Rev. Bras. Hematol. Hemoter. vol.38 número3

w w w . r b h h . o r g

Revista

Brasileira

de

Hematologia

e

Hemoterapia

Brazilian

Journal

of

Hematology

and

Hemotherapy

Original

article

A

new

index

to

discriminate

between

iron

deficiency

anemia

and

thalassemia

trait

Januária

F.

Matos

,

Luci

M.S.

Dusse

,

Karina

B.G.

Borges

,

Ricardo

L.V.

de

Castro

,

Wendel

Coura-Vital

_,

_Maria

_das

_G.

_Carvalho

a_{InstitutoFederaldeMinasGerais(IFMG),OuroPreto,MG,Brazil}

b_{UniversidadeFederaldeMinasGerais(UFMG),BeloHorizonte,MG,Brazil}

c_{HospitalOdilonBehrens,BeloHorizonte,MG,Brazil}

d_{UniversidadeFederaldeOuroPreto(UFOP),OuroPreto,MG,Brazil}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received12March2016 Accepted19May2016 Availableonline22June2016

Keywords:

Newindex

Irondeficiencyanemia Thalassemiatrait

a

b

s

t

r

a

c

t

Background:Themostcommonmicrocyticandhypochromicanemiasareirondeficiency

anemiaandthalassemiatrait.Severalindicestodiscriminateirondeficiencyanemiafrom thalassemiatraithavebeenproposedassimplediagnostictools.However,someofthebest discriminativeindicesuseparametersintheformulasthatareonlymeasuredinmodern countersandarenotalwaysavailableinsmalllaboratories.

Thedevelopmentofanindexwithgooddiagnosticaccuracybasedonlyonparameters derivedfromthebloodcellcountobtainedusingsimplecounterswouldbeusefulinthe clinicalroutine.Thus,theaimofthisstudywastodevelopandvalidateadiscriminative indextodifferentiateirondeficiencyanemiafromthalassemiatrait.

Methods:Todevelopandtovalidatethenewformula,bloodcountdatafrom106(thalassemia

trait:23andirondeficiency:83)and185patients(thalassemiatrait:30andirondeficiency: 155)wereused,respectively.Irondeficiency,␤-thalassemiatraitand␣-thalassemiatraitwere confirmedbygoldstandardtests(lowserumferritinforirondeficiencyanemia,HbA2>3.5%

for␤-thalassemiatraitandusingmolecularbiologyforthe␣-thalassemiatrait).

Results:Thesensitivity,specificity,efficiency,Youden’sIndex,areaunderreceiveroperating

characteristiccurveandKappacoefficientofthenewformula,calledtheMatos&Carvalho Indexwere99.3%,76.7%,95.7%,76.0,0.95and0.83,respectively.

Conclusion:Theperformanceofthisindexwasexcellentwiththeadvantageofbeingsolely

dependentonthemeancorpuscularhemoglobinconcentrationandredbloodcellcount obtainedfromsimpleautomaticcountersandthusmaybeofgreatvalueinunderdeveloped anddevelopingcountries.

©2016Associac¸ ˜aoBrasileiradeHematologia,HemoterapiaeTerapiaCelular.Published byElsevierEditoraLtda.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense

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

∗ _{Correspondingauthorat:InstitutoFederaldeMinasGerais,CampusOuroPreto,RuaPandiáCalógeras,898,Bauxita,35400-000Ouro}

Preto,MG,Brazil.

E-mailaddress:januaria.matos@ifmg.edu.br(J.F.Matos).

http://dx.doi.org/10.1016/j.bjhh.2016.05.011

Introduction

Anemiaaffectsabout800millionchildrenandwomen world-wide.Themostcommoncausesofanemiaareirondeficiency anemia(IDA)andthalassemiatrait(TT).1–3_{AccordingtoWorld}

HealthOrganization(WHO)estimatesin2004,IDAresultedin 273,000deathsandthelossof19.7milliondisability-adjusted lifeyears,accountingfor1.3%ofthe globaltotal,with97% occurringinlow-andmiddle-incomecountries.3,4

Differentialdiagnosisofmicrocyticanemiasisofgreat clin-icalimportancesinceprognosisandtreatmentare distinct. Thefirststeptodiagnosemicrocyticanemiasisbythe micro-scopicanalysisofbloodfilmanddeterminationofredblood cell(RBC)indicesusingcell counters.Consideringthegreat similaritybetweenIDAandTT,complementarylabmethods areneededbesidestheroutinebloodexam.Currently, diag-nosisofIDAisobtainedbyevaluatingtheironmetabolism, includingserumiron,totalserumironbindingcapacityand serumferritinmeasurements.Diagnosisofthe␤-thalassemia trait(␤-TT)isusuallymadebyhemoglobinelectrophoresisand HbA2 levels (>3.5%).5,6 Onthe other hand,diagnosis ofthe

