,perinealtears,shoulderdystocia andhypoxic-ischemicencephalopathy[1,2].

Theultrasound(US)evaluationofthefetalweighteitherwith two-dimensional (2D) ultrasound (US) [3] or with the more recentlydescribedthree-dimensional(3D)UStechniques[4–8]is currentlyconsideredthemostaccuratemethodfortheantenatal detectionoffetalmacrosomia,eventhoughUSestimationofthe fetal size is known to become increasingly imprecise at the extremesof the distribution of theestimated fetal weight and particularlyinthesettingofsuspectedfetalovergrowth[9].

Fetalmacrosomiahasbeencommonlyreferredtoabirthweight above the threshold of 4000g [10]. However, it is widely acknowledged that fetal macrosomia does not necessarily lead tocephalo-pelvicdisproportion(CPD),which isnotdetermined onlybythefetalsizebutalsobythesizeandtheanatomicfeatures ofthematernalpelvis[11,12].

Based on theassumption that fetal growth pattern maybe subjected to substantial variability related to the race and constitutional characteristics of the parents, over the last two decadestheuseofcustomizedantenatalgrowthchartshasbeen proposedinclinicalpracticewiththeaimofassessingthespecific growth potential of each fetus and to tailor on this the management of pregnancy. However, the clinical usefulness of such approach in improving the pregnancy outcome in still debated.Inparticular,thereislimiteddataontheperformanceof customizedfetalgrowth chartswithinthecontextofsuspected fetalovergrowthanditsrelatedcomplications[13–18].Theaimof thisstudywastocomparetheaccuracyofrecentlyimplemented antenatalultrasound-basedcustomizedgrowthcharts [19] with that of the growth charts developed by the World Health Organization (WHO) [20] in the identification of LGA fetuses undergoingcaesareansection(CS)duetointrapartumdystocia. Methods

The study involved five Italian Tertiary Maternity Units (University Hospitals of Parma, Brescia, Rome Tor Vergata and NaplesFedericoIIandBurloGarofoloHospital,Trieste)between January2017andJuly2018.Intheinvolvedcentres,allpatients withnon-anomaloussingletonpregnancyandacknowledgedrisk factorsforfetalmacrosomia–whichincludedamedicalhistoryof pregestationaldiabetesmellitus,obesity,asdefinedasbodymass index above 30kg/m [2], or obstetric risk factors such as gestationaldiabetes(diagnosedfollowingtheGuidelinesbythe WorldHealthOrganization)[21],previoushistoryoffetal macro-somiaorsuspectedfetalmacrosomiaatUSorsymphysis-to-fundal height assessment of the fetal growth performed in the third trimester–werereferredtoa dedicatedantenatalclinicforthe antenatalUS estimation ofthefetalweightbetween35and 38 weeksofgestation.Pregnancydatingwasderivedbyfirsttrimester ultrasound. Patients were considered eligible for the study purposes in the absence of any fetal genetic and structural abnormalitydiagnosedeitherantenatallyorpostnatallyandinthe caseofavailabilityoflaborandpostnataloutcomes.

Aspercommonpracticeinalltheparticipatingreferral Fetal MedicineUnitsprenatalUSexaminationsfollowthe2015 Guide-lines of the Italian Society on Ultrasound in Obstetrics and Gynecology(SIEOG)[22].Inallcasestheantenatalestimationof the fetal weight was performed by maternal-fetal medicine specialistsusingtwo-dimensionaltransabdominallowfrequency probesandtheEFWwascomputedaccordingtotheHadlockIV

Model[3],whichreliesonthecombinationofthemeasurements of the biparietal diameter, head circumference, abdominal circumferenceandfemurlength.Datawereprospectively collect-ed.TheEFWpercentilewascomputedaccordingtolocalcharts whichwereusedfortheclinicalmanagementofeachpatient.

For the study purposes, the EFW data obtained from each Centrewereplottedontheantenatalstandardgrowthchartsfor theestimationof thefetalweightdeveloped bytheWHO[20], whilebiometrydataaswellasmaternalandpaternal character-isticswereusedtocalculatetheEFWaccordingtothecustomized fetalgrowthchartspreviouslypublishedbytheSIEOG[19].Such charts are based on cross-sectional US measurements from uncomplicated singleton pregnancies, accurately dated by crown-rump length in the first trimester, delivering after 37 weeks’gestationwithknownoutcomeandinformationavailable onmaternalandpaternalheightandweight,parity,andethnicity [19]. In details, maternal and paternal characteristics were includedin the algorithm usedto obtaincustomized measure-ments,whilethecustomizedEFWwascomputedbymeansofthe Hadlock III model as previously described [23]. Of note, the absenceof any of thecovariates includedin thecustomization algorithm precludes to obtain the customized EFW. The 95th percentileofthecustomizedgrowthchartwasderivedusingthe casesfromouroriginalmanuscriptonantenatalultrasound-based customizedgrowthchartsforsingletons[19]and,togetherwith the95th percentileof theWHO growth chart, representedthe threshold for theretrospective identification of large-for-gesta-tionalagefetuses(LGA)inthestudypopulation.Thisevaluation wasusedonlyforthestudypurposesandhadnoinfluenceonthe clinicalmanagementofthepatients.Aspertheaimofthestudy, wecomparedlaborandpostnataloutcomesinfetusesidentifiedas LGAaccordingtotheWHOchartsandthosesodefinedaccordingto customized charts. Inparticular, the occurrence of intrapartum dystocia leadingtounplanned caesareansectionwascompared betweenLGAfetusesaccordingtoWHOvscustomizedcharts.

