www.jped.com.br
ORIGINAL
ARTICLE
Lower
waist
circumference
in
mildly-stunted
adolescents
is
associated
with
elevated
insulin
concentration
夽
Ana
Paula
Grotti
Clemente
a,∗,
Carla
Danusa
da
Luz
Santos
b,
Vinicius
J.B.
Martins
c,
Maria
Paula
Albuquerque
c,
Mariana
B.
Fachim
d,
Ana
Lydia
Sawaya
eaUniversidadeFederaldeAlagoas(UFAL),Maceió,AL,Brazil
bCiênciasAplicadasàPediatria,UniversidadeFederaldeSãoPaulo(UNIFESP),SãoPaulo,SP,Brazil
cProgramadePós-Graduac¸ãoemEndocrinologiaClínica,UniversidadeFederaldeSãoPaulo(UNIFESP),SãoPaulo,SP,Brazil dUniversidadeFederaldeSãoPaulo(UNIFESP),SãoPaulo,SP,Brazil
eDepartamentodeFisiologia,UniversidadeFederaldeSãoPaulo(UNIFESP),SãoPaulo,SP,Brazil
Received25September2013;accepted3January2014 Availableonline25June2014
KEYWORDS
Waistcircumference; Height;
Insulin
Abstract
Objective: Augmentedwaistcircumference(WC)isassociatedwithnon-communicablediseases andcouldrepresentavaluablemarkerinscreeningformetabolicdysfunctionsinsubjectswith insufficientlineargrowth.Theobjectiveofthepresentstudywastodeterminewhether bio-chemicalandhemodynamicparametersandwaistcircumferencevarybetweenmildly-stunted andnon-stuntedadolescentsfromimpoverishedcommunitiesofSãoPaulo,Brazil.
Methods: Thecross-sectionalstudyinvolved206subjects,agedbetween9and19yearsand livinginimpoverishedareasofSãoPaulo,Brazil.Thesamplepopulationwasdividedaccording toheight-for-ageZ-score(HAZ)intostunted(-1>HAZ≥-2)andnon-stunted(HAZ≥-1)groups, andwassub-dividedaccordingtogender.Logisticregressionanalysiswasemployedtocompare individuals with elevated(> 75th percentile) insulin concentrations.The receiver operating
characteristiccurveswereconstructedtodetermineWCcut-offpointsthatcouldbeusedto identifystuntedandnon-stuntedindividualswithelevatedinsulinconcentrations.
Results: WCcut-offpointsof58.25cmand67.2cmallowedforcorrectclassificationof90.7% ofstuntedand88.7%ofnon-stuntedindividualsinthestudiedpopulation.Whilethesensitivity ofthemodelwashighforstuntedandnon-stuntedsubjects(98.8%and97.2%,respectively), thespecificitywasmodest(57.1%and41.2%,respectively).
夽
Pleasecitethisarticleas:ClementeAP,SantosCD,MartinsVJ,AlbuquerqueMP,FachimMB,SawayaAL.Lowerwaistcircumferencein mildly-stuntedadolescentsisassociatedwithelevatedinsulinconcentration.JPediatr(RioJ).2014;90:479---85.
∗Correspondingauthor.
E-mail:anagrotticlemente@gmail.com(A.P.G.Clemente). http://dx.doi.org/10.1016/j.jped.2014.01.015
Conclusion: Theresultspresentedhereinsuggestthatanincreaseinplasmainsulinisoneof theprimarymetabolicmodificationsinstuntedindividuals,andthatthisalterationcouldbe identifiedatalowerWCcut-offpointthaninnon-stuntedcounterparts.
©2014SociedadeBrasileiradePediatria.PublishedbyElsevierEditoraLtda.Allrightsreserved.
PALAVRAS-CHAVE Circunferênciada cintura;
Altura; Insulina
Menorcircunferênciadacinturaemadolescentesdebaixaestaturaleveestá relacionadaàconcentrac¸ãoelevadadeinsulina
Resumo
Objetivo: Acircunferênciadacintura(CC)aumentadaestárelacionadaadoenc¸asnão trans-missíveis e pode representar um indicador valioso no exame de verificac¸ão de disfunc¸ões metabólicasem indivíduoscomcrescimento linearinsuficiente. Oobjetivodeste estudofoi determinarseosparâmetrosbioquímicosehemodinâmicoseacircunferênciadacinturavariam entreadolescentesdebaixaestaturaleveedeestaturanormaldecomunidadespobresdeSão Paulo,Brasil.