␣-thalassemiatrait(␣-TT)isconfirmedbyinvestigating muta-tionsinthe␣gene.Despitetheirgreatutility,goldstandard testsforthediagnosisofthesemicrocyticandhypochromic anemiasinvolvetime-consumingmethodologies,highcosts andareinaccessibletopoorerpopulations.5–8

Inanattempttosimplifythedifferentialdiagnosisbetween IDA and TT, several indicesusing blood cell count param-eters have been suggested.9–16 _{The formula developed by}

Green&King{[(MCV2₎

×RDW]/(Hb×100)whereMCVismean

corpuscular volume, RDW is RBC distribution width and Hb ishemoglobin}has, accordingto many studies,agood performance.16–18_{However,thisformulaisdependentonthe}

RDW,aparameterthatisnotprovidedbyallautomatic coun-ters.Furthermore,thecurrentlyexistingindexesreportedin theliteraturehavenotbeendevelopedusingmolecular biol-ogytodiagnose␣-thalassemia,whichisnecessarytoexclude concomitantdiseases.

Theaimofthisstudywastodevelopandvalidateanew indexthatdiscriminatesbetweenIDAandTTemployingvery simpleparametersthatareprovidedbyallautomaticblood counters.Thisstudyusedmoleculartechniquestodiagnose TTandexcludethepresenceofthisdisorderinIDApatients.

Methods

Populationevaluatedtodevelopthenewindex

Twogroupsofpatientswereselectedtodevelopthisindex, namelyover18-year-oldindividualswithIDAorTT.For inclu-sionintheIDAgroup,patientshadhemoglobinlevelsbelow 12g/dLforwomenand13g/dLformen,19_MCVandmean

cor-puscularhemoglobin (MCH) below80fL and 27pg for both genders,respectively,20_{andferritin<6ng/mLforwomenand}

28ng/mLformen,therebycharacterizingIDA.Exclusion crite-riafortheIDAgroupwerethepresenceofmalignanciesand inflammatory/infectious diseases or presenceof mutations associatedwith␣-TT(3.7,4.2,20.5,MED,SEA,THAI,FILand

Hph).Patientswiththesemutationswereexcludedsothatthis groupdidnothaveindividuals presentingthe twodiseases simultaneously.

For inclusion in the TT group, patients had the MCV below 80fL. ␤-TT carriers were confirmed by hemoglobin electrophoresisatpH8.6andHbA2 levels>3.5%.5 ␣-TTwas

confirmed by the presenceofmutations. All patients with malignancies orinflammatory/infectious diseasesbased on clinicaldataandpersonalinformationobtainedfrom medi-calrecordswereexcludedfromthisgroup,aswerepatients sufferingfromirondeficiencyidentifiedbymeasuringferritin serumlevels.

Samplecollection,hematologicalandbiochemical evaluations

Peripheral blood (5mL) was collected in tubes containing ethylenediaminetetraaceticacid(EDTA)atafinal concentra-tionof1mg/mL.Bloodcellcountresultswereobtainedusing automaticcell counters(ABXPentraDX-120,HoribaMedical DF® _{in IPSEMG and Sysmex}® _{XE-2100 in Hospital}

Univer-sitário).HemoglobinelectrophoresiswasperformedatpH8.6 using asemi-automated system(Hydrasys® _ofSebia®_{) and}

HbA2wasmeasuredbytheelutionmethodusingacellulose

acetatestrip.5_{Inaddition,thesesampleswereusedto}

investi-gateeightmutationsthatcause␣-TT(3.7,4.2,20.5,MED,SEA, THAI,FILandHph).Anothersampleoftheperipheralblood (5mL)wascollectedwithoutanticoagulanttoevaluateferritin values.SerumferritinlevelsweredeterminedbyImmulite®

DPC® _{usingthechemiluminescentimmunoassaytechnique}

andImmulite 2000DPC® _{kits(DiagnosticProducts}

Corpora-tion, LosAngeles,CA, USA)withreference valuesbetween 28and397ng/mLformenandbetween6and159ng/mLfor women.

