The Effects of Iron Deficiency and Anemia on Mental and ... Documents/Bibliography/The effects of... · iron deficiency anemia on developmental out-come at age 2 years. J Pediatr

The Effects ofIron Deficiency and Anemia onMental and Motor Performance,Educational Achievement, andBehavior in Children

An Annotated Bibliography

Catherine Nokes

Claire van den Bosch

Donald A.P. Bundy

A Report of the International NutritionalAnemia Consultative Group

The Effects of Iron Deficiencyand Anemia on Mental andMotor Performance, EducationalAchievement, and Behavior in Children:An Annotated Bibliography

Catherine Nokes

Claire van den Bosch

Donald A.P. Bundy

INACG publications are available free of charge to developing countries and for $3.50 (U.S.) todeveloped countries. Copies can be ordered from the INACG Secretariat.

For additional information contact: International Nutritional Anemia Consultative Group (INACG),c/o ILSI Human Nutrition Institute, 1126 Sixteenth Street, NW, Washington, DC 20036-4810,USA. Phone: 202-659-0789; Fax: 202-659-3617; Email: [email protected]

This report is the summary of the presentations and discussions that took place at the consultativemeeting at the University of Oxford, on September 16–17, 1996, and does not necessarily reflectthe scientific recommendations or views of INACG, USAID, or the International Life SciencesInstitute.

The publication of this report is made possible by support from Opportunities for MicronutrientInterventions (OMNI), a project of the Office of Health and Nutrition, Bureau for Global Pro-grams, Field Support and Research, U.S. Agency for International Development, under ContractHRN-5122-C-00-3025-00, Project 936-5122 and Partnership for Child Development. The contri-bution of the Partnership for Child Development was made possible by a grant from the James S.McDonnell Foundation.

Printed April 1998 in the United States of America.

��

��

CONTENTS

ALPHABETICAL LIST OF INCLUDED STUDIES ........................................................................................................................ i

INTRODUCTION ........................................................................................................................................................................ 1Studies Included in the Bibliography ................................................................................................................................... 1Structure of the Bibliography ............................................................................................................................................... 1

Age ................................................................................................................................................................................. 1Experimental Design ...................................................................................................................................................... 2Date ................................................................................................................................................................................ 3

Other Features of the Studies Highlighted in the Bibliography ......................................................................................... 3Location of Study .......................................................................................................................................................... 3Iron Status of Children .................................................................................................................................................. 3Length of Treatment ...................................................................................................................................................... 3Developmental, Cognitive, or Educational Outcome Measures Used ........................................................................ 4

How to Use the Bibliography to Find Specific References: the Numbering System ......................................................... 4Glossary of Terms and Key to Abbreviations ........................................................................................................................ 5Summary of experimental design by studies examining the effects of iron deficiency ......................................................... 6

INFANTS AND YOUNG CHILDREN (6–24 MONTHS) ........................................................................................................... 7Summary of Findings ............................................................................................................................................................ 7Short-term Effects: Observational Studies ........................................................................................................................... 8Short-term Effects: Intervention Trials ............................................................................................................................... 10Short-term Effects: Preventative Trials .............................................................................................................................. 18Long-term Effects: Observational Studies ......................................................................................................................... 19Long-term Effects: Follow-up of Intervention Trials ......................................................................................................... 21Long-term Effects: Preventative Trials ............................................................................................................................... 22

PRESCHOOL CHILDREN (2–5 YEARS) .................................................................................................................................... 23Summary of Findings .......................................................................................................................................................... 23Intervention Trials ............................................................................................................................................................... 23

SCHOOL-AGE CHILDREN AND ADOLESCENTS (5–16 YEARS) .......................................................................................... 27Summary of findings ........................................................................................................................................................... 27Observational Studies ......................................................................................................................................................... 27Intervention Trials ............................................................................................................................................................... 28

SATELLITE ISSUES ................................................................................................................................................................... 33Animal Studies and Biological Mechanisms ..................................................................................................................... 33Adult Iron Deficiency ......................................................................................................................................................... 33Other Outcomes of Iron Deficiency ................................................................................................................................... 34

Physical Fitness ........................................................................................................................................................... 34Infections ...................................................................................................................................................................... 34

Confounding and Covarying Factors ................................................................................................................................. 34

References ............................................................................................................................................................................. 35

READING LIST: OTHER REVIEWS AND RELATED PAPERS .................................................................................................. 37

APPENDIX A: Tabulated summary of studies investigating the effects of iron deficiency (with and without anemia) on thedevelopment of infants and young children, preschool children, and school-age children and adolescents ...................... 39

APPENDIX B: List of Studies Used in the Bibliography, by category .................................................................................... 59

APPENDIX C: Description of psychometric and development tests used in studies investigating the effects of ID and IDAon these functions ...................................................................................................................................................................... 60

ALPHABETICAL LIST

OF INCLUDED STUDIES

Aukett MA, Parks YA, Scott PH, Wharton BA (1986)Treatment with iron increases weight gain andpsychomotor development. Arch Dis Child61:849-857. Reference 12

Bruner AB, Joffe A, Duggan AK, et al (1996)Randomised study of cognitive effects of ironsupplementation in non-anemic iron-deficientadolescent girls Lancet 348:992-996. Reference42

Cantwell RJ (1974) The long term neurological se-quelae of anemia in infancy. Pediatr Res 8:342.Reference 26

Deinard AS, Gilbert A, Dodds M, Egeland B (1981)Iron deficiency anemia and behavioural deficits.Pediatrics 68:828-833. Reference 1

Deinard AS, List A, Lindgren B, et al (1986) Cogni-tive deficits in iron deficient and iron deficientanemic children. J Pediatr 108:681-689. Refer-ence 29

Grindulis H, Scott PH, Belton NR, Wharton BA(1986) Combined deficiency of iron and vita-min D in Asian toddlers. Arch Dis Child 61:843-848. Reference 3

Groner JA, Holtzman NA, Charney E, Mellitts E(1986) A randomized trial of oral iron on tests ofshort term memory and attention span in youngpregnant women. J Adolesc Health Care 7:44-48. Reference 37

Heywood A, Oppenheimer S, Heywood P, JolleyD (1989) Behavioral effects of iron supplemen-tation in infants in Mandung, Papua NewGuinea. Am J Clin Nutr 50:630-640. Reference17

Hurtado EK (1995) Cross linking a third data set:long term effects of iron deficiency during in-fancy and early childhood [abstract]. Biennialmeeting of the Society for Research in ChildDevelopment, Indianapolis, March-April 1995,217. Reference 22

Idjradinata P, Pollitt E (1993) Reversal of develop-mental delays in iron-deficient anemic infantstreated with iron. Lancet 341:1-4. Reference 15

Johnson DL, McGowan RJ (1983) Anemia andinfant behavior. Nutr Behav 1:185-192. Refer-ence 2

Kayshap P, Gopaldas T (1987) Impact of hemaniticsupplementation on cognitive function in un-derprivileged school girls (8-15 years of age).Nutr Res 7:1117-1126. Reference 38

Lozoff B, Brittenham GM, Viteri FE, Urrutia JJ(1982a) Behavioural abnormalities in infants withiron deficiency anemia. In Pollitt EL (ed), Brainbiochemistry and behaviour. Raven Press, NewYork, pp 183-195. Reference 8

Lozoff B, Brittenham GM, Viteri FE, et al (1982b)Developmental deficits in iron deficient infants:effects of age and severity of iron lack. J Pediatr101:948-952. Reference 7

Lozoff B, Brittenham GM, Viteri FE, et al (1982c)The effects of short-term oral iron therapy ondevelopmental deficits in iron-deficient anemicinfants. J Pediatr 100:351-357. Reference 6

Lozoff B, Wolf AW, Urrutia JJ, Viteri FE (1985) Ab-normal behavior and low developmental testscores in iron-deficient anemic infants. J DevBehav Pediatr 6:69-75. Reference 11

Lozoff B, Klein NK, Prabucki KM (1986) Iron-defi-cient anemic infants at play. J Dev Behav Pediatr7:152-158. Reference 4

Lozoff B, Brittenham GM, Wolf AW, et al (1987)Iron deficiency anemia and iron therapy effectson infant developmental test performance Pe-diatrics 79:981-995. Reference 13

Lozoff B, Jiminez E, Wolf AW (1991) Long-termdevelopmental outcome of infants with iron de-ficiency. N Engl J Med 325:6876-6894. Refer-ence 25

Lozoff B, De Andraca L, Walter T, Pino P (1996a)

Does preventing iron-deficiency anemia (IDA)improve developmental test scores? [personalcommunication]. Reference 19

Lozoff B, Wolf AW, Jimenez E (1996b) Iron defi-ciency anemia and infant development: effectsof extended oral iron therapy. J Pediatr 129:382–389. Reference 16

Moffatt MEK, Longstaffe S, Besant J, Dureski C(1994) Prevention of iron deficiency and psy-chomotor decline in high risk infants throughuse of iron-fortified infant formula: a random-ized clinical trial. J Pediatr 125:527-534. Refer-ence 18

Oski TE, Honig AS (1978) The effects of short termtherapy on the developmental scores of irondeficient infants. Pediatrics 92:21-25. Reference5

Oski FA, Honig AS, Helu B, Howanitz P (1983) Ef-fect of iron therapy on behaviour performancein non-anemic iron deficient infants. Pediatrics71:877-880. Reference 10

Palti H, Pevsner B, Adler B (1983) Does anemia ininfancy affect achievement on developmentaland intelligence tests? Hum Biol 55:189-194.Reference 23

Palti H, Meijer A, Adler B (1985) Learning achieve-ment and behavior at school of anemic andnon-anemic infants. Early Hum Dev 10:217-233.Reference 20

Pollitt E, Greenfield D, Leibel R (1978) Behaviouraleffects of iron deficiency among pre-school chil-dren inCambridge, Mass. Fed Proc 37:487. Ref-erence 27

Pollitt E, Leibel RL, Greenfield D (1983) Iron defi-ciency and cognitive test performance in pre-school children. Nutr Behav 1:137-146. Refer-ence 28

Pollitt E, Yunis F, Salem S, Scrimshaw NS (1985)Reported in Pollitt E, Soemantri AG, Yunis F,Scrimshaw NS, Cognitive effects of iron defi-ciency anemia. Lancet 19:158. Reference 35

Pollitt E, Saco-Pollitt C, Leibel RL, Viteri FE (1986)Iron deficiency and behavioral development ininfants and pre-school children. Am J Clin Nutr43:555-565. Reference 30

Pollitt E, Hathirat P, Kotchabhakdi NJ, et al (1989)

Iron deficiency and educational achievementin Thailand. Am J Clin Nutr 50:687-697. Refer-ence 41

Seshadri S, Hirode K, Naik P, Malhotra S (1982)Behavioural responses of young anemic Indianchildren to iron-folic acid supplements. Br J Nutr48:233-240. Reference 34

Seshadri S, Gopaldas T (1989) Impact of ironsupplementation on cognitive functions in pre-school and school aged children: the Indianexperience. Am J Clin Nutr 50:675-686. Refer-ence 39

Soemantri AG, Pollitt E, Kim I (1985) Iron deficiencyanemia and educational achievement. Am J ClinNutr 42:1221-1228. Reference 36

Soemantri AG (1989) Preliminary findings on ironsupplementation and learning achievement ofrural Indonesian children. Am J Clin Nutr 50:698-702. Reference 40

Soewondo S, Husaini MA, Pollitt E (1989) Effectsof iron deficiency on attention and learning pro-cesses in pre-school children: Bandung, Indo-nesia. Am J Clin Nutr 50:667-674. Reference 31

Walter T, Kovalskys J, Stekel A (1983) Effect of mildiron deficiency on infant mental developmentscores. Pediatrics 102:519-522. Reference 9

Walter T, De Andraca I, Chadud P, Perales CG(1989) Iron deficiency anemia: adverse effectson infant psychomotor development Pediatrics84:7-17. Reference 14

Walter T, de Andraca I, Castillo M, et al (1990) Cog-nitive effect at five years of age in infants whowere anemic at 12 months: a longitudinal study[abstract]. Pediatr Res 28:295. Reference 24

Wasserman G, Graziano JH, Factor-Litvak P, et al(1992) Independent effects of lead exposure andiron deficiency anemia on developmental out-come at age 2 years. J Pediatr 121:695-703.Reference 21

Webb TE, Oski F (1973a) Iron deficiency anemiaand scholastic achievement in young adoles-cents. J Pediatr 82:827-830. Reference 32

Webb TE, Oski FA (1973b) The effect of iron defi-ciency anemia on scholastic achievementbehavioural stability and perceptual sensitivity[abstract]. Pediatr Res 28:294. Reference 33

ii The Effects of Iron Deficiency on Child Development: An Annotated Bibliography

INTRODUCTION

Iron deficiency (ID) is a major public health prob-lem affecting more than 2000 million personsworldwide. Iron is an essential nutrient not only

for the normal growth, health, and survival of children,but also for their normal mental and motor develop-ment and cognitive functioning. Iron deficiency withanemia (IDA) is associated with significantly poorer per-formance on psychomotor and mental developmentscales and behavioral ratings in infants, lower scores oncognitive function tests in preschool children, and lowerscores on cognitive function tests and educationalachievement tests in school-age children.

A large number of studies have been conducted to ex-amine the effects of ID and IDA on mental outcomes,yet the evidence is not always complete nor conclusive.A need was identified, therefore, to catalog each of thestudies so that they could be easily accessed and con-sulted, allowing researchers to review the available data,draw their own conclusions from the available evidence,and verify any data reviewed or summarized elsewhere.

This bibliography is intended to be used in conjunctionwith and as a supplement to the Oxford Brief: ChildDevelopment and Iron Deficiency (by Alizon Draper,ILSI Press, May, 1997).

Studies Included in theBibliography

Included in this bibliography are studies that have lookeddirectly at the relationship between iron status and somemental or developmental outcome in humans. Becausethis is the main focus of the bibliography, as far as pos-sible all of the papers published in this area have beenincluded and individually described and summarized. Allreviews on this topic are listed in the Reading List sec-tion.

At the end of the bibliography, a description of selectedpapers on related topics (Satellite Issues) is included suchas studies investigating the relationship between iron sta-tus and mental and motor development in animals. Pa-

pers were selected if they were considered to have useda strong experimental design, were a good or representa-tive example of the type of research in the area, and gavea comprehensive review of the subject.

Structure of the Bibliography

The bibliography consists of 5 parts:

1. Detailed description of each experimental study

2. Detailed tabulation/summary of each study

3. Brief tabulated summary of studies

4. A list of reviews given in the reading list

5. A summary of studies on related/satellite issues

The following description of the structure of the bibli-ography refers only to those papers that have directlyexamined the relationship between iron status in hu-mans and its effect on mental function and development.

The bibliography has been structured to focus on theage of subjects studied, the experimental design, and thedate of publication. Specific details on each of these de-scriptors are given below.

Age. The papers are described and catalogued first ac-cording to the three main age groups being studied. Adiscussion and overview of the results for each age groupis given at the beginning of each section. The age groupsare as follows:

� Infants and young children. Subjects studied werechildren between ages 6 and 24 months; the out-come measures were mental and motor develop-ment and behavior.

� Preschool children. Subjects studied were childrenages 2 to 5 years; the outcome measure was cog-nitive function.

� School-age children and adolescents. Subjectsstudied were children between ages 5 and 16years; the outcome measures were cognitive func-tion and educational achievement.

2 The Effects of Iron Deficiency on Child Development: An Annotated Bibliography

Age is emphasized because it seems to determine theobserved level and type of developmental or cognitiveresponse to iron treatment. The age of children who mayrequire preventative or supplementary iron treatment alsohas important public health implications because childrenin different age groups are accessible through differentchannels; thus, age also determines the most appropriatedelivery system.

Experimental Design. Within each age group, studiesare further classified under subheadings according to ex-perimental design. Because of the complex range of in-tervention trials that have been undertaken (particularlyin studies with infants), studies have been grouped ac-cording to the following subcategories:

� Observational.

� Intervention. These are further divided into stud-ies assessing short-term or long-term effects.Short-term studies looked for an improvement inmental or motor outcomes from 1 week to 8 monthsafter iron supplementation. Iron supplementationin these short-term studies was either brief or ex-tended (with or without a placebo group). Withbrief supplementation, children were supplementedwith iron for 5 to 10 days only; with extendedsupplementation, children were supplemented withiron for more than 2 months. Long-term studieslooked at the effects of ID and IDA or the ben-efits of iron supplementation more than 2 yearslater. These studies also include prospective stud-ies (no treatment given).

� Preventative.

An understanding of the experimental design is importantfor interpreting the results from each study and for draw-ing valid conclusions. Thus, a brief description of the ad-vantages and disadvantages of each experimental designis given below.

In observational, cross-sectional, or case-control studies,iron and psychometric or developmental measurementsare taken from experimental groups at one time pointonly. These types of studies can show that one condition(developmental status) is associated with another (ironstatus) but they cannot determine the causality of therelationship. For example, other factors associated withiron status (such as socioeconomic conditions, mater-

nal education, and other nutritional deficits) could beresponsible for the observed developmental level of chil-dren.

In intervention studies, the groups with ID and IDA aregiven iron supplements after the initial baseline develop-mental or psychometric assessment and then the devel-opmental and psychometric assessments are repeated todetermine the degree of improvement. Intervention stud-ies may be double-blind, randomized, and treatment orplacebo-controlled, but those that include all of these fea-tures will have the strongest experimental design and thebest ability to infer causality between the putative cause(iron status) and the outcome (cognition or development).This is because each feature is designed to ensure equal-ity among the groups in all respects (for example, in so-cioeconomic status (SES), maternal education, and nutri-tional status) other than the treatment they were given sothat any confounding variables are equally distributedbetween the control and experimental groups. If this isensured, the experimental groups can be directly com-pared and the unique effect of iron treatment on children’sdevelopment determined.

In a double-blind trial, both the subjects and the personwho assesses the response are unaware of which groupthe child has been assigned to and what treatment hasbeen given. This means that both groups will be dealtwith by the experimenters in the same way and also thatone group of children will not feel more or less privilegedthan another.

In a randomized trial, children are assigned randomly tothe treatment groups (treatment and no treatment or treat-ment and placebo). Random assignment should ensurethat all the characteristics of the two groups are similarso that any differences among individuals, for example,differences in SES, are equally distributed between thetwo groups.

In a placebo-controlled trial, an effective control is pro-vided by the placebo, an inert preparation formulated toappear indistinguishable from the iron supplement. It isimportant to give a placebo to the control group becauseit accounts for unspecific differences in developmentalor cognitive responses that arise simply from the childrenbeing given something rather than nothing. Without a pla-cebo group it is not possible to determine whether im-provement in development or cognition observed after

Introduction 3

iron supplementation is due to improvement in ironstatus or to something related to iron but not iron itself,such as low birth weight, lack of breast-feeding, or envi-ronmental disadvantage.

The inclusion of a nonanemic, iron-sufficient controlgroup (which may or may not be matched with the IDor IDA group), although not essential to a treatment-or placebo-controlled trial, is important for two rea-sons. First, the inclusion of an iron-sufficient group canhelp determine whether there is any initial deficit in de-velopmental or cognitive outcome between the childrenwith ID or IDA and the iron-sufficient children. If nodeficit exists at baseline, an improvement following treat-ment would not be expected. Second, an iron-sufficientgroup can help determine whether the adverse effects ofID or IDA on performance are fully or only partiallyreversible following treatment, that is, whether the im-provement in performance of children with ID or IDAwho received iron treatment is large enough for them tocatch up to the iron-sufficient control group by the endof the study.

In conditions in which benefits from treatment have notbeen commonly observed, it is difficult to determinewhether this is because there are no direct effects ofIDA on performance or whether there are effects ofIDA that are not correctable with treatment. If the latteris considered the case, this could be studied in a preven-tative trial in which children who are supplemented (byreceiving iron-fortified milk formula, for example) andprotected from becoming iron deficient with anemia arecompared with children who are not supplemented. Al-though preventative trials have not been commonly un-dertaken in this area of research, they are considered tohave a strong experimental design.

Date. Studies are presented within each of the subsec-tions (e.g., under observational studies, intervention trials,or preventative trials) according to the year in which theywere published. The date of publication is consideredimportant for putting the studies into a historical perspec-tive and for observing how the research methods andquestions progressed over time.

Other Features of the StudiesHighlighted in the Bibliography

Certain key features of each paper presented have beenemphasized in the text and in the tabulated summary

(Appendix A) because they are considered importantfor interpreting results and for comparing results acrossstudies. These characteristics include the location of thestudy, the iron status of children at baseline, the lengthof treatment, and the developmental, cognitive, or edu-cational outcome measures used.

Location of Study. It is conceivable that baseline differ-ences and responses to iron therapy could vary accord-ing to the environment in which the child has grown up.For example, some environments may put children at in-creased risk of anemia or ID whereas others may pro-vide children with psychological buffers that prevent suchhematologic impairment impeding development.

Iron Status of Children. Whether children recruited tothe study have ID or IDA is clearly specified becauseevidence suggests that the effects of these conditions ondevelopment or psychometric function may differ. Usu-ally, only IDA is associated with lower developmental orcognitive scores, but this is not exclusively so and it is noteasy to differentiate whether the effect on developmentor cognition is due to the severity of the deficiency in iron,the age of onset of deficiency, or the duration of the defi-ciency.

A distinction has also been made between severe, mod-erate, and mild IDA. Where possible, the authors’ defini-tions have been used; otherwise, the mean hematologicvalues within each group as specified in the paper areused. This is because, while there are international crite-ria for defining IDA, (Hb 110g/L for children between 6months and 6 years of age (WHO, 1972), the studieshave not applied these criteria consistently. For example,according to Lozoff et al. (1987), a hemoglobin concen-tration less than 10.0 g/dL (or 100 g/L; to convert con-ventional values in g/dL to SI values in g/L, multiply by10) signifies moderate anemia and greater than or equalto 10.0 g/dL and less than 11.0 g/dL signifies mild ane-mia. On the other hand, according to Deinard et al. (1981),values of hemoglobin greater than 11.0 g/dL and a serumferritin less than or equal to 9 ng/mL (or 9 :g/L; to convertconventional values in ng/mL to SI values in :g/L, multiplyby 1) signify severe ID, and values for hemoglobin greaterthan 11.0 g/dL and serum ferritin between 10 and 19 ng/mL signify mild ID. When a study includes subjects frommore than one category of severity, it is classified ac-cording to the most severely iron-deficient group.


Length of Treatment. The length of treatment has beenspecified because it is important for determining the de-gree of improvement in iron status and the replenish-ment of iron stores. It is supposed that very-short-termtherapy of less than or equal to 10 days (termed “briefsupplementation”) will not result in any significant risein hemoglobin levels but will reduce the level of ID.Longer-term iron supplementation of 2 to 8 months(termed “extended supplementation”) will result in a risein hemoglobin and should therefore correct anemia ifpresent.

Developmental, Cognitive, or Educational Outcome Mea-sures Used. The tests used need to be sensitive to changesin iron status, culturally acceptable, reliable, and validfor the study population. Meeting all these criteria maybe extremely difficult, which may account for the failureof some studies to show improvement in the outcomesmeasured after iron supplementation.

The outcomes used to measure the effects of ID and IDAmust vary according to the age of the children beingstudied. Most studies with infants looked at differenceson the Bayley Scale of Infant Development (BSID), whichprovides indicators of mental development, motor devel-opment, and, indirectly, behavior. By being a global mea-sure of development, BSID is often criticized as beinginsensitive in the context of ID because of its inability toisolate effects on specific functions and its low test-retestreliability in children younger than age 18 months. How-ever, in the absence of good alternative tests of develop-ment, BSID is the most commonly used outcome mea-sure in infants. Thus, its wide usage facilitates compari-son of results across studies. Indeed, when drawing com-parisons across studies, the term “patterning” has beenused to describe situations in which a group consistentlyfails certain items or subtasks of a developmental mea-sure. Studies with preschool children have tended to fo-cus on trying to isolate the specific cognitive functionswhich are affected by ID and IDA. In practice, however,the range of functions which can be measured in any onestudy is limited and the cultural validity of the tests them-selves may be questionable.

Studies with school-age children have tended to mea-sure educational and/or cognitive function. While edu-

cational achievement is an important outcome becauseeffects on school performance have potential to influ-ence policy, it is a global measure and children’s perfor-mance will be heavily influenced by environmental andsocio-emotional conditions at home and at school. Forexample, improvements in educational achievement fol-lowing iron treatment may be difficult to measure if chil-dren are not attending school regularly or if the class-room is not condusive to learning.

