Neck circumference cutoff points to identify excess android fat

(1)

www.jped.com.br

ORIGINAL

ARTICLE

Neck

circumference

cutoff

points

to

identify

excess

android

fat

夽

Mariana

De

Santis

Filgueiras

∗

,

Fernanda

Martins

de

Albuquerque,

Ana

Paula

Pereira

Castro,

Naruna

Pereira

Rocha,

Luana

Cupertino

Milagres,

Juliana

Farias

de

Novaes

UniversidadeFederaldeViçosa(UFV),DepartamentodeNutriçãoeSaúde,Viçosa,MG,BrazilReceived25April2018;accepted19 November2018

Received25April2018;accepted23November2018

KEYWORDS Adiposity; Anthropometry; Riskfactors; Nutritional assessment; ROCcurve Abstract

Objective: Toevaluatetheabilityofneckcircumferencetoidentifyexcessandroidfatandto proposecutoffpointsforBrazilianchildren.

Method: Thiswas a cross-sectionalstudy with376 children aged 8and9years enrolled in publicandprivateschoolsintheurbanareaofthemunicipalityofVic¸osa,MinasGerais.A semi-structuredquestionnairecontainingsociodemographicandlifestyleinformationwas applied. Thefollowingwerecollected:neckcircumference,weight,andheightforthecalculationof bodymassindex.Thepercentageoffatintheandroidregionwasdeterminedbydualenergy X-rayabsorptiometry.Linearregressionanalysiswasusedtoevaluatetheassociationbetween neckcircumferenceandandroidfat,adoptingasigniﬁcancelevelof5%.Receiveroperating characteristiccurveswereusedtoevaluatethecapacityofneckcircumferencetodetermine theexcessandroidfat,aswellastoestimatethecutoffpointsofneckcircumferenceaccording togender.

Results: Multiple linear regressionshowed an associationbetween neck circumference and androidfat(ˇ:2.94,95%CI:2.41,3.47).Neckcircumferencewasabletoidentifyexcessandroid fatingirls(AUC:0.909,95%CI:0.999,0.945)andboys(AUC:0.938,95%CI:0.892,0.968).The proposedcutoffpointsshowedsatisfactorysensitivity,speciﬁcity,andpredictivevalues.

Conclusions: Neckcircumferenceiscapableofidentifyingexcessandroidfatinchildrenand canbeusedinclinicalpracticeandinpopulationstudiestodeterminecentraladiposity.The proposedcutoffpointsweresatisfactory,butshouldbevalidatedforotherpopulations. ©2019SociedadeBrasileiradePediatria.PublishedbyElsevierEditoraLtda.Thisisanopen accessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/ 4.0/).

夽 _Please_cite_this_article_as:_Filgueiras_MS,_Albuquerque_FM,_Castro_AP,_Rocha_NP,_Milagres_LC,_Novaes_JF._Neck_{circumference}_cutoff_points

toidentifyexcessandroidfat.JPediatr(RioJ).2019.https://doi.org/10.1016/j.jped.2018.11.009

∗_{Corresponding}_author.

E-mail:mariana.ﬁlgueiras@ufv.br(M.D.Filgueiras).

https://doi.org/10.1016/j.jped.2018.11.009

license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

(2)

PALAVRAS-CHAVE

Adiposidade; Antropometria; Fatoresderisco; Avaliac¸ãonutricional; CurvaROC

Pontosdecortedoperímetrodopescoc¸oparaidentiﬁcaroexcessodegordura androide

Resumo

Objetivo: Avaliaracapacidadedoperímetrodopescoçoemidentificaroexcessodegordura androideeproporpontosdecorteparacriançasbrasileiras.

Método: Estudotransversalcom376criançasde8e9anosdeidade,matriculadasemescolas públicaseprivadasdaáreaurbanadeViçosa,MinasGerais.Foiaplicadoumquestionário semi-estruturado contendo informações sociodemográficasede estilo de vida. Foramcoletados: perímetrodopescoço,pesoeestaturaparaocálculodoíndicedemassacorporal.Opercentual degorduranaregiãoandroidefoideterminadopelaabsorciometriaderaios-Xdeduplaenergia. Aanálisederegressãolinearfoiutilizadaparaavaliaraassociaçãoentreoperímetrodopescoço eagorduraandroide,adotando-seoníveldesignificânciade5%.PormeiodascurvasReceiver OperatingCharacteristicavaliou-seacapacidade doperímetrodopescoçoem determinaro excessodegorduraandroide,bemcomoestimou-seospontosdecortedeperímetrodopescoço segundoosexo.

Resultados: Aregressãolinearmúltiplademonstrouassociaçãoentreoperímetrodopescoçoe agorduraandroide(␤:2,94;IC95%:2,41;3,47).Operímetrodopescoçofoicapazdeidentificar oexcessodegorduraandroideemmeninas(AUC:0,909;IC95%:0,859;0,945)emeninos(AUC: 0,938;IC95%:0,892;0,968).Ospontosdecortepropostostiveramvalores desensibilidade, especificidadeevalorespreditivossatisfatórios.