Molecularanalysis

Genomic DNA was extracted from peripheralblood leuko-cytesusingtheGentraPuregeneBloodKit(Qiagen®_,Germany)

accordingtothemanufacturer’sinstructions.Investigationof the deletionalmutations (3.7, 4.2, MED,20.5, SEA, FIL and THAI)wasperformedbymultiplexpolymerasechainreaction (PCR)accordingtothemethodologydescribedbyTanetal.21

OligonucleotidesweresynthesizedbyPromega® _(Promega®_,

USA),whosesequenceswerealsodescribedbyTanetal.21_A

control,theLIS1genelocatedonchromosome17,wasusedin thereactiontodetectthesemutationsinordertoverifythe successofamplification.Afteramplification,thePCRproduct wassubjectedtoelectrophoresison1.5%agarosegel.

Developmentofthenewindex

A mathematical formula to discriminate betweenIDA and TT wasdevelopedwiththe supportofastatistician ofthe Universidade Federalde MinasGerais, Brazil, using values of the hematological parameters of 23 patients with ␤-TT and83patientswithIDAconfirmedusinggoldstandardtests (HbA2>3.5%in␤-TTandlowserumferritininIDA).5

employedtocreatethisnewindex.Basedonthelargestarea underthecurve,theformulawiththebestperformancewas theonethattookintoaccounttheRBCcountandmean cor-puscularhemoglobinconcentration(MCHC)parameters.

The adjusted formula of the new index developed in this study is Matos & Carvalho Index (MCI)=(1.91×RBC)+

(0.44×MCHC).

AccordingtotheROCcurve,theMCIpresentedacut-off pointwithavalueof23.85todiscriminatebetweenIDAandTT. Iftheindexis<23.85,thepatientisclassifiedasanIDApatient, whilevalues>23.85classifytheindividualasaTTcarrier.

ValidationoftheMatos&CarvalhoIndex

In order to validate the MCI, a cross-sectional study was conductedfrom2009to2011intwohospitals(Hospital Gov-ernadorIsraelPinheiro–IPSEMGandHospitalUniversitário) inBeloHorizonte,MinasGerais,Brazil.Atotalof227 outpa-tientswereincludedinthisstudy.Classificationofpatientsas havingIDAorTTwascarriedoutaccordingtothe aforemen-tionedcriteria.From227patients,197patientspresentedIDA and30patientswereTTcarriers.Ofthe197patientswithIDA, 42hadmutationsrelatedto␣-thalassemiaandwereexcluded

(Figure1).Thus,theIDAgroupwascomposedof155patients.

IntheTTgroup,24patientshad␤-TTandtheothersixpatients

had␣-TT(Figure1).

Statisticalanalysis

StatisticalanalysiswasperformedusingtheGraphPadPrism 5.0software(Graph-PadSoftwareInc.,LaJolla,CA).Normality oftheCBCdatawasassessedusingtheKolmogorov–Smirnoff test.Consideringthe nonparametricnatureofalldatasets, theMann–Whitneytestwasusedtoinvestigatedifferences betweenthetwogroups.Differenceswereconsidered signifi-cantwhenthep-value<0.05.

Thesensitivity,specificity,efficiencyandYouden’sIndex werecalculatedtoinvestigatetheperformanceoftheMCIto differentiatebetweenIDAandTT.Besidesthecalculationof theseparameters,theROCcurveofdiscriminantformulaswas plottedanditsareadeterminedusingGraphPadPrism5.0.The largestareaunderthecurveindicatestheindexmostlikelyto correctlydiscriminatepatientswithoneoftheseanemias.

ThevalidityoftheMCIwascomparedtotheresultsfrom moleculartechniques(␣-TT),HbA2measurement(␤-TT)and ferritin measurement (IDA) asreference standards, respec-tively.Inordertoinvestigatethedegreeofagreementbetween the results, theKappa coefficient wascalculated using the softwareOpenEpiversion2.3.1.InterpretationofKappawas according to the following scale: excellent 1.00–0.81; good 0.80–0.61;moderate0.61–0.40;weak0.40–0.21;andabsenceof agreement0.20–0whencomparedtothegoldstandardsinthe diagnosisofIDAandTT.22

Ethicalstatement

ThisstudywasapprovedbyboththeEthicsCommitteesofthe UniversidadeFederaldeMinasGerais(protocoln◦_344/09)and

theHospitalGovernadorIsraelPinheiro(protocoln◦_361/09)in

BeloHorizonte,Brazil.Patientswereinformedoftheobjectives ofthe research and signedinformed consentformsbefore bloodsamplinganddatacollection.