Métodos: Oestudotransversalenvolveu206indivíduoscomidadesentre9e19anosquemoram emáreaspobresdeSãoPaulo,Brasil.Apopulac¸ãodaamostrafoidividida,deacordocomo escorezdeestaturaporidade(HAZ),emumgrupodebaixaestatura(-1>HAZ≥-2)eumde estatura normal(HAZ ≥-1), esubdivididade acordocomogênero.A análisederegressão logística foi empregada para comparar indivíduos com concentrac¸ões elevadas de insulina (>75◦percentil).Ascurvasdecaracterísticadeoperac¸ãodoreceptorforamconstruídaspara
determinarospontosdecortedeCCquepoderiamserusadosparaidentificarosindivíduosde baixaestaturaedeestaturanormalcomconcentrac¸õeselevadasdeinsulina.
Resultados: OspontosdecortedeCCde58,25e67,2cmpermitiramaclassificac¸ãocorretade 90,7%deindivíduosdebaixaestaturae88,7%deindivíduosdeestaturanormalnapopulac¸ão estudada.Emboraasensibilidadedomodelofossealtaparaindivíduosdebaixaestaturaede estaturanormal(98,8%e97,2%,respectivamente),aespecificidadefoipequena(57,1%e41,2%, respectivamente).
Conclusão: Osresultadosapresentadosnesteinstrumentosugeremqueumaumentonainsulina plasmáticaéumadasprincipaismodificac¸õesmetabólicasemindivíduosdebaixaestatura,e queessaalterac¸ãopodeseridentificadaemumpontodecortedeCCmenorqueemparesde estaturanormal.
©2014SociedadeBrasileiradePediatria.PublicadoporElsevierEditoraLtda.Todososdireitos reservados.
Introduction
Irrespectiveofbirthweight,childrenandadolescentswith moderate/severe stunting are more prone to increased bodyfat (especially abdominalfat),1---4 diminishedrate of fat oxidation,5 reduced resting and postprandial energy expenditure,1 higher systolic and diastolic arterial pres-sure(SAPandDAP,respectively),6---8andlowerproductionof insulinbythepancreas.9,10 Alterationsintheseparameters arealsoaggravatedbythe presence ofobesity.5 Arecent studyonpre-adolescentsandadolescentswithmildstunting (-2<HAZ<-1)showedsimilarincreaseinSAPincomparison withtheirnon-stuntedpeers.11
In addition, overweight adolescents with mild stunt-ingpresentedsignificantlyhigherconcentrationsofplasma insulin,elevatedglycemia,increasedinsulinresistance,and diminishedpancreatic productionof insulin in comparison withindividualsofnormalbodymassindex(BMI).12Inthese subjects,abdominalfatandwaistcircumference(WC) val-ues were significantly higher in the mildly-stunted group thaninthenon-stunted controlgroup.13 Itis evidentfrom
these findings that individuals with mild stunting present physiological alterations that are very similar to those described earlier for individuals with moderate or severe stunting.
ItiswidelyacceptedthatelevatedWCvaluesarestrongly associated with increased risk of non-communicable dis-eases(NCDs).14---16 The objectiveof thepresent studywas to analyze physiological alterations, including SAP and DAP levels and concentrations of glucose, insulin, high-andlow-densitylipoproteincholesterol(HDL-CandLDL-C, respectively),andtriglycerides,in relationtotheWC val-uesofchildrenandadolescentswithmildstuntingcompared withtheirnon-stuntedcounterparts.
Subjectsandmethods
guardianswhereappropriate,priortothe commencement ofthestudy.