Informationconcerningmaternalage,ethnicity,parity, gesta-tionattheonsetoflaborandbodymassindex(BMI)atbookingand atdeliverywerecollectedfrompatientnotesandrecorded.After delivery,intrapartumandneonataloutcomedatawerecollected frompatientcasenotes.Outcomemeasuresincludedbirthweight, themodeof deliveryanda seriesofvariablessuchasepidural, augmentationrate,lengthoflabor,postpartumhemorrhageand thirdandfourth-degreetearinthemotherandbirthweightand birthweightcentilecorrectedforgenderandparityaccordingto theItaliangrowthchartspublishedbyBertinoetal.[24],rateof shoulderdystocia,APGARscoreat5min,cordarterialandvenous pH, admission to Neonatal Intensive Care Unit (NICU) for the neonate.Thedecisionastowhethertodeliverduetosuspected intrapartumdistresswassubjectivelydefinedbythephysicianin charge for the patient care based on abnormal CTG tracing accordingtoFIGOclassi_ficationsystem[25],whilethedecisionto optforoperativedeliveryduetointrapartumdystociawasbased ontherecommendationsoftheAmericanCollegeofObstetricians andGynecologistsforthesafepreventionoftheprimarycaesarean delivery[1].Deliverieswerecategorizedaccordingtothemodeof delivery in spontaneous vaginal delivery (SVD) and obstetric interventionbyvacuumextractor(VE)orCS,whichwerefurther sub-classified based on theprimary indication(i.e. dystocia or distress).

CasessubmittedtoelectiveCSforanyindicationortoinduction oflaborbefore39+0_{weeksduetosuspectedfetalmacrosomiawere}

excludedfromdataanalysis,aswerethosediagnosedwithfailed induction,whosediagnosticcriteriadifferedacrossthe participat-ingCentres.

Asamplesizeof238patientspergroup(n=476intotal)was plannedtocomparetheprimaryoutcomebetweenthetwogroups.

Thesamplesizeestimationwasbasedonrecentretrospectivedata comparingtheincidenceofemergencyCSinfetusesidentifiedas LGAbycustomizedvsreferencecharts[26].Weassumedthatthe incidenceofunplannedCSforlabordystociainthecontrolgroup (i.e.LGAdefinedbyreferencecharts)wouldbe42.7%andthatthe useof customizedantenatalgrowthchartswouldbeassociated withanemergencyCSrateincreasedupto55.5%.Thesamplesize wascomputedusingPowerandSampleSizeCalculator (Biostatis-tics Department, Vanderbilt University, Nashville, TN, USA) consideringan80%powerandaP-valueof0.05.

AsperNationalregulations,ethicsapprovalforthisstudywas granted by the local ethics committee in all the participating centers.

StatisticalanalysiswasperformedusingStatisticalPackagefor Social Sciences (SPSS) version 20 (IBMInc., Armonk, NY,USA). Comparisonofnormallyandnon-normallydistributedcontinuous variablesincludedtheTtestforindependentsampleand2-tailed t-test and the Mann-WhitneyU-test, respectivelyand data were shown as mean + standard deviation or as median (range) accordingly. Categorical variables were reported as number (percentage)andcomparedusingtheChi-squareorFisherexact test.Sensitivity,specificity,positiveandnegativepredictivevalues (PPVandNPV,respectively),positiveandnegativelikelihoodratios (LR+andLR-,respectively)fortheidentificationofcasesofCSdue tointrapartumdystociainLGAfetusesidentifiedatreferenceand customized charts wereevaluated and compared. p<0.05 was

considered as statistically significant. This study was reported accordingtotheSTROBEguidelines[27].

Results

Overall, 814 womenwere eligible over the study period, of whom562wereincludedfortheanalysisaftertheevaluationof theexclusioncriteria(Fig.1).

Table1

AgreementanddisagreementintheidentificationofLGAfetusesatWHOandcustomizedgrowthcharts.

Customized<95thpercentile Customized>95thpercentile

n(%) n(%)

WHO<95thpercentilen(%) 349(62.1%) 19(3.4%) WHO>95thpercentilen(%) 23(4.1%) 171(30.4%) Fig.1.Flowchart(accordingtoSTROBEguidelines)[27]forinclusionofcases.

Table2

Demographicfeaturesoftheincludedcases. Allcases N562 Age(years)mean+SD 32.4+5.8

Ethnicityn(%) White(Caucasian,Arabic)454 (80.5%)

African59(10.7%) Asian38(6.9%)

Other(Caribbean,SouthAmerican, Mixed)11(1.9%)

Riskfactorforfetalmacrosomian(%) DM/GDM+suspectedLGAfetus 124(22.3%) DM/GDM+polyhydramnios4(0.7 %) DM/GDM+BMI>30kg/m2 61 (10.7%) HistoryofGDM+fetal macrosomia5(0.8%) BMI>30kg/m2_179(32.0%)

SuspectedLGAfetus(atUSorSFH assessment)155(27.5%) Historyoffetalmacrosomia34 (6.0%)

Parityn(%) Nulliparae249(44.9%) BMIatconception(kg/m2_{)mean+SD 28.2+5.8}

TermpregnancyBMI(kg/m2_{)mean+SD 32.2+5.2}

GestationalageatUSexamination (weeks+days

)mean+SD

36+5

+0+6

EstimatedfetalweightcentileatUS examination(localchart)mean+SD

76.2+20.2 Gestationalageatdelivery(weeks+days₎

mean+SD 39+3₊₁+1 Inductionoflaborn(%) 249(45.9%) Modeofdeliveryn(%) SVD435(77.4%) VE31(5.5%) CS96(17.1%) Obstetricinterventionduetodystocian

