PublicHealthNutrition:12(4),444–454 doi:10.1017/S1368980008002401
Worldwideprevalenceofanaemia,WHOVitaminandMineral NutritionInformationSystem,1993–2005
ErinMcLean1,MaryCogswell2,InesEgli3,DanielWojdyla4 andBrunodeBenoist1,*
1DepartmentofNutritionforHealthandDevelopment,
WorldHealthOrganization,20AvenueAppia, CH-1211,Geneva27,Switzerland: 2DivisionofNutritionandPhysicalActivity,CentersforDiseaseControl andPrevention(CDC),USA: 3InstituteofFoodScienceandNutrition,SwissFederalInstituteofTechnology, Zurich,Switzerland: 4EscueladeEstadistica,UniversidadNacionaldeRosario,Argentina
Submitted28May2007:Accepted6March2008:Firstpublishedonline23May2008
Abstract
Objective: To provide current global and regional estimates of anaemia pre- valenceandnumberofpersonsaffectedinthetotalpopulationandbypopulation subgroup.
Settinganddesign:WeusedanaemiaprevalencedatafromtheWHOVitaminand Mineral Nutrition Information System for 1993–2005 to generate anaemia pre- valenceestimatesforcountrieswithdatarepresentativeatthenationallevelorat the?rstadministrativelevelthatisbelowthenationallevel.Forcountrieswithout
eligible data, we employed regression-based estimates, which used the UN Human Development Index (HDI) and other health indicators. We combined countryestimates,weightedbytheirpopulation,toestimateanaemiaprevalence atthegloballevel,byUNRegionsandbycategoryofhumandevelopment.
Results:Surveydatacovered48?8%oftheglobalpopulation,76?1%ofpreschool- agedchildren,69?0%ofpregnantwomenand73?5%ofnon-pregnantwomen.The estimatedglobalanaemiaprevalenceis24?8%(95%CI22?9,26?7%),affecting1?62 billionpeople(95%CI1?50,1?74billion).Estimatedanaemiaprevalenceis47? 4% (95%CI45?7,49?1%)inpreschool-agedchildren,41?8%(95%CI39?9,43?8%)in
pregnant women and 30?2% (95% CI 28?7, 31?6%) in non-pregnant women. In numbers,293million(95%CI282,303million)preschool-agedchildren,56million (95%CI54,59million)pregnantwomenand468million(95%CI446,491million) non-pregnantwomenareaffected.
Conclusion: Anaemia affects one-quarter of the world’s population and is concentratedinpreschool-aged children andwomen, making itaglobal public health problem. Data on relative contributions of causal factors are lacking, however,whichmakesitdif?culttoeffectivelyaddresstheproblem.
Keywords
Anaemia Haemoglobin Nutritionalstatus Ironde?ciency
Anaemia,oneofthemostcommonandwidespreaddis- de?cienciesofothervitaminsandminerals,suchasfolate,
orders in the world, is a public health problem in both vitaminsAandB12,andcopper(2). industrialised and non-industrialised countries. In 2002, Becauseironde?ciencymakesalargecontributionto
the WHO estimated that anaemia resulting from iron anaemia, global efforts to reduce the anaemia burden de?ciency was one of the ten most important factors have largely been directed towards increasing intake of contributing to the global burden of diseases and that iti ron through supplementation, food forti?cation and increases morbidity and mortality in preschool-aged diversi?cationofthediet.Toassesstheironstatusofthe childrenandpregnantwomen(1).Anaemiaisde?nedasa populationortheresponsetoaninterventiontoprevent
decrease in the concentration of circulating red blood and control iron de?ciency, haemoglobin concentration cells or in the haemoglobin concentration and a con- has often been used in surveys as a proxy indicator for comitant impaired capacity to transport oxygen. It has iron status under the assumption thatanaemia is always multiple precipitating factors that can occur in isolation associated with iron de?ciency, even if many other but more frequently co-occur(2). These factors may be possible causes are present. These surveys have rarely genetic,suchashaemoglobinopathies;infectious,suchas measuredironde?ciencyoranyoftheotherfactorsthat malaria, intestinal helminths and chronic infection; or contributetothedevelopmentofanaemiaandtherefore nutritional, which includes iron de?ciency as well as thecontributingfactorsfrequentlyremainunknown.