How to Use the Bibliography toFind Specific References: TheNumbering System

A numbering system has been used so that the papersdescribed in detail in the main body of the text can beeasily located. Each reference has been assigned a num-ber, which is given each time the reference is listed. Ex-amples of how to use the numbering system to find areference are as follows: If the name of the first author isknown, look in the section Alphabetical List of IncludedStudies (on pages i and ii) and find the reference numberof the paper (given at the end of the citation). A writtensummary of the paper can be found within the text, inwhich papers are discussed in numerical order.

Alternatively, to find, for example, all the observationalstudies conducted with infants several options are avail-able:

(a) for a detailed description, look in the Infants andYoung Children (6–24 months) section of the textunder the subheadings Short-term Effects: Ob-servational Studies (page 8), and Long-term Ef-fects: Observational Studies (page 19);

(b) for a list, Table 1 in the Infants column and theObservational studies row (page 6); or in Ap-pendix A in both the Age and Study design col-umns (page 39).

(c) for a detailed tabulation, studies are listed in Ap-pendix B according to the age of the subjects and,for infants and young children, according towhether the studies looked at short-term or long-term effects (page 59).

Introduction 5

Hematology Developmental Measures Study design

Hb = hemoglobin

ID = iron deficiencywithout anemia

IDA = iron deficiencywith anemia

BSID = Bayley Scale of InfantDevelopment

IBR = Infant Behavior Record(subscale of BSID)

MDI = Mental Development In-dex (subscale of BSID)

PDI = Psychomotor DevelopmentIndex (subscale of BSID)

WISC = Wechsler IntelligenceScale for Children

Difference—used when reporting results; refers to astatistically significant difference (at least p < 0.05)

Intervention—the IDA and ID groups are given ironsupplementation, usually after baseline assessment

Nonanemic control—a nonanemic group is included,usually matched to the IDA or ID group by age as aminimum, and given the same developmental assess-ments; in these studies the nonanemic group is oftengiven the same iron or placebo interventions as theIDA and ID children

Observational—only baseline hematologic and de-velopmental measurements are taken from groupsto be compared

Placebo control—a subgroup of the IDA and IDchildren are given a placebo according to the sameprocedure as the administration of the iron supple-ment

Preventative—anemia in subjects from an at-riskpopulation is prevented early in life with prophy-lactic supplementation, and development of thisgroup is compared with the development of subjectsfrom the same group not given supplements

Randomized—subjects are randomly assigned totreatment and placebo groups

* Studies that report a baseline positive associationbetween hematologic impairment and low develop-mental test scores

† Studies that report an association between iron treat-ment and developmental improvement

Glossary of Terms and Key toAbbreviations

Each citation in the text is annotated at the beginningaccording to its importance, experimental design,baseline results, post-intervention and follow-up results,and conclusions. The other subsection includes detailsof any noteworthy features of the paper (e.g., if the pa-per usefully discusses biological mechanisms or the roleof potentially confounding variables).

A complementary key word section is included witheach annotation. Studies reporting a baseline posi-tive association between hematologic impairmentand low developmental test scores are indicated byan asterisk (*). Studies reporting an association be-tween iron treatment and developmental improve-ment are indicated by a dagger (†). A value of p <0.05 is considered a statistically significant result.

Several terms and abbreviations are used through-out. These are defined as follows:

Appendix C gives a summary of the all the psychometric, development and educational tests which have beenused in studies of ID and IDA.


Table 1. Summary of experimental designs by studies examining the effects of iron deficiency with (IDA) orwithout (ID) anemia on development or cognition

Infants Preschool School age

Observational studies: Deinard et al., 1981 (1) Webb and Oski, 1973performance of IDA/ID Johnson and McGowan, (31,32)compared with IS; cross- 1983 (2)sectional design Grindulis et al., 1986 (3)

Lozoff et al., 1986 (4)

Short-term effects oftreatment for ID/IDA onperformance

• Treatment, no placebo Walter et al., 1983 (9)in IDA group Oski et al., 1983 (10)brief supplementation(5–10 days)

• Extended Lozoff et al., 1987 (13) Pollitt et al., 1978 (27) Seshadri and Gopaldas,supplementation Lozoff et al., (1996a) (16) Pollitt et al., 1983 (28) 1989 (Study 1) (34, 39)(2–8 month) Deinard et al., 1986 (29)

Pollitt et al., 1986 (30)

• Treatment and placebo, Oski and Honig, 1978 (5)brief supplementation Lozoff et al., 1982 (6-8)(5–10 days) Lozoff et al., 1985 (11)

Lozoff et al., 1987 (13)Walter et al., 1989 (14)

• Extended Aukett et al., 1986 (12) Soewondo et al., 1989 (31) Soemantri et al., 1985 (34)supplementation Idjradinata and Pollitt, Pollitt et al., 1985 (35)(2–8 month) 1993 (15) Groner et al., 1986 (37)

Soemantri, 1989 (40)Seshadri and Gopaldas, 1989

(Studies 2,3,4) (34,39)Pollitt et al., 1989 (41)Bruner et al., 1996 (42)Kayshap and Gopaldas,1987

(38)

Long-term effects of ID/IDAon performance

• >2 years follow-up after Palti et al., 1983 (23)treatment Palti et al., 1985 (20)

Walter et al., 1990 (24)Lozoff et al., 1991 (25)Hurtado, 1995 (22)

• Prospective studies (no Wasserman et al., 1992 (21)treatment orsupplementation)

Effects on performance of Cantwell, 1974 (26)preventing IDA and ID Heywood et al., 1989 (17)

Preventative trials (iron Moffatt et al., 1994 (18)supplementation) Lozoff et al., personal

communication, 1996 (19)

IS, iron sufficient. Reference number of each study given in parentheses.

Infants and Young Children (6–24 Months) 7

INFANTS AND

YOUNG CHILDREN

(6–24 MONTHS)

Summary of Findings

There is good evidence that IDA is associated with poorerperformance on developmental ratings of infants, but thedata are inconclusive as to the causal nature of theseeffects. This is because most evidence comes from ob-servations of baseline differences between infants withIDA and nonanemic infants. Despite efforts to carefullycontrol for confounding factors during recruitment and insubsequent statistical analyses, these studies cannot re-move the possibility that another factor related to ID isresponsible for the differences observed.

What is certain from the evidence is that lower develop-mental test scores are most likely seen in infants andyoung children with anemia rather than ID alone; fur-thermore, although precise cutoffs cannot be specified,the general observation is that the more severe the ane-mia, the more pronounced the effect on test scores. Wherelow developmental test scores have been found in chil-dren who are iron deficient without anemia, a small amountof evidence suggests that it is severe deficiency that re-sults in poorer performance.

Infants and young children with IDA often show diffi-culty with language, poor motor coordination and balance,and perhaps most evident, poorer ratings on attention,responsiveness, and mood assessments. In some studies,assessments of poorer attention and responsiveness andlower motor test scores were found in the absence oflower mental development test scores. It has been postu-lated, therefore, that poorer attention, poorer performanceon motor tasks, or both may mediate lower scores ontests of mental development.

Most intervention studies in which oral or intramusculariron treatment was administered for less than 2 weeksused a double-blind, randomized treatment, placebo-con-

trol design. The causal nature of the association betweena change in iron status (without correcting the anemia)and poorer performance on developmental tests was de-termined. However, in none of these studies was a sig-nificant improvement in mental or motor developmentfound that could be attributed to iron. Only in the study byOski and Honig (1978) [Ref 5], was a benefit of treat-ment on behaviour observed.

Studies with infants that looked at the effects on develop-ment of extended iron supplementation have also mostlyfound no significant improvement. The best evidence insupport of a causal link between IDA and developmentcomes from Idjradinata and Pollitt (1993) [Ref 15], whofound a reversal of poor mental and motor developmentaltest scores after 4 months of therapy. There may be manyreasons why most studies have not found an improve-ment in development after treatment of iron-deficient oranemic infants: the anemia may not have been particu-larly severe or chronic; the treatment period was not longenough to correct the anemia; the developmental testswere not sensitive enough; parental stimulation or an im-proved socioeconomic environment may be required inaddition to any treatment for children’s development toprogress; the association was not causal, because someother factor associated with iron but not iron itself mayhave led to the developmental deficits observed at baseline;or because the effects of IDA on development are notreversible with treatment.

Although these reasons are all plausible, it has generallybeen supposed that the failure to find a significant benefi-cial treatment effect is because the effects of anemia ondevelopment in infants and young children are not re-versible. One way to test this hypothesis is to conduct apreventative trial that is double-blind and placebo- con-trolled. If low developmental test scores are the result of


some other factor in the infant’s environment and not aresult of IDA, then preventing anemia from an early ageshould not prevent infants from scoring lower than theirnonanemic counterparts on tests of development. Threepreventative trials have been conducted and different re-sults were obtained (Heywood et al. 1989 [Ref 17],Moffatt et al. 1994 [Ref 18], Lozoff et al. 1996 [Ref 19]).It is therefore difficult to draw any conclusion; however,the evidence seems to suggest that IDA in infants doesaffect mental and motor development directly but onlyunder certain underprivileged conditions - such as in poorhomes or communities, where the prevalence of anemiais high, among low birth weights babies or wherebreastfeeding is uncommon (Moffatt et al. 1994 [Ref 18]).If this is the case, then under these conditions preventionrather than treatment of IDA in infants may offer thegreatest benefit. Further evidence from preventative stud-ies with infants with IDA is needed to clarify this issue.

Investigations of the long-term effects of anemia in in-fancy have generally shown there to be a significant ad-verse effect on development or cognition and educationalachievement in later life. This relationship appears to ex-ist even in those children who, after receiving iron treat-ment in infancy, scored similarly on tests of developmentto nonanemic control children. The long-term effects ondevelopment are seen on tests of motor and mental de-velopment and behaviour. Although infant assessmentscales such as BSID may not predict later intelligencequotients (IQs), evidence suggests that the same childrenperforming less well on these scales during a period ofinfant anemia show poorer IQs at age 5 years and older.In one study, care was taken to assess the extent to whichthe long-term poorer test performance was attributableto confounding variables, such as maternal education,rather than hematologic status earlier in development (Paltiet al. 1983) [Ref 23]. The authors found that at age 5years, the contribution of anemia in earlier childhood tolower developmental scores was significant even afterconfounding variables such as maternal education andsocioeconomic status were controlled for.

There are two excellent reviews of the effects of ID andIDA on infants and children. The first by Lansdown andWharton (1995) includes all age groups and is notable forits clear tabulation of studies cited, its comprehensivereference list, and its systematic discussion of methodol-ogy and issues of theoretical importance. The second

review by Lozoff (1990) deals exclusively with studiesconcerning infants and young children, and starts byclarifying whether any association can be concluded fromthe available evidence before moving on to discusswhether such an association has been shown to be causal.Both reviews discuss the problem of confounding andcofactors and highlight areas of the field yet to be stud-ied.

Short-term Effects:Observational Studies

1. Deinard et al. (1981) Key Words: USA; age15 months; ID of various severities; cross-sectional comparison only; BSID, an habitua-tion measure, and the Uzgiris and HuntOrdinal Scales of Psychological Development;design: observational.

Importance: This study sought to investigate the effectsof ID on development. It is of particular interest given itsassessment of independent effects as a function of thespecific ID level. A specific habituation measure is used,as well as the more global BSID.

Design: An observational design was used to comparethe various ID groups with one another and an nonanemiccontrol group.

Baseline: There was no difference in the overall scoreson any of the measures in any of the iron-deficient groupsand the nonanemic group. However, isolated and pat-terned differences were found in the severely iron-defi-cient group on items in the Infant Behaviour Record (IBR;a subscale of BSID). These children were rated as morefearful, less auditorily and visually attentive, and morevocal. This patterning is similar to that reported in otherstudies (see Oski and Honig 1978 [Ref 5], Lozoff et al.1982a [Ref 8], Walter et al. 1983 [Ref 9], Lozoff et al.1985 [Ref 11], Walter et al. 1989 [Ref 14]).

Conclusions: A number of difficulties with this paperhave been raised, notably the fact that the hematologicmeasures used were unreliable, thus making it possiblethat some of the children described as iron-deficient mayactually have been anemic. The observational design ofthe study means that a causal relation between iron de-ficiency and assessments of poorer behavior cannot be


inferred, although it should be noted that the papershowed a baseline association of ID with lower develop-mental test scores that was restricted to the most severelyiron-deficient children. However, the use of the unreli-able hematocrit measure to determine absence of ane-mia means that the behavioral effects may still resultspecifically from anemia rather than more generally fromiron deficiency.

Other: There is a good discussion of the possible meth-odological and theoretical reasons for their overall non-significant finding and a comment on the involvement ofenvironmental risk.

2. Johnson and McGowan (1983) Study 1.Key Words: USA; 12 months; IDA; mother-child interaction tests of activity, reactivity,emotional tone, and attention span; design:observational.

Importance: This study was conducted for two reasons;first, the authors wanted to continue to test the associa-tion between anemia and the behavior of infants, giventhat “based on previous research, it was hypothesizedthat anemic babies would show disturbed activity levels,be more irritable, less attentive and less responsive totheir mothers.” Second, the authors wanted to comparein Study 1 these behaviours under “low-demand” condi-tions with relatively high-demand conditions in study 2(below). The authors made a particular effort to deter-mine the role of potentially intervening variables. Underlow demand conditions, children’s behavior was observedduring normal play activity with their mothers.

Design: This was an observational comparison of IDAchildren with nonanemic control children.

Baseline: Under these low-demand test conditions, therewere no significant differences between the groups onany of the behavioral measures, even when the authorscompared only the most severely anemic infants with themost clearly nonanemic. Interestingly, this lack of differ-ence between the groups may be explained by the factthat these infants were also not different in maternal edu-cation, family income, birth order, language preference,or scores on the Home Observation for Measurement ofthe Environment (HOME) inventory.

Conclusions: The hypothesis of baseline association be-

tween anemia and behavioral disturbance is not sup-ported, perhaps because, as with the Deinard et al. (1986)[Ref 29] differences in development between ID/IDA andnonanemic children are a result of intergroup differ-ences on intervening variables with which anemia fre-quently co-occurs. Alternatively, the so-called low-de-mand situation may have been such that “latent” differ-ences between the groups were not manifested.

2. Johnson and McGowan (1983) Study 2. KeyWords: USA; 12 months; IDA; BSID (includingIBR); design: observational, double blind.

Importance: In contrast to Study 1, children’s develop-ment and behavior were measured in a relatively highdemand situation in which the BSID was given to chil-dren formally to replicate the test conditions under whichthey are usually assessed.

Design: An observational study comparing IDA infantswith nonanemic control infants. Assessment was con-ducted blind to the infants’ iron status.

Baseline: Under the high demand conditions, no differ-ences between groups were found on MDI, PDI, or IBR.Also, no differences were found between groups formaternal education, family income, birth order, languagepreference, or scores on the HOME inventory.

Conclusions: The demand of the environment cannotbe invoked as an explanation for the lack of intergroupdifferences observed in the first of these two studies; inthis second observational study, the hypothesis of asso-ciation of anemia with behavioral disturbance still receivesno support.

Other: There is good comparison of this with similar stud-ies.

3. Grindulis et al. (1986)* Key Words: UK;age 22 months; IDA; Sheridan developmen-tal sequences for psychomotor develop-ment; design: observational, double blind.

Importance: This study, primarily aimed to investigatethe relation between IDA and vitamin D deficiency, alsoassessed the children’s psychomotor development.

Design: The authors used a double-blind, observationaldesign.


Baseline: Although no association was noted betweenlower psychomotor test scores and vitamin D deficiency,children who were anemic showed significantly poorerperformance on fine motor and social development items.

Conclusions: No conclusions can be drawn regardingthe nature of the association between IDA and psycho-motor development, given the observational design. Therewere baseline differences between the IDA children andthe other groups for SES and maternal education, how-ever, no statistical analysis was conducted to assess thecontribution of these intervening variables to the between-group differences found. The Sheridan Test has not beenused elsewhere.

4. Lozoff et al. (1986)* Key Words: Guate-mala; age 6–24 months; IDA; behavior ratingon BSID; design: observational, double blind.

Importance: This report on the Guatemala study de-scribed below (Lozoff et al. 1982a,b,c) [Refs 6, 7, 8],assessed the behavior of infants and their mothers as afunction of the infant’s iron status. It aimed to determinethe extent to which behavioral differences are manifestedin free play (behavior rating of BSID) rather than in morestressful structured developmental testing environments.The study can be compared with that of Johnson andMcGowan (1983) [Ref 2], which found no differences inmother-child interaction during play between IDA andnonanemic matched control groups.

Design: A double-blind, observational comparison of in-fants with mild to moderate IDA and nonanemic controlinfants and their mothers. No behavioral analysis wasconducted after iron treatment. Subjects were recruitedfrom a socially homogenous, impoverished area of Gua-temala.

Baseline: The infants were not more irritable, distract-ible or apathetic than the nonanemic control infants. Theywere, however, more likely to seek body contact withtheir mother. The mothers did not differ in the extent towhich they initiated contact with and responded to con-tact from their children. The mothers of IDA infants spentless time beyond arm’s length of their children and wereless likely to break close contact. When infants movedaway from their mothers, the mothers of the IDA infantswere more likely to reinitiate the contact.

To determine whether these differences were attribut-able to intergroup differences other than iron status,several other confounding variables were measured. Theinfants did not differ in terms of parental age, education,or occupation or in terms of anthropometry. The IDAinfants came from larger families and had lower scoreson the MDI but not the PDI (differences on the PDI inthe larger sample (Lozoff et al. 1982 a,b,c) were found).The lack of difference between IDA and nonanemic con-trol infants on the PDI is attributed to the smaller size ofthe sample (fewer videotapes of infants were suitable foranalysis). Play behavior remained significantly differentbetween groups after controlling for differences in theirbaseline characteristics.

Conclusions: The authors concluded that infants respondto iron deficiency, as to many other insults, with an in-crease in proximity seeking and that this is both a mani-festation of affective and activity disturbance and an ef-fective compensatory and protective mechanism. In thelong term, increased attachment may interfere with in-fant development because of reduced exploratory behav-ior, which may mediate the acquisition of fundamentalcognitive skills such as depth perception. Although thestudy recruited infants from a socioeconomically, homog-enous population and statistically controlled for other vari-ables, the assessment of SES and other variables mayhave been too crude to detect critical between-group dif-ferences. This possibility cannot be rejected without com-paring play behavior after a placebo-controlled iron treat-ment.

Other: There is good discussion of attachment theory(the perspective on which the hypotheses of the studywere formulated).

Short-term Effects: InterventionTrials

5. Oski and Honig (1978) * H Key Words:USA; IDA; age 9–26 months; 1 week of intra-muscular (IM) therapy; BSID: brief interven-tion, randomized, double blind with placebocontrol

Importance: This was a pioneering study that motivatedmany other researchers to investigate the associationsbetween iron and cognition.


Design: The study was a double-blind, randomized, pla-cebo-controlled intervention study looking at the ef-fects of anemia on cognitive development in infants andyoung children. A nonanemic control group was notincluded. IDA infants and children were randomly as-signed to receive either intramuscular treatment or pla-cebo, which they received for 5–8 days.

Baseline: At baseline, there were no significant differ-ences on either the MDI or the PDI between the treat-ment groups. All anemic children demonstrated lowerscores in fine and gross motor coordination and on IBR.The IBR patterning reported agrees with that seen in anumber of other studies, with poorer ratings in attentionand reactivity (see summary of Deinard et al. 1981[Ref1], for list of studies showing similar behavioral pattern-ing).

Follow-up: A significant beneficial effect of iron treat-ment on MDI scores, motor skills, and responsiveness(an IBR item) was reported after only 1 week but thebenefit in the treatment group was only significantly greaterthan in the placebo group in the IBR rating. It is notewor-thy that a positive correlation was found between initialabsolute hemoglobin (Hb) status and extent of MDI im-provement, consistent with other studies reporting differ-ential baseline poorer performance and differential de-velopmental improvement as a function of severity ofhematologic impairment (see Lozoff et al. 1987 [Ref 13]).

Because other studies have failed to find a treatment ef-fect at 1 week, it has been suggested elsewhere that thetreatment effect here is due to the treatment having beenintramuscular rather than oral. This suggestion is contra-dicted by Lozoff et al.’s 1987 study [Ref 13], describedbelow.

Conclusions: Because no nonanemic control group wasincluded, the study cannot confirm that lower develop-mental test scores at baseline are to be found in anemicchildren. Furthermore, despite the significant MDI scoreimprovement for iron-treated children, the unique effectof iron therapy on cognition cannot be concluded. Oskiand Honig state that although there was a significant im-provement in the MDI scores of the iron-treated IDAchildren, this improvement was not statistically greaterthan the improvement in the placebo-treated group.The improvement in scores in the treatment group canbe attributed to the effect of repeating the same test within

a short time. The study confirms the baseline associa-tion of IDA with lower PDI scores and poorer IBR as-sessments but provides no evidence that the IDA infants’performance was poorer than that of nonanemic infants.The study offers no confirmation of the contention thatiron supplementation is beneficial to performance on themeasures used.

Other: There is very good discussion of biochemical theo-ries of the effects of iron deficiency on cognition and alsoof possible mechanisms associated with IBR and PDImeasures that may mediate poor cognitive performance.

6. Lozoff et al. (1982c)* Key Words:Guatemala; age 6–24 months; IDA; 1 week oftherapy; BSID; design: brief intervention,randomized, double blind with placebo andnonanemic controls.

Importance: The authors aims were to assess the effects ofIDA on development, using a controlled and random-ized placebo controlled design to enable firmer conclu-sions to be drawn than were previously possible.

Design: A double-blind, randomized, placebo-con-trolled intervention was used in IDA and nonanemiccontrol children. Although the authors describe the IDAsubjects as mildly anemic, at least some of them had Hblevels below 10.0 g/dL; in the Lozoff et al. 1987 paper[Ref 13], children with Hb below 10.0 g/dL were classi-fied as moderately anemic.

Baseline: At baseline, the IDA children scored signifi-cantly lower than the nonanemic control children on MDIand nonsignificantly lower on PDI.

Follow-up: After treatment, Hb levels in the treated IDAchildren improved significantly, but the developmentalscores of these children showed no effect of treatment.All groups showed an improvement in MDI score, andthe IDA group’s scores remained significantly lower thanthose of the nonanemic control group.

Possible explanations for this failure of developmentalmeasures to improve are discussed extensively. It is alsosuggested that, given the range of Hb levels within theanemic group, the anemia of the less severely anemicchildren had more reversible effects and that their re-sponse to therapy was concealed in these data. The real-ity of this suggestion cannot be assessed.


Conclusions: Thus again the baseline association of IDAwith lower cognitive test scores is confirmed, but there isno evidence of a causal association, given the failure oftherapy to improve performance. If the association iscausal, then either longer treatment is needed or the ef-fects are irreversible. In support of a causal hypothesis,the authors did go to some trouble to assess the involve-ment of potentially confounding variables and found thatthe experimental groups were not different in birth his-tory, SES, or general nutritional status. Maternal educa-tion was not assessed.

Other: There is reference to animal studies and how thesemay help to explain a failure to improve development af-ter therapy.

7. Lozoff et al. (1982b)* Key Words:Guatemala; age 6–24 months; IDA; 1 week oftherapy; BSID; design: brief intervention,randomized, double blind with placebo andnonanemic controls.

Importance: This second report of the Lozoff et al. 1982c[Ref 6] study described above is important in that it de-tails analysis performed to determine whether lower de-velopmental test scores and therapy responses are thesame across age subgroups within their original sample,which spanned a large age range.

Baseline: At baseline, lower PDI scores were presentand the same across all age subgroups, however, thepoorer performance on MDI only emerged in the 19–24months group (the oldest subgroup). Furthermore, andvery importantly, the authors observed that within thissubgroup the more severe the anemia, the poorer theperformance on the developmental tests.

Further analysis revealed patterning on the developmen-tal scales: the anemic infants had disproportionate diffi-culty with language items, a trend observed by Walter etal. 1989 [Ref 14], and the 19–24 month-old childrenconsistently failed 11 of the 12 items taken to predictlater IQ.

Follow-up: No treatment effect, specific to an age sub-group, was found within PDI. All infants with lowerscores failed to improve after one week of treatment.The iron-treated infants with lower scores at baseline onMDI also did not improve after treatment.