Conclusões: Operímetrodopescoçoéumamedidacapazdeidentificaroexcessodegordura androideemcrianças,podendoserutilizadonapráticaclínicaeemestudospopulacionaispara determinaraadiposidadecentral.Ospontosdecortepropostosforamsatisfatórios,entretanto devemservalidadosparaoutraspopulações.

Introduction

Childhoodobesityis consideredaworldwidepublichealth problem, asit is associated withdifferent comorbidities, suchasdyslipidemias,type2diabetesmellitus,andarterial hypertension.1_In_Brazil,_the_Family_Budget_Survey_(Pesquisa

deOrc¸amentos Familiares, POF2008---2009) indicated the presenceofobesityin14.3%ofchildrenaged5to9years, while33.5%wereoverweight.2 _However,_it_is_important_to

considerthebodyfat distribution,sinceit playsa crucial roleinthedevelopmentofcardiometabolicrisk.3

Therearedifferenttechniquestoevaluatebodyfat,with dual energyX-ray absorptiometry (DXA) being considered a referencemethod toevaluate central adiposity in chil-drenofallages,sinceitestimatestheamountoffatinthe androidregion.4_However,_its_use_in_{epidemiological}_studies

andclinicalpracticeisalsolimitedduetoitshighcostand complexperformance,requiringsimpleandlessexpensive methods.

Inthiscontext,neckcircumference(NC) hasbeen pro-posed as a useful tool to identify young individuals with excessweight andcardiometabolicrisk, asit is a simple, low-costmethodandshows asmallervariationthroughout theday compared tothe waistcircumference.5 _However,

todate,nostudieshave beenfound intheliteraturethat evaluatedtheassociationandtheabilityofNCtoidentify excesscentraladiposityestimatedbyareferencemethod, such as fat content in the android region determined byDXA.

ThelackofthedeﬁnitionofNC-speciﬁccutoffpointsfor Brazilianchildrenmaybealimitationtoitsroutineuse.In theliterature,tothebestofourknowledge,thereis only onecross-sectionalstudywithBrazilianadolescentsenrolled inpublicschoolsinthecityofSãoPaulo,Brazil,which pro-posedNCcutoffpointstoidentifyoverweightandobesity.6

Therefore,studiesusingthisapproachatearlierageranges arescarce.

Thus,theobjectiveofthisstudywas:(1)toevaluatethe abilityofNCtoidentifyexcessandroidfat;(2)topropose NCcutoffpointstodetermineexcessandroidfatinBrazilian children.

Methods

Studypopulationanddesign

Thisisacross-sectionalstudy,carriedoutwitha representa-tivesampleof8-and9-year-oldchildrenenrolledinpublic andprivateschoolsintheurbanareaofthemunicipalityof Vic¸osa,state of MinasGerais,Brazil, fromMayto Decem-ber2015.The municipalityofVic¸osaislocatedinZonada

Mata Mineira, 227km from the capital city of Belo

Hori-zonte.Accordingtothe2010Census,Vic¸osahasalandarea of 299km2 _and _72,244 _inhabitants,_of _whom _93.2%_live _in

urbanareas.Thepercapitagrossdomesticproduct(GDP)is R$19391.087_{(US$5272.82).}

(3)

The study participants came from the Schoolchildren HealthAssessmentSurvey(PesquisadeAvaliac¸ãodaSaúde doEscolar,PASE),across-sectionalinvestigationthataimed to evaluate the cardiovascular health of this pediatric population in the municipality of Vic¸osa.8 _In _2015, _the

municipality had17publicschools andsevenprivateones intheurbanarea,attendedbychildrenaged8and9years, totalling1464enrolledchildren.

Samplecalculationandsamplingprocess

ThesamplewascalculatedusingEpiInfosoftware(version 7.2,Atlanta,GA,UnitedStates),consideringthetotal num-berofstudentsaged8and9yearsintheyear2015(n=1446), aprevalenceof10.7%ofobesity,9_tolerated_error _of_3%,_a

95%conﬁdencelevel,andanincreaseof20%tocoverlosses, totalling384childrentobestudied.

The children were selected by stratiﬁed random sam-pling, and the sample of each school reached the proportionalityofstudentsenrolledbyageandgender.The children’sselectionwasperformedbysimplerandomdraw untilthenumberneededforeachschoolwasreached.

Aftertheselection,atelephonecontactwasmadewith the children’sparents or guardians,inviting them to par-ticipateinthestudy.Thestudyaimandmethodologywere explainedtothe children’sparents or guardians,the ﬁrst meetingwasscheduled,andthosewhoagreedtoparticipate signedthefreeandinformedconsentformafterreceiving explanationsabouteachstepofthestudy.