(%)

VE13(41.9%) CS76(79.2%) Shoulderdystocian(%) Yes4(0.7%) FetalGendern(%) Male316(56.2%) Birthweight(grams)mean+SD 3737+474 Birthweightpercentilemean+SD 73.0+24.5 UmbilicalarterypH 7.26+0.08 mean+SD n=464 UmbilicalveinpH 7.34+0.07 mean+SD n=356 Apgarat5mins 9(4–10) median(range) n=554

NICUadmissionn(%) Yes8(1.4%) Lengthofneonataladmissionmedian

(range)

2(1–17)

Amongtheincludedcases,194(34.5%)fetuseshadEFW>95th percentileaccordingtoWHOchartsand198(33.8%)accordingto customizedcurves. In7.5%ofcases(42/562)theWHOandthe customizedchartsyieldeddiscordantresultsintheidentification ofLGA(Table1),withacomparablenumberofcasesthat were classifiedasLGAonlybyonechart(4.1%[23/562]forWHOand3.4 %[19/562]forcustomizedcharts,respectively).

Demographicfeaturesandperinataloutcomesoftheincluded casesaresummarizedinTable2.Withinourpopulationof 562 fetusesat risk ofmacrosomia, vaginaldelivery occurredin466 cases(82.9%),ofwhich435(77.4%)wereSVDand31VE(5.5%), whileCSwasperformedintheremaining96cases(17.1%).Inthis latter group, intrapartum dystocia represented the leading indication76cases(79.2%).

Thebirthweightwas>95thcentilein103(18.3%)neonates.Of these,66(64.1%)and62(60.2%)hadbeenantenatallyclassifiedas LGAatWHOandcustomizedcharts,respectively(p=0.57).

TheperinataloutcomesoffetusesidentifiedasLGAaccordingto WHOandcustomizedantenatalchartsaresummarizedinTables3 and 4, respectively. Fetuses identified as LGA at WHO charts showedsignificantlylowergestationalageatscan(p=0.01)andat birth (p<0.001) and had lower augmentation rate (p=0.002), whilethefrequencyofobstetricinterventionandCSduetolabor dystociaweresignificantlyhighercomparedtofetuseswithEFW <95th percentile (p<0.001 for both), as werethe birthweight (p<0.001),thebirthweightpercentile(p<0.001)andthelength

of NICU admission (p=0.003). In fetuses with EFW >95th percentileaccordingtocustomizedgrowthchartsasignificantly lower gestational age at scan was also noted (p<0.001); additionally, a significantly lower gestational age at birth and augmentationratewererecorded(p<0.001forboth);finally,the frequencyofobstetricinterventionandCSduetolabordystocia wassignificantlyhighercomparedtonon-LGAfetuses(p=0.001 andp<0.001,respectively),aswerethebirthweight(p<0.001), the birthweight percentile (p<0.001) and the length of the neonataladmissiontoNICU(p0.04).

Theresultsofthepairedcomparisonoftheperinataloutcomes offetusesidenti_fiedasLGAatWHOandcustomizedgrowthcharts, respectively, is shown in Table 5. No differences were found betweenthetwogroupsforanyoftheevaluateddemographicand perinataloutcomes,therateofobstetricinterventionandCSdueto intrapartum dystocia. Sensitivity, specificity, PPV and NPV and LR+andLR-fortheidentificationofcasesofCSduetointrapartum dystociainLGAfetusesidentifiedatWHOandcustomizedcharts aresummarizedinTable6.Bothchartsshowedlowandsimilar sensitivity[0.57,95%CI(0.45–0.68)forWHOchartsand0.52,95% CI(0.40–0.64)forcustomized,respectively]andspecificity[0.72, 95%CI(0.68–0.76))forWHOchartsand0.73,95%CI(0.69–0.77) forcustomizedcharts],withlowPPV[0.24,95%CI(0.18–0.31)for WHOchartsand0.23,95%CI(0.17–0.30)forcustomizedcharts] andLR+[2.01,95%CI(1.57–2.56)and1.90,95%CI(1.47–2.47)for WHOandcustomizedcharts,respectively].

Table3

Perinataloutcomesinnormalsizevslarge-for-gestationalagefetusesaccordingtoWHOcharts.

EFW<95thpercentileWHOchart EFW>95thpercentileWHOchart p

N368 N194

GestationalageatUSexamination(weeks+days

)mean+SD 36+6

+0+5

36+5

+0+6

0.01 Gestationalageatdelivery(weeks+days

)mean+SD 39+5

+1+0

39+0

+1+2 _<0.01

FetalGendern(%) Male200(54.3%) Male116(59.8%) 0.23 Inductionoflaborn(%)n=561 Yes167(45.5%) Yes82(42.3%) 0.46 Lengthoflabor(minutes) 283+196 307+224 0.24 mean+SD

n=446

Modeofdeliveryn(%) SVD303(82.3%) SVD132(68.0%) <0.01 VE18(4.9%) VE13(6.8%)

CS47(12.8%) CS49(25.2%)

Modeofdelivery–Obstetricinterventionduetofetaldistressexcludedn(%) SVD303(88.1%) SVD132(73.3%) <0.01 VE8(2.3%) VE5(2.8%)

CS33(9.6%) CS43(23.9%)

Birthweight(grams)mean+SD 3617+468 3957+405 <0.01 Birthweightpercentilemean+SD 65.9+25.9 86.3+14.4 <0.01 Shoulderdystocian(%) Yes3(0.8%) Yes1(0.5%) 0.69 Augmentation Yes147(48.7%) Yes58(33.3%) <0.01 n(%)

n=476

Epidural Yes124(52.8%) Yes59(50.4%) 0.68

n(%) n=352

III-IVdegreetearn(%) Yes6(1.6%) Yes4(2.1%) 0.71 Episiotomy Yes65(18.1%) Yes46(24.5%) 0.08 n(%)

n=548

APGAR5<7 Yes2(0.6%) Yes1(0.5%) 0.96 n(%) n=554 UApH 7.26+0.08 7.25+0.08 0.20 mean+SD n=464 UApH<7.10 Yes13(4.3%) Yes6(3.7%) 0.75 n(%) n=464

NICUadmission Yes5(1.4%) Yes3(1.7%) 0.82 n(%)

N=523

Lengthofneonataladmission(days)median(range) 2(2–15) 3(1–17) <0.01 US:ultrasound.