Conclusions: There are two interesting implications ofthis study: First, the results agree with the observation byLozoff et al. 1987 [Ref 13] in that PDI scores will sufferwith less severe anemia. An interesting implication is thatpoorer motor performance may mediate poorer cognitiveperformance, having occurred earlier during the less se-vere stages of anemia. Second, either the absolute timingof the hematologic deficit or the duration and severity ofthe anemia seem to be crucial factors in determiningwhether children will show lower cognitive test scores.The older IDA children showed signs of being under-nourished, suggesting that their anemia may have lastedlonger. Comparing this study with the Lozoff et al. 1987[Ref 13] study, where all the subjects were of an agesimilar to the differentially lower-scoring subgroup de-scribed here and only the moderately anemic childrenshowed poorer performance on MDI, the suggestionseems to be that it is the nature of the anemia rather thanthe absolute age of the child that is important.

8. Lozoff at al. (1982a)* Key Words: Guate-mala; age 6–24 months; IDA; 1 week of therapy;BSID; design: brief intervention, randomized,double blind with placebo and nonanemiccontrols.

Importance: This is the third report of the Lozoff et al.1982c study [Ref 6]. The report deserves particular at-tention because the IBR ratings of the subjects are dis-cussed.

Baseline: The IDA subjects were more fearful, showedincreased body tension and decreased gross body move-ment, were less responsive to the examiner, were lessreactive to ordinary stimuli, and tended to be less persis-tent. These patterns were primarily in the 19–24 monthage group, and they are very similar to the behavioraleffects reported elsewhere (see the review by Deinardet al. 1981 [Ref 1]).

Follow-up: These behavioral differences disappearedafter treatment except for the elevated fearfulness. Thechange could not be attributed to treatment, however,because iron- and placebo-treated infants showed thesame reversal.

Conclusions: The study demonstrated an adverse effectof IDA on infant behaviour, but in contrast to the study


by Oski and Honig (1978 ) [Ref 5] where parenteral ironwas given, the effect was not reversible after short-termoral iron treatment. At this time, they suggest the differ-ent results might be because the response times to oraland parenteral iron therapy differ. The possibility thatthe observed differences in behaviour at baseline weredue not to iron deficiency but to some other interveningvariable is rejected because there were no characteristicswhich differed between the groups which could explainthe results.

Other: The authors offer a short, clear discussion of thepossibly mediating relation between behavioral effectsand poorer performance on test of development.

9. Walter et al. (1983) * † Key Words:Chile; age 15 months; IDA and ID; 10 daysof therapy; BSID; design: brief interven-tion, double blind, with nonanemic control.

Importance: This study is important in its direct com-parison of the effects of IDA with those of ID.

Design: A double-blind iron-treatment-only interventionwas used in IDA, ID, and nonanemic control children.

Baseline: The IDA children scored significantly loweron MDI and were rated as more unhappy than thenonanemic control group; the IDA children did not scorelower on PDI. Although the authors classify the anemiain the IDA children as mild, this classification is possiblyincorrect because at least some of the children had Hblevels below 10.0 g/dL. There were no significant differ-ences in MDI or PDI at baseline between children withID and non anemic controls.

Follow-up: After 10 days of iron treatment, IDA infantsimproved significantly on MDI, and this improvementcoincided with an improvement in cooperativeness andattention span. No similar improvement was found in thenonanemic group, and improvement was only seen in IDAchildren with two or more abnormal biochemical mea-sures of hematology at baseline. None of the groupsshowed improvement on PDI.

Conclusions: This study confirms the baseline associa-tion of impaired hematology and lower developmentalscores and suggests that IDA, but not ID, will result indetectably lower developmental scores. This contradicts

the findings of Deinard et al. (1981) [Ref 1]. It may besuggested that ID was shown to be associated with poorerperformance on tests of development, because infantswith two or more abnormal biochemical measures im-proved in developmental measures even though they didnot show relatively lower scores at baseline.

If these data are valid, the implications of the scorechanges, after only 1 week of treatment, are that thelower scores are attributable to iron deficiency ratherthan specifically to anemia and that the locus of the ef-fect is probably at least attentional or behavioral. Thislatter suggestion receives support from the coincidentbehavioral effects and the behavioral response to therapy(see the review by Deinard et al.(1981) [Ref 1]) for a listof papers reporting similar IBR patterning).

No placebo group was included in the study design. Theshort-term improvement observed may be attributed to apractice effect rather than a therapeutic effect. The lackof improvement in the control group may be due to aceiling effect. Because no follow-up assessment of he-matology was performed, it is impossible to conclude thatiron intervention and subsequent hematologic correctionwas responsible for the developmental improvements.

Therefore, although a baseline association between IDAand lower developmental test scores (and possibly be-tween ID and developmental test scores) is confirmed,the unique effect of iron treatment is not shown.

10. Oski et al. (1983) † Key Words: USA;age 9–12 months; ID; 7 days of therapy; BSID;design: brief intervention, double blind, withnonanemic control.

Importance: Because of Oski and Honig’s 1978 study [Ref5], which demonstrated a causal relation between IDAand lower cognitive test scores, the authors conductedthis study to assess the extent to which ID is also associ-ated with these lower scores. The results should be com-pared with those of Deinard et al. (1981) [Ref 29] andthose of Walter et al. (1983) [Ref 9], who directly com-pared the effects of IDA and ID.

Design: The study was a double-blind intervention, withiron-depleted and nonanemic ID subjects, matched forcolor and sex, and who received IM treatment for only 7days. There was no placebo control.


Baseline: At baseline, there were no significant differ-ences between the groups, suggesting that neither irondepletion nor ID is sufficient to cause infants to scorelower on developmental tests than nonanemic controlsubjects. This contradicts Walter et al.’s (1983) [Ref 9]findings of a baseline association but agrees with Walteret al.’s negative finding for ID children.

Follow-up: After 7 days of treatment, however, MDIscores for ID children improved significantly and sug-gested, that this group had been performing suboptimallyat baseline. Because no placebo group was included, it ispossible that this improvement was due to a practice ef-fect.

The authors note that their findings regarding the short-term effects of iron treatment disagree with similar stud-ies by Lozoff et al. (1982b,c) [Refs 7,6] in Guatemala.They suggest that their positive treatment effect afteronly 1 week may have been attributable to the use ofparenteral iron administration. As discussed above, thispossibility was subsequently rejected by Lozoff et al.(1987) [Ref 13]. In the Lozoff et al. (1982) [Refs 6-8]studies, the children were ID rather than IDA.

Conclusions: The implication of this study is that treat-ment of ID may lead to improved performance on MDI.

Other: There is a good discussion of biochemicalmechanisms.

11. Lozoff et al. (1985) * Key Words:Guatemala; 6–24 months; IDA; IBR of BSID(focusing on Test Affect and Task Orientationfactors); design: brief intervention, double blind,randomized, with placebo and nonanemiccontrols

Importance: This is a fifth report on the study conductedin Guatemala by Lozoff and collaborators. This reportexplores the association between poor developmental testscores and abnormal behaviors documented in childrenwith IDA. Specifically, the authors investigated whetherpoor developmental test performance in infants withIDA may be mediated by disturbances in affective be-havior.

Design: This study was a double-blind, intervention trialcomparing IDA infants with nonanemic control infants

recruited from a socioeconomically homogenous, im-poverished area of Guatemala. The infants in this studywere randomly assigned to receive treatment or placebo,although the analysis documented in this report focusedmainly on baseline data. Infants were assessed on themental and motor scales of the BSID as well as the IBR,and the extent to which individuals manifesting abnor-mal behaviors, with lower MDI or PDI scores, was de-termined.

Baseline: A greater proportion of anemic thannonanemic infants showed abnormal affective responsesto testing; they were more fearful, tense, restless, with-drawn from the examiner, and unhappy. The infantsdemonstrating abnormal affective behaviors were alsothe infants with profoundly lowered MDI scores; ane-mic infants who did not rate abnormally on the IBR didnot show significantly lower MDI scores.

Despite the age effect described by Lozoff et al. (1982b)[Ref 7] for these infants, affectively abnormal infants inall age groups tended to receive low MDI scores. Lozoffet al. describe a severity effect such that the greater thenumber of abnormal affective behaviors demonstratedby the infant, the lower the MDI score.

Lower PDI scores were found in IDA infants demon-strating abnormal affect, but this was completely accountedfor by the infants who also demonstrated abnormal taskorientation. Interestingly, although Hb level was directlyrelated to the extent of behavioral disturbance, it was notsimilarly directly related to MDI score

Follow-up: Because of the small number of infants withabnormal behavior who also received iron treatment, aconclusion cannot be drawn from this study regarding theeffects of short-term oral iron therapy on behavior. Theindication is that there is no treatment effect after 1 weekand that repetition of testing leads to improved behaviorin the affectively disturbed IDA infants and a greaterimprovement than found in the affectively disturbednonanemic control subjects. Also interesting is that theinfants who improved their IBR ratings also improvedtheir MDI scores. No such improvement was found inPDI scores.

Conclusion: This study provides strong evidence thatIDA is associated with IBR abnormalities and lower de-velopmental test scores . Furthermore, the close covariation


of abnormal behaviors and lower test scores, the directrelation between abnormal behavior and lower Hb lev-els, and the differential MDI improvements after 1 weekof treatment indicate that behavioral disturbances maymediate poor developmental performance. Althoughstatistical investigation revealed no involvement of po-tentially intervening variables, without finding a treat-ment effect it cannot be concluded that abnormal be-havior is caused by anemia.

Other: There is good discussion of the similarities be-tween the findings of this study and of others (principallythose studies listed in the review of Deinard et al. (1981)[Ref 1]).

12. Aukett et al. (1986)* † Key Words: UK;age 17–19 months; IDA; 2 months of therapy;Denver developmental screening test forpsychomotor development; design: extendedintervention, randomized, double blind, withplacebo control.

Importance: The study assessed whether the well-repli-cated baseline association of IDA with lower develop-mental test scores was causal.

Design: A double-blind intervention design with placebocontrol was used. IDA children were treated for 2 months.

Baseline: No nonanemic control group was included andbaseline scores on the Denver Developmental Test arenot reported so this study cannot confirm the hypothesisthat IDA infants show poorer test performance than dotheir nonanemic counterparts.

Follow-up: After 2 months of treatment, 58% of thechildren with hematologic improvement of greater than 2g/dL did not increase their developmental score by theexpected amount. However, more of the children whoshowed hematologic improvement increased their scoresby the expected amount than did children who did notshow such hematologic improvement.

Conclusions: Although the study provides some evidencethat the association between IDA and cognitive devel-opment is causal, the evidence is not conclusive becauseneither the developmental test nor the use of a 2 g/dLresponse as a criterion for effective treatment has beenused in any other study.

Other: This study also considers weight velocity as anoutcome measure, and there is a very useful table in-cluded that compares this study with previous studies.

13. Lozoff et al. (1987)* † Key Words: CostaRica; age 12–23 months; IDA, ID, and irondepletion; 1 week and 3 months of therapy;BSID; design: brief and extended periodintervention, randomized, double blind,partially placebo controlled, with nonanemiccontrol.

Importance: This is an important study because it inves-tigates how much the severity of anemia alters the devel-opmental outcome. In particular, the results should becompared with those of Walter et al. (1989) [Ref 14],who also investigated the existence of a differential de-velopmental effect as a function of the specific level ofanemia. Other studies described above have also com-pared anemia with ID and with various severities of ane-mia (e.g., see Deinard et al. (1981) [Ref 1] and Walter etal. (1983) [Ref 9]). This paper is also important for twoother reasons. First, it compares IM with oral irontherapy. Oski et al. (1983) [Ref 10] found a treatmenteffect after 1 week with IM therapy, in comparison withstudies such as that by Lozoff et al. (1982c) [Ref 6] whereno oral iron effect was found after the same amount oftime. Oski et al. (1983) [Ref 10] suggested that this dif-ference may have been due to differing methods of ironadministration. Second, the authors undertook to as-sess the role of coexisting variables, such as low SES, inthe etiology of the lower developmental test scores asso-ciated with IDA.

Design: The study was a double-blind intervention, withIDA, ID, iron depletion, and nonamemic control childrenrandomly assigned to receive treatment or placebo forthe first week, with some children assigned to IM treat-ment and others to oral treatment. After the first week,all previously IM-treated children and all nonanemic con-trol children were given a placebo for 12 weeks. All pre-viously iron-deficient children given oral iron were con-tinued on oral iron, with no placebo-control, for thesame time.

Baseline: At baseline, only the IDA children demon-strated poorer developmental test performance. Withinthis group, lower PDI scores were found in the entire


group (Hb 10.5 g/dL), regardless of the severity of theanemia. However, lower MDI scores were only to befound in the moderately anemic children (Hb 10.0 g/dL).

There was patterning on the MDI and PDI scales com-parable with that found by Walter et al. (1989) [Ref 14].IDA children consistently failed the showing-shoes item;in the 12–14-month age group, the IDA children hadparticular difficulty walking alone, standing from sitting orbeing supine, and standing on left foot.

Follow-up: After 1 week of treatment, there were nodifferences between the orally and parenterally treatedchildren, in contrast to Oski et al. (1983) [Ref 10]. After1 week, all children improved on MDI scores regardlessof their iron status or whether they received iron or pla-cebo.

After 12 weeks of therapy, the Hb level of most of themoderately anemic children was not corrected, and thesechildren did not improve on either MDI or PDI. The chil-dren whose Hb level was corrected (primarily the mildlyanemic children) had no MDI improvement, but therewas a PDI improvement.

Conclusions: The study provides evidence that com-pared to nonanemic controls, children with mild anemia(Hb 10.5 g/dl) will perform poorly on tests of psychomo-tor development whereas more severe or longer-termanemia (Hb 10.0 g/dl) is required before effects on men-tal functioning will be seen. This may point to a mediatingrole of psychomotor disturbances in cognitive develop-ment; IM iron does not differ from oral iron treatment inits effects after 1 week; improvements after 1 week areprobably attributable to a practice effect; whereas a ben-eficial hematologic and developmental effect of longertherapy relies on the initial anemia being mild; when ane-mia is moderate, mental functioning as well as psycho-motor functioning is affected and does not seem to bereversible.

Unfortunately, because of the lack of an appropriate pla-cebo control, after the first week, for comparisons of theperformance of the IDA orally treated infants and be-cause no intergroup comparisons are reported, theunique effect of iron treatment still cannot be concludedfrom this study.

Other: The paper offers a good discussion of the relatedfindings of other authors.

14. Walter et al. (1989) * Key Words: Chile;ages 12 and 15 months; IDA and ID; 3 monthsof therapy; BSID; design: brief and extendedintervention, double blind, with partial placebocontrol and nonanemic controls.

Importance: This study is to be compared with those ofLozoff et al. (1987 and 1982b) [Refs 13 & 7] because itinvestigates the existence of a differential effect on de-velopment as a function of the specific severity or dura-tion of anemia. Furthermore, certain specific areas ofdifficulty on MDI and PDI reported elsewhere are repli-cated and, again, there is clear evidence of IBR pattern-ing.

Design: IDA, ID, and nonanemic control children wereassigned to receive treatment or placebo for the first 10days, after which time all children received iron treat-ment for 3 months. Hematologic assessment was at ages9 and 12 months and the first developmental assessmentwas at 12 months. Treatment then recommenced for 3months and the final developmental assessment was at15 months. All developmental assessments were reviewedas a function of hematology at 9 months, thus it was pos-sible to distinguish children who had been anemic for atleast 3 months from those who had become anemic inthe last 3 months.

Baseline: At the first developmental assessment, IDAchildren had lower MDI and PDI scores than did thenonanemic control children, with no similar lowering ofscores evident in the ID children. Specifically, infants haddifficulty on items requiring language comprehension (asin Lozoff et al. (1982b) [Ref 7]); vocalization of bisyllabicwords; showing shoes, clothing, or own toy; sitting fromstanding; standing and walking alone; standing from sit-ting; and standing on the left foot (similar to Lozoff etal. (1987) [Ref 13]). The IDA infants also differed fromthe others on responsiveness to examiner, other people,and mother; general emotional tone; test affect and taskorientation; goal directedness; attention span; activity;vocalization; and body motion. This patterning was simi-lar to that found in the studies listed in the review byDeinard et al. (1981).

Importantly, the authors found that MDI and PDI scoreswere lower for all children with Hb less than 10.9 g/dL,and that the children with the lowest Hb values also had


the lowest MDI scores, which were also lower than thoseof other less anemic children. Children who had beenanemic for at least 3 months were generally more se-verely anemic and were also the children whose testscores were affected the most. The hemoglobin level atwhich effects on mental development are observed areremarkably similar between this study (Hb 10.5 to 10.9g/dL for MDI and PDI g/dL) and that of Lozoff et al.(1987) [Ref 13] where effects on MDI and PDI were ob-served in children whose Hb was less than 10.0 g/dL and10.5 g/dL, respectively.

Follow-up: As in the Lozoff et al. (1987) study [Ref 13]and others, after only 10 days of treatment, all groupshad improved their MDI and PDI scores. After 3 monthsof treatment, there were no changes in any of thechildren’s BSID scores (including IBR), with IDA chil-dren still scoring low on the language and psychomotoritems listed above.

Conclusions: The baseline association of iron and devel-opment was again confirmed, with further support forthe view that the size of the difference in test scoresbetween IDA and nonanemic control children reflectsthe duration or severity of the anemia. However, the causalhypothesis receives no support from this study perhapsbecause the association between IDA and developmentis not causal, or because 3 months of therapy is not longenough, or because the effect is irreversible.

15. Idjradinata and Pollitt (1993) * † KeyWords: Indonesia; age 12–18 months; IDA andID; 4 months of therapy; BSID; design: extendedintervention, randomized, double blind, withplacebo and nonanemic controls.

Importance: This is a very important study because it isthe first to use a placebo-control for an extended periodof treatment.

Design: This was a double-blind intervention trial withIDA, ID, and nonanemic control subjects randomly as-signed to receive treatment or placebo for 4 months.The authors made a concerted effort to match their sub-jects for confounding variables.

Baseline: At baseline, the well-documented associationbetween IDA (not ID) and lower scores on tests of devel-opment was replicated, with IDA children scoring lower

than ID and nonanemic control children on MDI andPDI.

Follow-up: After 4 months of treatment, the iron-treatedIDA children improved significantly more than thenonanemic group in MDI and PDI. Baseline differencesbetween these groups were eliminated.

Conclusions: The hypothesis that the association be-tween iron and development is causal receives strongsupport and, again, it seems that the only detectable ef-fect is to be found in IDA and not ID children. This is thefirst study to show that pooerer PDI and MDI test per-formance in infants is reversible following 4 months ofiron treatment. There is a need for the results to be repli-cated.

Other: The authors offer a short discussion of the waysin which poorer psychomotor performance may mediatepoorer cognitive development.

16. Lozoff et al., (1996b). * Key Words: Costa-Rica; age 12–23 months; IDA; 6 months oftherapy; BSID & IBR; design: extended double-blind intervention, nonanemic controls ran-domly assigned treatment or placebo; IDAtreated.

Importance: This study was conducted to replicate thefindings of Idjradinata and Pollitt (1993) [Ref 15]. How-ever, for ethical reasons, the authors did not randomlyassign the IDA group to treatment or placebo. There-fore, the two study designs are not directly comparable.

Design: This was a double-blind intervention trial in whichthe IDA group received oral iron treatment for 6 monthsand the nonanemic controls were randomly assigned totreatment or placebo. BSID and IBR were assessed atbaseline and repeated 3 and 6 months after the start oftreatment.

Baseline: IDA infants scored significantly lower (mean6.1 points) than the nonanemic controls on the MDI.There was no significant difference on the PDI. On theIBR, IDA infants were significantly more fearful, un-happy and hesitant with the examiner.

Follow-up: Children who received treatment showed agood hematological response. The differences observedat baseline in MDI between IDA and nonanemic con-


trols were still present at the 3 and 6 months follow-up.There was no change in PDI. The IBR of IDA childrenimproved significantly more than that of the nonanemiccontrols such that the initial differences were no longerpresent after 6 months of treatment.

Conclusions: IDA children scored significantly worse atbaseline on the IBR rating and MDI but not PDI. Irontreatment did not significantly improve the MDI of chil-dren but it did lead to a signficant improvement in the IBRrating. However, in this study, the improvement in IBRcould not be considered a direct result of treatment be-cause there was no IDA placebo group with which tocompare results.

Other: The paper provides a good discussion of the ef-fects of IDA on MDI, PDI and IBR and gives severalhypothesises as to why benefits from treatment are notconsistently observed.

Short-term Effects: PreventativeTrials

17. Heywood et al. (1989) † Key Words:Papua New Guinea; age 1 year; IDA; oneiron dextran injection at 2 months; totaland mean fixation time, habituation, anddishabituation tested; design: extendedintervention, randomized, double blind,with placebo control.

Importance: The primary importance of this study is thatit is the first trial to investigate the benefits to infants frompreventing IDA. Preventative trials permit the investiga-tion of the causality of the relationship between IDA anddevelopment in circumstances where benefits of iron treat-ment have not been observed. The study also measuredspecific functions within attention rather than more globalmeasures of development.

Design: This study was a double-blind randomized pre-ventative trial of iron supplementation and placebo inIDA infants, with no nonanemic control group included.At age 2 months, the children were randomly assignedto receive IM either iron dextran or placebo, and theeffects of this supplementation on hematology and at-tention were assessed at 1 year.

The number of fixations and total fixation time wereassessed, as well as rates of habituation and disha-bituation. A higher rate of habituation is thought toreflect the rate at which an internal model of a stimulusis acquired. Dishabituation rates are thought to mea-sure schema acquisition.

Follow-up: The results of this study are confused by thepresence of malaria infection, which interacted with ironstatus in the treatment group when attention scores wereassessed. The main result of this study was thataparasitemic IDA infants who received iron treatmenthad higher total fixation times than did aparasitaemic IDAplacebo-treated infants, suggesting that iron treatmentleads to longer visual exploration of stimuli. Iron treat-ment did not eliminate anemia, although it did significantlyimprove the iron status of the treated children in theaparasitemic group. There was no evidence of a relation-ship between socioeconomic variables and this measureof attention.

Conclusion: There is some evidence that preventing IDAin infants can result in a beneficial effect on attention.However, the results are confounded by the concurrenteffects of malaria on the iron status of children in thisstudy. The results are interesting in the light of reportsfrom other studies, using more global measures, of diffi-culties on attentional items (Review by Deinard et al.(1976) ).

18. Moffatt et al. (1994)† Key Words:Canada; 2–15 months; at risk of IDA; 13months of iron-fortified infant formula; BSID;design: preventative trial, double-blind, ran-domized, with placebo-control.

Importance: This was a well designed preventative trialwhere children’s iron status was manipulated from age 2months. All children came from the same disadvantagedbackground and did not differ sociodemographically or inthe amount of stimulation available to them as measuredby the HOME Inventory. An association was demon-strated between iron deficiency and PDI without any ef-fect on cognition. Thus, it is possible that motor distur-bance precedes and may cause cognitive disturbance.

Design: This was a double-blind, randomized, prospec-tive iron status manipulation, where infants at age 6 monthswere randomly assigned to receive either iron fortified


infant formula or regular formula. The fortification wascontinued until age 15 months, with hematologic and de-velopmental assessments at ages 6, 9, 12, and 15 months.

Results: There were significant hematologic differencesbetween groups at all ages. No difference on MDI orIBR was found between the groups at any age. At ages6 and 15 months, there were no intergroup PDI differ-ences, but at ages 9 and 12 months there was a decline inthe PDI scores of the regular-formula group, making theirscores significantly lower than those of the iron fortified-formula group.

Conclusions: The prospective nature of this study andthe randomizing procedure indicate that strong support isleant to the hypothesis that the association between IDAand motor disturbance is causal. The lack of differencebetween these groups on potentially intervening variablesalso supports this hypothesis. The fact that, at age 15months, the 9- and 12-month PDI score differences haddisappeared suggests that the developmental effect as-sociated with IDA may be transitory.

19. Lozoff et al. (personal communication,1996a) Key Words: Chile; 6–12 months; IDA; 6months of supplemental iron; BSID; design:preventative trial, double blind, randomized,with placebo control.

Importance: This abstract describes another preventa-tive trial that, like Moffet et al. (1994) [Ref 18], used astrong experimental design to investigate the nature ofthe relationship between iron and cognition in infants. Thestudies are comparable except that Moffatt et al. (1994)[Ref 18] looked at infants in a very impoverished area ofCanada, whereas in this study, the infants were from lessimpoverished homes in Chile.

Design: This was a randomized, double-blind, placebo-controlled preventative trial, in which 944 healthy Chilean6-month-old infants, with normal Hb levels from simi-lar socio-economic backgrounds and with similar an-thropometric status, were randomly assigned to receivesupplemental iron or no added iron until age 12 months.At age 12 months, the infants’ hematologic status wasassessed for a second time and all infants completed theBSID.