Thiscross-sectionalstudyinvolved206subjects(9to19 yearsofage)whowereattendingschoolsor other institu-tionslocatedinimpoverishedareasofthesouthernareaof thecityofSãoPaulo.Sincethepurposeofthestudywasto detect earlychanges causedby mild stunting,thesample population wasdivided intotwogroups according toHAZ, namely,stunted(HAZ<-1and≥-2)andnon-stunted(HAZ
≥-1)withstandardreferencevaluesbasedontheCenters forDiseaseControlandPrevention(CDC)2000growthcharts fortheUnitedStatesforchildrenandadolescents.17
This study is part of a largerhealth surveythat inves-tigated pre-adolescents and adolescents with stunting, conducted in the city of São Paulo, with approximately 400 individuals. The sample size of this study was calcu-lated using odds ratio (OR), with significance ␣ level of 0.05,power of0.80, andasampling ratioof1:1between thoseexposedandunexposedtothepredictor.AnORinthe sourcepopulationequalto2.5andanexpectedfrequency of statureequal to0.08 amongtheunexposedpopulation wereassumed.Duetohighcostsanddifficultyofobtaining bloodsamples,itwasconsideredthatasample of approx-imately50%ofthesurveypopulationwassufficienttotest thehypothesisofthestudy.Themaximumacceptablebeta errorwas0.20.Subjectswereselectedforthestudy accord-ingtotheirnutritionalstatus.
Prior to the commencement of the study, participants underwent clinical examinations,as well asblood, urine, andparasitologicaltests.Individualsdiagnosedwithgenetic and neurological syndromes, dementia, or cardiovascular, respiratory,ormetabolicdisorderswereexcludedfromthe study,as were those usinganti-inflammatory medications and those with physical limitations. Subjects presenting infectious or parasitic diseases weretreated according to normalprotocolsandsubsequentlyincludedinthestudy.
Socioeconomic and environmental data were obtained fromparentsorlegalguardiansbyapplicationofaspecific questionnaire.Theweightofeachparticipant(wearinglight clothesandwithoutshoes)wasobtainedbysingle measure-mentusingaCountryTechnologiesmodelSD-150platform scale(GaysMills,WI,USA)withacapacityof150kgandan accuracy of 100g.Stature wasassessed using an AlturEx-ata (TBW,São Paulo, Brazil) portable stadiometer with a precision of0.1cm.BMIvalueswere calculatedasweight (kg)/height squared (m2). For the determination of waist circumference,subjectswereplacedinastandingposition withtheabdomenandarmsrelaxedalongsidethebody,and a flexiblemeasuring tape (1mm accuracy) washeld hori-zontallyat themidpointbetween thebottomedgeof the lastribandtheiliaccrest.Thewaistcircumferencedeciles werecalculatedfortheentirepopulationandcomparedto thevaluesofstature(cm).
Blood pressure wasmeasured using a standard clinical sphygmomanometer. Subjectswereseatedand allowedto restfor10min,followingbythreepressuremeasurements conducted at 5min intervals. SAP wasdetermined at the onset of the first Korotkoff sound, while DAP was deter-minedafterthedisappearanceofthefifthKorotkoffsound.18 Fastingplasmaglucoseconcentrationsweredeterminedby aBeckman Coulter UniCell DXI800 (CA,USA) spectropho-tometer, while specific insulin (without C peptide) levels
wereassessedusinganenzymeassayandaMedcorpAdvia 2400/Kovalent(RN, Brazil) analyzer.Serumlevels oftotal cholesterol,LDL-C,HDL-C,andtriglyceridesweremeasured usinganAdvia2400/Kovalentanalyzer.Elevated concentra-tionswereconsideredthoseabovethe75thpercentileofthe studiedpopulation.
Calculationsof nutritionalstatuswereperformedusing Epi-Info 2000 (Centers for Disease Control, GA, USA) software.ParticipantsweredistributedbyBMI-for-age per-centilesaccording tostandard referencevalues ofthe US Centers for Disease Control and Prevention 2000 growth charts.17
All participants were examined by a trained physician andclassifiedregardinggenderdevelopmentinaccordance withtherecommendationsofTanner.19Individualswhohad reached breast-stage 2 for females and genitalia-stage 3 formalesaccordingtoWHOcut-offpointswereconsidered pubertal.20StatisticalanalyseswereperformedusingPASW Statisticsversion19(SPSSInc.,Chicago,IL,USA) withthe levelofstatistical significancesetat p<0.05.Meanvalues ofage, height, weight,HAZ, BMI, andWC of the stunted andnon-stuntedgroups,stratifiedaccordingtogender,were compared using the Student’s t-test, and assumptions of homoscedasticitywereverifiedusingtheLevenetest. Anal-ysisofcovariance(ANCOVA)wasusedtoestablishsignificant differencesinplasmainsulin,totalcholesterol,HDL-C, LDL-C, triglycerides, SAP, and DAP values(adjusted according toage andweight)between the stuntedand non-stunted groups.