Discussion

Thisstudyhasdemonstratedthat,withinaselectedpopulation of pregnant women at risk for fetal macrosomia, the use of customizedgrowth charts comparedtostandard growthcharts doesnotimprovetheidentificationofLGAfetuseswhoeventually undergounplannedcaesareansectionduetointrapartumdystocia. Inourcohort,theperformanceofWHOandcustomizedgrowth chartswassimilarintermsofidentificationofLGAfetusesandthe perinataloutcomewasalsocomparablebetweenthetwogroups. Itisacommongroundinobstetricstoassumethatthegreater thematernal size thelower thelikelihood of obstructed labor, irrespective of the actual fetal size. Indeed, the available data suggestsadirectrelationshipbetweenmaternalheightandsizeof the birth canal [28–30]. Furthermore, the use of customized growth chartsallows tocorrelatethefetal sizetomaternal (or parental)characteristics[31–34]andissupposedtodefinewhich istheappropriatebiometryofaspeci_ficfetus inrelationtothe maternalanthropometricfeatures.

On this basis, it seems tempting to speculate that a fetus appearing large after having adjusted its growth trajectory for maternal (and/or parental) characteristics is expectedto be at higher risk of caesarean delivery due toobstructed labor. The background concept is that infants whose size is

disproportionately large for their mothers are at risk of intra-partumand/orperinatalcomplications.

On the other hand, the results from our multicentricwork suggestthatcustomizedgrowthchartsareunlikelytobehelpfulin the prediction of obstructed labor due to CPD nor in the antepartumidentificationofcasesathigherriskofmajorperinatal complications.Differentlyfromwhatweexpectedindeedinthe late3rdtrimestertheuseofItaliancustomizedgrowthchartsbuilt onalgorithmswhich includea seriesofparentalcharacteristics [19] showed identical performances compared to the recent standardchartsdevelopedbytheWHO[20].Thedecisiontoadopt theselatterchartsasthestandardreferenceforcomparisonwas sharedamongthemainstudyinvestigatorsandwasbasedonthe robustnessandexternalvalidityofthedatacollectedbyKiserud et al. [20]. It is unknown whether the comparison of the customized Italian charts with different standard growth chart at 35–37weeks wouldhaveledtodifferentresultsinterms of identi_fication of LGA newborns and prediction of CS due to obstructedlabor.

The evaluation of the role of the customized charts in the identificationoffetalovergrowthleadingtoobstructedlaborand itsassociatedperinatalmorbidityhassofaryieldedcontradictory results innon-selectedaswell asinpopulationsat riskoffetal macrosomia.Fourstudiesconductedonnon-selectedpopulations

Table4

Perinataloutcomesinnormalsizevslarge-for-gestationalagefetusesaccordingtocustomizedcharts.

EFW<95thpercentilecustomizedchart EFW>95thpercentilecustomizedchart P

N372 N190

Gestationalageatscan(weeks+days

)mean+SD 36+6

+0+5

36+4

+0+6

<0.01 Gestationalageatdelivery(weeks+days

)mean+SD 39+5

+1+0

38+6

+1+2 _<0.01

FetalGendern(%) Male206(55.5%) Male110(57.9%) 0.59 Inductionoflabor Yes166(44.7%) Yes83(43.7%) 0.81 n(%)

n=561

Lengthoflabor(minutes) 291+202 284+226 0.81 mean+SD

n=446

Modeofdeliveryn(%) SVD305(82.0%) SVD130(68.4%) <0.01 VE19(5.1%) VE12(6.3%)

CS48(12.9%) CS48(25.3%)

Modeofdelivery–Obstetricinterventionduetofetaldistressexcluded SVD305(86.9%) SVD130(75.6%) <0.01 VE10(2.9%) VE3(1.7%)

CS36(10.2%) CS39(22.7%) n(%)

Birthweight(grams)mean+SD 3627+470 3931+415 <0.01 Birthweightpercentilemean+SD 66.0+25.9 85.8+14.5 <0.01 Shoulderdystocian(%) Yes4(1.1%) Yes0(0.0%) 0.15 Augmentation Yes152(49.5%) Yes53(31.4%) <0.01 n(%)

n=476

Epidural Yes123(52.8%) Yes60(50.4%) 0.67

n(%) n=352

III-IVdegreetearn(%) Yes8(2.2%) Yes2(1.1%) 0.35 Episiotomy Yes67(18.4%) Yes44(23.9%) 0.13 n(%)

n=548

APGAR5<7 Yes1(0.3%) Yes2(1.1%) 0.23 n(%) n=554 UApH 7.26+0.09 7.26+0.08 0.18 mean+SD n=464 UApH<7.10 Yes14(4.6%) Yes5(3.1%) 0.43 n(%) n=464