Results: At age 12 months, fewer of the iron supplementedchildren were anemic and fewer were iron deficient.However, the supplemented children did not have higherBSID scores.

Conclusions: There is no conclusive evidence of a benefitto children’s development resulting from preventingIDA. These results are in contrast to those of Moffatt etal. (1994) [Ref 18] where the population was more socio-economically disadvantaged. Lozoff et al. state that thelower BSID scores observed in IDA infants may resultfrom factors other than IDA, including environmentaldisadvantage, lack of breast-feeding, birth weight of lessthan 3 kg, IDA onset before 6 months or after 12 months,or IDA lasting longer than 6 months.

A failure to find a protective advantage of prophylacticiron treatment may be only be in situations where infantsare already at risk of other biological, social, or environ-mental stressors that are manifested by ID or IDA.

Long-term Effects: ObservationalStudies

20. Palti et al. (1985) * Key Words: Israel;children 0 to 10–13 years; questionnaire basedon the Shaefer Classroom Behavior Inventory,supplemented with learning achievement items;design: observational, double blind.

Importance: This study assessed the effects of IDA experi-enced at age 9 months on the educational achievementand behavior of school children aged 10–13 years. Be-cause this was a longitudinal prospective study and nota preventative trial, the subjects who were anemic at age9 months are very likely to have differed from thenonanemic control group for other variables other thaniron status. However, as with the Palti et al. (1983) study[Ref 23], care was taken to assess the extent to whichany differences in test scores were attributable to inter-vening variables. Because the children were monitoredyearly after the first year of life, the possibility is greatlyreduced that educational achievement and behavior atschool is a function of some intervening period of anemiarather than anemia in the first year.

Design: Although infants with Hb less than 11 g/dL atage 9 months received iron treatment to correct their ane-


mia, this study investigated the association between ironstatus before treatment and school performance andbehavior at school. It did not assess differences betweentreated and untreated children when they were infantsor when they were of school-age. Therefore the study isclassed as observational.

Baseline: After maternal education and the sex of thechild were controlled for, IDA children scored lower thannonanemic control children on educational achievementand on positive task orientation, but there was no effecton negative task orientation or on mood. Maternal educa-tion and the sex of the child also contributed to learningachievement, task orientation, and mood score.

Conclusions: The authors conclude that although ma-ternal education is still the single best predictor of thechild’s achievement on development and intelligence tests,IDA in infancy is also significantly associated with poorerperformance in the school years. The hypothesis that ironstatus in infancy is associated with long-term effects onschool performance and behavior receives support. Thatthe association is causal is unclear for two reasons. First,although the involvement of potentially intervening vari-ables was assessed, the choice of maternal education asthe only indicator of SES may have been inadequate.Second, there was no analysis of the results of giving irontreatment.

Other: Biochemical mechanisms, the findings of studiesof IDA in animals, and clinical and public health implica-tions of the findings to date are discussed.

21. Wasserman et al. (1992)* Key Words:Yugoslavia; age 2 years; Hb levels, someIDA; prospective study; MDI subscale ofBSID; design: observational.

Importance: One of the key features of this study is itscomprehensive assessment of the role of various poten-tially confounding variables (including lead exposure) inthe development of infants and young children.

Design: In this prospective study MDI scores, lead mea-surements, and Hb measurements were taken at ages 6,12, 18, and 24 months. The number of children anemic ateach time and the extent of their anemia are not stated,but the authors noted that Hb values varied from below9.0 g/dL to greater than 13.0 g/dL. The study assessed

the extent to which a 2 g/dL decrease in Hb at each timepredicted the MDI decrement observed at 24 months.

Results: Ethnic group and HOME status accounted for agreat deal of the variance in MDI, but even after control-ling for these variables, the authors found that at 18 months,a 2 g/dL decrease in Hb was a statistically significantpredictor of 24-month MDI. Moreover, the authors re-port that Hb concentrations measured at ages 12 and 18months were better predictors of MDI decrements at 24months than was the Hb concentration at 24 months.This is interesting given the number of studies that con-centrate on assessing development as a function of cur-rent Hb levels.

Conclusions: This study implies that decreases in Hbare associated with decreases in MDI (although the ex-tent of the anemia in the children showing these decreasesis not given) and that this association is causal, given thecare taken to control for confounding variables. Further-more, decreases in Hb levels up to a year before testingwere better predictors of MDI decrements than wereHb decreases at the time of testing.

More information about the Hb values for children whoseHb decreases predicted MDI decreases is needed be-fore a definite statement can be made about the nature ofthe association being observed.

Other: In general this paper is confusing in its report ofthe findings; in particular it is unclear about the mean Hblevels at each time, about when the children became ane-mic, and about the uniformity and extent of the decline inHb.

22. Hurtado (1995)* Key Words: USA; 0–10years; ID in infancy; children assessed with testof cognitive development; design: observational.

Importance: This abstract reports on a study where adose-response relationship between hemoglobin and testscore and an interaction between test score, iron statusand maternal education was observed.

Design: Subjects were enrolled at birth in a nutritionalsupplementation program and followed until age 10 years.Children’s cognitive performance was tested at age 10years to determine differences in development betweenchildren who were ID and children who were iron suffi-


cient in infancy and early childhood. The precise designof the study is not given; it is not clear whether all chil-dren were supplemented or whether there was a placebogroup. No report of a randomization procedure is madeand no information on the SES of the subjects is given.

Follow-up: At age 10 years, children were assessed with atest of cognitive performance. No levels of statistical sig-nificance are reported in the abstract, but the authorstates that there were indications of an increased risk ofdisability resulting from ID, especially when it was se-vere, and that additive interactions with maternal educa-tion were also observed.

Conclusions: This is a potentially crucial study, but moreinformation is needed on the exact design and the dataanalysis. Without this information, it is still possible thatthe children showing disability at age 10 years did so as aresult of factors other than the presence of ID. It is un-clear whether the children in this study had ID or IDA.They are described as ID, but the dose-response relationreported is “between hemoglobin and disability.”

Long-term Effects: Follow-up ofIntervention Trials

23. Palti et al. (1983)* Key Words: Israel;ages 2, 3, and 5 years; IDA and intermedi-ate at 9 months; 3 months of therapy;Brunet-Lezine’s Developmental Quotient,MILI IQ test, and Wechsler Scale; design:follow-up of previous intervention trial.

Importance: The authors distinguished between childrenwho were moderately anemic and children who weremildly anemic. The long-term outcomes for these twogroups of anemic children are of particular interest, giventhe studies of Lozoff et al. (1987) [Ref 13] and Walter etal. (1989) [Ref 14].

Design: Developmental assessments were carried outat ages 2, 3, and 5 years as a function of Hb status at age9 months. Children who had been anemic at age 9months were given iron treatment, with no placebo con-trol, whereas the intermediate and nonanemic childrenreceived no treatment.

Results: At the first developmental assessment and thenat 3 years, the initially, moderately IDA children scored

lower than did nonanemic control children on the Bru-net-Lezine and the MILI, respectively. However, thesedifferences were not significant after maternal educa-tion and birth weight were controlled for. At age 5 years,however, the lower scores in the moderately IDA chil-dren persisted, and this time the difference was signifi-cant even after the confounding variables were controlledfor. Because this was not a double-blind study, it is pos-sible that measurements were biased in the direction ofthe experimental hypothesis.

Conclusions: The strong implication is that anemia ininfancy, in particular moderate anemia, will lead to a poorerperformance on tests of cognitive development in the longterm.

24. Walter et al. (1990)* Key Words: Chile;age 5 years; previously moderately IDA, IDbefore and after therapy, or ID correctedby therapy; Stanford-Binet IQ test, anassessment of fine and gross motor co-ordination, an assessment ofpsycholinguistic ability, a test of visualmotor integration, and a preschool educa-tional scale; design: follow-up of previousintervention.

Importance: This abstract reports a study that is a fol-low-up of Walter et al. (1989) [Ref 14]. No separateanalysis was conducted as a function of the severity andduration of anemia at age 12 months as was done in theearlier study.

Results: Anemic infants, at 12 months, scored signifi-cantly lower on IQ, motor scores, psycholinguistic items,visuomotor coordination, and the educational preschoolassessment even though hematology problems had beencompletely corrected earlier in childhood.

Conclusions: Anemia was corrected in infancy yet lowerdevelopmental test scores persisted. Because no assess-ment is made of whether or not the children included inthe follow-up had been anemic between the two devel-opmental assessments, the lower scores here may havebeen due to a further period of anemia later in childhood.It is at least likely that, if a child’s environment early oninfluences whether or not the child will become anemic,continuing to live in such an environment means con-tinued risk of anemia. Therefore firm conclusions can-


not be drawn about the long-term developmental riskto children anemic in infancy, although this study doesat least suggest that these risks are to be taken as a seriouspossibility.

25. Lozoff et al. (1991)* Key Words: Costa Rica;age 5 years; IDA in infancy; psychoeducationalmeasures, a test of visual motor integration, theDraw a Man IQ test, a motor proficiency test,and the Wechsler Preschool and Primary Scaleof Intelligence; design: follow-up of a previousintervention.

Importance: This paper reports a follow-up of Lozoff etal.’s Costa Rica studies, which compared the develop-mental status of children at age 5 years who were eithermildly anemic, moderately anemic, or nonanemic atage 12–23 months.

Results: At age 5 years, the authors reported all childrento be comparable in terms of their hematology regardlessof their hematologic status at the end of the previous short-term study. Children who were moderately anemic in theearlier study and children who were mildly anemic butexperienced no hematologic correction scored lower ontests of development at age 5 years.

At age 5 years, children scored lower on all measuresexcept verbal IQ, visual auditory subtests of the Wood-cock test, and the Goodenough Draw a Man test. Gener-ally, between-group score differences were greatest ontasks requiring nonverbal skills, visuomotor coordination,and motor coordination and less pronounced on tasks re-quiring purely verbal skills.

The effects on developmental test scores were stronglyinfluenced by intervening variables. Another problem withthis study, as for Walter et al. (1990) [Ref 24],was thefailure of the authors to investigate the hematologic his-tory of the children between the end of the previous studyand the long-term follow-up. The reason why childrenhad persistently lower scores at 5 years may not only bebecause of their hematologic status in early childhood,but also because they were subsequently anemic betweenstudies. As noted above in the review of Walter et al.

(1990) [Ref 24], it is very likely that the factors predis-posing children to hematologic impairment persist.

Long-term Effects:Preventative Trials

26. Cantwell (1974)* Key Words: USA; age 0–7years; IDA at 6–18 months; neurologic evalua-tion and the Stanford-Binet IQ test; design;preventative.

Importance: This study is important in its assessment ofthe long-term developmental consequences of anemia inearlier childhood.

Design: This abstract describes a prospective iron statusmanipulation study, where the developmental assessmentat 6–7 years was blind to the iron status of the child ininfancy.

Results: The group of children who had become anemicbetween ages 6 and 18 months showed difficulty at 6–7years with clumsiness in balancing on one foot, problemswith tandem walking, and difficulty with repetitive handand foot movements. These motor coordination problems,thought to be a long-term developmental outcome of ane-mia in earlier childhood, may be seen as paralleling thedifficulties experienced by the anemic children describedby Lozoff et al. (1987) [Ref 13] and Walter et al. (1989)[Ref 14]. As with many of the studies of infants (seethose listed in the review by Deinard et al. (1976), thechildren in this study were also rated as more inattentive.

Conclusions: These results cannot be taken as conclu-sive proof of the causal relation between infant IDA andlong-term poorer performance on tests of developmentbecause no placebo group was included; the develop-mental protection apparently afforded the IM treated in-fants may have resulted from nonspecific factors. Fur-thermore, no assessment of other differences betweenthe groups was reported, either at baseline or follow-up, and without knowing whether group assignmentwas randomized, intervening variables such as SES ormaternal education may account for the differential ef-fect reported. No statistical comparison of the IQ mea-sures was reported.

Preschool Children (2–5 Years) 23

PRESCHOOL

CHILDREN

(2–5 YEARS)

Summary of Findings

The cognitive benefits of iron treatment in preschoolchildren are more apparent than in infants. The reasonsfor this are not understood but it may be that the testsavailable for use in older children are inherently moresensitive or that preschool children have passed the criti-cal age at which IDA can have long-lasting effects.

Only five studies have looked at the effects of ID or IDAon cognition in preschool children and they each havetwo notable features. First, all are intervention studieswith an extended period of intervention (2 6 months) com-pared with the studies of infants in the previous section.Second, in each study there were repeated attempts tomeasure cognitive function as opposed to developmentand to determine the locus of a cognitive effect resultingfrom ID or IDA.

Although it is difficult to draw firm conclusions from theresults of so few studies, the evidence suggests that chil-dren with IDA benefit from iron treatment in terms ofimproved performance on tests of discrimination and odd-ity learning/concept acquisition. There may also be be-havioral differences in that IDA children are observed tobe more unhappy and less responsive than iron-sufficientcontrol children (Deinard et al. 1986 [Ref 29]). The ma-jor influence of IDA on preschool children is thought tobe on attention, arousal, and motivation rather than onbasic cognitive abilities.

In three of the five studies, a battery of tests were de-signed to distinguish between problems with attention andthose with concept acquisition. Distinguishing betweenattention and concept acquisition problems is of par-ticular interest because attentional difficulties may con-ceal cognitive abilities in children. Such difficulties may,in the longer term, result in poorer performance on testsof cognitive function. It is also possible that poorer at-

tention and irritability are children’s reaction to tasksbeyond their capabilities. This issue was also relevant inthe previous section on infants and young children. Thetype of attentional disturbances seen in preschool chil-dren have clear parallels in the behavioral disturbancesfound in infants and young children, in whom differ-ences in attention, reactivity, and mood on BSID wereconsistently shown in children with ID and IDA.

As with studies in infants, with preschool children thebenefits of treatment on attention and cognition are seenonly in children who are initially iron deficient with ane-mia and are not apparent in children who are initially irondeficient but not anemic (see reasons given in infantssection).

Evidence to date is limited but has shown that improve-ments in attention and cognition result from iron treat-ment of children with IDA suggesting that the relation-ship is causal and that a program in which iron supple-ments are given to preschool children could be of benefitto children’s attention and cognitive function.

Intervention Trials

27. Pollitt et al. (1978)* H Key Words:USA; age approximately 4 years; ID; 4months of therapy; attention, learning, andmemory measured; design: extendedintervention, double blind, with nonanemiccontrol.

Importance: This abstract describes a study of particu-lar interest because it assesses the effects of ID on cog-nition, comparing ID preschool children with nonanemic,control preschool children before introducing iron therapy.The tests are designed to distinguish between control pro-cesses and structural capacity.


Design: This study used a double-blind, iron-treatment-only intervention design and looked at the effects of IDon cognition. Both ID children and nonanemic controlchildren were assigned to receive iron treatment.

Baseline: At baseline, the ID children performed morepoorly than did their nonanemic counterparts on tests ofattention and memory control processes. The results in-dicate that the effect is at the control level rather than inthe capacity of the structures themselves. This contra-dicts Soewondo et al.’s (1989) paper [Ref 31] which con-cludes that the effect is also in the capacity of the struc-tures themselves and only in children with IDA, not ID.

Follow-up: After 4 months of therapy, the relativelypoorer performance of the ID group compared to thenonanemic control group was no longer apparent.

Conclusions: This study indicates that ID is associatedwith lower scores on cognitive development tests in chil-dren of approximately 4 years of age and that perfor-mance returns to an optimal level after at least 4 monthsof therapy. However, the paper does not report that anysteps were taken to assess the role of potentially inter-vening variables that may have distinguished the twogroups at baseline. This, combined with the lack of inclu-sion of a placebo group, means that the unique effects ofiron deficiency and iron therapy on cognition cannot bedetermined.

28. Pollitt et al. (1983)* † Key Words: USA;age 3–6 years; IDA; 3 months of therapy;discrimination learning tests; design: extendedintervention.

Importance: This study’s importance probably lies in itscontradiction of the findings of the studies by Pollitt et al.(1986) [Ref 30] and Soewondo et al. (1989) [Ref 31]with respect to whether infants are scoring lower onattentional or concept acquisition tasks.

Design: This study was an intervention trial where IDAchildren were given treatment for 3 months whereas thenonanemic control children received no treatment. Therewas no placebo control.

Baseline: As with the Pollitt et al. (1978) study [Ref 27],the IDA children had lower scores than nonanemic con-trols on tests of discrimination learning, thought to tapattentional processes, and also on the attention based items

of the memory task. However, there was no difference intests of concept acquisition.

Follow-up: There was a significant improvement aftertreatment in the IDA group compared to the nonanemicplacebo group in the tests of discrimination but not in thetests of concept acquisition.

Conclusion: This study suggests that treatment can leadto improvements in attention. However, children didnot experience structural difficulties in contrast to thestudies of Soewondo et al. (1989) [Ref 31] and Pollitt etal. (1986) [Ref 30].

29. Deinard et al. (1986) † Key Words: USA; age18–60 months; IDA and ID; 6 months oftherapy; BSID or Stanford-Binet IQ; design:extended intervention, double blind, partiallyplacebo controlled.

Importance: This study was very well designed yet in con-trast to most studies, no differences between IDA, IDand nonanemic controls were observed at baseline inperformance on the MDI or Standford Binet. There wasalso no improvement in the IDA or ID group after treat-ment. The study was also important in its direct compari-son of the effects on development of ID and IDA.

Design: This was a double-blind intervention study witha nonanemic control group matched to IDA and ID groupsfor maternal education and, separately, for baseline MDIscores. The IDA group received iron treatment for 6months, the ID group received treatment or placebo, andthe nonanemic control group received a placebo.

Baseline: Overall, the authors reported no differences inMDI or Stanford-Binet between the ID, IDA andnonanemic control groups at baseline. The authors didreport that the IDA group was more unhappy and lessresponsive than the nonanemic control group, again con-firming the emerging pattern of behavior found in IDAand ID infants, young children, and preschool children(see Deinard et al. 1981 [Ref 1]).

Follow-up: After 3 months of therapy, neither the IDAnor ID group improved on MDI or Stanford-Binetwhereas the nonanemic control children tended to in-crease their scores, thus creating significant differencesbetween the nonanemic control children and the IDAand ID children.

Preschool Children (2–5 Years) 25

After 6 months of therapy, the pattern was the same,with neither the treated IDA nor the treated ID childrenimproving their MDI or Standford-Binet scores, and thecontrol children improving as before.

Conclusions: Despite the failure to detect baseline cogni-tion differences between groups, this study supports ahypothesis of an association between iron deficiency (withor without anemia) and poorer performance on tests ofcognitive development, given the failure of the IDA andID groups to match the improvements on MDI andStanford-Binet seen in the nonanemic control group. Theauthors suggest the lack of improvement in the IDA groupcould reflect the fact that these children were less test-able despite repeated testings (as reflected in the behav-ioral differences between groups) and perhaps this is be-cause of some irreversible behavioral deficit or becausethey were more susceptible to adverse environmentalconditions such as stress.

The absence of a developmental improvement in the IDAand ID groups after 6 months of iron treatment suggeststhat either the association between IDA and develop-ment is noncausal or that the effect is irreversible. Be-cause the IDA and ID children were matched tononanemic control children for maternal education andthe control group improved on MDI at 3 and 6 months,the failure of the IDA group to improve does not appearto be a function of maternal education but rather becausethe effect was not reversible with treatment.

Other: The authors offer an excellent discussion of thefindings of other studies.

30. Pollitt et al. (1986) * † Key Words:Guatemala; age 3–6 years; IDA; 3 months oftherapy; three discrimination learning tests, twoshort-term memory tests, and four odditylearning tasks; design: extended interventionwith nonanemic control.

Importance: This paper provides further evidence thatIDA children experience difficulties in the structural orcontrol processes involved in cognition. See also Pollittet al. (1978) [Ref 27] and Soewondo et al. (1989) [Ref31].

Design: An intervention trial comparing the cognitivefunction of ID children with nonanemic controls. The

ID group received treatment for 11-12 weeks. Cognitivefunction was assessed at baseline and immediately aftertreatment in both groups. There was no placebo con-trol.

Baseline: At baseline, the IDA children needed more tri-als to reach criterion on the discrimination learning tests.No differences were found between the groups on odd-ity learning / concept acquisition or on memory. The re-sults indicate a difficulty in IDA children in attending torelevant information. This agrees with the Pollitt et al.(1978) [Ref 27] and Pollitt et al. (1983) [Ref 28] studiesand also with the studies in the Infants and Young Chil-dren category, where IBR patterning has been found(Lozoff et al. 1985 [Ref 11], Oski and Honig 1978 [Ref5], Walter et al. 1983 [Ref 9]).

Follow-up: After treatment, the baseline difference indiscrimination / attention in the IDA children relative tothe control children was eliminated. However, after treat-ment, the IDA children made more errors than thenonanemic control group on the oddity learning task, andthis task is thought to tap conceptual acquisition.

Conclusions: This study provides further evidence insupport of an effect of IDA on attention. Because noplacebo control group was included and post hoc statisti-cal controls focused on possible anthropometric and nu-tritional differences between groups and not on, for ex-ample, SES or maternal education, the baseline differ-ences between ID/IDA children and nonanemic childrenmay be attributable to intervening variables and thechanges at follow-up to nonspecific therapy effects.

The effects of IDA on concept acquisition in this studyare only suggestive. In the Pollitt et al. (1978 and 1983)studies [Refs 27 and 28], there was no evidence thatchildren were experiencing concept acquisition difficul-ties whereas in the study of IDA children by Soewondoet al. (1989) [Ref 31], concept acquisition was adverselyaffected. Why the group treated with iron showed rela-tively poorer scores at follow-up but not at baseline in thisstudy is unclear, but it is possible, as with the Deinard etal. (1986) study [Ref 29], that the IDA children weremore sensitive to the stressful testing conditions and there-fore less able to improve with practice than nonanemiccontrol children.

Other: There is an extensive review of the field: in par-ticular, an excellent discussion of the findings concerning


ID and a comparison of these findings with the thoseconcerning IDA.

31. Soewondo et al. (1989)* † Key words:Indonesia; age approximately. 4.5 years;IDA and ID; 2 months of therapy; 2 x 2choice discrimination learning tests, 3oddity learning tests, and the PeabodyPicture Vocabulary Test; design: extendedintervention, double blind, randomized,with placebo and nonanemic controls.

Importance: This investigation is complementary to thestudies of Pollitt et al. (1978, 1986) [Refs 27, 30] in itsattempt to determine whether lower developmental testscores associated with iron deficiency are to be attrib-uted to structural or control process difficulties.

Design: The study was a double-blind intervention trialwith IDA, ID, and nonanemic control children randomlyassigned to receive either treatment or placebo for 8weeks.

Baseline: At baseline, of the small number of childrenwho reached the learning criterion on reversal discrimi-nation color tasks, the IDA children learned more slowly

than did the nonanemic control children. On the odd-ity learning task, the IDA children’s performance wasworse than that of the nonanemic control children onthe twice- and thrice-repeated versions. Thus detectablylower developmental test scores were found in IDA chil-dren on tests of attention and conceptual ability. Theattentional deficits in the IDA children, but not the ef-fects on concept acquisition, confirm findings of Pollittet al. (1978, 1983 and 1986) [Refs 27, 28, 30].

Follow-up: After 8 weeks of iron treatment, the iron-treated IDA children learned more quickly on the colordiscrimination and performed better on the twice- andthrice-repeated items of the oddity learning task than ei-ther the nonanemic control children treated with iron orthe IDA children who received placebo. No benefits oftreatment were observed in the ID group.

Conclusions: The strong implication of this study is thatIDA but not ID is associated with lower developmentaltest scores. This is the first study to shows effects of IDAon both attentional control processes and on conceptacquisition structural processes. Improvement in cog-nitive function of IDA children after iron treatment sug-gests that iron supplementation programs would ben-efit children of pre-school age.

School-age Children and Adolescents (5–16 Years) 27

Summary of Findings

There is strong evidence that among school-age chil-dren, initially lower scores on tests of cognition or schoolachievement due to IDA can be improved and in someinstances even reversed after iron treatment. One rea-son for this evidence might be the large number of pla-cebo-controlled trials, which are more able to pick uptreatment effects. Another reason might be the increasedsensitivity of the tests used. Alternatively, it could be thatthe effects of IDA in school-age children are more tran-sitory than in infants and thus more responsive to theeffects of iron treatment.