Childrenwhohadanyhealthproblemsthatalteredtheir nutritionalstatusorbodycomposition(physical,cognitive, ormultipledeﬁciencies,thyroiddisorders,etc.),aswellas thosewithchronicuseofmedicationsthatinterferedwith glucose and/or lipid metabolism (corticosteroids, statins, hypoglycemicagents,etc.),andthoseusingvitaminor min-eralsupplementsduringthestudyorinthepreviousthree months (vitamin D and zinc supplementation) were not included, aswell ascasesin which it wasnotpossible to contactparents or guardian after threeattempts. A pilot studywascarriedoutinFebruary2015with10%ofthe sam-ple,includingchildrenaged8and9enrolledinarandomly selectedpublicschool,whichwerenotpartoftheﬁnalstudy sample.Throughthepilotstudy,itwaspossibletotestthe applicationofthequestionnaires.

Demographic,socioeconomic,andlifestyledata

A semi-structured questionnairecontaining socioeconomic and demographic information, such as age, gender, type of school(publicor private),and per capitaincome (cat-egorized according to the median of the total sample, equivalenttoR$500.00-US$135.96)wasapplied.

Aquestionnaireonlifestylewasappliedtoestimatethe screen time in hours per day (time spent sitting for long periodsinfrontofthetelevision,videogame,orcomputer). Sedentarybehaviorwasclassiﬁedasscreentime>2h/day.10

Anthropometryandbodycomposition

Datacollectionwasperformed in an appropriate environ-mentbyatrainedteammember,withbarefootvolunteers, whowerewearinglightclothing.Theweightwasmeasured onanelectronicdigitalscale,withacapacityof150kgand aprecisionof0.1kg(Tanita® ModelIronmanBC553;Tanita Inc.,ArlingtonHeights,IL, UnitedStates)andstature was determinedusing a stadiometer,divided into centimeters and subdivided into millimeters (Alturexata®, Belo Hori-zonte,MG,Brazil).

Usingbodyweightandheightdata,thebodymassindex (BMI) was calculated, and the BMI cutoff point for age (BMI/A)Z-scoreswereusedfortheclassiﬁcationofthe chil-dren’s nutritional status, as established by World Health Organization.11_To_assess_the_mothers’_nutritional_status,_the

criteriafor BMI classiﬁcation recommended by the World Health Organization were used.12 _Overweight _and _obese

children and mothers were grouped in the excess weight group.

The NCwasevaluatedwiththe child standingupright, withtheheadinthehorizontalplaneofFrankfurt,usingan inextensiblemeasuringtape(TBW®,SãoPaulo,Brazil),just belowthelaryngealprominenceintheneck,perpendicular tothelargestneckaxis.Minimumpressurewasexertedto allowfullcontactofthetapewiththeskin.The measure-mentwasperformedinduplicate,consideringtheaverage ofthetwomeasurements.Fortheclassiﬁcation,theauthors usedthecutoffsobtainedfromthereceiveroperating char-acteristic(ROC) curveswiththe bestvaluesof sensitivity, speciﬁcity, and predictive values for excess android fat, accordingtogender.

Thepercentageof fatintheandroidregionwas deter-mined by dual energy X-ray absorptiometry (DXA; Lunar ProdigyAdvance, GE Medical SystemsLunar--- Milwaukee, WI, United States) using the equipment software for the analysisof bodycomposition. The androidfat was identi-ﬁed in the region between the ribs and the pelvis, with theupperdemarcationbeingmadeat20%of thedistance oftheiliaccrestandtheneck,whereasthelower demar-cationwasmadeabovethepelvis.4 _No_cutoff_points_were

foundintheliteratureforthediagnosisofelevatedandroid fatinchildren.Therefore,the85thpercentileofthe sam-pleaccordingtogenderwasconsideredasincreasedvalues, whichcorresponded to34.0%and28.9%forgirlsandboys, respectively.

Biochemicalandbloodpressureassessments

Bloodcollectionwasperformed aftera12-hfastingatthe ClinicalAnalysisLaboratoryoftheHealthDivisionof Univer-sidadeFederaldeVic¸osa(UFV).Thesampleswerecollected byvenipunctureinappropriatetubes,centrifugedfor15min at 3500rpm, and then separated into 1.5mL-Eppendorf tubesandstoredinanultra-freezerat−80◦C.

Thelipidproﬁle(totalcholesterol,high-density lipopro-tein[HDL-c], low-density lipoprotein [LDL-c], and triglyc-erides) and fasting glycemia were determined by the colorimetricenzymaticmethodusingacommercialBioclin® kit(BeloHorizonte,Brazil)andmeasured in anautomatic analyzerdevice (BS-200Mindray®, Nanshan,China)at the

(4)

Laboratory of Clinical Analyses (LAC) of the Department of Nutrition and Health of UFV. The classiﬁcation of the lipidproﬁle components wasperformed according to the updatedBrazilianDyslipidemiaandAtherosclerosis Preven-tionGuidelines.13

Fasting insulin was determined utilizing the chemilu-minescent immunoassay method at the Diagnósticos do Brasillaboratoryandquantiﬁedbytheinsulinassay(Elecsys Insulin®,Roche,Basel,Switzerland)withadetectionlimit of0.200---1000␮U/mL.AHOMA-IRvalue≥3.1614_was_used_to

classifyinsulinresistance.