NICUadmission Yes4(1.1%) Yes4(2.3%) 0.31 n(%)

N=523

Lengthofneonataladmission(days) 2(2–9) 3(1–17) 0.04 median(range)

[13,15,16,26]foundthatcustomizedmodelsbasedonapreviously described algorithm [34] can improve the recognition of LGA populationsatriskofintrapartummorbidity,whileSjaardaetal. [18]couldnotdemonstrateinthisrespectadecisivesuperiorityof customized charts compared with population-based curves. Similarresultswerefoundinstudiesspecificallyfocusedonhigh risk populations. Within a selected cohort of pregnancies

complicated by diabetes, Gonzalez et al. [14] found that the identificationofLGAusingcustomizedchartswasassociatedwith a higher incidence of caesarean sections performed due to intrapartumdystocia,whileanearlierstudyconductedonmothers affectedbygestationaldiabetescouldnotdemonstratea better identification of LGA neonates at risk of adverse perinatal outcomescomparedtopopulationcurves[17].Itisimportantto

Table5

Comparisonoftheperinataloutcomesoffetuseslarge-for-gestationalagefetusesaccordingtoWHOversuscustomizedgrowthcharts.

EFW>95thpercentileWHOchart EFW>95thpercentilecustomizedchart p

N194 N190

Gestationalageatscan(weeks+days

)mean+SD 36+5

+0+6

36+4

+0+6

0.54 Gestationalageatdelivery(weeks+days

)mean+SD 39+0

+1+2

38+6

+1+2

0.59 FetalGendern(%) Male116(59.8%) Male110(57.9%) 0.70 Inductionoflabor Yes82(42.3%) Yes83(43.7%) 0.78 n(%)

n=561

Lengthoflabor(minutes) 307+224 284+226 0.45 mean+SD

n=446

Modeofdeliveryn(%) SVD132(68.0%) SVD130(68.4%) 0.99 VE13(6.8%) VE12(6.3%)

CS49(25.2%) CS48(25.3%)

Modeofdelivery–Obstetricinterventionduetofetaldistressexcludedn(%) SVD132(73.3%) SVD130(75.6%) 0.77 VE5(2.8%) VE3(1.7%)

CS43(23.9%) CS39(22.7%)

Obstetricinterventionduetodystocian(%) Yes48(24.7%) Yes42(22.1%) 0.54 Caesareansectionduetointrapartumdystocian(%) Yes43(22.2%) Yes39(20.5%) 0.70 Birthweight(grams)mean+SD 3957+405 3931+415 0.72 Birthweightpercentilemean+SD 86.3+14.4 85.8+14.5 0.93 Shoulderdystocia Yes1(0.5%) Yes0(0.0%) 0.32 n(%)

Augmentation Yes58(33.3%) Yes53(31.4%) 0.70 n(%)

n=476

Epidural Yes59(50.4%) Yes60(50.4%) 0.99

n(%) n=352

III-IVdegreetear Yes4(2.1%) Yes2(1.1%) 0.43 n(%)

Episiotomy Yes46(24.5%) Yes44(23.9%) 0.90

n(%) n=548

APGAR5<7 Yes1(0.5%) Yes2(1.1%) 0.55

n(%) n=554 UApH 7.25+0.08 7.26+0.08 0.97 mean+SD n=464 UApH<7.10 Yes6(3.7%) Yes5(3.1%) 0.79 n(%) n=464

NICUadmission Yes3(1.7%) Yes4(2.3%) 0.69

n(%) N=523

Lengthofneonataladmission(days) 3(1–17) 3(1–17) 0.64 median(range)

US:ultrasound. UA:umbilicalartery. UV:umbilicalvein.

NICU:neonatalintensivecareunit.

Table6

Sensitivity,specificity,positiveandnegativepredictivevalues(PPVandNPV),positiveandnegativelikelihoodratios(LR+andLR-)fortheidentificationofcasesofcaesarean sectionduetointrapartumdystociainlarge-for-gestationalage(LGA)fetusesidentifiedatWHOandcustomizedcharts.

EFW>95thpercentileWHOchart EFW>95thpercentilecustomizedchart Sensitivity 0.57,95%CI(0.45–0.68) 0.52,95%CI(0.40–0.64) Specificity 0.72,95%CI(0.68–0.76) 0.73,95%CI(0.69–0.77) PPV 0.24,95%CI(0.18–0.31) 0.23,95%CI(0.17–0.30) NPV 0.91,95%CI(0.88–0.94) 0.91,95%CI(0.87–0.93) LR+ 2.01,95%CI(1.57–2.56) 1.90,95%CI(1.47–2.47) LR- 1.65,95%CI(1.27–2.15) 1.51,95%CI(1.19–1.93)

notethatinalltheaforementionedstudiesthecustomizedmodels weredevelopedbasedonapreviouslydescribedalgorithmbased onbirthweight[34],whiletheSIEOGcustomizedchartshavebeen developedfromcross-sectionalultrasoundmeasurements[19].

Theuseofcustomizedfetalgrowthchartsinclinicalpracticeis still a matter of controversy. Some have suggested a better performanceofthecustomizedapproachintheidentificationof smallfetuses atriskofadverseoutcomescomparedtodifferent local,national orinternational standards[35–38]. Nevertheless, the rationale behind customized growthcharts has been chal-lengedbytherecentlyimplementedinternationalgrowth stand-ardspublishedbytheIntergrowthconsortiumandbytheWorld HealthOrganization[20,39]andasystematicreviewof20studies failed to demonstrate the superiority of either method in the identification of fetuses at risk for adverse perinatal outcome includingmortality[40].