Results

Hematologicalparameters

ThemedianandinterquartilerangesofCBCparametersfor the 185patients (TT and IDA groups) enrolled in the vali-dationofMCIwerecalculated.RBC,Hb,hematocritandMCHC valuesforTTpatients weresignificantly higherthan those observedforIDApatientswhereasMCV,MCH,RDWandthe

IDA (n=197) (Ferritin: <6; <28)

Ferritin (ng/mL)

Investigated mutations

Positive

(IDA and α-TT, n=42)

Negative (IDA n=155)

Eligible (IDA) Eligible (TT)

Posit., HbA2<3.5 g/dL

(α-TT; n=6)

Negat., HbA2>3.5 g/dL

β-TT (n=24)

Investigated

mutations and HbA2

Not eligible

Samples (n=227)

(Hb: <12, <13 g/dL; MCV<80 fL; MCH<27pg)

TT (n=30) (Ferritin: >6; >28)

Table1–Hematologicalparametersofpatientswithirondeficiencyanemiaandthalassemiatraitusedtovalidatethe Matos&CarvalhoIndex.

Parameter Irondeficiencyanemia(n=155) Thalassemiatrait(n=30) p-value

RBC(106_{/␮L) 4.4(4.2–4.7) 5.4(5.1–5.7) <0.0001}

Hb(g/dL) 10.0(9.3–10.7) 10.9(10.5–11.8) <0.0001

Ht(%) 32.2(30.3–33.8) 34.6(33.2–36.2) <0.0001

MCV(fL) 73.0(67.0–76.0) 63.8(61.6–68.5) <0.0001

MCH(pg) 22.7(20.8–24.1) 20.4(19.7–21.6) <0.0001

MCHC(%) 31.2(30.4–31.8) 31.8(31.2–32.6) 0.0003

RDW(%) 17.9(16.6–19.4) 15.9(15.3–16.9) <0.0001

PLT(103_{/␮L) 319.0(257.0–390.0) 233.5(197.5–275.3) <0.0001}

RBC:redbloodcount;Hb:hemoglobin;Ht:hematocrit;MCV:meancorpuscularvolume;MCH:meancorpuscularhemoglobin;MCHC:mean corpuscularhemoglobinconcentration;RDW:redbloodcelldistributionwidth;PLT:plateletcount.

Thedataarepresentedasmediansandinterquartilerange.

Table2–PerformanceofMatos&CarvalhoIndexinthediscriminationbetweenirondeficiencyanemiaandthalassemia trait.

Group IDA TT

Sensitivity%(95%CI) 99.3(96.4–99.9) 76.7(59.1–88.2)

Specificity%(95%CI) 76.7(59.1–88.2) 99.3(96.4–99.9)

Accuracy(%) 95.7(91.7–97.8)

Youden’sIndex 76.0

AUC(95%CI) 0.95(0.90–1.00)

Kappacoefficient(95%CI) 0.83(0.68–0.97)

IDA:irondeficiencyanemia;TT:thalassemiatrait;95%CI:95%confidenceinterval;AUC:areaunderreceiveroperatingcharacteristiccurve.

plateletcountweresignificantlylowerthanthoseobservedfor IDApatients.Hematologicaldataofthetwostudygroupsare showninTable1.

Validationoftheindex

MCIvalueswerecalculatedforallpatientsusingCBCdata. MCIcorrectlydiagnosed154(99.3%)ofthe155patientswith IDAand23(76.7%)ofthe30patientswithTT.MCIperformance wasanalyzedbycalculatingsensitivity[99.3;95%confidence interval(95%CI):96.4–99.9],specificity(76.7;95%CI:59.1–88.2), accuracy(95.7;95%CI:91.7–97.8),areaunderROCcurve(0.95; 95%CI:0.90–1.00),Youden’sIndex(76.0)andthekappa coeffi-cient(0.83;95%CI:0.68–0.97)(Table2).Figure2showstheROC curvefortheMCI.Thisindexshowedhighaccuracyandhas revealedexcellentagreementwiththegoldstandard diagnos-tictechniquesaswellasgoodclinicalapplicabilityasamethod ofscreening.

Discussion

Inthisstudy,anewindex(MCI)wasdevelopedusingdetailed statistical analysiswith the aimof accurately discriminat-ing between IDA and TT. This formula produced excellent diagnosticaccuracyandtheparameters,suchassensitivity, efficiency,Youden’sIndex,andareaundertheROCcurvewere quitehigh.MCI showed excellentagreement withthegold standardmethodsofdiagnosisfortheseanemias, showing goodapplicabilityasascreeningtoolintheclinicalpractice.

