Iron deficiency and the immune response

Iron deficiency and the immune response13

Am J C/in Nuir 1987;46:329-34. Printed in USA. © 1987 American Society for Clinical Nutrition 329

Peter R Dailman, MD

ABSTRACT The importance ofiron deficiency as a public health problem is based ultimately

on the seriousness ofits consequences on health. The most extensively investigated consequencesof iron deficiency involve work performance and immune function. The significance of the

effects on work performance are generally accepted. In contrast, data on the influence of irondeficiency on immune function are often perceived as being confusing and contradictory. From

reexamination ofrelevant literature, it seems safe to conclude that abnormalities in cell-mediatedimmunity and ability of neutrophils to kill several types of bacteria are well established underexperimental conditions in iron-deficient patients. It remains uncertain whether these abnor-malities result in an increased incidence and duration of infections. This area still requires

careful study. Am J C/in Nutr 1987;46:329-34.

KEY WORDS Iron deficiency, infection, immunity

Introduction

In this report, I will first describe the laboratory ab-normalities in immune function that appear most con-vincing and distinguish these from other components of

the immune system about which there is uncertainty orwhich seem to be normal. The emphasis will be on studiesperformed in man. Next, I will try to categorize the epi-demiologic studies aimed at determining whether there is

an increased prevalence of infection in individuals withiron deficiency and indicate the reasons why carefully de-signed studies dealing with this issue are still needed.

There have been many commentaries about iron de-ficiency and immune function (1-1 1). Some reviews fo-cused on the risk of infection with iron excess and raisedthe question of whether a paucity of iron can actuallyconfer benefits (2, 7, 9, 10, 12). The purpose for reex-amining the topic ofiron deficiency and immune functionis that it has been a puzzling one with often conflictingand apparently paradoxical conclusions. It may be helpful

to scrutinize the literature and the problem itself from anadditional perspective.

Defects of the immune response in iron deficiency

A briefoutline ofthe major components ofthe immuneresponse and the laboratory tests by which they are eval-uated is a useful starting point for organizing the resultsof studies of immune function in patients with iron de-ficiency.

Components ofthe immune response

The immune response involves a close interactionamong three major components: cell-mediated immunity,

antibody-mediated immunity, and phagocytosis. Cell-mediated immunity depends on the thymus-derived lym-phocytes (T cells) in the blood and peripheral lymphoidtissues. When an individual is reexposed to a specific type

of infection, T cells multiply and release lymphokines,

substances that amplify the immune response. The sim-plest way to evaluate cell-mediated immunity is to deter-

mine the number of lymphocytes in the blood and thepercentage that are T cells. The in vitro function of Tcells can be tested by the extent to which they multiplywhen exposed to various mitogens in a culture medium.The function of T cells within the body is evaluated by abattery of skin tests. A positive skin test, consisting ofredness and swelling after 24-48 h, depends on T-cellproliferation and is referred to as a delayed hypersensitivityresponse.

Antibody-mediated immunity involves the productionof circulating antibodies by bone marrow-derived lym-

1 From the Department of Pediatrics, University of California, SanFrancisco, CA.

2 Supported in part by NIH grant AM 13897.

3 Address reprint requests to Peter R Dailman, MD, University ofCalifornia, San Francisco, Dept of Pediatrics, Room M-650, San Fran-cisco,CA 94143.

ReceivedAugust 15, 1986.Accepted for publication November 1 1, 1986.

at TO

KY

O IK

A S

HIK

A U

NIV

on February 13, 2012

ww

w.ajcn.org

Dow

nloaded from

330 DALLMAN

phocytes (B cells). IgG, 1gM, and IgA, the three majorclasses of antibodies, can be readily measured in plasma.Even ifthese values are normal, the antibody response toa specific antigen may be defective. For example, mdi-viduals who have a subnormal quantity ofcirculating an-titetanus antibody despite having been immunized againsttetanus would be suspected of having abnormal antibody-mediated immunity.

Phagocytosis or ingestion ofbacteria or fungi is a func-tion ofthe neutrophils in the blood (neutrophils and lym-phocytes are the two most abundant white blood cells)and of so-called fixed macrophages in the spleen, liver,and other tissues. Normal neutrophils exhibit a suddenincrease in oxygen consumption after phagocytosis. Thisincrease is referred to as an oxidative burst and is a pre-requisite for the destruction of many types of bacteria.The oxidative burst depends on an iron-containing b cy-tochrome. The ability ofphagocytes to kill bacteria is mi-tially evaluated by determining the number of neutrophilsin the blood and by the nitroblue tetrazolium (NBT) test,which reflects the capacity for generating an oxidativeburst. The direct quantitation of the oxidative burst re-quires isolation ofneutrophils from relatively large bloodsamples and measurement of oxygen consumption afterthe cells are stimulated to phagocytosis. Another special-ized and difficult test of phagocytic function involvesmeasuring the rate at which neutrophils can ingest andkill specific varieties ofbacteria and fungi. The neutrophiliron enzyme, myeloperoxidase, destroys certain micro-organisms by a different mechanism and can be measured

either histochemically or by a more quantitative deter-mination of enzyme activity.