*Correspondingauthor:Emaildebenoistb@who.int
rTheAuthors2008
Worldwideanaemiaprevalence 445
Previousestimatesoftheprevalenceofanaemiainthe Weexcludedsurveysthatmeasuredonlyclinicalsigns
world were reported on population subgroups in 1982 of anaemia or the haematocrit and contacted study and1992(3,4) andonallpopulationgroupsin1985(5) and authors for clari?cation or additional information when 2001(6).Withtheexceptionofthemostrecentestimates, necessary.Theadministrativelevelofasurveyisnational however,whichincludeddataupto1995(6),thesereports when the sample is nationally representative, or sub- did not include data collected after 1990. Moreover, the national when the sample is representative of a given 1985reportbyDeMaeyerandTegman(5) didnotinclude administrative level, namely, region, state (?rst adminis- data for China, which represents 20% of the global trativeboundary),district (secondadministrativebound- population. ary)orlocal.Infrequently,surveyscouldbenationaleven
though some regions had to be left out for security or Theobjectiveofthepresentpaperistoprovidecurrent
otherreasons. globalandregionalestimatesoftheprevalenceofanae-
mia and of the number of persons affected based on surveys conducted between 1993 and 2005 for the 192
Member States of the WHO. As a result of the vastly Dataselection
different methodologies used, these estimates are not Forthisanalysis,weusedthefollowingfourvariablesin quantitativelycomparabletopreviousestimates. selectingdatafromtheWHOGlobalDatabaseonAnaemia
onhaemoglobinconcentrationand/ortheprevalenceof anaemia:thetimeframeofthesurvey,theadministrative
Methods
levelforwhichthesurveywasrepresentative(nationalor subnational),thesamplesizeandthepopulationgroups
Datasource
surveyed.
We based the current estimates on data available in the
ThetimeframefortheestimateswasfromJanuary1993
WHOGlobalDatabaseonAnaemia,apartoftheVitamin
to December 2005, and surveys that took place during
and Mineral Nutrition Information System (VMNIS)
this time period and were published by 31 December
(http://www.who.int/vmnis).Thisdatabaseincludesdata
2005 were eligible. As of that date, 696 surveys that
on haemoglobin concentration and the prevalence of
reportedondatacollectedbetween1993and2005were
anaemia presented by country in a standardised, easily
available.Weusedthepublicationdatewhentheperiod
accessibleformat.
ofdatacollectionwasnotspeci?ed.
To establish the WHO Global Database on Anaemia,
We useddatafromthemostrecentnationalsurvey in
we systematically searched and collected data from the
preferencetosubnationalsurveysofmorerecentvintage.
scienti?c literature (Medline and WHO regional data-
For one country, where an area had been left out of a
bases)andthroughabroadnetworkofpartners,includ-
national survey because of security concerns, available
ingWHOregionalandcountryof?ces,UNorganisations,
data from the missing region (weighted by the general
ministries of health, research and academic institutions
population estimate for that area) were pooled with the
and non-governmental organisations. We augmented
nationaldatatoprovideanestimateforthecountry.The
theseresourcesbymanualsearchingofarticlespublished
estimatewasdeterminedbyusingthemostrecentcensus
in non-indexed medical and professional journals and
data from the country. The surveys were conducted
reportsfromprincipalinvestigators.
within1yearofeachotherandaddingthemissingregion
To include data in the WHO Global Database on
changed the overall estimate by only 0?1%. If two
Anaemia, werequired a complete original survey report
nationalsurveyswereconductedinthesameyear,aswas
with details of the sampling method used. In a few
thecasefortwopopulationgroupsfromonecountry,we
cases, we accepted data provided in writing directly by
pooledthesurveyresultsintoasinglesummarymeasure,
ministries of health with detailed methodology, even
weighted by the sample size of the two surveys. The
without a formal published report. We included
differencebetweentheestimatesinthetwosurveyswas
surveysrepresentativeofanyadministrativelevelandany
5–15%, depending on the population group. In the
population group in the WHO Global Database on
absence of national data, we used surveys that were
Anaemiaifthey:
representative at the ?rst administrative-level boundary f two or more surveys at this level were available for > were population based or facility based (for pregnant i
women,newborns,preschool-andschool-agedchildren), thepopulationgroupandcountryconcernedwithinthe
acceptable time frame. We pooled the results into a > were cross-sectional or baseline values from an
single summary measure, weighted by the total general interventionprogramme,
> measured haemoglobin concentration from capillary, population for that region or state, based on the most venous or cord blood using quantitative photometric recentandavailablecensusdatabetween1993and2005,
withoutconsideringtheagerangecoveredbythesurvey. methodsorautomatedcellcountersand
strict-level surveys in these > reported the prevalence of anaemia or mean haemo- We did not use local or di
estimatesbecausetheyhavethepotentialformorebias. globinconcentrations.