Customization per se represents a mathematic algorithm designed in order to adjust the fetal growth trajectory on the basisoftheanthropometriccharacteristicsoftheparents[19,31– 34].Regardingthepossibleusefulnessofacustomizedmodelin predictingtheoccurrenceofcephalopelvicdisproportionamong LGA infants, the contribution of paternal characteristics is debatable when not difficult to ascertain. Additionally, the currentlyavailablecustomized growthmodelsdo not takeinto account specifically maternal pelvimetryparameters which are relatedtothewomansizebutmayhaveanindependentmajor impactonthechanceofdystocialeadingtoobstetricintervention. Amongthese,anarrowwidthofthesubpubicarchangle(SPA)asa surrogateofanarrowbirthcanalhasbeendemonstratedbysome investigatorsofourgrouptobeindependentlyassociatedwiththe riskof obstructedlaborwithina selectedcohort ofnulliparous womenwithLGAfetuses[12];ofnote,inarecentworkbyRizzo etal.bothfetal HCandSPAwerefoundtobeindependentrisk factorsfor intrapartumdystocialeading toemergencyobstetric intervention [41]. On this ground, we do envisage that it is reasonabletohypothesizethattheperformanceofthe customiza-tioninpredictingtheriskofobstructedlaboramongLGAfetuses can be improved by including pelvimetric parameters in the customization method [18]. Finally, intrapartum fetal head malpositions and malpresentations also represent a major determinant of dystocia leading to caesarean section and this factorisnotincluded–andcannotbeincluded–inanyantepartum customizedmodelaimingatthedetectionofcasesatriskofCPD [42–45].

Toourknowledge,thisisthefirststudyevaluatingtherole of newly developed ultrasound-derived customized growth charts[19] intheidentificationof LGAwithinapopulationat high risk of fetal overgrowth and to compare their perfor-mance tothat of recently implemented standardcharts [20]. Theprospectivedesignofthedatacollectiontogetherwiththe wide patient sample collected and the strict criteria for the inclusionof thepatientsare themajor strengthsof ourwork. Moreover,in allincluded units theUSestimationof the fetal biometrywasperformedbyFetalMedicinespecialists.Finally, ithastobeacknowledgedthatalltheparticipatingcentersare tertiaryreferral hospitalsthat use sharedand internationally acknowledgedprotocol forthemanagementof labor progres-sion [1]. On the other hand, the fact that each center used localgrowthstandardsforthemanagementofthecasesatrisk of fetalmacrosomia andthatthe policy ofinduction of labor for suspected macrosomia also differed in the participating unitsmayrepresentalimitation,howeverdifferencesinterms of management policy across different Centres are not uncommon in the routine clinical practice. On this basis, 88 cases were excluded due to elective induction of labor for suspectedmacrosomiabefore39+0_{weeks,howeverwebelieve}

that their inclusion would have biased the validity of our results. Within such context, the paired comparisonbetween the customized and the WHO standards did not yield any “subclinical” difference worth to be investigated on a wider number ofcases.Anotherlimitationmaybeaccountedbythe method adopted to obtain the EFW from customized US measurements, which may have impacted on the effect of customisation. Furthermore, the decision to use the 95th percentileasthecut-offvalueforLGAwasarbitraryandbased on the fact that we aimed to identify the fetuses at highest risk of obstructedlabor. Of note, such cut-offvalue hasbeen extensively used by several Authors and research groups [10,46–49] and has recently been suggested to reduce the likelihood of false-positivecases [47].

Inconclusion,ourdataonaselectedcohortofwomenatriskfor fetal macrosomia suggest that the use of newly developed ultrasound-derived customized antenatal growth charts does not improvetheidentification ofLGA infantsundergoing intra-partum caesarean section due to suspected cephalo-pelvic disproportion.Furtherresearchiswarrantedinordertoassessif theuseofcustomizedgrowthchartsspecificallyfocusedonthe identificationoffetalovergrowthmayrefinethepredictionofthe riskofobstetricinterventionduetointrapartumdystocia. DeclarationofCompetingInterest

The Authors state no financial disclosures nor conflict of interestrelatedtothecontentofthiswork.

Funding None.

Acknowledgements None.

References

[1]AmericanCollegeofObstetriciansandGynecologists(College),Societyfor Maternal-FetalMedicine,CaugheyAB,CahillAG,GuiseJM,RouseDJ.Safe preventionoftheprimarycesareandelivery.AmJObstetGynecol2014;210 (March3):179–93.

[2]BarberEL,LundsbergLS,BelangerK,PettkerCM,FunaiEF,IlluzziJL.Indications contributingtotheincreasingcesareandeliveryrate.ObstetGynecol2011;118 (July1):29–38.

[3]HadlockFP,HarristRB,SharmanRS,DeterRL,ParkSK.Estimationoffetal weightwiththeuseofhead,body,andfemurmeasurements-aprospective study.AmJObstetGynecol1985;151:333–7.

[4]CatalanoPM,TyzbirED,AllenSR,McBeanJH,McAuliffeTL.Evaluationof growth by estimation of neonatal body composition. Obstet Gynecol 1992;79:46–50.

[5]AkSood,YanceyM,RichardsD.Predictionoffetalmacrosomiausinghumeral softtissuethickness.ObstetGynecol1995;85:937.

[6]LeeW,DeterR,EbersoleJ,etal.Birthweightpredictionbythree-dimensional ultrasonography:fractionallimbvolume.JUltrasoundMed2001;20:1283–92. [7]LeeW,BalasubraniamM,DeterRL,etal.Fractionallimbvolume-asofttissue parameteroffetalbodycomposition:validation,technicalconsiderationsand normalrangesduringpregnancy.UltrasoundObstetGynecol2009;33:427–40. [8]FavreR,BaderAM,NidansG.Prospectivestudyonfetalweightestimation using limb circumferences obtained by three-dimensional ultrasound. UltrasoundObstetGynecol1995;6:140–4.