Logistic regression analysis (forward LR method) was employed to compare individuals with elevated insulin concentrations (> 75th percentile) using the Wald test to determinewhich factors shouldbeemployedaspredictor variables in the final model. Pubertal stage, gender, and WCweredefinedasindependentvariablesintheregression model.Areceiveroperatingcharacteristic(ROC)curvewas constructedinordertoestablishWCcut-offpointsfor indi-vidualsin the stuntedand non-stunted groups that could be used to predict insulin concentrations above the 75th percentile.
Results
The study population comprised206 children and adoles-cents (53.5% males and 46.5% females), most of whom (74.8%)wereclassifiedaspre-pubertal.Thefamiliesofthe majorityoftheparticipantswereconsideredpoor,withan averageof6.0±3.6individualsperhouseholdandamonthly familyincome of US$ 484±328, which is equivalent toa dailypercapita incomeofapproximatelyUS$4±2.7.The prevalenceofilliteracywashigheramongmothers(10.6%) than fathers(5.6%), and 10% of the dwellings were inad-equateshacksconstructed of woodor a mixof woodand brick.
Table1 Anthropometricalandbiochemicalcharacteristicsofthestudiedpopulationandprevalencesofmetabolicalterations observed.
Characteristicsand prevalences
Females Males
Stunteda
(n=35)
Non-stuntedb
(n=67)
pvalue Stunteda
(n=36)
Non-stuntedb
(n=68)
pvalue
Pubertalstagec
Pre-pubertal 26 50 0.83 27 51 0.23
Pubertal 9 17 9 17
Anthropometricalparameters (mean±standarddeviation)d
Age(years) 11.09±2.19 10.49±2.55 0.154 11.21±2.45 10.26±2.50 0.017
Weight(kg) 36.63±11.46 45.85±17.56 0.001 33.45±10.81 41.42±16.65 0.001
Height(cm) 137.13±11.07 145.75±12.00 <0.001 138.25±12.83 144.52±14.95 0.006
Height-for-agez-score -1.50±0.25 0.33±0.72 <0.001 -1.49±0.28 0.18±0.80 <0.001
Bodymassindex(BMI;kg•m−2) 19.02±3.65 20.94±5.36 0.020 17.08±2.88 19.22±4.54 0.001
BMI-for-agepercentile 55.34±30.06 67.98±30.06 0.016 36.07±30.26 61.24±32.49 <0.001
Waistcircumference(WC;cm) 61.82±10.66 66.75±12.69 0.045 61.27±9.71 64.53±9.16 0.081
WC/heightratio 0.45±0.06 0.46±0.07 0.750 0.43±0.04 0.45±0.07 0.242
WC/BMIratio3.32±0.27 3.35±0.32 0.653 3.53±0.28 3.48±0.30 0.404
Biochemicalandhemodynamic parameters
(mean±standarderror)e
Glucose(mg•L−1) 85.36±0.98 86.38±0.74 0.435 89.32±0.92 89.16±0.67 0.895
Insulin(pmol•L−1) 10.70±0.97 7.62±0.73 0.018 7.68±0.62 5.42±0.45 0.006
Totalcholesterol(mg•dL−1) 163.18±4.98 157.47±3.73 0.387 159.01±4.25 154.67±3.10 0.440
HDL-C(mg•dL−1) 57.67±2.00 54.10±1.50 0.180 54.82±1.65 54.04±1.20 0.719
LDL-C(mg•dL−1) 91.92±3.84 89.12±2.88 0.582 91.54±3.31 88.45±2.41 0.480
Triglycerides(mg•dL−1) 75.85±4.84 74.77±3.63 0.866 68.95±4.11 64.35±2.99 0.397
SAP(mmHg) 105.23±2.19 109.72±1.68 0.126 106.85±1.86 106.04±1.42 0.746
DAP(mmHg) 64.37±1.19 63.83±0.82 0.794 64.33±1.51 63.30±1.16 0.617
Prevalenceofmetabolic alterations(%)f
Elevatedglucose (>91mmol•L−1)
3.5 14.6 0.028 7.2 18.1 0.124
Elevatedinsulin(>8U•mL−1) 11.1 14.6 0.463 6.0 10.8 0.735
Elevatedtotalcholesterol (>176.75mmol•L−1)
10.4 63.2 0.416 7.8 12.0 0.912
ElevatedHDL-C (>61mmol•L−1)
11.1 16.7 0.782 9.0 8.4 0.109
ElevatedLDL-C (>104mmol•L−1)
9.7 14.6 0.801 7.2 12.0 0.892
Elevatedtriglycerides (>81.75mmol•L−1)
8.3 19.4 0.209 5.4 11.4 0.445
ElevatedSAP(>117mmHg) 9.0 18.8 0.383 7.2 15.7 0.314
ElevatedDAP(>70mmHg) 9.0 20.1 0.251 9.6 19.9 0.312
HDL-C,high-density lipoproteincholesterol;LDL-C,low-densitylipoproteincholesterol;SAP,systolicarterialpressure;DAP,diastolic arterialpressure.