100

50

0

0 50

100%-specificity%

Sensitivity

, %

100

Figure2–Receiveroperatingcharacteristiccurveforthe Matos&CarvalhoIndex.

Furthermore,theMCIwasdevelopedusingamolecular biol-ogytechniquetodiagnose␣-thalassemia,whichisnecessary to exclude concomitant diseases. The indexes currently described in the literature were not developed using this technique.

thosethatcorrectlydiagnoseTT.Thisisbecauseonemonth ofiron supplementation to a patient mistakenly classified ashavingIDAwhoinfacthasTT,causeslessdamagethan thelackofthissupplement toapatientwithIDAwhowas mistakenlyclassifiedasaTTcarrier.Ironanderythropoietin arerequiredfortheformationofhemoglobin.5_Furthermore,

iron isa component requiredin critical cellularprocesses suchasthetransportandutilizationofoxygen,productionof adenosinetriphosphate(ATP),DNAsynthesis,metabolismof catecholamines,mitochondrialelectrontransportandother physiologicalprocesses.23–25_{Thus,biologicalsystems}

includ-ingtheimmuneandneurologicalsystems,areallaffectedby thelackofiron.23_{Forthesereasonsthelackoftreatmentfor}

IDAwouldbeveryharmfultotheindividual.

Despite the moderate sensitivity for detecting patients withthalassemia,applicationoftheMCIisstilladvantageous ifthereissuspicionofthisdisorder.Inthiscase,thepatient whopresentedavalueofMCI>23.85shouldbereferredfor confirmationbyconventionalmethods.Nevertheless,in prac-ticaltermsthis isanadvantage,becauseforthis group,an investigationofironlevelscanbeconsideredasecondoption. The MCI represents a breakthrough in discriminating between IDA and TT, with potential for wide application considering its advantage ofbeing dependent only on the numberofRBCsandtheMCHC.Itshouldbenotedthatthese twoparametersareobtainablefrommostsimplecellcounters, thereforesophisticatedautomaticcountersarenotrequired. Hence,theMCIcanbeappliedinareaswhereadvanced tech-nologiesarenotavailableinclinicallabs.

Despitetheadvantagesandsimplicityofthe implementa-tionoftheMCIinthelaboratorypractice,thereisalimitation ofMCIandotherdiscriminatingformulassincetheyarenot abletodifferentiateallcasesofIDAfromTT.Inlightofthis, twosituationscanoccurthatdeservespecialattention:(i)the indexindicatedTT,butthepatienthadIDAand(ii)the appli-cationoftheindexindicatedIDA,butthepatientwasaTT carrier.Inthe firstcase,patientfollow-upisnecessaryand willindicate,overtime,asignificantreductionincirculating hemoglobinlevels,promptingthephysiciantorequest inves-tigativetestsofironmetabolism.Inthelattersituation,patient follow-upisalsoneededwhichwillshowtheneedfor medi-calprocedures.Inthiscase,theprescriptionofironwouldnot increasethehemoglobinlevelduetothegeneticdisorder.In casesofconcomitantdiseases,monitoringcanalsoclarifythe bestmedicalapproach.Therefore,MCIisausefultoolin guid-ingthephysicianregardingtheinitialconducttobeadopted; however,itdoesnoteliminatetheneedofafollow-upthat eventuallymayrequireconfirmatorytests.

Conclusion

TheMCIhas anexcellentperformanceand may be poten-tiallyusefulforscreeningpatientswithmicrocyticanemias. TheroutineuseoftheMCIprovidedgreatdiagnosticaccuracy, andcould contributedecisivelytoguidethe choiceof con-firmatorylaboratorytestsleadingtoacorrectdiagnosisand treatment.Thiswouldresultinasignificantcostsavingforthe health system, especially advantageous in underdeveloped anddevelopingcountrieswithlimitedfinancialresources.

Conflicts

of

interest

Theauthorsdeclarenoconflictsofinterest.

r

e

f

e

r

e

n

c

e

s

1.KassebaumNJ,JasrasariaR,NaghaviM,WulfSK,JohnsN,

LozanoR,etal.Asystematicanalysisofglobalanemia

burdenfrom1990to2010.Blood.2014;123(5):615–24.

2.PasrichaSR.Anemia:acomprehensiveglobalestimate.Blood.

2014;123(5):611–2.