Eleven studies have examined the effects of iron supple-mentation on the cognitive function or educational achieve-ment of school-age children with ID or IDA. Of these,nine used placebo-controlled experimental designs, thuspermitting the causal effect of iron on cognition to beinvestigated. All but one study (Pollitt et al. 1989) [Ref41] showed significant improvements in the cognitive func-tion or educational achievement of children who receivediron supplementation compared with those who receivedplacebo. Furthermore, the one study (Bruner et al. 1996)[Ref 42] that investigated the effects of treating adoles-cent girls with ID also found significant cognitive ben-efits. These results are in contrast with those observedwith infants where benefits of treatment on developmentare rarely observed. As with studies with infants and pre-school children, there seems to be an indication of distur-bances in attention and behavior in children with IDA.

Note that none of the studies with school-age childrendocumented the hematologic history of the children. Itis possible, therefore, that the lower scores at baseline incognitive functioning in children with ID and IDA werea result of hematologic impairment earlier in life. In-deed, if the factors that predispose school-age children

SCHOOL-AGE CHILDREN

AND ADOLESCENTS

(5–16 YEARS)

to ID or IDA are not recent, there is an increased likeli-hood that these children had been anemic before, as in-fants and as young children.

The adverse effects on cognitive and educational testperformance due to IDA in preschool and school-agechildren appear more transitory in nature than the effectson development in infants and imply that treatment ofIDA in preschool and school-age children through ironsupplementation programs may be beneficial and haveimmediate effects. This is in contrast to the effects ofIDA on infants, for whom poorer performance on devel-opmental tests may not be reversible with treatment andwhere programs aimed at the prevention of IDA may bethe most appropriate action.

Observational Studies

32. Webb and Oski (1973a) * Key Words:USA; age 12–14 years; IDA; cross sectional;Iowa Basic Skills, levels A-F, form F; design:observational.

Importance: This study was carried out to assess the rela-tion between the presence of anemia and school perfor-mance, as measured by the Iowa Basic Skills test.

Design: A cross-sectional comparison of IDA withnonanemic control children.

Baseline: Although all subjects performed poorly rela-tive to national norms, the anemic children scored signifi-cantly lower than did the nonanemic control children.The anemic males demonstrated a “progressive depar-ture” in performance with increased age.

Conclusion: It is not clear whether the assessment wasdouble blind, and no attempt was made to match the sub-


jects for intervening variables. Thus, although IDAschoolchildren scored lower on school achievement teststhan the nonanemic control group, this may have beendue to some other variable related to iron but not ironitself.

Other: A short discussion of possible biochemical mecha-nisms is given.

33. Webb and Oski (1973b) *

This abstract reports on the same Webb and Oski (1973a)study [Ref 32], giving the results of a teachers’ behav-ioral assessment of the subjects. Again, attentional, be-havioral, and perceptual mechanisms are implicated aspossible factors mediating effects on cognitive perfor-mance (See discussions in previous sections and bySoemantri et al. (1985) [Ref 36]).

The teachers’ assessment is not reported to have beenblind to the children’s hematologic status, and althoughthe children were described as all coming from a sociallyhomogenous background, the behavioral disturbances inthe IDA group may not have been due to differences iniron status alone.

Intervention Trials

34. Seshadri et al. (1982) Study 1.† KeyWords: India; age 5–8 years; IDA; 2 months oftherapy; verbal and performance tests from theWechsler Intelligence Scale for Children(WNAC); design: extended intervention,randomized.

Design: This intervention study stratified subjects by ageand randomly assigned them to treatment or no treatmentbefore their iron status was determined. No placebo con-trol was included and it is not clear that assessment wasdouble blind.

Baseline: The between-groups comparisons at baselineare not reported.

Follow-up: After treatment, iron-treated children im-proved their WNAC scores more than did the nontreatedchildren for all iron status groups. Disaggregation by thepresence and absence of anemia showed that IDA chil-dren only improved more than the nonanemic control chil-

dren in the 7–8-year age group, and this was across treat-ments.

Conclusion: The lack of placebo control means that im-provements in cognition at follow-up in the IDA groupcould be due to a practice effect.

34. Seshadri et al. (1982) Study 2: † KeyWords: India; age 5–6 years (boys); childrenIDA; 2 months of therapy; WISC verbal andperformance tests; design: extended intervention,double blind, randomized, with placebo control.

Design: This was a double-blind intervention study withanemic male subjects only. Subjects were pair-matchedat baseline for Hb, age, height, weight, Draw a Man IQ,WISC, per capita income, and maternal education. Aftermatching, one child from each pair was randomly as-signed to receive iron treatment or placebo. A nonanemiccontrol group was not included.

Follow-up: After 2 months of treatment, the iron-treatedchildren showed WISC score improvements, and thesescores were significantly higher than those of the pla-cebo group. The iron-treated group also had significantimproved Hb levels compared to the placebo group.

The verbal IQ improvements in the iron-treated groupwere seen in the information, similarities vocabulary, arith-metic, and digit-span subtests. However, improvementswere also seen in the arithmetic and digit-span subtests inthe placbo group.

Conclusions: The initial matching procedure and the ran-domization of treatment and placebo means that a hy-pothesis of causal association between iron status andschool performance is strongly supported in this study.

35. Pollitt et al. (1985)* † Key Words:Egypt; average age 9.5 years; IDA; 4months of therapy; matching familiarfigures test; design: extended intervention,double blind, with placebo and nonanemiccontrols.

Importance: This study is notable for its strong experi-mental design and assessment of specific cognitive ef-fects of ID and IDA in school children.


Design: This was a double-blind intervention study inwhich IDA and nonanemic control children were as-signed to receive treatment or placebo. No randomiza-tion procedure is reported.

Baseline: IDA and ID children were less efficient and lessaccurate than nonanemic control children on thematching familiar figures test. However, it is not clear ifthe groups were similar in terms of SES, maternal educa-tion or other potential confounding variables.

Follow-up: After treatment, the iron-treated anemic chil-dren were more efficient than the placebo group, andtheir scores were similar to those of the nonanemic, con-trol children.

Conclusion: The study supports the hypothesis that IDand IDA adversely affect learning and problem solvingcapacity in school-age children and that the effect is re-versible with treatment.

Other: It is not clear that the test, as used in this study,has cultural validity.

36. Soemantri et al. (1985) * † Key Words:Indonesia; age 10–11 years; IDA; 3 months oftherapy; educational achievement, concentra-tion, and Ravens Progressive Colored MatricesIQ assessment; design: extended intervention,double blind, randomized, with placebo andnonanemic controls.

Importance: The study was conducted after the Pollitt etal. (1985) [Ref 35] study in Egypt to determine whetherthe effects on learning and problem solving had long-termimplications to a child’s educational achievement.

Design: This double-blind intervention randomly assignedIDA and nonanemic control children to receive treat-ment or placebo.

Baseline: At baseline, the IDA children did not differfrom the nonanemic control children on the Ravens IQtest. They did, however, perform more poorly on mea-sures of educational achievement. The results of the as-sessment of concentration are unclear.

There were no differences between groups in SES ormaternal age, but because maternal education was notassessed it is possible that differences at baseline weredue to an intervening variable.

Follow-up: After treatment, the iron-treated group im-proved their educational achievement score significantlymore than did the placebo group. However, the improve-ment was not enough to catch up with the scores of thenonanemic control group.

Conclusions: There is evidence that IDA adversely af-fects school achievement and that iron treatment can leadto significant benefits.

Other: The authors discuss the role of concentration andattentional disturbances in mediating poor performanceon educational achievement tests and how these distur-bances may be influenced by physiological arousal fac-tors. This paper also offers a good discussion of otherstudies of school-aged children.

37. Groner et al. (1986)*† Key Words:USA; pregnant women, age 14–24 years; IDrisk; 1 month of therapy; arithmetic, total digitspan, digit symbols, vocabulary, consonanttrigrams, Rey auditory verbal learning test;design: extended intervention, double blind,randomized, with placebo control.

Importance: This study is notable for its examination ofpregnant adolescent girls and young women who aredeemed to be at risk for iron deficiency. The study doesnot address the issue of the implications of maternal IDfor birth outcome, but rather for the cognitive performanceof the mother herself. The finding of an apparently ben-eficial effect of iron therapy is interesting given the he-matologic status of the subjects as nonanemic or, at most,only mildly iron deficient.

Design: This was a double-blind intervention study, withsubjects randomly assigned to receive treatment or pla-cebo.

Baseline: At baseline, there were no differences in cog-nition or iron status between the experimental and pla-cebo control group.

Follow-up: After 1 month of therapy, the hemoglobinlevel of both the treated and the placebo group declinedbut the decline was less in the treated group. The de-creases in hemoglobin were probably because all sub-jects were pregnant.


After treatment, the treated group improved on the digitsymbol test, one subsection of the consonant trigramstest, and two subsections of the Rey test. The placebo-control group decreased on arithmetic and increased onlyon one subsection of the Rey test. The treated group’shaving greater test-retest improvements than did the con-trol group was due to decreasing scores in the placebo-control group. Attentional effects are consistent withfindings reported elsewhere.

Conclusions: The study provides evidence that ID affectscognitive functioning and that iron supplementationwill prevent a decline in cognitive performance duringpregnancy.

38. Kayshap and Gopaldas (1987)† KeyWords: India; age 8-15 years (girls); IDA; 4months of therapy in two 2-month sessions;visual recall, digit span, mazes, and clericaltask; design: extended intervention, doubleblind, randomized, with placebo andnonanemic controls.

Importance: This study investigated the effects of ironsupplementation on cognition in anemic and nonanemicschoolgirls.

Design: The study was a double-blind intervention; thegirls were pair matched for age, Hb, and individual totalcognitive function test scores, after which one memberof each pair was randomly assigned to receive treatmentor placebo for 8 months. No double-blind procedure wasdescribed. Cognitive assessments were made at baseline,4, 8 and 12 months after treatment.

Baseline: There were no baseline differences on cogni-tive function, even when the subjects were disaggregatedfrom the original experimental and control groups by pres-ence or absence of anemia.

Follow-up: After 8 months but not after 4 months oftherapy, disaggregation of the original groups revealedthat the IDA girls given iron had improved significantlymore than the placebo group on clerical task, digit span,mazes, and overall score]. After 4 months, both the pla-cebo and the iron-treated IDA groups had improvedequally.

At the 12-month follow-up, 4 months after the comple-tion of supplementation, there was evidence of a sus-

tained beneficial effect of iron supplementation on allbut the mazes index. Hb had returned to near baselinelevels, and only the mazes score had dropped signifi-cantly. This study is interesting in combination withSoemantri’s (1989) [Ref 40] preliminary findings paperreviewed below, because Soemantri also included anassessment after the termination of therapy, which indi-cated that the benefits of therapy were sustained.

Conclusions: The selective improvement in the IDA iron-treated group over the placebo group suggests that cog-nitive improvement is enhanced with iron supplementa-tion in a group performing suboptimally and the benefit issustained at least 4 months after treatment has finished.

Because there were no differences in cognitive scoresas a function of iron status at baseline yet the IDA groupimproved in cognition with treatment, suggests that bothgroups were scoring lower at baseline than their potentialbut for different reasons.

Other: There is a good discussion of the possible bio-chemical mechanisms supporting the sustained improve-ment above baseline and of the general biochemical ef-fects of iron deficiency and how these may be associatedwith poor cognitive performance.

39. Seshadri and Gopaldas (1989) Study 1.This is a re-report of Seshadri et al. (1982)[Ref 34]. The reporting of the results forthis study is particularly unclear andinconsistent with the tabulation.

39. Seshadri and Gopaldas (1989) Study 2:This is a re-report of Seshadri et al. (1982)[Ref 34].

39. Seshadri and Gopaldas (1989) Study 3:† Key Words: India; 8–15 years (boys);IDA; 2months of therapy; WISC, design: extendedintervention, double blind, randomized, withplacebo and nonanemic controls.

Importance: A notable feature of this study is the inclu-sion of two iron therapy groups given different doses ofthe iron supplement.


Design: This was a double-blind intervention study; boysages 8–15 years (including anemic and nonanemic con-trol subjects) were triplet-matched for age, Hb, andbaseline cognitive function before randomly assigning themembers of each triplet to receive placebo, 30 mg iron,or 40 mg iron. Post hoc disaggregation of the groups wasperformed to compare IDA subjects with nonanemiccontrol subjects.

Baseline: No baseline comparison between scores wasmade for the anemic and nonanemic control groups cre-ated post hoc.

Follow-up: Post hoc analysis found that iron supplemen-tation improved the individual and overall scores on theWISC in IDA children only. There was a dose-effectwith the 40-mg treatment group showing improvementon more measures. It was reported that the WISCbaseline scores of the 30-mg group were relatively higherthan those of the 40-mg group. Therefore it cannot beconcluded that the association between iron status andWISC scores is causal.

Conclusions: No attempt was made to assess whetherthese post hoc groups were distinguishable from one an-other on other intervening variables. Such differencesbetween groups may have accounted for any baselinedifferences in cognition but not necessarily for the differ-ential response to iron therapy over placebo.

Thus, this study provides strong evidence that IDA iscausally associated with lower scores on tests of cogni-tive functioning and that performance can improve withiron treatment.

39. Seshadri and Gopaldas (1989) Study 4. Thisis a re-report of Kayshap and Gopaldas (1987)[Ref 38].

40. Soemantri (1989)*† Key Words: Indonesia;average age 10.5 years; IDA; 3 months oftherapy; educational achievement; design:extended intervention, double blind, random-ized, with placebo and nonanemic controls.

Importance: This study assessed cognitive performance 3months after completion of therapy, as in the Kayshapand Gopaldas (1987) study [Ref 38].

Design: This was a double-blind intervention study withIDA and nonanemic control subjects randomly assignedto receive 3 months of treatment or placebo. Cognitivefunction was assessed at baseline and 3 and 6 monthsafter treatment.

Baseline: There is no statistical analysis of the prelimi-nary data. The trends seem to support the hypothesis thatchildren with IDA have lower scores on tests of cogni-tive performance at baseline than the nonanemic con-trols.

Follow-up: The trends also support the existence of abeneficial effect of iron supplementation at 3 monthsand that the benefit persists at least 3 months aftertherapy is withdrawn. This sustained effect is similar tothat found by Kayshap and Gopaldas (1987) [Ref 38].No report is made of attempts to control for the effectsof intervening variables such as maternal education orSES, which may have accounted for baseline differences.

Conclusions: The indication is that IDA is causally re-lated to poorer performance on tests of cognitive devel-opment and that performance is responsive to therapy.Statistical assessments are obviously required before anyconclusions can be drawn from this study.

41. Pollitt et al. (1989)* Key Words: Thai-land; age 9–11 years; IDA and ID; 4 months oftherapy; Ravens IQ test and educationalachievement; design: extended intervention,double blind, randomized, with placebo andnonanemic controls.

Importance: This study is important because it includesan ID group with whom IDA and nonanemic control chil-dren are compared and because it does not replicate thefindings of Soemantri et al. (1985) [Ref 36].

Design: Before iron status was determined, subjects wererandomly assigned to receive treatment or placebo. Thereare similarities between this study and that of Soemantriet al. (1985) [Ref 36].

Baseline: At baseline, IDA children had lower IQ scoresthan did nonanemic control children in contrast to Soemantriet al. (1985) [Ref 36] and IDA and ID children scoredlower than nonanemic children on the Thai LanguageTest.


Follow-up: In contrast to Soemantri et al. (1985) [Ref36], there were no significant benefits of treatment in theIDA group compared to the placebo or nonanemic con-trols.

Compared with the Soemantri et al. (1985) [Ref 36], thetreatment period was longer and the supplement waslarger, however, the children were of a similar age, atleast some children were moderately anemic, and thenature of the treatment was the same. Unlike theSoemantri et al. (1985) study [Ref 36], where only thechildren infected with intestinal worms were given an-thelmintic drugs, the children in this study were all givenanthelmintic drugs at the start. Therefore the failure todemonstrate a treatment effect may have been a resultof a beneficial effect, across all subjects, resulting fromtreating for intestinal worm infections.

Conclusions: IDA and ID are associated with poorerscores on tests of educational achievement and cogni-tion, but the effect is not reversible with treatment. Inother studies of this age group, reversibility has been readilydemonstrated. The inability to replicate the study bySoemantri et al. (1985) is not fully understood.

Other: There is an excellent and systematic discussionof the possible reasons for this negative finding and of themediators of poor cognitive performance. Also includedis an incisive comment by Frank Oski concerning theimportant questions raised by research into iron deficiencyin schoolchildren.

42. Bruner et al. (1996)† Key Words: USA;age 13–18 years (girls); ID; 8 weeks of ferroussulfate therapy; Brief Test of Attention (BTA),Symbol Digit Modalities Test (SDMT), VisualSearch and Attention Test (VSAT), and HopkinsVerbal Learning Test; design: extended periodintervention, double blind, randomized, withplacebo control.

Importance: This study assessed the effects of ID aloneand has adolescent girls as its subjects, a group particu-larly at risk of iron deficiency. It also documents a differ-ential improvement with treatment in learning and memorybut not in attention. This is interesting given the sugges-tion from several other studies that attentional disturbancesprecede and mediate subsequent cognitive disturbances.

Design: This was a randomized, placebo-controlled in-tervention. There was no nonanemic control group. Allgirls were tested on three attentional measures and onelearning and memory test before treatment and after 8weeks of iron therapy.

Baseline: At baseline, there were no hematologic ornonhematologic differences between the two groups, in-dicating that the randomizing procedure had been suc-cessful.

Follow-up: After 8 weeks of iron or placebo, the treatedgroup had higher Hb and serum ferritin levels than did theplacebo-treated group, and they showed an improvementon the total recall test. However, no improvement wasseen on the attentional tests compared with the placebogroup.

Conclusions: The lack of inclusion of an nonanemic con-trol group means that this study offers no support for thehypothesis that ID children perform poorly on devel-opmental tests relative to nonanemic control children.However, in the absence of baseline developmental dif-ferences before treatment, the difference between groupsin the amount of improvement on the learning test indi-cates that ID girls are indeed performing suboptimallyand can only fulfill their potential when this deficiency iscorrected. This study is unusual in its finding that IDalone may be sufficient to result in cognitive disturbance.

Other: There is a brief but comprehensive discussion ofcausal mechanisms by which iron deficiency may alterbrain function.


SATELLITE ISSUES

vide an insight into the biochemical mechanisms at workin iron deficiency and a cautious consideration of howthe direct effects of iron deficiency on emotion and cog-nition may be a consequence of the altered central neu-rotransmitter metabolism. Sections of the review dealwith body iron distribution, a model of body iron storesand their respective susceptibilities to iron depletion,the association of iron with disease states, and the rolesof iron in brain function (oxidation, reduction, elec-tron transport, synthesis, packaging, uptake, and deg-radation of neurotransmitters). Beard et al. (1993) lookat all the major hypotheses concerned to explain theprobably interrelated biochemical effects of iron defi-ciency; specifically they consider the association of ironwith norepinephrine, dopamine, monoamine oxidaseactivity, (-aminobutyric acid metabolism, endogenousopiate system alterations, and the diurnal cycle. Theconclusion discusses the nature of the tests of functioncurrently used and how the sensitivity of these tests mightbe improved. Beard et al. (1993) cite 193 references, whichreaders can use to pursue experimental papers first hand.

Adult Iron Deficiency

Although this review focuses on the effects of ID andIDA on the cognitive, psychomotor, and behavioral statusof infants, preschool children, and schoolchildren, it can-not be forgotten that IDA is a life cycle issue. Adult irondeficiency has direct and indirect effects on children, aswell as on adults. The anemia frequently observed in preg-nant women is considered to be a normal physiologicalchange; however, severe anemia seems to affect not onlythe physiological status of the mother, but also the fetusduring pregnancy and the infant after birth (Achadi etal. 1995, Felt and Lozoff 1995, Gebre Medhin andBirgegard 1981, Godfrey et al. 1991, Larkin and Rao1990, Morgane et al. 1993, Tojyo 1983). At risk are theinfant’s normal growth and possibly also levels of activ-ity and early emotional development.

If infant IDA is truly a threat to the cognitive, psycho-

IDA is a common nutritional disorder that affects anestimated 25% of the world’s infant population(DeMaeyer and Diels-Tegman 1985). This review neces-sarily focused on a very narrow set of issues associatedwith IDA. However, this section has been included topresent a broad overview of satellite fields of research;references are cited throughout, and at the end of thesection is a reading list intended to help direct prelimi-nary inquiries into these related issues.

Animal Studies and BiologicalMechanisms

As Dallman (1987) notes, the literature on this topic hasbecome too large to encompass in a single review. Someof the literature documents the poorer performance onbehavioral, cognitive, and psychomotor indexes in iron-deficient rats, paralleling the lower test scores describedin the studies of humans. Various theories have been putforward as to what biochemical effects ID and IDA have,and numerous hypotheses have been suggested for howthese biochemical effects are expressed as the lowercognitive, motor, and behavioral test scores described inthe literature and reviewed here. Work has also beendone in an attempt to explain not just the nature of theseeffects, but also the reasons behind the age and severitydependency of the effects in humans.

For each hypothesis there is a huge and ever-increasingbody of literature of experimental studies (primarily ani-mal).

To even select central papers for each hypothesis is toogreat a task to be performed here in the context of areview primarily documenting the effects of ID and IDAon cognitive, psychomotor, and behavioral develop-ment. Two papers (Chen et al. 1995a,b) and two reviewsof the field (Beard et al. 1993, Beard et al. 1995) aresuggested for further reading.

Beard et al. (1993) present a long, comprehensive, andintegrative review of the literature that attempts to pro-


motor, and behavioral development of the individual,then maternal iron status during pregnancy and lacta-tion needs to be addressed to make programs compre-hensive for the prevention of IDA in infants. The youngwomen studied in the papers discussed above in the sec-tion School-Age Children and Adolescents are them-selves of child-bearing age, as becomes apparent inGroner et al. ‘s (1986) study [Ref 37]. Pregnancy in ado-lescence, when menstruation has only recently begun,puts the individual at even greater risk of anemia (Nelson1996).

Adults with IDA are at risk of failure to fulfill their cog-nitive potential, and because many new parents are them-selves in their late school years, they are also subject tothe developmental risks associated with iron deficiency.If it is accepted that parental schooling is an importantdeterminant of a child’s cognitive development (reviewedin bibliography: Deinard et al. 1986, Palti et al. 1983), thesuccessful prevention and treatment of IDA in all thoseof a child-rearing age are highly important. Iron deficiencyin parents must be treated and prevented for the sake ofthe cognitive and physical productivity of both parentsand children.

Other Outcomes of IronDeficiency

Physical Fitness: Adult iron deficiency may also detri-mentally affect physical fitness and productivity (Davies1973, Viteri and Torin 1974, Basta et al. 1979, Ohira et al.1979, Bhati and Seshadri 1987). The repercussions ofsuch effects are manifold. When parents are fatigued,the child-parent interaction will be impeded, and this initself could affect children’s emotional, motor, behavioral,and cognitive health. Adult physical fitness itself is also aconcern; not only may iron deficiency be preventing indi-viduals from fulfilling their individual potential, but when anation of adults suffers physical effects of iron deficiency,work productivity is compromised and the economic po-tential of the nation goes unfulfilled.

Infections: There is extensive experimental literature de-scribing the effects of iron deficiency on processes in-volved in host defence mechanisms, and it is commonlythought that infants and children who have moderate tosevere iron deficiency tend to have more infections thanthose who do not. However, although this widespread

belief exists, there is some clinical and much experimen-tal evidence to the contrary. For a comprehensive list ofexperimental studies, readers are directed to the refer-ence sections of these reviews (Pearson and Robinson1976, Brock and Mainou-Fowler 1986, Dallman 1987).

Not only may iron deficiency increase the risk of infec-tion, but certain infections (specifically worm infectionssuch as schistosomiasis) may also result in IDA. Dallman(1987) begins by describing the laboratory abnormalitiesin immune function that appear the most convincing anddistinguishes these from other components of the immunesystem about which there is uncertainty or that seem tobe normal during iron deficiency. He then categorizes theepidemiologic studies aimed at determining whether thereis an increased prevalence of infection in individuals withiron deficiency. He looks at cell- and antibody-mediatedimmunity and phagocytosis. Although most of the 59 stud-ies reviewed deal with human subjects, there is also ashort section on the findings of animal studies

Pearson and Robinson (1976) review studies indicatingthat an excess of iron facilitates growth and multiplicationof a number of microorganisms in various biogenic fluids,specifically studies in which sickle-cell anemia, over-whelming sepsis, and salmonellosis are associated withhyperferremia. They also review the literature concernedwith the connection between iron deficiency and increasedsusceptibility to infection. There are subsections on ph-agocytic function, cell-mediated immunity and iron defi-ciency, and iron deficiency as associated with clinical in-fection. Pearson and Robinson cite 119 references.