aHeight-for-agez-score<-1and≥-2. b Height-for-agez-score≥-1.
c Significantdifferencesbetweenthegroupsweredeterminedusingthe
2test(p<0.05).
d SignificantdifferencesbetweengroupsweredeterminedusingStudent’st-test(p<0.05).
Table2 Correlationsbetweenwaistcircumference(WC),andbiochemicalandhemodynamicparameters.
Correlations Females Males
Stunteda(n=35) Non-stuntedb(n=67) Stunteda(n=36) Non-stuntedb(n=68)
Spearman test(r)
p-value Spearman
test(r)
p-value Spearman
test(r)
p-value Spearman
test(r)
p-value
InsulincxWC 0.507 0.002 0.684 <0.001 0.565 <0.001 0.602 <0.001
GlucosedxWC 0.375 0.027 0.262 0.032 0.270 0.111 0.289 0.017
TotalcholesterolxWC 0.024 0.890 0.133 0.282 0.154 0.371 0.057 0.647
HDL-CexWC 0.045 0.797 0.266 0.030 0.388 0.019 0.155 0.208
LDL-CfxWC 0.014 0.935 0.028 0.824 0.084 0.624 0.122 0.322
TriglyceridesgxWC 0.060 0.733 0.401 0.001 0.401 0.015 0.169 0.167
SAPhxWC 0.212 0.207 0.101 0.390 0.418 0.008 0.231 0.049
DAPixWC 0.388 0.018 0.287 0.013 0.532 <0.001 0.450 <0.001
a Height-foragezscore<-1and≥-2. b Height-foragezscore≥-1. c Insulin(pmol•L−1). d Glucose(mg•L−1).
eHDL-C,high-densitylipoproteincholesterol(mg•dL−1). f LDL-C,low-densitylipoproteincholesterol(mg•dL−1). g Triglycerides(mg•dL−1).
h SAP,systolicarterialpressure(mmHg). i DAP,diastolicarterialpressure(mmHg).
stunted females, but there was no significant difference between the two groups with respect to the WC/height ratioandWC/BMIratio.Insulinconcentrationswere signif-icantlyhigherinstuntedmalesandfemalesincomparison with their non-stunted counterparts, but the prevalences of elevated (> 75th percentile) plasma insulin concentra-tionsweresimilarinallgroups.Theprevalencesofelevated (>75th percentile)totalserumcholesterol, HDL-C,LDL-C, triglycerides, SAP, and DAP within the stunted and non-stuntedgroups of bothgenders weresimilar.There wasa significantlyhigherprevalenceofincreasedplasmaglucose amongnon-stuntedfemalesincomparisonwiththeirstunted counterparts.
ThereweresignificantpositivecorrelationsbetweenWC andthevariablesinsulin,glucose,andDAPforstuntedand non-stunted females, and between WC and the variables HDL-Candtriglyceridesfor non-stuntedfemales(Table2). Significantpositivecorrelationswerealsoobservedbetween WCandthevariablesinsulin,SAP,andDAPforstuntedand non-stuntedmales,whereasWCwascorrelatedwithglucose formalesonly.
Fig. 1 shows the prevalences of elevated (> 75th percentile)insulin concentrations in stunted (A)and non-stunted(B)individualsdistributedaccordingtoWCdeciles. Within thestunted group,elevated insulin concentrations wereobservedwithprevalencesof10%andhigherfromthe secondWCdecileonwards,whileinthenon-stuntedgroup, increasesininsulin concentrationscouldonlybedetected fromthefifthWCdecileonwards.Approximately70%to80% ofstuntedindividualsshowedraisedinsulinconcentrations fromtheeighthWCdecileandabove.Incontrast,the preva-lenceofelevatedinsulinconcentrationsamongnon-stunted individualsreachedamaximumof60%onlyinthetenthWC decile.