A digital electronic device wasused tomeasure blood pressureinthearm,withautomaticairinﬂationand deﬂa-tion,and cuffsizethatwasappropriateforthepatient.15

Bloodpressurewas assessedconsidering gender, age,and heightpercentile,alsoaccording totheclassiﬁcation rec-ommendedbytheBrazilianSocietyofCardiology.15 Statisticalanalyses

StatisticalanalyseswereperformedusingSPSSforWindows, version 23.0 (SPSS Inc., Chicago, IL, United States). The Kolmogorov---Smirnovnormalitytestwasusedtodetermine whether the continuous quantitative variables hada nor-maldistribution.Thesigniﬁcancelevelwassetat5%forall hypothesistests.

Descriptiveanalysesofthedatawerecarriedoutthrough relativefrequencies,means,(standarddeviation)or medi-ans(interquartilerange).Forthe comparisonofmeans or medians between two independent samples, Student’s

t-testortheMann---Whitneytestwasperformed,respectively. TheassociationbetweentheNC(independentvariable)and androidfat (dependentvariable)wasassessedbymultiple linearregression.

The adjustment variables that constituted the ﬁnal modelwere:gender, schooltype(publicorprivate),total cholesterol,HDL-c, HOMA-IR,and maternal BMI.The ﬁnal model was evaluated according to the assumptions of linearity,homoscedasticity, normality,independence, mul-ticollinearity,andstandardizedresiduals.

Thereceiveroperatingcharacteristic(ROC)curves deter-mined the ability of NC to identify excess android fat. Additionally,NCcutoffswithbettervaluesofsensitivity(S) andspeciﬁcity(Sp)wereestimated,aswellaspositive(PPV) andnegative(NPV)predictivevalues.

Ethicalaspects

This study wascarried out in accordancewith the guide-linesestablishedintheDeclarationofHelsinkiandapproved bytheHumanResearchEthicsCommitteeof Universidade FederaldeVic¸osa(OpinionNo.663.171/2014). Itwasalso approved by the Municipal Secretariat of Education, the RegionalSuperintendentofEducation,andschoolprincipals.

Results

Theﬁnalsampleconsistedof376children,andtherewasa lossof2.1%oftheparticipants,astheydidnotcompleteall theresearchphases(theydidnothavebloodcollectedfor

biochemicaltests).Basedontheproposedcutoffpointsin theROCcurves,itwasobserved that20.2%(n=76)ofthe childrenhadelevatedNC(Table1).

In the general sample, the NC and the percentage of android fat showed median values of 27.0cm and 14.9%,respectively.Higherpercentagesofandroidfatwere observed in female children, those from private schools, withexcessweight,undesirablelipidproﬁles,insulin resis-tance,highbloodpressure,andmaternalexcessweight.The sameresultswerefoundforthehighestNCvalues,except forschooltype,totalcholesterol,andLDL-cconcentrations (Table1).

Intheﬁnalmultiplelinearregressionmodel, apositive associationwasobservedbetweenNCandandroidfat per-centage.An increase of1cminNC wasassociated witha 2.94% increase in android fat, when adjusted for gender, schooltype,totalcholesterol,HDL-c,HOMA-IR,and mater-nalBMI(Table2).

TheanalysisoftheROCcurveshowedthatNCwasable toidentifyexcessandroidfatingirls(AUC:0.909,95%CI: 0.859,0.945)andboys(AUC:0.938,95%CI:0.892,0.968). The proposedcutoffpointsfor female(28.3cm)andmale individuals(28.6cm)showedsatisfactorySvalues(female: 76.1%,male: 85.2%),Sp(female:92.7%,male: 87.6%)and predictive values (female: PPV, 76.1%; NPV, 92.7%; male: PPV,56.4%;NPV,97.1%;Fig.1).

Discussion

To date,thisis theﬁrstBrazilianstudy carriedoutwitha representative sample of children that demonstrated the ability of NCtoidentifyexcess androidfat. Moreover,NC showedadirectassociationwithandroidfat,regardlessof gender,schooltype,serumtotalcholesterol,HDL-c, HOMA-IR,andmaternalBMI.Itshouldbenotedthattheproposed cutoffvaluesshowedsatisfactoryS,Sp,andpredictive val-ues,makingitpossibletouseNCinepidemiologicalstudies, cardiometabolicriskscreenings,andambulatorycare.