3.WorldHealthOrganization.Theglobalprevalenceofanaemia in2011.Availablefrom:http://www.who.int/nutrition/

publications/micronutrients/globalprevalenceanaemia

2011/en/.

4.MathersC,StevenG,MascarenhasM.Globalhealthrisks:

mortalityandburdenofdiseaseattributabletoselectedmajor

risks.Geneva,Switzerland:WorldHealthOrganization;2009.

5.RodakBF,FritsmaGA,KeohaneEM.Hematology:clinical

principlesandapplications.4thed.ElsevierSaunders;2011.

6.MartinA,ThompsonAA.Thalassemias.PediatrClinNorth

Am.2013;60(6):1383–91.

7.CaoA,GalanelloR.Beta-thalassemia.GenetMed.

2010;12(2):61–76.

8.HandinRJ,LuxSE,StosselTP,editors.Blood:principlesand

practiceofhematology.2nded.Philadelphia:Lippincott

Williams&Wilkins;2003.

9.MentzerWCJr.Differentiationofirondeficiencyfrom

thalassaemiatrait.Lancet.1973;1(7808):882.

10.SrivastavaPC.Differentiationofthalassaemiaminorfrom

irondeficiency.Lancet.1973;2(7821):154–5.

11.ShineI,LalS.Astrategytodetectbeta-thalassaemiaminor.

Lancet.1977;1(8013):692–4.

12.RicercaBM,StortiS,d’OnofrioG,ManciniS,VittoriM,Campisi

S,etal.Differentiationofirondeficiencyfromthalassaemia

trait:anewapproach.Haematologica.1987;72(5):409–13.

13.GreenR,KingR.Anewredcelldiscriminantincorporating

volumedispersionfordifferentiatingirondeficiencyanemia

fromthalassemiaminor.BloodCells.1989;15(3):481–91,

discussion492–5.

14.JayaboseS,GiavanelliJ,Levendoglu-tugalO,SandovalC,

OzkaynakF,VisintainerP.Differentiatingirondeficiency

anemiafromthalassemiaminorbyusinganRDW-based

index.JPediatrHematolOncol.1999;21(4):314.

15.EhsaniMA,ShahgholiE,RahiminejadMS,SeighaliF,Rashidi

A.Anewindexfordiscriminationbetweenirondeficiency

anemiaandbeta-thalassemiaminor:resultsin284patients.

PakJBiolSci.2009;12(5):473–5.

16.SirdahM,TaraziI,AlNajjarE,AlHaddadR.Evaluationofthe

diagnosticreliabilityofdifferentRBCindicesandformulasin

thedifferentiationofthebeta-thalassaemiaminorfromiron

deficiencyinPalestinianpopulation.IntJLabHematol.

2008;30(4):324–30.

17.ShenC,JiangYM,ShiH,LiuJH,ZhouWJ,DaiQK,etal.

Evaluationofindicesindifferentiationbetweeniron

deficiencyanemiaandbeta-thalassemiatraitforChinese

children.JPediatrHematolOncol.2010;32(6):e218–22.

18.JanelA,RoszykL,RapatelC,MareynatG,BergerMG,

Serre-SapinAF.Proposalofascorecombiningredbloodcell

indicesforearlydifferentiationofbeta-thalassemiaminor

fromirondeficiencyanemia.Hematology.2011;16(2):123–7.

19.WorldHealthOrganization.Guidelinesfortheuseofiron

supplementstopreventandtreatirondeficiencyanemia;

20.WorldHealthOrganization.Irondeficiencyanaemia:

assessment,preventionandcontrol.Aguideforprogramme

managers;2001.p.114.

21.TanAS,QuahTC,LowPS,ChongSS.Arapidandreliable

7-deletionmultiplexpolymerasechainreactionassayfor

alpha-thalassemia.Blood.2001;98(1):250–1.

22.SzkloM,NietoJ.Epidemiology:beyondthebasics.2nded.

Jones&BartlettLearning;2007.

23.NobiliB,PerrottaS,MatareseSMR,ConteML,GiudiceEM.

EvaluationofbodyironstatusinItaliancarriersof

beta-thalassemiatrait.NutrRes.2001;21(1/2):55–60.

24.BeardJL,DawsonH,PineroDJ.Ironmetabolism:a

comprehensivereview.NutrRev.1996;54(10):295–317.

25.CohenAR,GalanelloR,PennellDJ,CunninghamMJ,Vichinsky

E.Thalassemia.HematologyAmSocHematolEducProgram.