In the logistic regression analysis, the predictor-dependent variable was represented by elevated insulin
level,whiletheindependentvariablesweregender,stageof pubertaldevelopment,andWC.Accordingtotheregression model,thefirsttwoindependentvariablesmentionedhad noinfluenceonthelikelihoodofpresentingelevatedinsulin concentrations, as indicated by p values for the stunted and non-stunted groups of 0.518 and 0.491, respectively, for gender; and of 0.541 and0.752, respectively, for the pubertalstage.Incontrast,thevariableWCwassignificantly associated with elevated insulin concentrations in both groups (stunted group - B=0.271, X2
Wald (1)=6.239, p=0.012;non-stuntedgroup-B=0.119,X2
Wald (1)=14.386, p<0.001).Thus,theriskofstuntedandnon-stunted individ-ualspresentingelevated insulinconcentrations wasraised by 31.1% and 12.7%, respectively, for each additional increaseinWCof1cm.
Analysis of the ROC curves revealed cut-off points of 58.25cm for stunted individuals and 67.20cm for non-stuntedsubjects(Fig.2).SincetheareaundertheROCcurve forthestuntedgroupwas84.2(p=0.001)andthatfor non-stuntedgroupwas85.9%(p<0.001),thedifferencebetween therespectiveWCcut-offpointswasstatisticallysignificant.
Discussion
A
B
0% 20% 40% 60% 80% 100%
1 (14)
2 (10)
3 (7)
4 (4)
5 (9)
6 (5)
7 (5)
8 (4)
9 (6)
10 (7)
0% 20% 40% 60% 80% 100%
1 (9)
2 (10)
3 (19)
4 (10)
5 (16)
6 (14)
7 (15)
8 (15)
9 (15)
10 (12)
WC deciles
WC deciles
Figure1 Distributionofstunted(A)andnon-stunted(B) indi-vidualsaccordingtowaistcircumference(WC)decilesandtheir respectiveprevalencesofelevatedinsulinconcentrations:() >75th percentile; ()≤75th percentile. The WC deciles
cor-respondtothefollowingabsolutevaluesofstatureofstudied population:1)53cm;2)55.90cm;3)57.50cm;4)59.50cm;5) 62cm;6)65cm;7)68cm;8)71cm;9)76.74cm.
The numbers between parenthesesrepresent the number of individualsofthesampleineachdecileofWC.
nutritionalriskandfor monitoringvulnerablechildrenand adolescentsinordertoensuretheirlong-termhealth.
Thepresentresultsareinlinewithpreviousstudies.11---13 Theyconfirm the original hypothesis that individuals with mild stunting show metabolic alterations similar to mod-eratelyor severely undernourishedindividuals. They have morefatinthetrunk,asdescribedpreviously,comparedto moderately/severelystuntedadolescents.3
Itiswellknownthatindividualswithincreased abdom-inal fat are more susceptible to metabolic dysfunctions, andthat such alterations develop during childhood.15,16,23 On this basis, the identification of threshold values for WCin childrenand adolescentsis a crucialcomponent in developing a strategy for the prevention of NCDs in sub-jectswithinsufficientlineargrowth,andmostparticularly in developing countries where the prevalence of stunting is high.24 A few studies, however, have focused on body composition,andabdominal adiposity inadolescents with poorlineargrowth,andthepossibleassociationwithriskof metabolicalterations.3,13 The use ofsuch markerin clini-calpracticewouldbeveryvaluable,sinceanthropometrical measurements are inexpensive and straightforward. The implementationofpreventive measures amongvulnerable
1.0
A
B
1.0 0.8
0.8 0.6
0.6 0.4
0.4
1 - specificity
Sensitivity
0.2
0.2
Figure 2 Receiver-operating characteristic (ROC)curve for theidentificationofmetabolicalterationsbasedonthewaist circumference(WC)ofnon-stunted(—)andstunted(····) chil-dren andadolescents. The area under theROC curve ofthe non-stuntedgroupwas0.85(95%CI=0.79---0.92),whilethatof thestuntedgroupwas0.84(95%CI=0.75---0.93).TheWCcut-off pointforthenon-stuntedgroupwas67.2cm(A),whilethatfor thestuntedgroupwas58.25cm(B).
populationswouldensureabetterqualityoflifeandwould servetominimizefuturespendingbyhealthcaresystems. Forthisreason,the present studyexamined the hypothe-sisthat,incomparisonwiththeirnon-stuntedcounterparts, subjects withmild stuntingwouldbe at riskof NCDs at a lowerWCvalue.Insupportofthishypothesis,theseresults revealedthatindividualswithmildstuntingpresented ele-vatedinsulin concentration at lowerWCvaluescompared withnon-stuntedsubjects.