TheuseofNCinadultshasbeenextensivelystudied,with a direct association being observed with visceral adipose tissueandHOMA-IRin obeseindividuals.16 _Furthermore,_a

population-basedstudywithBrazilianadultshasshownthat NCcanbeusedtoidentifymetabolicsyndromeandinsulin resistance.17

Inthissense,theuseoftheNCinthepediatricpopulation becomesimportant,andmoststudiesinthispopulationhave compareditwithdoublyindirectmethodsusedtoestimate adiposity,suchasBMI,18 _waist_{circumference,}19 _and

waist-to-heightratio.20

In individuals aged 9---13 years from Greece, NC was directlyassociatedwithserumtriglyceridesandinsulin lev-els,HOMA-IR, andblood pressure;and inverselytoserum levelsofHDL-c,quantitativeinsulinsensitivitycheckindex (QUICKI),andfastingglucose-insulinratio(FGIR).Thesame study demonstrated that other anthropometric measures usedinchildrenandadolescents,suchasBMIZ-score,waist and hip circumference, waist-to-height ratio, and waist-to-hipratio,showedassociationswithcardiometabolicrisk markerssimilartoNC.21

NCis considered an alternativetoidentifycentral adi-posity, as it has some advantages in relation to waist

(5)

Table1 Values ofneckcircumference andandroid fat according to sociodemographic,anthropometric, biochemical, and clinicalvariables.Vic¸osa,MG,Brazil2015.

Variable n(%) Neckcircumference(cm) Androidfat(%)

Total 376(100.0) 27.0(26.0;28.5) 14.9(7.1;26.4) Gender Female 196(52.1) 26.7(25.5;28.1) 17.1(8.6;20.0) Male 180(47.9) 27.4(26.4;28.5) 9.7(5.9;22.7) p-value 0.001d _<0.001d Age 8years 182(48.4) 26.6(25.5;28.0) 13.1(6.8;23.9) 9years 194(51.6) 27.5(26.2;28.5) 16.30(7.8;27.8) p-value <0.001d _0.101 Schooltype Public 266(70.7) 27.2(2.0) 12.5(6.9;25.7) Private 110(29.3) 27.4(1.9) 19.3(8.7;27.1) p-value 0.448 0.018d

Percapitaincome

≥R$500.00 191(50.8) 27.4(1.9) 17.0(7.3;27.2) <R$500.00 185(49.2) 27.1(2.0) 12.8(7.1;24.6) p-value 0.085 0.115 Screentime/day >2h 178(47.3) 27.3(26.0;28.5) 15.5(7.3;27.1) ≤2h 198(52.7) 27.0(25.7;28.1) 13.2(7.1;25.6) p-value 0.119 0.377 Excessweight Yes 123(32.7) 29.2(1.6) 31.3(9.7) No 253(67.3) 26.3(1.2) 11.1(6.7) p-value <0.001c _<0.001c Neckcircumferencea Increased 76(20.2) 29.9(29.0;31.0) 32.7(25.3;39.3) Adequate 300(79.8) 26.5(25.5;27.5) 10.5(6.2;19.8) p-value --- <0.001d Androidfatb ≥p85 73(19.4) 29.5(28.5;30.9) 35.9(32.2;43.3) <p85 303(80.6) 26.5(25.5;27.7) 10.5(6.3;18.6) p-value <0.001d

---Maternalexcessweight

Yes 168(57.3) 27.8(2.0) 18.3(8.4;30.8) No 125(42.7) 26.5(1.8) 11.0(6.2;19.9) p-value <0.001c _<0.001d Totalcholesterol ≥170mg/dL 86(22.9) 27.5(1.9) 20.5(7.5;29.4) <170mg/dL 289(77.1) 27.2(1.9) 12.8(7.1;24.1) p-value 0.278 0.011d HDL-c ≤45mg/dL 131(34.9) 27.6(2.1) 18.4(8.0;32.1) >45mg/dL 244(65.1) 27.1(1.8) 12.9(6.9;23.0) p-value 0.006c _0.005d LDL-c ≥110mg/dL 58(15.5) 27.7(1.9) 22.0(12.8) <110mg/dL 316(84.5) 27.2(1.9) 16.8(12.0) p-value 0.064 0.003c

(6)

Table1(Continued)

Variable n(%) Neckcircumference(cm) Androidfat(%)

Triglycerides ≥75mg/dL 176(46.9) 27.7(2.1) 19.9(8.2;30.7) <75mg/dL 199(53.1) 26.9(1.7) 10.6(6.5;19.8) p-value <0.001c _<0.001d HOMA-IR ≥3.16 9(2.4) 30.1(1.1) 35.6(13.2) <3.16 362(97.6) 27.2(1.9) 17.3(11.9) p-value <0.001c _<0.001c Bloodpressure Increased 25(6.7) 29.1(1.8) 29.0(13.6) Adequate 350(93.3) 27.1(1.9) 16.9(11.8) p-value <0.001c _<0.001c

HDL-c,highdensitylipoprotein;LDL-c,lowdensitylipoprotein;HOMA-IR,homeostaticmodelassessmentforinsulinresistance. Datapresentedasmean(standarddeviation)ormedian(interquartilerange).

a_{Classiﬁcation}_according_to_the_cutoff_points_obtained_from_the_receiver_operating_{characteristic}_curves. b _Girls_(p85₌_34.0%);_boys_(p85₌_28.9%).

c _Student’s_t-test._p<0.05 d _{Mann---Whitney}_test._p<0.05

Table2 Associationbetweentheneckcircumference(independentvariable)andthepercentageofandroidfat(dependent variable)inchildren.Vic¸osa,MG,2015.