[9]HoopmannM,AbeleH,WagnerN,etal.Performanceof36differentweight estimation formulae in fetuses with macrosomia. Fetal Diagn Ther 2010;27:204–13.

[10]BoulvainM,SenatMV,PerrotinF,etal.GroupedeRechercheenObstétriqueet Gynécologie(GROG).Inductionoflabourversusexpectantmanagementfor large-for-datefetuses:arandomisedcontrolledtrial.Lancet2015;385(June 9987):2600–5.

[11]CunninghamFG,LevenoKJ,BloomSL,etal.Williamsobstetrics.24th

edition NewYork:McGraw-HillEducation;2014.

[12]GhiT,Dall’AstaA,SupraniA,etal.Correlationbetweensubpubicarchangle andmodeofdeliveryinlarge-for-gestational-agefetuses.FetalDiagnTher 2018;44(3):221–7.

[13]ChaHH,KimJY,ChoiSJ,OhSY,RohCR,KimJH.Canacustomizedstandardfor largeforgestationalageidentifywomenatriskofoperativedeliveryand shoulderdystocia?JPerinatMed2012;40:483–8.

[14]González-GonzálezNL,González-DávilaE,CabreraF,etal. Applicationof customizedbirthweightcurvesintheassessmentofperinataloutcomesin infantsofdiabeticmothers.FetalDiagnTher2015;37:117–22.

[15]LarkinJC,SpeerPD,SimhanHN.Acustomizedstandardoflargesizefor gestational age to predict intrapartum morbidity. Am J Obstet Gynecol 2011;204(499):e1–10.

[16]PasupathyD,McCowanLM,PostonL,KennyLC,DekkerGA,NorthRA.Perinatal outcomes in large infants using customised birthweight centiles and conventional measures of high birthweight. Paediatr Perinat Epidemiol 2012;26:543–52.

[17]CostantineMM,MeleL,LandonMB,etal.EuniceKennedyShriverNational InstituteofChildHealthandHumanDevelopmentMaternal-FetalMedicine UnitsNetwork,Bethesda,Maryland.Customizedversuspopulationapproach forevaluationoffetalovergrowth.AmJPerinatol2013;30(August7):565–72. [18]Sjaarda LA,Albert PS, Mumford SL,Hinkle SN, Mendola P,Laughon SK. Customizedlarge-for-gestational-agebirthweightattermandtheassociation withadverseperinataloutcomes.AmJObstetGynecol2014;210(January1)63 e1-63.e11.

[19]GhiT,CarielloL,RizzoL,etal.Customizedfetalgrowthchartsforparents’ characteristics, race, and parity byquantile regression analysis:a cross-sectionalmulticenteritalianstudy.JUltrasoundMed2016;35(January1):83– 92.

[20]KiserudT,PiaggioG,CarroliGetal.TheWorldHealthOrganizationFetal GrowthCharts:AMultinationalLongitudinalStudyofUltrasoundBiometric Measurements and Estimated Fetal Weight. PLoS Med. 2017;14(Jan 1): e1002220.10.1371/journal.pmed.1002220eCollection2017Jan.Erratumin: PLoSMed.2017Mar24;14(3):e1002284.Erratumin:PLoSMed.2017Apr 20;14 (4):e1002301. PubMed PMID: 28118360; PubMed Central PMCID: PMC5261648.

[21]WorldHealthOrganization.Definition,diagnosisandclassificationofdiabetes mellitusanditscomplications.Geneva:WorldHealthOrganization;1999. [22]ItalianSocietyonUltrasoundinObstetricsandGynecology(SIEOG).

Guide-lines. Availableat.2015.https://www.sigo.it/wp-content/uploads/2015/12/ LineeGuidaSieog_2015.pdf.

[23]GhiT,PrefumoF,FicheraA,etal.Developmentofcustomizedfetalgrowth chartsintwins.AmJObstetGynecol2017;216(5)514,doi:http://dx.doi.org/ 10.1016/j.ajog.2016.12.176e1–514.e17.

[24]BertinoE,SpadaE,OcchiL,etal.Neonatalanthropometriccharts:theItalian neonatalstudycomparedwithotherEuropeanstudies.JPediatrGastroenterol Nutr2010;51(September3):353–61.

[25]Ayres-de-Campos D, Spong CY, Chandraharan E. FIGO intrapartum fetal monitoringexpertconsensuspanel. FIGOconsensus guidelineson intra-partumfetalmonitoring:Cardiotocography.IntJGynaecolObstet.2015;131 (October1):13–24.

[26]Pritchard N,Lindquist A, Hiscock R, Diksha P, Walker SP, Permezel M. Customised growth charts in large-for-gestational-age infants and the associationwithemergencycaesareansectionrate.AustNZJObstetGynaecol 2019;59(June3):380–6.

[27]VonElmE,AltmanDG,EggerM,etal.TheStrengtheningtheReportingofthe ObservationalStudiesinEpidemiology(STROBE)statement:guidelinesfor reportingobservationalstudies.Lancet2007;370:1453–7.

[28]LiseleleHB,BoulvainM,TshibanguKC,MeurisS.Maternalheightandexternal pelvimetry to predict cephalopelvic disproportionin nulliparous African women:acohortstudy.BJOG2000;107(August8):947–52.