Nodifferenceswereobservedbetweenstuntedand non-stuntedpubertalstages.Themeanvaluesofweight,height, andBMIof stuntedsubjectswerelowerthanthoseof the non-stunted group, although the WC/height and WC/BMI ratiosweresimilarinbothgroups, independentofgender. Insulin concentrationsof stuntedmales andfemales were higherincomparisonwiththoseoftheirnon-stunted coun-terpartsand,whengroupsweredistributedaccordingtoWC deciles,ahigherprevalenceofelevatedinsulin concentra-tions (> 75th percentile) was observed withinthe stunted group.Inconfirmationofearlierreports,14,16,23itwasshown that WC measurementsprovide a simple and appropriate marker of elevated insulin concentrations among stunted children and adolescents. The 75th percentile cut-off for increasedinsulinwaschoseninordertoidentifyearly alter-ationsinhormoneconcentrations,sincethisvariableshowed thebestassociationwithWCingroupsbasedonstatureand gender.
WCcut-offpointsidentified(58.25cmforthestuntedgroup and 67.20cm for the non-stunted group) were based on themostappropriatecombinationofthetwocriteria,and allowedthecorrectclassificationof90.7%ofstunted indi-vidualsand88.7%ofthenon-stuntedgroup.Thestatistical power of theadjusted model employedwas excellent,as demonstratedbythehighsensitivityvaluesobtainedforthe stuntedandnon-stunted groups(98.8%and97.2%, respec-tively).Incontrast,thespecificityoftheadjustedmodelwas onlymodest,i.e.,57.1%forthestuntedgroupand41.2%for thenon-stuntedgroup.
In conclusion, it was determined that the WC cut-off pointforthepredictionofmetabolicalterationsinchildren andadolescentswithmild stuntingwassmaller compared withnon-stuntedcounterparts.Theresultsalsosuggestthat anincreaseinplasmainsulinisoneoftheprimarymetabolic deviationsthatoccurinstuntedindividuals.Special atten-tionmustbeprovidedtoindividualswithmildstunting.
Funding
ThepresentstudywasupportedbyFundac¸ãodeAmparoa PesquisadoEstadodeSãoPaulo(FAPESP;Proc.06/56218-0).
Conflicts
of
interest
Theauthorsdeclarenoconflictsofinterest.
Acknowledgments
The authorsthanktheFundac¸ãodeAmparo aPesquisado EstadodeSãoPaulo(FAPESP;Proc.06/56218-0)forfinancial support.
References
1.HoffmanDJ,SawayaAL,CowardWA,WrightA,MartinsPA,de NascimentoC,etal.Energyexpenditureofstuntedand non-stuntedboysandgirlslivingintheshantytownsofSãoPaulo. BrazilAmJClinNutr.2000;72:1025---31.
2.MartinsPA,HoffmanDJ,FernandesMT,NascimentoCR,Roberts SB,SessoR,etal.Stuntedchildrengainlessleanbodymassand morefatmassthantheirnon-stuntedcounterparts:a prospec-tivestudy.BrJNutr.2004;92:819---25.
3.Hoffman DJ, Martins PA, Roberts SB, Sawaya AL. Body fat distributioninstuntedcomparedwithnormal-heightchildren fromtheshantytownsofSãoPaulo.Brazil.Nutrition.2007;23: 640---6.
4.KrugerHS,PretoriusR,SchutteAE.Stunting,adiposity,and low-gradeinflammationinAfricanadolescentsfromatownshiphigh school.Nutrition.2010;26:90---9.
5.HoffmanDJ, Sawaya AL, Verreschi I,Tucker KL, Roberts SB. Why are nutritionally stunted children at increased risk of obesity?Studiesofmetabolicrateandfatoxidationin shanty-townchildrenfromSãoPaulo,Brazil.AmJClinNutr.2000;72: 702---7.