Variable _ˇ(95%CI) p-value

Neckcircumference(cm) 2.94(2.41;3.47) <0.001 Femalegender 5.31(3.33;7.29) <0.001 Privateschool 2.95(0.75;5.15) 0.009 Totalcholesterol(mg/dL) 0.08(0.04;0.11) <0.001 HDL-c(mg/dL) −0.16(−0.27;−0.05) 0.004 HOMA-IR 3.05(1.78;4.33) <0.001 MaternalBMI(kg/m2₎ _0.29_(0.10;_0.47) _0.002

Multiplelinearregression;95%CI,95%conﬁdenceinterval;HDL-c,highdensitylipoprotein;HOMA-IR,homeostaticmodelassessmentfor insulinresistance;BMI,bodymassindex.

Adjustmentbygender,schooltype,serumconcentrationsoftotalcholesterol,HDL-c,HOMA-IR,andmaternalBMI.

circumference: good inter- and intraobserver agreement; notinﬂuenced by the timeof measurement (pre-prandial andpost-prandialperiods);andeasiertomeasure,sinceit does not require the person being assessed towear light clothing,andthus ismoresociallyacceptable.19 _However,

thereisnointernationalconsensusoncutoffstobeusedin children.

ThereareNCcutoffsthathavebeenproposedbasedon ROC18_curves_and_percentiles22_for_children_and_adolescents

according togender and age in different locations. How-ever,mostofthemarecutoffpointscapableofdetermining overweight/obesitythroughBMI.18,22_Therefore,_the_present

studyproposedNCcutoffpointsthatcanbeusedtoidentify excessandroidfatthroughDXA,whichiscapableof estimat-ingcentraladiposity,amethodconsideredasreference.23

Inacross-sectionalstudywithBrazilianadolescentsfrom Campinas, in the state of São Paulo, the NCcutoffs that weresuggested todetermine insulinresistance were32.0 and30.3cm,forprepubertalgirlsandboys,respectively.19

The cutoff points proposed in a cross-sectional study carried out by Naﬁu et al.18 _to _identify _excess _weight _in

Americanchildrenaged8and9yearswere29.0and30.5cm (boys), 27.9 and 29.3cm(girls), respectively, being close to thevalues found in the present study.Therefore, it is recommended to validate the cutoff points proposed in thisstudy forother populations,aimingtoidentifyexcess centraladiposityandothercardiometabolicriskfactors.

Studies that address the presence of excess central adiposity in the pediatric population are of great rele-vance, since the anthropometric measures that estimate centraladiposityshowedagreaterassociationwiththe car-diometabolicrisk,inrelationtotheBMIinchildren.24 _The

fatintheandroidregioncomprisesthevisceraladipose tis-sue, which is more lipolytic compared to the peripheral regions. Agreater hepatic releaseof LDL-c, triglycerides, andotherfatty acidsintheportal circulation,stimulation ofhepaticgluconeogenesis,andhepaticinsulinuptake inhi-bitioncanbeobserved,causinghyperinsulinemiaandinsulin resistance.25

Consideringthehighprevalenceofexcessweight, unde-sirablelipidproﬁle,andsedentary behaviorin thissample andothersimilarstudies,26,27_it_becomes_necessary_to_attain

(7)

Girls Girls Gender n 196 0.909 (0.859; 0.945) 0.938 (0.892; 0.968) 28.3 76.1 (61.2; 87.4) 92.7 (87.3; 96.3) 76.1 92.7 97.1 56.4 87.6 (81.3; 92.4) 85.2 (66.3; 95.7) 28.6 180

AUC (95% CI) S (95% CI)

1.00 1.00 0.80 0.80 0.60 0.60 0.40 0.40 0.20 0.20 0.00 0.00 0.00 0.00 0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00 Cutoff (cm) Sp (95% CI) PPV NPV 1 - Specificity 1 - Specificity Sensitivity Sensitivity Boys Boys

Figure1 Areasunderthereceiveroperatingcharacteristic(ROC)curve(AUC)andcutoffpointsforneckcircumferencewiththe bestsensitivity,speciﬁcity,andpredictivevaluesforexcessandroidfat,accordingtogender.Vic¸osa,MG,2015.