Confounding and CovaryingFactors

That iron deficiency can impede cognitive developmenthas been a very controversial hypothesis, especially inthe light of the data concerning the effects of iron defi-ciency on infants, where the causal nature of the associa-tion postulated cannot be concluded because of the lackof iron treatment effects on infant cognition and behav-ior. It has been frequently suggested that some otherfactor closely related to or covarying with iron deficiencyis responsible for the documented poorer developmen-tal test scores.

As mentioned above in the context of the implicationsof adult iron deficiency, parental schooling may well be

Satellite Issues 35

a stronger determinant of infant cognitive developmentthan iron deficiency. Parents with less education weresuggested to covary with an environment in which nu-tritional iron is not readily available. The elements hereare inextricably entangled: low SES and home environ-mental stress may be responsible for pressure to leaveschool early or for lack of adequate education account-ing for lower educational achievement among childrenwho go on to become parents themselves (Willerman etal. 1970, Lieblichet al. 1972, Czajka-Narins et al. 1978,Escalona 1982). Low-SES parents will have a lower in-come, less per capita to spend on food, and less time tointeract in a positive way with their children and so couldhave children who are nutritionally, behaviorally, andeducationally disadvantaged.

School absence has been suggested to be more respon-sible than nutritional status for lower cognitive test scoresin children (Powell and Grantham-McGregor 1980). How-ever, although there may be many determining factors ofincreased absenteeism among children, including factorsagain associated with a low-SES background, one factormay well be poor health resulting from nutritional defi-ciencies. School absence is particularly worrisome be-cause children’s cognitive and social development dependslargely on the provision of adequate educational stimula-tion.

With all of these factors, it is probably misleading to sug-gest that any one potential cause of lower scores on testsof development confounds the involvement of another.Such disadvantageous factors almost necessarily coexistand likely have an additive and even possibly multiplica-tive effect. For example, when undernourished childrengo to school without breakfast, the adverse effects oncognitive function are greater than in well nourished chil-dren (Simeon and Grantham-McGregor 1989, Simeon etal. 1994, Chandler et al. 1995, Pollitt 1995).

Given the complex interrelation of low SES, poor nutri-tion, lack of educational and psychosocial stimulation,and increased illness it may be that iron treatment ornutritional supplementation is not enough. Rather, anall-embracing program of environmental enrichmentmay be more appropriate (Lozoff 1990, Grantham-McGregor et al. 1991, Lansdown and Wharton 1995,Brown and Pollitt 1996).

References

Achadi EL, Hansell MJ, Sloan NL, Anderson MA (1995)Women’s nutritional status iron consumption and weightgain during pregnancy in relation to neonatal weight andlength in West Java, Indonesia. Int J Gynaecol Obstet48:S103-S119

Basta S, Soekirman S, Karayadi D, Scrimshaw NS (1979)Iron deficiency anemia and the productivity of adult malesin Indonesia. Am J Clin Nutr 32:916-925

Beard JL, Connor JD, Jones BC (1993) Brain iron: locationand function. Prog Food Nutr Sci 17:183-221

Beard JL, Zhan CS, Brigham DE (1995) Growth in ID rats.Proc Soc Exp Biol Med 209:65-72.

Bhati D, Seshadri S (1987) Anaemia undernutrition and physi-cal work capacity of young boys. Indian J Paediatr 24:133-139

Brock H, Mainou-Fowler T (1986) Iron and immunity. ProcNutr Soc 45:305-315

Brown L, Pollitt E (1996) Malnutrition, poverty, and intellec-tual development. Sci Am February:26-31

Chandler AM, Walker SP, Conolloy K, Grantham-McGregorSM (1995) School breakfast improves verbal fluency inundernourished Jamaican children. J Nutr 125:894-900

Chen Q, Beard JL, Jones BC (1995) Abnormal rat brainmonoamine metabolism in iron deficiency anemia. J NutrBiochem 6:486-493

Chen Q, Connor JR, Beard JL (1995) Brain iron transferrinand ferritin concentrations are altered in developing iron-deficient rats. J Nutr 125:1529-1535

Czajka-Narins DM, Haddy TB, Kallen DJ (1978) Nutrition andsocial correlates in iron deficiency anemia. Am J Clin Nutr31:955-960

Dallman PR (1987) Iron deficiency and the immune response.Am J Clin Nutr 46:329-334

Dallman PR (1986) Biochemical basis for manifestation ofiron deficiency. Annu Rev Nutr 6:13-40

Davies CTM (1973) Physiological responses to exercise inEast African children. II. The effects of schistosomiasisanaemia and malnutrition. Environ Child Health 19;2:115-119

DeMaeyer E, Diels-Tegman M (1985) The prevalence ofanaemia in the world. World Health Stat Q 38:302-316

Escalona SK (1982) Babies at double hazard:early develop-ment of infants at biologic and social risk Pediatrics 70:670-676

Felt BT, Lozoff B (1995) Brain iron and behavior of rats are notnormalized by treatment of iron deficiency anemia duringearly development. J Nutr 126:693-701

Gebre Medhin M, Birgegard G (1981) Serum ferritin in Ethio-pian mothers and their newborn infants. Relation of ironintake and socio-economic conditions. Scand J Haematol27:247-252

Godfrey KM, Redman CWG, Barker DJP, Osmond C (1991)The effect of maternal anaemia on the ratio of fetal weightto placental weight. Br J Obstet Gynaecol 98:886-891


Grantham-McGregor SM, Powell CA, Walker SP, Himes JH(1991) Nutritional supplementation, psychosocial stimu-lation, and mental development of stunted children: theJamaican study Lancet. 338:1-5

Lansdown R, Wharton B (1995) Iron and mental and motorbehaviour in childhood iron: nutritional and physiologicalsignificance. Chapman and Hall, New York, pp 65-180

Larkin EC, Rao GA (1990) Importance of fetal and neonataliron:adequacy for normal development of central nervoussystem. In Dobbing J (ed), Brain behaviour and iron in theinfant diet. Springer-Verlag, London, pp 43-62

Lieblich A, Ninio A, Kugelmass S (1972) Effects of ethnicorigin and parental SES on WPPSI performance of pre-school children in Israel. J Cross Cult Psychol 3:159-168

Lozoff B (1990) Has iron deficiency been shown to causealtered behavior in infants? In Dobbing J (ed), Brainbehaviour and iron in the infant diet. Springer-Verlag, Lon-don, pp 107 - 131

Morgane PJ, Austin-LaFrance A, Bronzino J, et al (1993)Prenatal malnutrition and development of the brain.Neurosci Biobehav Rev 17:91-128

Nelson M (1996) Anaemia in adolescent girls:effects on cog-nitive function and activity. Proc Nutr Soc 55:359-367

Ohira Y, Edgerton VR, Gardner GW, et al (1979) Work capac-ity heart rate and blood lactate responses to iron treat-

ment Br J Haematol 41:365-72Pearson HA, Robinson JE (1976)The role of iron in host re-

sistance. Adv Pediatr 23:1-3Pollitt E (1995) Does breakfast make a difference in school?

J Am Diet Assoc 95:1134-1139Powell C, Grantham-McGregor SM (1980) The associations

between nutritional status school achievement and schoolattendance in twelve year old children at a Jamaicanschool. W I Med J 29:247-253

Simeon DT, Grantham-McGregor S (1989) Effects of missingbreakfast on the cognitive functions of school children ofdiffering nutritional status. Am J Clin Nutr 49:646-653

Simeon DT, Callender J, Wong M, et al (1994) School perfor-mance nutritional status and Trichuriasis in Jamaicanschool children. Acta Paediatr 83:1188-1193

Tojyo H (1983) Effect of different intensities of iron deficientanemia in pregnant rats on maternal tissue iron and fetaldevelopment. J Nutr Sci Vitaminol (Tokyo) 29:339-351

Viteri F, Torin B (1974) Anaemia and physical work capacity.Clin Haematol 3:609-626

Willerman L, Broman SH (1970) Infant development pre-school IQ and social class. Child Dev 41:69-77

World Health Organization (1972) Nutritional anaemia re-port of a WHO group of experts. Technical Report Seriesno. 503. Geneva: WHO

Satellite Issues 37

READING LIST

Reviews on the effects of ID and IDA on mental andmotor performance, cognition, behaviour and schoolachievement in children

Avramidis L (1983) Developmental deficits in iron-deficientinfants. J. Pediatr 103: 339-340

Beard J (1995). One person’s view of iron deficiency, devel-opment and cognitive function. Am J Clin Nutr 62: 709-710

Brown, JL, Pollitt, E (1986) Malnutrition, poverty and intellec-tual development. Scientific American Feb 26-31.

Deinard AS, Murray MJ, Egeland B (1976) Childhood irondeficiency and impaired attentional development or scho-lastic performance: Is the evidence sufficient to establishcausality? J Pediatr 88: 162

Galler JR, ed (1984) Nutrition and Behavior. Plenum Press,New York

Grantham-McGregor SM (1990) Malnutrition, mental func-tion and development. In Suskind RM, Lewinter-Suskind L(eds), The Malnourished Child, Vol. 19 Raven Press, NewYork, pp 197-212

Horowitz FD (1989) Using developmental theory to guide thesearch for the effects of biological risk factors on the devel-opment of children. Am J Clin Nutr 50:509-597

Lansdown R, Wharton B, (1995) Iron and mental and motorbehaviour in childhood. In Iron: Nutritional and Physiologi-cal Significance, Clapham and Hall, pp 65-180.

Leibel RL, Pollitt E, Kim I, Viteri FE (1982) Studies regardingthe impact of micronutrient status on behaviour in man:iron deficiency as a model. Am J Clin Nutr 35:1211-1221

Lozoff B (1989) Methodologic issues in studying behaviouraleffects of infant iron-deficiency anemia. Am J Clin Nutr(suppl) 50:641-654

Lozoff B (1990) Has iron deficiency been shown to causealtered behaviour in infants? In Dobbing J (ed), Brain,Behaviour and Iron in the Infant. Springer Verlag, Londonpp 107-131

Lozoff B (1994) Iron deficiency and infant development JPediatr 125:577-578

Nelson M (1996) Anemia in adolescent girls: effects on cog-nitive function and activity. Proc Nutr Soc 55:359-367

Oski FA (1993) Iron deficiency in infancy and childhood NewEng J Med 329:190-193

Pollitt E, Leibel RL (1976) Iron deficiency and behaviour JPediatr 88:372-281

Pollitt E, Gorman K (1990) Long term developmental implica-tions of motor maturation and physical activity in infancy ina nutritionally at risk population. In Schurz B, ScrimshawNS (eds), Activity energy expenditure and energy require-ments of infants and children. Proceedings of InternationalDietary Energy Consultative Group. Lausanne, Switzer-land (IDECG), pp 279-296

Pollitt E, Metallinos-Katsaras E (1990) Iron deficiency andbehaviour: constructs, methods, and validity of the find-ings. In Wurtman RJ, Wurtman JJ (eds), Nutrition and brain,Vol 8. Raven Press, New York.

Read MS (1975) Anaemia and behaviour: nutrition growthand development. Mod Probl Pediatr 14:189-202

Scrimshaw NS (1984). Functional consequences of iron de-ficiency in human populations. J Nutr Sci Vitaminol 30:47-63

Simeon DT, Grantham-McGregor S (1990) Nutritional defi-ciencies and children’s behaviour and mental develop-ment. Nutrition Research Reviews 3:1-24

Appendices

39

APPENDIX A

Tabulated summary of studies investigating the effects of iron deficiency (with and without anemia) on the development of infantsand young children, preschool children, and school-age children and adolescents

Studies of the effects of iron deficiency (with and without anemia) on cognitive, motor, and behavioral indexes in infants and young children (6–24 months)

Author Age Study design Groups; n Length of Measures Results at baseline Results at(hematologic selection treatment and follow-up periodcriteria) [hematologic follow-upgroup mean and mildvs. moderateclassification]

SHORT-TERM EFFECTS: OBSERVATIONAL STUDIES

1. Deinard et al.1981

USA

15 months double blindobservational

-ID severe; n=34(SF≤9 ng/mL)[mean SF 8.85±1.58]-ID mild; n=21(SF 10-19 ng/mL)[mean SF 16.10±2.91]-NA; n=157(SF≥20 ng/mL)[mean SF 50.72±22.8]-nonanemic=Hct >34%

BSIDhabituationU/HSPD

-None noted overall.-ID severe group shows asystematic pattern ofdifficulties on the IBR.

2. Johnson andMcGowan, 1983

Study 1

USA

12 months observational

NA control

-IDA; n=31(Hb<10.5 g/dL)[mean Hb 8.7±0.9-moderate]-NA; n=31(Hb≥11.5)[mean Hb 12.2 g/dL]

tests of mother-child interaction:-activity,-reactivity-emotional tone-attention span

No significant groupdifferences on any of themeasures.

2. Johnson andMcGowan, 1983

Study 2

USA


-IDA; n=25(Hb<10.5 g/dL)[mean Hb 8.7±0.9-moderate]-NA; n=25 (Hb≥11.5)[mean Hb12.2 g/dL

IBR of BSID No significant differencesnoted between groups.

40B

ibliography of Studies of the Effects of Iron D

eficiency on Developm

ent



SHORT-TERM EFFECTS: OBSERVATIONAL STUDIES

3. Grindulis etal. 1986

UK


NA control

IDA; n=80(Hb < 11.0 g/dL)[not given]

NA; n=54(Hb > 11.0 g/dL)

Sheridandevelopmentalsequences,testing psycho-motor develop-ment

IDA children showedsignificant difficulties onfine motor and socialdevelopment items.

4. Lozoff et al.1986

Guatemala

6-24 months double blindobservational

NA control

-IDA; n=21(Hb<10.5 g/dL, plus 2of 3 abnormalbiochemical measures)[mean Hb 9.6 g/dL]-NA; n=21(Hb≥12.0 g/dL)[mean Hb 12.6 g/dL]

-assessment ofbehaviorfrequencies andquality by use ofa computer-compatible evenrecording system-BSID

After other variables werecontrolled for-the anemic infantsinitiated and maintainedmore contact with theirmothers than did the NAcontrol infants, and-the mothers of IDAinfants spent less timebeyond arm’s length oftheir children, were lesslikely to break close contactwith their children, andwere more likely toreestablish close contact.

SHORT-TERM EFFECTS: INTERVENTION TRIALS

5. Oski andHonig 1978

USA

9-26 months double blindintervention

IDA subjectsrandomlyassigned totreatment orplacebo; no NAcontrol

IDA; n=24(Hb<10.5 g/dLplus two measures ofiron deficiency)[iron-treated group Hbrange 6.2-10.3; Hbmean = 8.85±0.86;

placebo-treated groupHb range 7.6-10.2; Hb[mean 8.73 ± 1.09]

5-8 days of IMiron; dosecalculatedaccording tochild’s specifichematologicstatus

BSID -No significant differencesnoted between treatmentand placebo group on MDIor PDI.-All subjects showedsimilar difficulties on IBRand poorer gross and finemotor coordination.

-Iron-treated IDA childrenimprove on MDI, but thisimprovement was notsignificantly different fromthe improvement in theplacebo-treated group.-Initial Hb (not otherhematologic measures)correlated with amount ofMDI improvement withinthis group.- Nonsignificant changenoted on PDI.- Iron-treated IDA childrenimprove on reactivity(IBR) and on fine andgross motor coordination.

Appendices

41

6 and 7. Lozoffet al. 1982 (band c)

Guatemala


IDA and NAcontrol childrenrandomlyassigned toreceive treatmentor placebo

IDA; n=28(Hb≤10.5 g/dlplus three abnormalbiochemical measures)[mean Hb 9.5 ± 0.9 -nonspecifiable]

NA; n=40(Hb≥12.0 g/dLplus three normalbiochemical measures)

1 week of ferrousascorbate; dosecalculatedaccording toeach child’sspecifichematologicstatus

BSID -IDA children’s MDI scoreswere significantly lower in19-24-month-olds thanthose of the NA children ofthe same age.-Severity of IDA waspositively associated withthe extent of the difficultyexperienced on motortasks.-IDA children’s PDI scoreswere nonsignificantly lowerin all age subgroups thanthose of NA children.-Compared with infantsaged 6-18 months, IDAchildren at 19-24 monthsshowed disproportionatedifficulty on languageitems and on 11 of 12items shown to predictlater IQ.

-Previous deficitspersisted in iron statusgroups.-All groups improvedtheir MDI scores, with nodifferences in amount ofimprovement betweeniron- and placebo-treatedinfants

8. Lozoff et al.1982a(same study asLozoff et al. 1982b and c above)

see above see above see above see above BSID IBR -IDA children show aconsistent pattern ofdifficulties on the IBR.

-Behavioral differencesbetween IDA and NAinfants disappeared atfollow-up in both iron-treated and placebo-treatedIDA subjects.

9. Walter et al.1983

Chile

15 months double blindintervention

IDA, ID, and NAcontrol childrengiven treatment(no placebocontrol)

-IDA; n=10(Hb <11.0 g/dL, >8.5 g/dL plus two abnormalbiochemical measuresand treatmentresponse)[mean Hb 9.8 ± 0.8 -nonspecifiable]-ID; n=12(Hb≥11.0 g/dL and atleast one abnormalbiochemical measurewith treatmentresponse)NA; n=15

10 days of3-4 mg ferroussulfate/kg perday

BSID -IDA children scored loweron MDI than did controlchildren.-No differences notedbetween groups on PDI.-IDA children were moreunhappy than were othergroups.

-IDA infants improvedsignificantly on MDI.-NA control subjects showno significant improve-ment on MDI.-Improvement on MDI inthe IDA group coincidedwith IBR improvements incooperativeness andattention span.-Only ID children with twoor more abnormalbiochemical measuresshowed improvement.- No improvement notedon PDI.

42B



ent




10. Oski et al.1983

USA


ID, Idepl, andNA subjects,matched on sexand color, giventreatment; noplacebo control

-ID (biochemical andcellular); n=8(≥11.0 g/dL + abnormalbiochemical measuresand abnormal MCV)[mean SF = 9.8±2.0 -nonspecifiable severity]-ID(biochemical); n=10(≥11.0 g/dL + 3abnormal biochemicalmeasures)[mean SF = 10.2±1.3 -nonspecifiable severity]-Idepl; n=10(≥11.0 g/dL +abnormal SF)-NA; n=10

7 days of 50 mgiron dextran IM

BSID -No differences notedbetween groups on MDI.

-MDI scores increasedsignificantly in all IDchildren but not in NAchildren or Idepl children.-No patterning of areas ofimprovement noted onMDI or IBR.


Guatemala


IDA and NAinfants randomlyassignee to ironor placebo

-IDA; n=28(Hb<10.5 g/dL, plustwo of three abnormalbiochemical measures)[not given]-NA; n=40(Hb≥12.0 g/dL)

iron or placebotwice daily for 1week

BSID (particularfocus on the testaffect and taskorientationfactors of theIBR)

-IDA infants showed morebehavioral disturbance andlower MDI and PDI scoresthan the did NA infants.-The lower MDI scoreswere found in the infantswith abnormal ratings ontest affect.-The lower PDI scores werefound in the infants withabnormal ratings on testaffect and task orientation-Hb was directly related toabnormal IBR ratings, andonly indirectly related toMDI scores.

-There was no evidence ofa treatment effect after 1week.-The affectively disturbedIDA infants tended toimprove their behavioralratings regardless oftreatment group.-Affectively disturbed IDAinfants whose IBR ratingsimproved also improvedtheir MDI scores.-There was no differentialimprovement in PDIscores.

Appendices

43

12. Aukett et al.1986

UK


DA subjectsrandomlyassigned toreceive treatmentor placebo; noNA control

-IDA + iron andvitamin C; n=48(Hb < 11.0 g/dL)[nonspecifiableseverity]-IDA + vitaminC; n=49(Hb<11.0)

2 months oftherapy with 24mg ferroussulfate, 10 mgvitamin C, orboth

Denverdevelopmentalscreening test forpsychomotordevelopment

-Baseline Denver scoreswere not assessed.

-Significantly more of thechildren whose Hb valuesrose more than 2 g/dLachieved the standardchange in Denver scorethan did children whoseHb values did not rise.


Costa Rica


randomizedassignment ofIDA, ID, Idepl,and NA controlchildren toreceive treatmentor placebo(partial)

-IDA; n=52(Hb≤10.5 g/dLplus 2 abnormalbiochemical measures)[mean Hb 9.5 ± 0.1 -moderate IDA]-Intermediate ID; n=45(Hb >10.6<11.9+ 2 abnormal bio-chemical measures)-ID; n=21 (Hb≥12.0 +two abnormalbiochemical measures)-Idepl; n=28 (Hb≥12.0+ abnormal serumferritin)-NA; n=35 (Hb≥12.0+ normal biochemicalmeasures)

-1 week of FeSO4

oral orparenteralintervention;IDA andintermediatechildrenassigned toreceive oral orparenteral ironor placebo; NAchildrenassigned toreceive oral ironor placebo-12 weeks ofFeSO

4 oral iron

for all iron-deficientchildrenpreviously givenoral iron; allparenterallytreated or iron-sufficientchildren givenplacebo

BSID -within the anemic group,the authors distinguishedbetween moderately andmildly anemic children:-moderately anemic infants(Hb≤10.0) scored lower onthe MDI than all othergroups-both mild (Hb 10.0 -10.5) and moderatelyanemic infants scoredlower than the othergroups on the PDI-anemic children consis-tently failed on certainitems of the MDI and PDIscales

At 1 week:-no differences noted onMDI and PDI betweenparenteral and orallytreated groups,-significant effect of irontherapy seen on iron status,and-all groups improved theirMDI scores, regardless oftheir iron status or whetherthey were given treatmentor placebo.

At 12 weeks:-there was no improve-ment in MDI scores in theanemic group (mild ormoderate),-moderately anemicchildren showing no Hbcorrection or persistence ofabnormal biochemicalmeasures still showedsignificantly lower PDIscores than did mildlyanemic children,-anemic children whoseHb value was correctedshowed PDI scoreimprovement and were nolonger different from theNA group,-mildly anemic childrenwere more likely to showhematologic correction,and-improvements in the IDAgroup could not becompared with effects in aplacebo group because nosuch group was included inthe design.

44B



ent




14. Walter et al.1989

Chile


IDA, ID, and NAchildren assignedto receivetreatment orplacebo for thefirst 10 days,after which timeall childrenreceivedtreatmentIDA; n=39

(Hb<11.0 g/dL+ two ormore abnormalbiochemical measures,an abnormal cellularindex and a response totherapy of≥1 g/dL)[mean Hb=10.0 ± 0.9nonspecifiable]

ID; n=127(Hb≥11.0 g/dL plusabnormal biochemicalmeasures)

NA; n=30(Hb≥11.0 g/dL plusnormal biochemicalmeasures with lack ofresponse to supple-mentation)

at age 9 monthshematologicassessment made

at age 12months,3 months of 0.6mL ferroussulfate suspen-sion given threetimes a day withassessmentsmade at 10 daysand 3 months

BSID At age 12 months:-IDA children had lowerMDI and PDI score thandid NA control and IDchildren;-the authors reported asigmoid distribution ofMDI scores as a functionof increasing Hb, with thelowest developmental testscores to be found ininfants with the lowest Hb;infants with Hb > 10.9showed no evidence ofpoorer performance;-there was patterning onMDI, PDI, and IBR scales;-there was a significanteffect of duration andseverity of anemia, i.e.,children who at initialdevelopmental assessmenthad been anemic 3 monthsearlier had much lowerscores than did IDAchildren who had becomeanemic in the past 3months;-at the age-12-monthsassessment, children whohad become anemic in thepast 3 months were lessseverely anemic thanchildren who were anemicat initial assessment

At the 10-day assesment:-all groups improved onMDI and PDI regardless ofiron status or whether theywere given placebo or irontreatment, and-no change in IBR

At the 3-months assess-ment:-there were no changes inany of the children’s BSIDscores (including IBRratings) between follow-up1 and follow-up 2 and-IDA children were stillscoring low on languageitems and on showingshoes, clothing, or own toyand were less able to standfrom sitting or stand onleft foot with help.