6.FernandesMT,SessoR,MartinsPA,SawayaAL.Increasedblood pressureinadolescentsoflowsocioeconomicstatuswithshort stature.PediatrNephrol.2003;18:435---9.
7.SessoR,BarretoGP,NevesJ,SawayaAL.Malnutritionis associ-atedwithincreasedbloodpressureinchildhood.NephronClin Pract.2004;97:c61---6.
8.FebbaA, SessoR,BarretoGP,Liboni CS,FrancoMC,Casarini DE. Stunting growth: association of the blood pressure levels and ACE activity in early childhood. Pediatr Nephrol. 2009;24:379---86.
9.MartinsPA,SawayaAL.Evidenceforimpairedinsulinproduction andhighersensitivityinstuntedchildrenlivinginslums.BrJ Nutr.2006;95:996---1001.
10.González-Barranco J, Ríos-Torres JM, Castillo-Martínez L, López-Alvarenga JC, Aguilar-Salinas CA, Bouchard C, et al. Effect of malnutrition during the first year of life on adult plasma insulin and glucose tolerance. Metabolism. 2003;52:1005---11.
11.ClementeAP,SantosCD,SilvaAA,MartinsVJ,MarchesanoAC, FernandesMB, et al.Mild stunting is associatedwith higher blood pressure in overweight adolescents. Arq Bras Cardiol. 2012;98:6---12.
12.da Luz Santos CD, Clemente AP, Martins VJ, Albuquerque MP, Sawaya AL. Adolescents with mild stunting show alter-ations in glucose and insulin metabolism. J Nutr Metab. 2010;2010:943070.
13.ClementeAP,SantosCD,MartinsVJ,Benedito-SilvaAA, Albu-querqueMP,SawayaAL.Mildstuntingisassociatedwithhigher bodyfat:studyofalow-incomepopulation.JPediatr(RioJ). 2011;87:138---44.
14.KotlyarevskaK,WolfgramP,LeeJM.Iswaistcircumferencea betterpredictorofinsulinresistancethanbodymassindexin U.S.adolescents?JAdolescHealth.2011;49:330---3.
15.GlässerN,ZellnerK, Kromeyer-Hauschild K.Validityof body massindexandwaistcircumferencetodetectexcessfatmass inchildrenaged7-14years.EurJClinNutr.2011;65:151---9. 16.FreedmanDS,SerdulaMK,SrinivasanSR,BerensonGS.Relation
ofcircumferencesandskinfoldthicknessestolipidandinsulin concentrationsinchildrenandadolescents:theBogalusaHeart Study.AmJClinNutr.1999;69:308---17.
17.OgdenCL,KuczmarskiRJ,FlegalKM,MeiZ,GuoS,WeiR,etal. CentersforDiseaseControlandPrevention2000growthcharts fortheUnitedStates:improvementstothe1977National Cen-terforHealthStatisticsversion.Pediatrics.2002;109:45---60. 18.NationalHighBloodPressureEducationProgramWorkingGroup
on High Blood Pressure in Children and, Adolescents. The fourth reportonthediagnosis,evaluation, and treatmentof high bloodpressure inchildrenand, adolescents.Pediatrics. 2004;114:555---76.
19.TannerJM.Growthatadolescence.2nded.Oxford:Blackwell; 1962.
20.World Health Organization (WHO). Physical status: the use andinterpretationofanthropometry.ReportofaWHO.Expert Committee.TechnicalReport Series,No.854.Geneva: WHO; 1995.
21.deOnisM,OnyangoAW,BorghiE,SiyamA,NishidaC,Siekmann J.DevelopmentofaWHOgrowthreferenceforschool-aged chil-drenandadolescents.BullWorldHealthOrgan.2007;85:660---7. 22.WorldHealthOrganization,(WHO).Measuringchangein nutri-tional,status.,Guidelinesforassessingthenutritionalimpact of supplementary feeding programs for vulnerable, groups. Geneva:WHO;1983.
23.FekaduS,YigzawM,AlemuS,DessieA,FieldhouseH,GirmaT, etal.Insulin-requiringdiabetesinEthiopia:associationswith poverty, early undernutrition and anthropometric dispropor-tion.EurJClinNutr.2010;64:1192---8.