[29]RozenholcAT,AkoSN,LekeRJ,BoulvainM.Thediagnosticaccuracyofexternal pelvimetryandmaternalheighttopredictdystociainnulliparouswomen:a studyinCameroon.BJOG2007;114(May5):630–5.

[30]ConnollyG,McKennaP.Maternalheightandexternalpelvimetrytopredict cephalo-pelvicdisproportioninnulliparousAfricanwomen.BJOG2001;108 (March3):338.

[31]Gardosi J, Francis A, Turner S, Williams M. Customized growth charts: rationale,validationand clinicalbenefits.AmJObstet Gynecol2018;218 (February2S):S609–18.

[32]FiguerasF,GardosiJ.Shouldwecustomizefetalgrowthstandards?FetalDiagn Ther2009;25(3):297–303.

[33]GardosiJ.Customizedchartsandtheirroleinidentifyingpregnanciesatrisk becauseof fetal growth restriction.J ObstetGynaecol Can 2014;36(May 5):408–15.

[34]GardosiJ,ChangA,KalyanB,SahotaD,SymondsEM.Customizedantenatal growthcharts.Lancet1992;339:283–7.

[35]ClaussonB,GardosiJ,FrancisA,CnattingiusS.PerinataloutcomeinSGAbirths definedbycustomisedversuspopulation-basedbirthweightstandards.BJOG 2001;108(August8):830–4.

[36]OdiboAO,CahillAG,OdiboL,RoehlK,MaconesGA.Predictionofintrauterine fetaldeathinsmall-for-gestational-agefetuses:impactofincluding ultra-soundbiometryincustomizedmodels.UltrasoundObstetGynecol2012;39 (March3):288–92.

[37]AndersonNH,SadlerLC,McKinlayCJD,McCowanLME.INTERGROWTH-21stvs customizedbirthweightstandardsforidentificationofperinatalmortalityand morbidity.AmJObstetGynecol2016;214(April4)509e1-509.e7.

[38]FrancisA,HughO,GardosiJ.CustomizedvsINTERGROWTH-21(st)standards fortheassessmentofbirthweightandstillbirthriskatterm.AmJObstet Gynecol2018;218(February2S):S692–9.

[39]PapageorghiouAT,Ohuma EO, AltmanDG, etal. InternationalFetaland NewbornGrowthConsortiumforthe21stCentury(INTERGROWTH-21st). Internationalstandardsforfetalgrowthbasedonserialultrasound measure-ments: the FetalGrowth Longitudinal Studyof the INTERGROWTH-21st Project. Lancet 2014;384(September 9946)869–79, doi:http://dx.doi.org/ 10.1016/S0140-6736(14)61490-2 Erratum in: Lancet. 2014 Oct 4;384 (9950):1264.PubMedPMID:25209488.

[40]ChiossiG, PedrozaC,CostantineMM,TruongVTT, GarganoG,SaadeGR. Customizedvspopulation-basedgrowthchartstoidentifyneonatesatriskof adverseoutcome:systematicreviewandBayesianmeta-analysisof observa-tionalstudies.UltrasoundObstetGynecol2017;50(Aug2):156–66. [41]RizzoG,AielloE,BosiC,D’AntonioF,ArduiniD.Fetalheadcircumferenceand

subpubicangleareindependent riskfactorsforunplannedcesareanand operativedelivery.ActaObstetGynecolScand2017;96(Aug8):1006–11. [42]MalvasiA,BarberaA,DiVagnoG,etal.Asynclitism:aliteraturereviewofan

oftenforgottenclinical condition.J Matern Fetal NeonatalMed 2015;28 (November16):1890–4.

[43]GhiT,BellussiF,PiluG.Sonographicdiagnosisoflateralasynclitism:anew subtypeoffetalheadmalpositionasamaindeterminantofearlylaborarrest. UltrasoundObstetGynecol2015;45(Feb2):229–31.

[44]GhiT,Dall’AstaA,KienerA,VolpeN,SupraniA,FruscaT.Intrapartumdiagnosis ofposterior asynclitismusing two-dimensional transperineal ultrasound. UltrasoundObstetGynecol2017;49(June6):803–4.

[45]Dall’AstaA,VolpeN,GalliL,FruscaT,GhiT.Intrapartumsonographicdiagnosis ofcompoundhand-cephalicpresentation.UltraschallMed2017;38(October 5):558–9.

[46]Wright D, Wright A, Smith E, Nicolaides KH. Impact of biometric measurement error onidentificationofsmall-andlarge-for-gestational-agefetuses.UltrasoundObstet Gynecol2020;55(2):170–6,doi:http://dx.doi.org/10.1002/uog.21909. [47]KhanN,CiobanuA,KarampitsakosT,AkolekarR,NicolaidesKH.Predictionof

large-for-gestational-age neonate by routine third-trimester ultrasound. Ultrasound Obstet Gynecol 2019;54(3):326–33, doi:http://dx.doi.org/ 10.1002/uog.20377.

[48]LipschuetzM,CohenSM,Ein-MorE,etal.Alargeheadcircumferenceismore stronglyassociatedwithunplannedcesareanorinstrumentaldeliveryand neonatalcomplicationsthanhighbirthweight.AmJObstetGynecol2015;213 (6)833,doi:http://dx.doi.org/10.1016/j.ajog.2015.07.045e1–833.e12. [49]Mazzone E,

Dall’AstaA,KienerAJO,etal.Predictionoffetalmacrosomiausingtwo-dimensionalandthree-dimensionalultrasound.EurJObstetGynecolReprodBiol 2019;243:26–31,doi:http://dx.doi.org/10.1016/j.ejogrb.2019.10.003.