Appendices

45

15. Idjradinataand Pollitt 1993

Indonesia


IDA, ID, and NAsubjectsrandomlyassigned toreceive treatmentor placebo

-IDA; n=50(Hb≤105 g/L+ two abnormalbiochemical measures)[iron-treated children’sHb mean =95.7±1.1;placebo-treatedchildren’s Hb mean =98.0±1.0 - moderateIDA]-ID; n=29(Hb≥120 g/L+ two abnormalbiochemical measures)-NA; n=47(Hb≥120 g/L+ normal biochemicalmeasures)

4 months of 3mg/kg per day oforal ferroussulfate

BSID -IDA children scored lowerthan did ID and NAchildren on MDI and PDI.-No significant differencenoted between ID and NAchildren.-IDA children’s mothershad achieved significantlylower maximum schoolgrade than had mothers inID or NA groups.

-Iron-treated IDA childrenshowed significant MDIand PDI improvements.-The improvements onMDI and PDI in the iron-treated IDA children werelarger than improvementsseen in NA and IDAplacebo-treated children.-The baseline differencesin MDI between iron-treated IDA children andID or NA children wereeliminated, thus the poorertest performance observedat baseline was reversed.-There were no significantdifferences in the amountof MDI and PDI improve-ment between iron- orplacebo-treated IDchildren.-Correction of Hb valuesnoted in iron-treated IDAand ID children.

16. Lozoff et al.(1996b) CostaRica

12–23 months Double blindintervention. IDA givestreatment.NA randomlyassigned toreceive treatmentor placebo

IDA n=32 (Hb ≤100 g/L+ two of threeabnormal biochemicmeasures)NA n=54 (Hb ≥12.5 g/L)

6 months of oralferrous sulphategiven 3mg/kgper dose, twice aday

BSID & IBR At baseline, IDA infantsscored significantly lower(mean 6.1 points) than NAon MDI. No significantdifference on PDI. IDAinfants on IBR weresignificantly more fearful,unhappy and hesitant withexaminer.

-Good hematologicalresponse to treatment.-Differences in MDIbetween IDA and NAremained at the 3 and 6months follow-up.-No change in PDI.-No differences in IBRbetween IDA and NA. Thesignificant improvement inIBR is not considered tobenefit directly from irontreatment.

46B



ent




17. Heywood etal. 1989

Papua NewGuinea

1 year double blindintervention

IDA childrenrandomlyassigned at age 2months toreceive treatmentor placebo; noNA control

IDA, malaria positive +iron; n=18[mean Hb 84.1 g/dL -range 70.5-97.6:moderate]

IDA, malaria negative +iron; n= 30[mean Hb 100.5 g/dL -range 92.3 - 108.9:unspecifiable]

IDA, malaria positive +placebo; n= 11[mean Hb 80.0 g/dL -range 65.9 - 94.1:moderate]

IDA, malaria negative +placebo; n=36[mean Hb 92.9 g/dL -range 80.2 - 105.5:unspecifiable]

1 x 3 mL IMdextrantreatment at age2 months, withtests of attentiongiven at age 1year

attention-total fixationtime-mean fixationtime-habituation-dishabituation

No baseline assessmentmade of attention.

Iron-treated IDA malaria-negative infants hadsignificantly longer totalfixation times than didplacebo-treated malaria-negative infants.

Appendices

47

SHORT-TERM EFFECTS: PREVENTATIVE TRIALS

18. Moffatt et al.1994

Canada

2-15 months double blindpreventative

infants randomlyassigned at age 2months toreceive iron-fortified orregular formula;,assessments at 6,9, 12, and 15months

-6 months:n=225regular [Hb 107.7±5.3 -nonspecifiable]fortified [Hb113.5±5.3]-9 months: n=204regular [Hb 111.6±5.6 -nonspecifiable]fortified [Hb116.4±5.6]-12 months: n=186regular [Hb 111.8±5.7 -nonspecifiable]fortified [Hb117.5±5.6]-15 months: n=154regular [Hb 115.1±5.7-nonspecifiable]fortified [Hb118.6±5.7]

13 months ofiron-fortifiedformula (12.8mg) or regularformula (1.1mg)

BSID (includingIBR); attentionpaid to the test-affect and task-orientationclusters of items

Overall:-all measures of iron statuswere significantly differentbetween the two groupsand-mental development andbehavior were not affected.

At 6 months:-PDI values were similar inall groups

At 9 months:-PDI scores for the regularformula group fellsignificantly and gave thisgroup a score significantlylower than that of thefortified-formula group

At 12 months:-intergroup PDI differ-ences persisted.

At 15 months:-the group means werecloser again and thedifference was nonsignifi-cant.

19. Lozoff et al.1996a

Chile

6-12 months double blindpreventative

infants randomlyassigned at age 6months toreceive supple-mental iron orno added ironuntil age 12months

-supplemental iron;n=625-no supplemental iron;n=319anemia [Hb< 110 g/L +2 of three abnormalbiochemical measures]

6 months ofsupplementaliron

BSID No significant differencesnoted between thesupplemented andnonsupplemented groups.

-At 12 months, thesupplemented infants hadless anemia and less irondeficiency than did thenonsupplemented groups.

-There were no significantdifferences in BSID scoresbetween the groups.

48B



ent



LONG-TERM EFFECTS : OBSERVATIONAL STUDIES

20. Palti et al.1985

Israel

10-13 years single blindobservational

childrenintermediate andNA at 9 monthsfollowed yearlyuntil 10-13 yearsand assessed

At 9 months:intermediate; n=20(Hb<10.5 g/dL)[not given]-NA; n=56(Hb>11.5 g/dL)[not given]

subsection of theShaeferClassroomBehaviorInventory witheducationalachievementitems

-After maternal educationand sex of the child werecontrolled for, childrenIDA at 9 months scoredlower on educationalachievement and positivetask orientation, with noeffect on negative taskorientation or mood.-Maternal education andsex of the child were alsosignificant predictors ofeducational achievement,task orientation, andmood.

21. Wasserman etal. 1992

Yugoslavia

2 years prospective

MDI decrementsat 24 monthsmeasured as afunction of Hbdecreases at 6,12, 18, and 24months

n = 392;Hb measurements atfour time periods notgiven

MDI -Significant MDI decre-ments at 18-24 monthspredicted by Hb decreaseof 2 g/dL at 18 months.-MDI decrement at 24months predicted better byHb decrease of 2 g/dL at 12and 18 months than by Hbdecrease at 24 months.

22. Hurtado 1995 0-10 years preventative

no note made ofgroups,randomizationprocedure, oruse of a placebo

n=5411

groups not described

dose or lengthof treatment notgiven

test of cognitivedevelopment

At 10 years:-iron deficiency wasassociated with poorerperformance on a cognitivetest, with a dose-responserelation between Hb andperformance, and-interaction with maternaleducation reported.

Appendices

49

LONG-TERM EFFECTS: FOLLOW-UP OF INTERVENTION TRIALS

23. Palti et al.1983

Israel

0-5 years intervention

children withIDA at 9-10months giventreatment; noplacebo control;IDA and NAchildren givenneither treat-ment norplacebo;developmentalassessmentsconducted at 2,3, and 5 years asa function ofhematology at 9months

-IDA moderate at 9months,3 years n=32,5 years n=29(Hb≤9.9) [moderate]-IDA mild at 9 months,2 years n=84,5 years n=52(Hb 10-10.9)[mild]-IDA at 9 months,3 years n=149,5 years n= 99(Hb 11-11.9)-iron sufficient at 9months,3 years n=123,5 years n= 59(Hb≥12.0)

2 years(baseline)

3 years

5 years

BLDQ at 2 years

Mili Israeli IQTest at 3 years

Wechsler WPPSI(Israeli version)

(Baseline is taken asdevelopmental measure-ments at age 2.)

Initially moderate IDAchildren scored lower onBLDQ than did other threegroups, but this differencewas not significant aftermaternal education andbirth weight werecontrolled for.

At 3 years the initiallymoderately IDA infantsstill scored lower than didall other groups, this timeon the Mili test, but again,the between-groupsdifference only approachedsignificance after maternaleducation and birth weightwere controlled for.

At 5 years the initially IDAinfants still scored lowerthan all other groups, butat 5 years these differencesbetween groups on IQ weresignificant even aftermaternal education andbirth weight werecontrolled for.

24. Walter et al.1990(abstract)

Chile

5 years see Walter et al.1989 above; notethat it is notclear whether theassessments ofiron status at 4years or thedevelopmentalassessments at 5years weredouble blind

IDA;n=43[severity of anemia ininfancy was notspecifiable]

NA; n=29

see Walter et al.1989 above

-SBIQ-BruininksOseretsky fine/grossmotor assess-ment-Illinoispsycholinguisticassessment-visual motorintegration-Woodcockeducationalpreschool scale

see Walter et al. 1989 above Infants with IDA at 12months of age (whoshowed poorer develop-mental test performance atthe time of originalassessment) showedpersistence of lower testscores at age 5 yearsregardless of the hemato-logic correction thatoccurred after 3 months ofsubsequent iron therapy.


Costa Rica

follow-up ofsubjects fromLozoff et al. 1987

5 years see Lozoff et al.1987 above

note that is notclear whether the5-year follow-upassessment wasdouble blind

-initially moderatelyIDA-initially iron deficientbefore and after therapy-initially iron deficientcorrected by therapy

5 years -Woodcock-Johnsonpsychoeducationalbattery-test of visualmotor integra-tion-GoodenoughHarris Draw aMan test-Bruininks-Oseretsky test ofmotor profi-ciency-WPPSI

see Lozoff et al. 1987 above Infants originally withmoderate IDA or mild IDAuncorrected by therapyshowed persistently lowerscores on developmentaltests at age 5 years.

50B



ent



LONG-TERM EFFECTS: PREVENTATIVE TRIALS

26. Cantwell 1974(abstract)

USA

0-7 years double blindpreventative

no note made ofrandomizationprocedure, noplacebo control

neonatal IM irondextran, NA at 6-18months; n=28

IDA; n=32(criteria not given)[Hb range = 6.1-9.5 -moderate]

IM dextrantreatment; doseand length oftreatment notspecified;developmentalassessmentsmade at6-7 years

neurologicalevaluation

SBIQ

At 6-7 years-children who were IDA asinfants showed clumsinessbalancing on one foot,tandem walking, andrepetitive hand and footmovements and were moreinattentive and hyperactive,and-the Stanford Binet meanscore for IDA children was92 and for NA children was98.

Appendices

51

Studies of the effects of iron deficiency (with and without anemia) on cognitive, motor, and behavioral indexes in preschool children (2–5 years)


INTERVENTION TRIALS

27. Pollitt et al.1978

USA


ID and NAcontrol giventreatment

ID; n= 23(criteria not given)[mean Hb =11.5,MCV=79.7, SI/TIBCratio =13.4,serum iron= 35.0 -severity not specifiable]

NA; n= 23(criteria not given)

4 months of irontherapy (dosenot specified)

-three discrimi-nation learningtests

-two short-termmemory tests

-four oddity-learning tasks

-ID children perform worseon tests of their attentionand memory controlprocesses but not on testsof the capacity of thestructures themselves.

Differences in scoresbetween groups at baselineeliminated.


USA


IDA childrengiven treatment,NA controlchildren nottreated; noplacebo control

IDA; n=15

NA control; n=15

3 months of oraliron

-discriminationlearning-short-termmemory-oddity learning

-IDA children performedworse than NA control onthree discriminationlearning tests.

After treatment thedifferences between theIDA and NA controlchildren had beeneliminated.

29. Deinard et al.1986

USA


IDA giventreatment only, IDgroup assigned totreatment orplacebo, and NAgroup givenplacebo only;NA controlmatched twice toexperimentalgroup: on1)maternaleducation, sex,and age and2) baseline MDIscore

-IDA; n=25(≤11.0 g/dL plus threeabnormal biochemicalmeasures and abnormalcellular index)[mean group Hb = 10.4g/dL-SD not given, butpotentially categorizableas mild anemia]-ID; n=45(≥11.0 g/dL plus twoabnormal biochemicalmeasures and abnormalcellular index)-NA; n=?(≥11.0 g/dL plus threenormal biochemicalmeasures and normalcellular index)

6 months of oraliron (6 mg/kgper day) withfollow-up

assessments at 3months and 6months

-BSID MDI forchildren underage 2 years-SBIQ forms L-Mfor children overage 2 years

-No significant differencenoted between the IDAchildren and their NAmaternal education controlon MDI and SBIQ.-There were no significantdifferences between IDAchildren and their NAcontrols matched onmaternal education on 8 of10 IBR measures.-The IDA group was moreunhappy and lessresponsive to theirenvironment than weretheir NA controls matchedon maternal education.-Similar patterns emergedbetween the IDA childrenand their controls matchedfor MDI baseline.

At 3 months:-the IDA group did notimprove on MDI and SBIQ;-neither of the ID groupsimproved on MDI and SBIQ;-overall, all NA control groupstended to increase their scoreson the MDI and SBIQ frombaseline, creating significantdifferences between the IDand IDA groups and their NAcontrols.

At 6 months:-IDA children showed noMDI and SBIQ increase;-ID treated children showedno increase over baseline inMDI and SBIQ;-ID children given placeboincreased on MDI SBIQ; and-again, NA control childrentended to significantlyincrease their scores.

52B



ent

Studies of the effects of iron deficiency (with and without anemia) on cognitive, motor, and behavioral indexes in preschool children (2–5 years)


INTERVENTION TRIALS


Guatemala


IDA and NAcontrol giventreatment; noplacebo control

IDA; n=25(Hb≤10 g/dLplus one abnormalbiochemical measureand Hb response totherapy >2 g/dL)[mean Hb = 8.96 ±0.98 - moderate IDA]

NA; n=25(Hb >11 g/dLplus one normalbiochemical measureand no Hb response totherapy)

11-12 weeks of30 mg/kg perday oral ferroussulfate

-three discrimi-nation learningtests-two short-termmemory tests-four oddity-learning tasks

Discrimination learning:-IDA children needed moretrials to criterion-no differences on odditylearning-no differences on memory

Discrimination learning:-IDA group improved ontrials to criterion aftertreatment eliminating thebaseline differencesbetween IDA and NAgroups on trials needed toreach criterion.

Oddity learning:-both groups improved,-the NA group improvedsignificantly more than theIDA group, and-there were no significantdifferences on memory.

31. Soewondo etal. 1989

Indonesia

NA mean age57.5 months

IDA mean age54.2months

double blindintervention

IDA, ID, and NAsubjectsrandomlyassigned toreceive treatmentor placebo

IDA; n=49 (Hb <110 g/L, plus two abnormalbiochemical measures)[group Hb mean =106.0 ±0.7 - mild]

ID; n=60(Hb≥110 g/L plus twoabnormal biochemicalmeasures)NA; n=70 (Hb≥110 g/Lplus two normalbiochemical measures)

8 weeks , 50 mgoral elementalFe

2 x two-choicediscriminationlearning tests

3 x odditylearning tasks

PPVT

-Among the small numberof children who reachedlearning criterion onreversal discriminationcolor tasks, there was asignificant differencebetween groups in numberof trials to criterion, withNA children learning faster.

-NA children performedbetter than IDA childrenon “repeated twice andthrice” variations of theoddity-learning task.

-Of children who reachedcriterion on trials tolearning on color discrimi-nation, the IDA+ironchildren learned morequickly than did the NA+iron children.-On the “repeated twiceand thrice” versions of theoddity-learning task, iron-treated IDA childrenperformed better thaniron-treated NA children.-There was no significantimprovement in theperformance of the IDAchildren treated withplacebo.-Iron-treated IDA childrenshowed significantimprovement in all ironmeasures, but theimprovement was smalland it is. probable that IDwas not fully reversed.

Appendices

53

Studies of the effects of iron deficiency (with and without anemia) on cognitive, motor, and behavioral indexes in school age children and adolescents (5–16 years)

Author Age Design Groups; n Baseline to Measures Baseline differences Differences at follow-up(hematologic selection follow-upcriteria) [hematologicgroup means]

OBSERVATIONAL STUDIES

32. Webb andOski 1973a

USA

12-14 observationaltrial with nodouble blindprocedure; datawere analyzed asa function of agesubclass (12, 13,and 14 years)

IDA; n=92(Hb 10.1-11.4 g/dL)[lowest Hb was 10.6 ±0.4 - mild anemia]

NA; n=101(Hb 14.0 - 14.9 g/dL)

Iowa Tests ofBasic Skills levelsA-F, form 3

-IDA children scored lowerthan did NA children.-The older anemic malesshowed a progressivedeparture from theircontrols.

33. Webb andOski 1973b

USA

abstract reportof Webb andOski 1973 above

same as Webb and Oski 1973 above - Peterson-QuayBehaviourProblemChecklist

-visual after-image task

- Anemic males showedmore conduct disturbancethan did their controls,including distractibility(see Soemantri et al. 1985above for a similarconcentration finding),overactivity, disruptiveness,and negativism.-IDA subjects showedlonger latency to after-image report than didcontrol subjects.

INTERVENTION TRIALS

34. Seshadri etal.1982: India

Study 1


before disaggre-gation into IDAand NA controlgroups, childrenstratified by ageand randomlyassigned toreceive treatmentor no treatment;no placebocontrol; blindstatus of the trialnot reported

Experimental groups:5-6 years; n=236-7 years; n-217-8 years; n=19

Control groups:5-6 years; n=126-7 years; n=127-8 years; n=10

IDA; n=39 (Hb <110 g/L)[means not given]NA; n=24 (Hb>110 g/L)

2 months of20 mg oralelemental Fe andfolic acid

WNAC:-six verbal tests-six performancetests

Between-groups compari-sons on initial develop-mental scores were notreported.

-Iron-treated childrenimprove in Hb and WNACand control shildren donot, regardless of hemato-logic status.-In children aged 7-8 yearsonly, WNAC scoreimprovement was higherfor IDA than NA groups(across treatments).

54B



ent

Studies of the effects of iron deficiency (with and without anemia) on cognitive, motor, and behavioral indexes in school age children and adolescents (5–16 years)



34. Seshadri etal.1982: India

Study 2

5-6 years double blindintervention

after these IDAchildren were pairmatched onweight, Hb, DAM-IQ, per capitaincome, andmaternaleducation, onechild from eachpair was randomlyassigned to receivetreatment orplacebo; no NAcontrol included

-treated; n=14[mean Hb = 96.2±2.7 -moderate]-placebo control; n=14(mean Hb = 98.2±2.3 -nonspecifiable]

(IDA Hb < 105 g/L +post hoc Hb increase >15 g/L)

2 months of40 mg elementalFe and folic acid

WNAC verbaland performancetests

No differences as matchedwere noted.

-Only the treated groupshowed WNAC scoreimprovements.-WNAC scores of thetreated group were higherthan those of the controlgroup.-The experimental groupshowed Hb increase of 23g/L whereas the controlgroup showed a decrease of2 g/L.

35. Pollitt et al.reported 1985

Egypt

average 9.5 years double blindintervention

IDA and NAcontrol childrengiven treatmentor placebo; noreport ofrandomizationprocedure

-IDA; n=28(Hb≤11.5 g/dLplus one abnormalbiochemical measureand response to therapy)[not specified]-NA; n=40(Hb >13 g/d plus anormal biochemicalmeasure and no Hbresponse to iron therapy)

4 months of 50mg oral ferroussulfate

form F of thematchingfamiliar figuretest (assessmentis of speed andaccuracy)

-NA children were fasterand more accurate atbaseline.

-Only the anemic childrenwho were treated with ironbecame more efficient.-Efficiency scores for thetreated anemic childrenwere similar to the scoresof the control children.

36. Soemantri etal. 1985

Indonesia


IDA and NAcontrolrandomlyassigned toreceive treatmentor placebo

IDA; n=78(Hb≤11.0 g/dL plus oneabnormal biochemicalmeasure)[mean Hb= 9.74 ± 1.31-nonspecifiable]

NA; n=41(Hb≥12.0 g/dL plusnormal measurementsin the biochemicalindex)

3 months of10 mg oralferrous sulfate

-educationalachievement:maths, biology,social science,and language;Bourden-Wisconsinconcentrationtest-RPCM

-No baseline differencesnoted between IDA andNA on RPCM.-NA children performedbetter on educationalachievement than did IDAchildren.-NA children were heavierand taller than IDAchildren.-The results of concentra-tion assessment were notclear.

-Iron-treated IDA children hadchanged Hb values more thandid other groups.-The results of concentrationassessment were not clear.-Iron-treated IDA childrenimproved their EA score morethan did placebo-treated IDAchildren-A difference on EA remainedbetween iron-treated IDAchildren and all NA children.-No hematologic differenceswere noted between treatedIDA and NA children.

Appendices

55

37. Groner et al.1986

USA

pregnant womenages 14-24 years

assumed to be atrisk of ID


randomizedassignment toreceive treatmentor placebo

-experimental; n=16(≤16 weeks pregnant,Hct≥31%)[mean Hb 12.2 g/dL -normal]

-control (just prenatalvitamin); n=9≤16 weeks pregnant,Hct≥31%)[mean Hb 12.3 g/dL -normal]

1 month of 90mg ferrousfumerate BID +prenatal vitamintreatment

-arithmetic-total digit span-digit symbol-vocabulary-consonanttrigrams-RAVLT

-No psychometricdifferences except on onesubsection of the RAVLT,with experimental subjectsperforming better thancontrol subjects.

-No correlation betweenhematologic status andpsychometric test scores.

Note that decreases in ironmeasures were expectedbecause of the dilutionaleffects of pregnancy.

-Control subjectsdemonstrated greaterhematologic decrease thandid experimental subjects.-Experimental groupimproved on digit symbol,on one subsection ofconsonant trigrams, andon two subsections ofRAVLT.-Control subjects decreasedon arithmetic andincreased only on onesubsection of RAVLT.-Experimental groupshowed greater test-retestimprovements than didcontrol subjects on threetests because of control-group score decreases-No correlation notedbetween hematologic statusand psychometric testscores.-No correlation notedbetween hematologicchanges and changes inpsychometric test score.

38. Kayshap andGopaldas 1987

India

girls ages 8-15years


after being pairmatched on age,Hb, and indi-vidual totalcognitive functiontest scores, onemember of eachpair was randomlyassigned to receivetreatment orplacebo

groups weredisaggregated intoIDA and NAcontrol groups forfurther analysis

experimental; n=65

control; n=65

IDA; n=81( Hb <105 g/L)[placebo: 95.0±1.3 -moderate;experimental:96.0±1.3 - moderate]

NA; n=20

assessmentsevery 4 monthsfor 1 year,includingassessment at 12months, 4months aftercompletion oftherapy;60 mg oralFeSO

4 treatment

as one tablet perday for 2months at thebeginning ofeach of two 4-month schoolterms

cognitivefunction tests:-visual recall-digit span-maze-clerical task

After disaggregation,anemic and NA groupsshowed no baselinedifferences in cognitivefunction.

-At 8 months iron-supple-mented IDA children hadimproved their Hb andclerical task, digital span,maze, and overall scores morethan had placebo-treated IDAchildren.-The only improvement in theNA children was in the iron-treated group and was onlyon mazes.-At 12 months there were nodifferences between placeboand treated groups except onmazes, with the iron-treatedgroup scoring higher still.-The only significant drop inscores in the treated groupwas on mazes.

56B



ent

Studies of the effects of iron deficiency (with and without anemia) on cognitive, motor, and behavioral indexes in school age children and adoles-cents (5–16 years)



39. Seshadri and Gopaldas 1989, India, Study 1re-report of Seshadri et al. 1982, Study 1, above

39. Seshadri and Gopaldas 1989, India, Study 2re-report of Seshadri et al. 1982, Study 2, above

39. Seshadri andGopaldas1989

Study 3

India

boys ages 8-15years


after beingtriplet matchedon age, Hb,and baselinecognitivefunctionscores, subjectswithin each setwere randomlyassigned toreceive 30 mgtreatment, 40mg treatment,or placebo,

groupsdisaggregatedinto IDA andNA controlgroups forfurther analysis

experimental 30 mgdose (both IDA andNA); n = 16

experimental 40 mgdose (both IDA andNA); n=16

placebo (both IDA andID); n=16

(IDA Hb<105 g/L)NA (Hb>115 g/L)

IDA + iron[98 ± 2.0 - moderate]

IDA+ placebo[placebo98 ±2.6 -nonspecifiable]

-4 months

-30 mg or 40 mgoral ferroussulfate therapyfor 2 months

cognitivefunction tests:-visual recall-digit span-maze-clerical task

After disaggregation:-30 mg and 40 mg iron-treated IDA and NAchildren improved theirindividual and overallcognitive scores except forthe 30 mg group on mazes.-The overall scores of theiron-treated groups werehigher than the placebo-treated group, with the 40-mg group showing moreindividual child improve-ment on visual recall, digitspan, maze, and clericaltask and the 30-mg groupshowing more individualchild improvement onclerical task and visualrecall.-These differences resultedfrom changes in theanemic iron-treated group,which improved more thandid the placebo-treatedanemic group and all NAgroups.-All subjects given ironresponded hematologically,with no dose effect.