AUC,area underthecurve;95%CI,95% conﬁdenceinterval;S,sensitivity;Sp,speciﬁcity; PPV, positivepredictivevalue; NPV, negativepredictivevalue.

an early diagnosis of excess centraladiposity inchildren. It is known that all these alterations can persist in adult life,inadditiontobeingrelatedtotheatherogenicprocess and cardiac damage, which may begin as early as during childhood.28,29 _Moreover, _the _evaluation _of _the _maternal

nutritionalstatus is important toidentifythe excess cen-traladiposityinchildren,asthemothersandtheirchildren share genetic, environmental (food availabilityat home), andbehavioral(dietaryandlifestylehabits)factors.30

The positive aspects of this study are the fact that the anthropometricevaluation wasperformed by a single researcher,whichcontributedtoreducemeasurementbias; andthefactthatthesampleconsistedexclusivelyof chil-dren,whichminimizesthepossiblehormonalinfluencesof adolescenceonbodycomposition.TheROCcurve analysis wasnotstratifiedbyagebecauseofthephysiological sim-ilaritybetween8-and9-year-oldchildren.AlthoughDXAis unabletodifferentiatebetweentypesofadiposetissue,the fatintheandroidregionassessedthroughthistestisstrongly correlated to visceral adipose tissue in children. Regard-inglimitations,notethelackofreferencevaluestoclassify androidfatinchildren,whichcompelledtheauthorstouse the85thpercentileofthesampleaccordingtogender. Con-sideringthis,theauthorssuggeststudiesthatproposecutoff pointsfor android fat andNC in allpediatric age groups, aimingtoestablishaninternationalconsensusinthefuture. Inconclusion,NCisameasurementcapableofidentifying excessandroidfatinBrazilianchildren.Theproposed cut-offpointscanbeusedinpopulationstudies,screening,and outpatientcare,providedtheyarevalidatedfor the pedi-atricpopulationtowhichtheywillbeapplied.Thus,NCcan bepartoftheevaluationroutineofchildren’shealthstatus, asitisasimpleandlow-costmeasure,helpingintheearly identificationofexcesscentraladiposity.

Funding

This study was funded by Conselho Nacional de

Desen-volvimento Cientíﬁco e Tecnológico (CNPq) (process no.:

478910/2013-4).

Conﬂicts

of

interest

Theauthorsdeclarenoconﬂictsofinterest.

Acknowledgements

The authors would like to thank the Conselho Nacional

de Desenvolvimento Cientíﬁco e Tecnológico (CNPq) for

their financial support, Coordenação de Aperfeiçoamento

dePessoalde NívelSuperior (CAPES)for the study grant,

Quibasa/BioClinfor the donationof biochemicalkits, and thechildrenandtheirparentsfortheirparticipationinthe study.

References

1.NCDRiskFactorCollaboration(NCD-RisC).Worldwidetrendsin body-massindex,underweight,overweight,andobesity from 1975to2016:apooledanalysisof2416population-based mea-surementstudiesin128.9millionchildren,adolescents,and adults.Lancet.2017;390:2627---42.

2.InstitutoBrasileirodeGeografiaeEstatística(IBGE).Pesquisa deOrçamentosFamiliares2008---2009.Antropometriaeestado nutricionaldecrianças,adolescenteseadultosnoBrasil.Riode Janeiro:IBGE;2010.

3.ChiarelliF,MarcovecchioML.Insulinresistanceandobesityin childhood.EurJEndocrinol.2008;159:S67---74.

(8)

4.MazessRB,BardenHS,HansonJ.Dual-energyX-ray absorptiom-etryfortotalbodyand regionalbone-mineraland soft-tissue composition.AmJClinNutr.1990;51:1106---12.

5.Naﬁu OO,ZepedaA, CurcioC,PrasadY. Associationofneck circumferenceandobesitystatuswithelevatedbloodpressure inchildren.JHumHypertens.2014;28:263---8.

6.FerrettiRL,CintraIP,PassosMA,FerrariGL,FisbergM.Elevated neckcircumferenceandassociatedfactorsinadolescents.BMC PublicHealth.2015;15:208.

7.InstitutoBrasileirodeGeograﬁaeEstatística(IBGE).Cidades@ --- Minas Gerais: Vic¸osa. Brasil: IBGE; 2018. Available from: http://cod.ibge.gov.br/690[cited14.11.18].

8.Albuquerque FM, Filgueiras MS, Rocha NP, Castro APP, Mila-gresLC,PessoaMC,etal.Associaçãodasconcentraçõesséricas dezincocomhipercolesterolemiaeresistênciaàinsulinaem criançasbrasileiras.CadSaudePublica.2018;34:e00175016.

9.NovaesJF,PrioreSE,FranceschiniSCC,LamounierJA.Doesthe bodymassindexreﬂectcardiovascularriskfactorsinBrazilian children?JTropPediatr.2013;59:43---8.

10.American Academy of Pediatrics. Children, adolescents, and themedia.CouncilonCommunicationsandMedia.Pediatrics. 2013;132:958---61.

11.World Health Organization. Growth reference data for 5---19 years.Geneva:WHO;2007.

12.World Health Organization. Obesity: preventing and manag-ing theglobal epidemic (WHO technical report series, 894). Geneva:WorldHealthOrganization;2000.

13.FaludiAA, Izar MC,Saraiva JF, ChacraAP,Bianco HT, Afiune Neto A, et al. Atualização da diretriz brasileira de dislipi-demiaseprevençãodaaterosclerose---2017.ArqBrasCardiol. 2017;109:1---76.