446 EMcLeanetal.
Asageneralrule,weexcludedprevalencedatabased these factors when provided by the survey. We did not
accept any other corrections. For severe anaemia, we onasampleoffewerthan100persons.Givenasampleof
100 and a con?dence level of 95%, the error around a includedsurveysthatusedahaemoglobincut-offof70g/l, prevalence estimate of 50% would be 610 percentage whichwasusedbyalmostallofthesurveysthatreported points;asmallersamplewouldhaveanevenlargererror. theprevalenceofsevereanaemia. Afewexceptionsweremade,however.Nationalsurveys For surveys thatclassi?ed persons by the WHOanae- withfewerthan100butmorethan?ftyparticipantswere mia threshold, we used the reported prevalence data accepted but only where the results were being extra- withoutanyadditionalcalculations.Ifprevalencewasnot polated to fewer than 500000 people or to pregnant reported, or was reported for a different threshold, women. we estimated the prevalence using one of the several
methods, all of which assumed a normal distribution Forthisanalysis,wede?nedpopulationsubgroupsas
follows: preschool children below 5 years, school-aged of haemoglobin concentrations. This would slightly childrenaged5?00–14?99years,pregnantwomenofany overestimate the prevalence of anaemia in populations age,non-pregnantwomenaged15?00–49?99years,men where it is high because population curves of hae- aged15?00–59?99yearsandtheelderly,aged601years. moglobinconcentrationswouldbeskewedtothelower
values. We used the following methods to estimate the Where possible, we excluded children below 0? 5 years
from preschool-aged children because an appropriate prevalenceofanaemiainorderofpreference: haemoglobin cut-off for this age group has not been determined(6).Wedidnotprovideaseparatepopulation 1. We used the mean and SD of the haemoglobin estimate for women aged 50?00–59?99 years, as these concentration to estimate the proportion of persons womenarerarelysurveyed.Wedid,however,includean fallingbelowtheappropriatehaemoglobincut-offfor estimate for women aged 50?00–59?99 years in our esti- thepopulationsubgroup(n20).Wevalidatedthisby mate of the global anaemia burden. The methods for assessing the correlation between the estimated and accomplishing this are detailed later. Infrequently, if data predicted prevalence of anaemia in surveys from the were not disaggregated, we included all women in the databasewhereamean,anSDandaprevalenceforthe estimatefornon-pregnantwomenevenifwedidnotknow WHO age- and sex-speci?c cut-off were provided. whether pregnant women were included. Where surveys This relationship was plotted (n 508), and for most provided data disaggregated by physiological status, lac- surveys, the two ?gures were extremely close tating women and non-pregnant women were combined (r250?95, P,0?001) for all four cut-offs (haemo-
globinconcentration,110,115,120,130g/l).Overall, forthepopulationsubgroupnon-pregnantwomen.
predictedprevalenceoverestimatedactualprevalence We used data disaggregated by the ages that were
by 3?8 percentage points. For 6?5% of the surveys, closest to the de?ned age ranges for the population
subgroups. If the age range overlapped two population estimatedprevalenceoverestimatedactualprevalence
by10percentagepointsormore,andinthesesurveys subgroups, we placed the survey with the subgroup
overestimationaveraged16? 3%. where there was a greater overlap in age. When the
age range was unavailable, we used the mean age of 2. Whenno SDwasprovided,butprevalenceforanon- thesampletoclassifythedata.Ifthiswasunavailableand WHO cut-off and mean haemoglobin concentration the age range equally spanned two population sub- were available (n 3), we used these two ?gures to groups, we used the population-speci?c haemoglobin calculatethe SD ofthehaemoglobinconcentrationby concentration threshold to classify the data. If data assuminganormaldistributionwithinthepopulation representedlessthan20%oftheagerangeofapopulation and deriving the Z-score for the prevalence in order group,wedidnotincludethesurvey. to back-calculate the SD
[SD5(provided cut-off2
mean haemoglobin)/Z-score for given prevalence]. Following this calculation, the mean and SD were Prevalenceofanaemiaforcountrieswith
usedasabovetoderivetheprevalencefortheWHO surveydata
cut-off. Normalhaemoglobindistributionsvarywithage,sexand
physiological status, for example, pregnant (varies by 3. Finally, for surveys (n 23) that did not present the trimester) and non-pregnant(7). We used WHO hae- mean and SD or the prevalence at the recommended moglobinthresholdstoclassifypersonslivingatsealevel threshold, we estimated the prevalence of anaemia as anaemic: children 0?50–4?99 years and pregnant from the prevalence at an alternative threshold. We women, 110g/l; children 5?00–11?99 years, 115g/l; chil- assumed that an average SD for the same population dren12?00–14?99yearsandnon-pregnantwomen$15?00 subgroupwouldbeclosetotheactualSDinthesurvey. years, 120g/l; men $15?00 years, 130g/l(6) . Statistical We calculated the mean SD of the haemoglobin and physiological evidence indicate that haemoglobin concentration for each population subgroup from distributions vary with smoking(8) and altitude(9) thesurveysincludedintheestimates,whichhaddata
and, availableforparticipantswithinthede?nedagerange
therefore,weusedtheprevalenceofanaemiacorrectedfor
E,etalPublicHealthNutr;()–