39. Seshadri and Gopaldas 1989, India, Study 4re-report of Kayshap and Gopaldas 1987, above

Appendices

57

40. Soemantri 1989

Indonesia

average 10.5years


IDA and NAcontrol childrenrandomlyassigned toreceive treatmentor placebo

IDA; n=58(Hb <110, plus oneabnormal biochemicalmeasure)[IDA + iron meanHb=97.7 ± 4.0;IDA + placeboHb=96.8±5.1 -nonspecifiable]NA; n=72(Hb > 120 plus onenormal biochemicalmeasure)

3 months of 10mg oral ferroussulfate treat-ment, withdevelopmentalassessments at 3months and at 6months(3 months aftertherapycompletion)

EA-maths-biology-social science-language

Note that in this prelimi-nary study, no statisticalanalysis was performed.

-No apparent IQ differ-ences noted betweengroups.- NA children appear toperform better than IDAchildren on all EAmeasures.- NA children were tallerand heavier than IDAchildren.

Note that in this prelimi-nary study, no statisticalanalysis was performed.-Iron-treated IDA childrenappeared to consistentlyshow improvement on allEA measures at all follow-ups, including 3 monthsafter completion oftherapy.-There was also a similarimproving trend in theplacebo-treated group,although this appeared tobe of a smaller size.-The changes in the NAchildren were not in anyconsistent direction or of alarge size.


Thailand


before iron statushad beenassessed, childrenwere randomlyassigned toreceive treatmentor placebo

-IDA; n=101(Hb <120 g/L plus twoabnormal measures)[non-specifiable]-ID; n=47(Hb≥120 g/L plus twoabnormal measures)-NA; n=1210(Hb≥120 g/L plusnormal biochemicalmeasures)

2 weeks’ 50 mgoral ferroussulfate and 14weeks’ 100 mgoral ferroussulfate

RPCM

EA-Thai language-maths

-IDA children scored loweron IQ did than NA controlchildren.

-IDA and ID childrenscored lower on Thailanguage than did NAcontrol children.

-All children improvedtheir scores at follow-up,regardless of iron status.-There were no differencesin improvement betweenthe iron- and placebo-treated children.

42. Bruner et al.1996

USA


ID girlsrandomlyassigned toreceive treatmentor placebo; noNA controlgroup

-ID + treatment; n=37(Hb >12.0 g/dL,SF <12.0 µg/L)[Hb 13.1±0.7; SF9.1±2.2]-ID + placebo; n=36(Hb >12.0 g/dL,SF <12.0 µg/L)[Hb 13.0±0.7; SF8.5±2.6]

8 weeks of 325mg ferroussulfate twice perday

cognitivefunctioning :-Brief Test ofAttention-Symbol DigitModalities Test-Visual Searchand AttentionTest-Hopkins VerbalLearning Test

Both groups were similaron hematologic andcognitive measures.

-Iron-treated group hadhigher mean SF and Hbvalues.-There was no effect ofiron treatment on any ofthe measures of attention.-Treated girls performedsignificantly better thanbaseline scores and theplacebo-treated group onthe total recall score of theHopkins Verbal LearningTest.

58B



ent

Legend

ID, iron deficiency (deficient); IDA, iron deficiency anemia (anemic); Idepl, iron depleted; IM, intramuscular; NA, nonanemic control; Hb, hemoglobin; Hct, hematocrit; MCV, meancorpuscular volume; SF, serum ferritin.

BLDQ, Brunet-Lezine’s Developmental Quotient; BSID, Bayley Scale of Infant Development; DAM-IQ, Draw a Man IQ Test; EA, Educational Assessment; IBR, Infant Behaviour Record(subscale of BSID); MDI, Mental Development Index (subscale of BSID); PDI, Psychomotor Development Index (subscale of BSID); PPVT, Peabody Picture Vocabulary Test; RAVLT, ReyAuditory Verbal Learning Test; RPCM, Raven Progressive Color Matrices; SBIQ, Stanford-Binet Intelligence Quotient; U/HSPD, Uzgiris and Hunt Ordinal Scales of Psychological Develop-ment; WNAC, Wechsler Intelligence Scale for Children; WPPSI, Wechsler Preschool and Primary Scale of Intelligence.

Observational: baseline hematologic and nonhematologic measurements only are taken from groups to be compared; intervention: the IDA and ID groups are given iron supplementation,usually after baseline hematologic and nonhematologic assessment; preventative: IDA and ID in a risk population are prevented with prophylactic supplementation from an early age, andthe development of the subjects is compared with the development of subjects from the same population not given supplements; placebo control: a subgroup of the IDA or ID children aregiven a placebo according to the same procedure as the administration of the iron supplement; randomized: subjects are randomly assigned to treatment or placebo groups; NA control: anNA group is included, usually matched to the IDA or ID group for age (as a minimum) and given the same developmental assessments. Here, the NA control group often receives the sameiron supplementation or placebo administration as do the IDA and ID groups.

To convert from conventional values to SI values, use the following factors: for ferritin, multiply by 1 to convert from ng/mL to µg/L; for hemoglobin, multiply by 10 to convert from g/dL tog/L; for hematocrit, multiply by 0.01 and delete the percent sign.

Appendices 59

APPENDIX B

List of studies used in thebibliography, by category

Studies of the Short-term Effects of IDA and ID: Infantsand Young Children (6B24 Months)

Aukett et al. (1986). Reference 12

Deinard et al. (1981). Reference 1

Grindulis et al. (1986). Reference 3

Heywood et al. (1989). Reference 17

Idjradinata and Pollitt (1993). Reference 15

Johnson and McGowan (1983). Reference 2

Lozoff et al. (1982a). Reference 8

Lozoff et al. (1982b). Reference 7

Lozoff et al. (1982c). Reference 6

Lozoff et al. (1985). Reference 11



Lozoff et al. (personal communication, 1996). Reference19

Moffat et al. (1994). Reference 18

Oski and Honig (1978). Reference 5

Oski et al. (1983). Reference 10

Walter et al. (1983). Reference 9

Walter et al. (1989). Reference 14.

Studies of the Long-term Effectsof IDA and ID: Infants and YoungChildren (6B24 Months)

Cantwell (1974). Reference 26

Hurtado (1995). Reference 22


Palti et al. (1983). Reference 23

Palti et al. (1985). Reference 20

Walter et al. (1990). Reference 24

Wasserman et al. (1992). Reference 21

Studies of the Effects of IDA andID: Preschool Children (2B5Years)

Deinard et al. (1986). Reference 29

Pollitt et al. (1978). Reference 27



Soewondo et al. (1989). Reference 31

Studies of the Effects of IDA andID: School-age Children andAdolescents (5B16 Years)

Bruner et al. (1996). Reference 42

Groner et al. (1986). Reference 37

Kayshap and Gopaldas (1987). Reference 38



Seshadri and Gopaldas (1989). Reference 39

Seshadri et al. (1982). Reference 34

Soemantri (1989). Reference 40

Soemantri et al. (1985). Reference 36

Webb and Oski (1973a). Reference 32

Webb and Oski (1973b). Reference 33


APPENDIX C

A brief description of the test instruments used in the study of ID/IDA on cognition, behavior, development, andeducational achievement

Tests are listed in alphabetical order. A description of each test is given and/or the name of the test publisher.

1. Tests of Development,Behavior, and Intelligence

Bayley Scale of Mental Development

Bayley N (1969) Bayley Scales of Infant DevelopmentManual. New York: Psychological Corporation

The scales give three complementary tools for assessinga child’s developmental status between the ages of 2months and 2½ years. The test is given individually.

1. The Mental Scale (MDI). Measures perception,memory, learning, problem solving, vocalization, thebeginnings of verbal communication, and rudimentaryabstract thinking.

2. The Motor Scale (PDI). Measures gross motor abili-ties, such as sitting, standing, walking, and stair climb-ing. Also measures manipulatory skills of hands andfingers. At the infant level, these abilities are consideredto play an important part in children’s interactions withthe environment and thus in the development of theirmental processes.

3. The Infant Behavior Record (IBR). This is a ratingscale completed by the examiner after the other two partshave been given. It assesses various aspects of personal-ity development such as emotional and social behavior,attention span persistence, and goal directedness.

U.S. norms were developed for children 2 to 30 monthsold. The MDI and PDI give separate developmentalindexes expressed as normalized standard scores with amean of 100 and a standard deviation of 16 (as in theStanford-Binet). Raw scores on individual subtests mayalso be used.

The scales are advised for assessing current developmen-tal status and not for predicting future abilities. Theymay be useful in the early detection of sensory and neu-rological defects, emotional disturbances, and environ-mental deficits.

Beery Test of Visual Motor Integration

Beery KE (1982) Administration, scoring, and teach-ing manual for the developmental test of visual-motorintegration. Cleveland, OH: Modern Curriculum Press

Bourden-Wisconsin Card Sorting Test

Heaton RK (1981) A manual for the Wisconsin CardSorting Test. Odessa, FL: Psychological AssessmentResources

The test was originally developed to measure concep-tual levels of the normal adult population. It has, how-ever, become very popular in clinical neuropsychologyand is regarded as sensitive in revealing behavioral ri-gidity, which is typically exhibited by patients with fron-tal lobe syndrome. Adult-level performance in the testemerges at the age of approximately 10 years.

The test is given individually. It consists of two identicalpacks of cards which the subject has to categorize withthe help of four stimulus cards and cued feedback givenby the experimenter. When a particular sorting crite-rion (color, form, or number) is established, the experi-menter changes it unexpectedly. The test is finished whensix categories have been identified or all (128) cards havebeen used. Successful performance requires reasoningability and flexibility to alter sorting strategies.

Appendices 61

Bruininks-Oseretsky Test of Motor Proficiency

Herrick VE (1959) The Iowa School Performance Test.In: Buros OK, ed. The fifth mental measurementshandbook. Highland Park, NJ: Gryphon, p. 311 (alsopublished by American Guidance Service in 1978)

Originally published in Russia in 1923. The most recentversion was published in 1978. Designed to cover allmajor types of motor behavior and used for testing men-tally retarded children (who frequently have motorimpairment) and children with motor handicaps, mini-mal brain dysfunction, or learning disabilities. Stan-dardized for ages 4½ to 14½ years. The complete bat-tery consists of 46 items grouped into eight subtests. Ittakes 45–60 minutes to complete and gives three scores:a Gross Motor Composite, which measures performanceof the large muscles (shoulders, trunk, and legs); a FineMotor Composite, which measures performance of thesmall muscles (fingers, hands, and forearm); and a totalBattery Composite Score. There is also a 14-item ShortForm, which gives a single index of general motor profi-ciency.

Brunet-Lezine’s IQ Test

Brunet O, Lezine I (1951) Le developpementpsychologique de l’enfant. Paris: Presse Universitairede France

A French version of Gesell’s developmental schedule.The test consists of 10 items for each age (birth, 3, 6, 9,12, 18, and 24 months). Psychomotor development ismeasured by four subscales: (1) motor development,(2) coordination development, (3) language develop-ment, and (3) social-personal development.

Denver Developmental Screening Test

Frankenburg WK, Dodds JB, Fandal AW (1973) Den-ver Developmental Screening Test: manual workbookfor nursing and paramedical personnel. Denver, CO:University of Colorado Medical Center

The test was designed for use with children from birth to6 years of age. It measures infant and child developmentand is divided into four parts: (1) personal-social, e.g.,ability to put on and remove clothes; (2) fine-motor–adaptive, e.g., ability to draw; (3) language, e.g., abilityto use plurals, opposite analogies, define words; and (4)gross motor skills, e.g., ability to throw a ball.

U.S. age-standardized scores are available. The test hasbeen validated against the Bayley Scale of Infant Devel-opment and the Stanford-Binet and also standardizedfor use in multiracial communities.

Goodenough-Harris Draw-a-Man Test

Harris DB (1962) Children’s drawings as measures ofintellectual maturity: a revision and extension of theGoodenough Draw-a-Man Test. New York: HarcourtBrace and World

Draw-a-Man Test was first standardized in 1929. Cur-rent version was restandardized by Harris in 1962. Thetest instructions are simply, “Make a picture of a man;make the very best picture that you can.” There are alsoinstructions for drawing a woman and for drawing one-self. It may be given individually or in groups. Empha-sis is placed on the child’s accuracy of observation andon the development of conceptual thinking, rather thanon artistic skill. Points are given for the inclusion ofindividual body parts, clothing details, proportion, per-spective, etc. The maximum number of points for theman or the woman is 73 and 71, respectively. These aretransformed into standardized scores based on U.S.norms with a mean of 100 and a standard deviation of15. An alternative, simplified scoring system called theQuality Scale may be used where the child’s drawing ismatched to the one it resembles most closely in a gradedseries of 12 sample pictures.

In older children, performance on the test is said to cor-relate well with tests of reasoning, spatial ability, andperceptual accuracy, whereas in pre-school-age childrenit correlates best with numerical aptitude. Performanceon the test is quite heavily influenced by cultural back-ground.


Illinois Test of Psycholinguistic Abilities: Re-vised (ITPA-R)

Champaign, IL: University of Illinois Press

This is an individually administered test for children 2to 10 years old. The test was designed to follow a three-dimensional model of the process of communication.The model has two channels (auditory-vocal and vi-sual-motor), three processes (receptive, organizing, andexpressive), and two levels (representational and auto-matic). The abilities covered by the test are located atthe intersections of the three dimensions. E.g., theManual Expression Test asks children to perform amanual gesture to “show what we do with” each pic-tured object such as a telephone. This task includes thefollowing dimensions: visual-motor channel, expressiveprocess, and representational level.

The ITPA is considered very culturally restricted in thatit was developed for U.S. middle- class children, so itsuse is not recommended for lower-socioeconomic orminority groups.

Iowa Tests of Basic Skills Levels A–F, Form 3

Chicago: Riverside Publishing Company

This is a test of educational achievement in U.S. childrenin kindergarten to grade 9. The battery gives a profile ofscores on the major academic areas of reading, writing,and arithmetic. The three domains tested are (1) lan-guage (vocabulary, reading comprehension, spelling,capitalization, punctuation, language usage), (2) work-study skills (visual materials, reference materials), and(3) mathematics (mathematical concepts, problemssolving, computation).

Performance on the test at different grades can be di-rectly compared.

Multi-Level Informal Language Inventory (MILI)Intelligence Scale for Preschool Children

Developed by C. Goldsworthy and W. Secord in 1982,the test measures language growth and development inIsraeli children 4–6 years old.

Peabody Picture Vocabulary Test: Revised (PPVT-R)

Dunn LM (1981) Peabody Picture Vocabulary Test—Revised. Circle Pines, MN: American Guidance Ser-vice

The test provides a quick measure of the “use” of vocabu-lary. It takes around 15 minutes to complete and con-sists of a series of 175 plates, each containing four pic-tures. As each plate is presented, the examiner providesa stimulus word orally. The test taker responds by point-ing to the picture on the plate that best illustrates themeaning of the stimulus word. It is especially useful fortesting vocabulary use among people who are not liter-ate, who are not able to vocalize well, or who are deaf.In studies investigating the effects of iron deficiency ane-mia on children’s cognitive/motor/mental development,it is sometimes used as a relative measure of the educa-tional level of parents. Performance in the United Statesand the United Kingdom correlates well with other vo-cabulary tests and also reasonably well with verbal in-telligence, educational achievement, and scholastic ap-titude.

Peterson-Quay Behavior Problem Checklist

Peterson DR, Quay HC (1967) Manual for the Be-havior Problem Checklist. Unpublished manuscript,University of Illinois

Ravens Coloured Progressive Matrices

London: HK Lewis & Co. Ltd.; San Antonio, TX: Psy-chological Corporation

This is an easier version than the Ravens ProgressiveMatrices (RPM). RPM was designed in 1983 as a mea-sure of intelligence, specifically Spearman’s g factor. Per-formance is based on an ability to analyze/interpret ab-stract items. The colored version has three parts, eachwith 12 items. Each item consists of a picture of a pat-tern at the top or a series of pictures arranged in a ma-trix from which a piece is missing. Children are asked topick from six choices the picture that completes the pat-tern or matrix. The easier items measure accuracy of

Appendices 63

discrimination, and the more difficult items measureanalogies and logical relations. It can be given individu-ally or in groups. There is no time limit, although chil-dren usually complete it in 10 to 40 minutes.

Shaefer Classroom Behavior Inventory

Shaefer ES, Droppleman LE, Kalverboer AF (1965)Development of classroom behavior checklist and fac-tor analyses of children’s behavior in the United Statesand the Netherlands. Proceedings of a Conference ofthe Society for Research in Child Development, Min-neapolis, MN

The Classroom Behavior questionnaire was designedfor use as a routine screening procedure for learningand behavioral problems of children. Children are as-sessed by their teachers. There are 60 questions on sec-tions including (1) learning achievement (assessed byreading, writing, spelling, arithmetic, and general knowl-edge), (2) task orientation—(2a) positive task orienta-tion (comprehension, persistence, and self-control) and(2b) negative task orientation (distractibility, lack ofdiscipline, and motor restlessness)—and (3) mood(anxiety, depression, aggressiveness, and mood swings).Each question is scored on a four-point scale (0–3). Thequestions have equal weight.

Sheridan Developmental Sequences

Sheridan M (1978) Children’s developmental progress.Windsor: NFER, pp 1–55 (currently out of print)

Stanford-Binet Intelligence Scale

Chicago: Riverside Publishing Company

Originally designed in 1905 as a measure of IQ/generalintelligence. The current (fourth) edition was publishedin 1986. It is administered individually. Fifteen testsmeasure four cognitive domains: (1) verbal reasoning,e.g., vocabulary, comprehension; (2) abstract/visualreasoning, e.g., number series, equation building; (3)quantitative reasoning, e.g., pattern analysis, copyingpaper folding, and cutting; and (4) short-term memory,e.g., bead, sentence, digit, or object memory. The itemsare given in a mixed order to retain a child’s interest andattention. Six of the tests are used for all ages, whereas

the remaining nine tests either begin or end at a higherlevel.

Although the overall score gives a useful measure of IQ,the scores on the individual subtests provide informa-tion on specific learning difficulties. U.S. normalizedage-standardized scores are available for individuals 2to 23 years old. The overall score has a mean of 50 and astandard deviation of 8, whereas the scores for each ofthe four cognitive domains have a mean of 100 and astandard deviation of 16. Raw scores on individualsubtests can also be used.

Uzgiris and Hunt Ordinal Scales of Psychologi-cal Development

Uzgiris IC, Hunt J McV (1975) Assessment in infancy:Ordinal Scales of Psychological Development. Urbana,IL: University of Illinois Press

Wechsler Intelligence Scales

Wechsler Adult Intelligence Scales—Revised (WAIS-R)

Wechsler Intelligence Scale for Children—Re-vised (WISC-R)

Wechsler Pre-school and Primary Scale of Intelli-gence—Revised (WPPSI-R)

New York: Psychological Corporation

There are three scale: (1) WAIS for adults (>16½ years)published in 1939, (2) WISC-R for school-age children(ages 6½ to 16½ years) published in 1974, and (3) theWPPSI-R for preschool children (ages 4 to 6½ years)published in 1989. Each scale is used as measures of gen-eral intelligence and sometimes as an aid in psychiatricdiagnosis, and is given individually.

Within each scale there are two main parts: (1) VerbalScale. In the WISC-R, this constitutes six subtests: In-formation, Similarities, Arithmetic, Vocabulary, Com-prehension, and Memory (Digit-Span). (2) Perfor-mance Scale. In the WISC this constitutes five subtests:Picture Completion, Picture Arrangement, Block De-sign, Object Assembly, and Coding (or Mazes). Perfor-mance on individual subtests can give information aboutspecific learning difficulties. Subtests that have been used


in studies on iron deficiency anemia and cognition aredescribed individually below.

The scores may be age standardized to U.S. norms, witha mean of 100 and a standard deviation of 15. However,raw scores or standardized scores on individual subtestscan also be used. Performance correlates highly (0.80)with the Stanford-Binet. The full battery takes around60 minutes to give.

Woodcock-Johnson Psycho-educational Battery

McGrew KS, Werder JK, Woodcock RW (1991) WJ-Rtechnical manual. Allen, TX: DLM

2. Tests of Cognitive Function

A Brief Test of Attention

Schretlen D, Bobholz J, Brandt J (1996) Developmentand psychometric properties of the Brief Test of Atten-tion. Clinical Neuropsychology 10:80–89

This is a measure of auditory divided attention in whichparticipants listen to a tape of letters and numbers (e.g.,4-A-8-G-3-2) and are asked to report how many lettersor numbers they hear.

Hopkins Verbal Learning Test

Brandt J (1991) The Hopkins Verbal Learning Test:development of a new memory test with six equivalentforms. Clinical Neuropsychology 5:125–142

This is a test of cognitive function. It is a 12-item seman-tically categorized word-list learning test. There arethree free recall trials, a delayed recall trial, and a yes/norecognition test. Participants are read the same list ofwords three times, and each time they are asked to re-peat as many words as they can recall. Twenty minuteslater they are asked to say which words they rememberand are read 24 words that include the original 12 wordsplus 12 semantically related and unrelated words.

Symbol Digit Modalities Test

Smith A (1973) Symbol Digit Modalities Test: manual.Los Angeles: Western Psychological Services

This is a timed measure of visual attention, motor speed,and rapid coding in which participants print the num-ber that corresponds to a written symbol listed at thetop of the test page. The task is then repeated with theparticipant saying the digits.

Visual Search and Attention Test

Psychological Assessment Resources (1990) VisualSearch and Attention Test: professional manual.Odessa, FL: Psychological Assessment Resources Inc.

This is a timed test of visual scanning, target detection,and cancellation in which participants locate and crossout letters or symbols that look like the target.

Test of Attention in Infants—Habituation/Fixa-tion

Test reported by Heywood et al. (1989)

The child was seated on its mother’s lap. Two meters infront of the mother and child was a screen with a peep-hole in the middle. The observer sat behind the screenout of sight of the child and observed and timed thevisual fixation of the child to the stimulus with a cumu-lative stopwatch and recorded the number of lookingepisodes. The stimuli were presented in front of thescreen. Two stimuli were used: a brightly colored blown-up plastic fish or doll. The fish was presented for fourtrials, each lasting 30 seconds, with a 15-second interval.The doll was presented on the fifth trial for 30 seconds.At the start of each trial, the stimulus was pressed so itwould squeak.

Consonant Trigrams

Test used in Groner et al. (1986). No reference pro-vided.

Appendices 65

Rey Auditory Verbal Learning Test

Test used in Groner et al. (1986). No reference pro-vided.

3. Discrimination Learning

(1) Two choices with three-dimensional objects

The child is presented two three-dimensional objects (toycar, toy whistle, etc.) mounted on wooden bases. A yel-low happy face is pasted on the bottom of only one ofthe bases. The child’s task is to discover which stimulushides the happy face underneath it. After each trial thestimuli are rearranged out of the child’s view, and theprocedure continues until a criterion of seven correctresponses in a row is met. The reverse problem is thenadministered with the same criterion.

(2) Two choices with two-dimensional objects

The task is identical to the three-dimensional task ex-cept that the happy face is pasted on the bottom of acardboard picture cut from a child’s book.

Short-Term Memory Task

A large number of two-choice visual discriminationlearning problems consisting of two-dimensional pic-tures are presented concurrently for a total of four trials

each. Trials 1 and 2 are consecutive. Trials 2 and 3 haveeither zero, four, or eight interpolated items separatingthem. Trial 4 occurs 24 hours later. A happy face isposted on the back of the correct stimulus.

Oddity Learning

Three two-dimensional pictures (stimuli) are presentedsimultaneously on a black card, and two of the stimuliare identical. The only instructions given are “to findthe winner.” The correct answer/winner is the picturethat is different from the other two (oddity). In the firsttrials, new stimuli are used each time. In the remainingtrials the same stimuli are repeated each time in an AAB,ABB manner. Thus, the specific character of a stimulusdoes not determine its correctness; rather, it is the rela-tionship of a character to other stimuli in the array.

Matching Familiar Figures Test—Form F

The child is shown two cards: one card, the standard,has a picture on it, and the other card has pictures of sixvariants of the standard. The child has to match thestandard and one of the alternatives. The time taken forthe first response and the errors made before a perfectmatch is achieved are recorded.

The remaining cognitive tests mentioned in the Bibliog-raphy are all tests within the Wechsler Scales. Theseinclude Mazes, Clerical Task, Visual Recall, Digit-Span,Digit-Symbol, Arithmetic, and Vocabulary

��

��