14.Keskin M, Kurtoglu S, Kendirci M, Atabek ME, Yazuci C. Homeostasis model assessment is morereliable than fasting glucose/insulinratioandquantitativeinsulinsensitivitycheck indexforassessinginsulinresistanceamongobesechildrenand adolescents.Pediatrics.2005;115:e500---3.

15.MalachiasMV,SouzaWK,PlavnikFL,RodriguesCI,BrandãoAA, NevesMF,etal.7aDiretrizbrasileiradehipertensãoarterial. ArqBrasCardiol.2016;107:1---83.

16.YangL,SamarasingheYP,KaneP,AmielSA,AylwinSJ.Visceral adiposityiscloselycorrelatedwithneckcircumferenceand rep-resentsasigniﬁcantindicatorofinsulinresistanceinWHOgrade IIIobesity.ClinEndocrinol(Oxf).2010;73:197---200.

17.Stabe C, Vasques AC, Lima MM, Tambascia MA, Pareja JC, YamanakaA,etal.Neckcircumferenceasasimpletoolfor iden-tifyingthemetabolicsyndromeandinsulinresistance:results fromtheBrazilianMetabolicSyndromeStudy.ClinEndocrinol (Oxf).2013;78:874---81.

18.NaﬁuOO,BurkeC,LeeJ,Voepel-LewisT,MalviyaS,TremperKK. Neckcircumferenceasascreeningmeasureforidentifying chil-drenwithhighbodymassindex.Pediatrics.2010;126:e306---10.

19.SilvaCC,Zambon MP, VasquesAC,Rodrigues AM, CamiloDF, AntonioMA,etal.Neckcircumferenceasanew anthropomet-ricindicatorforpredictionofinsulinresistanceandcomponents ofmetabolicsyndromeinadolescents:BrazilianMetabolic Syn-dromeStudy.RevPaulPediatr.2014;32:221---9.

20.KelishadiR,DjalaliniaS,MotlaghME,RahimiA,BahreynianM, AreﬁradT,etal.Associationofneckcircumferencewith gen-eraland abdominalobesity inchildrenand adolescents: the weightdisorders survey ofthe CASPIAN-IV study.BMJ Open. 2016;6:e011794.

21.AndroutsosO,GrammatikakiE,MoschonisG,Roma-Giannikou E,ChrousosGP,ManiosY,etal.Neckcircumference:auseful screeningtoolofcardiovascularriskinchildren.PediatrObes. 2012;7:187---95.

22.CoutinhoCA, Longui CA,Monte O, Conde W, Kochi C. Mea-surementofneckcircumferenceanditscorrelationwithbody compositioninasampleofstudentsinSãoPaulo,Brazil.Horm ResPaediatr.2014;82:179---86.

23.Tompuri TT, Lakka TA, Hakulinen M, Lindi V, Laaksonen DE, Kilpeläinen TO, et al. Assessment of body composition by dual-energyX-rayabsorptiometry,bioimpedanceanalysisand anthropometricsinchildren:thePhysicalActivityandNutrition inChildrenstudy.ClinPhysiolFunctImaging.2015;35:21---33.

24.SavvaSC,TornaritisM,SavvaME,KouridesY,PanagiA,Silikiotou N, et al. Waist circumference and waist-to-height ratio are betterpredictorsofcardiovasculardiseaseriskfactorsin chil-dren than body massindex. IntJ Obes RelatMetab Disord. 2000;24:1453---8.

25.BjörntorpP.Abdominalfatdistributionandthemetabolic syn-drome.JCardiovascPharmacol.1992;20:526---8.

26.KitBK,KuklinaE,CarrollMD,OstchegaY,FreedmanDS,Ogden CL.Prevalenceofandtrendsindyslipidemiaandbloodpressure amongUSchildrenandadolescents,1999---2012.JAMAPediatr. 2015;169:272---9.

27.Norman GJ, Carlson JA,Patrick K, Kolodziejczyk JK,Godino JG,Huang J, etal. Sedentarybehavior and cardiometabolic health associations in obese 11---13-year olds. Child Obes. 2017;13:425---32.

28.Berenson GS, Srinivasan SR, Barshop R, Xu J, Chen W. Adi-posity and cardiovascular risk factor variables in childhood are associated with premature death from coronary heart disease in adults: the BogalusaHeart Study. Am J MedSci. 2016;352:448---54.

29.PieruzziF,AntoliniL,SalernoFR,GiussaniM,BrambillaP, Gal-biati S, et al. The role of bloodpressure, bodyweight and fatdistributiononleftventricularmass,diastolicfunctionand cardiacgeometryinchildren.JHypertens.2015;33:1182---92.

30.Santiago-TorresM,AdamsAK,CarrelAL,LaRoweTL,Schoeller DA.Homefoodavailability,parentaldietaryintake,and famil-ialeating habitsinﬂuence thediet qualityofurbanHispanic children.ChildObes.2014;10:408---15.