Over a dozen research papers deal with defects in theimmune response in iron-deficient infants, children, andadults (1 3-28). Some of the earliest and most influentialreports included many individuals who were diagnosed

as iron deficient but who were brought to medical atten-tion because they had an infection (16, 17). However,most of the more recent papers make a point of havingexcluded subjects with infections or additional nutritionaldeficiencies because iron deficiency is difficult to diagnosein the presence of these conditions.

Cell-mediated immunity. The number of T cells wasfound to be depressed in four studies where this was eval-

uated (17, 21, 24, 26). The degree of depression was pro-portional to the severity ofiron deficiency and was �-20%in pregnant women (27) and 35% in children whose he-moglobin was < 80 g/L (21).

Iron-deficient patients have been found more frequentlyto have an absent or diminished skin-test response thando control subjects to a variety of antigens, includingCandida, mumps, diphtheria, trichophyton, and strep-tokinase-streptodornase (13, 16, 17). In one ofthese stud-ies, the skin test abnormalities were reversed 2-3 mo afterintramuscular treatment with iron dextran (16). Re-sponses to tuberculin testing (1 3, 17) and dinitrochioro-benzene (19) did not differ significantly from those incontrol individuals. However, the differences in responseto the majority of skin-test antigens in three studies and

the reversal of the abnormalities with iron treatment inone study constitute fairly convincing evidence for a defectin cell-mediated immunity in iron deficiency. One res-ervation is whether coexistence of infection could havebeen a confounding factor in two of the reports (16, 17).Additional studies to exclude this possibility would beworthwhile.

Reports disagree about the proliferative response ofisolated lymphocytes. Many investigators found no ab-normality in lymphocyte proliferation in iron-deficientpatients (14, 15, 19, 25, 27) whereas others described adepressed response (13, 16, 18, 20) that in some reportswas corrected with iron treatment. Recently a plausiblebasis for these inconsistent results has become evident.The iron-containing enzyme, ribonucleotide reductase, isrequired for cells to produce DNA and to divide. Presum-ably it is a deficiency of this enzyme that accounts forimpaired division of T cells in iron deficiency; however,there is still no information on the effects of iron statuson the activity of ribonucleotide reductase in human oranimal tissues. In most iron enzymes the iron is an in-trinsic part ofthe molecule and is not lost until the entireprotein is broken down. Ribonucleotide reductase is anexception because it requires a continuous supply of ironfrom its environment to replace its loosely held iron (29,30). For this reason the activity of the enzyme may bedecreased in the iron-poor environment of a patient (ac-counting for abnormal skin-test results) but be normal inlymphocyte culture if adequate iron is supplied by theculture medium. As a consequence the finding of abnor-mal skin-test responsiveness carries more weight than theinconclusive studies of lymphocyte proliferation in cul-ture. This previously unrecognized source of error inlymphocyte culture studies was delineated in a recentpaper (31).

Antibody-mediated immunity. Antibody-mediated im-munity has been examined in detail and appears to benormal in iron-deficient individuals. The concentrationsof IgG, IgA, and 1gM were either normal or elevated iniron-deficient patients (16, 17, 24, 26). Production of an-tibodies to specific antigens, such as tetanus toxoid, diph-theria, and Salmonella typhi, were also unimpaired. Onlyone report describes an ‘--20% decrease in the percentageofB cells in severe iron-deficiency anemia (26) but anothershowed no difference from control values in a somewhatless anemic group (24). There is, therefore, little or noevidence for impaired antibody-mediated immunity iniron-deficient humans. However, the evidence for suchan impairment in the iron-deficient rat is convincing (32).

Phagocytosis. The NBT dye test was distinctly abnormalin iron-deficient individuals in two studies reported byChandra (14, 17). The abnormality was reversed whenthe patients were reinvestigated 4-7 d after treatment withiron (14). Conversely, no abnormality in the NBT dyetest was found by either Macdougall et al (16) or Yetginet al (23). Yetgin et al (23) did find a highly significantdecrease in the magnitude of the oxidative burst (p< 0.001) by an assay of hexose monophosphate shuntactivity that is more quantitative than the NBT test. Re-

at TO

KY

O IK

A S

HIK

A U

NIV

on February 13, 2012

ww

w.ajcn.org

Dow

nloaded from

IM�N DfflCI�Nfl’ � INW�’J�331

suIts in respect to myeloperoxidase activity were discor-dant: normal activity in one report (23) but depressedactivity in another (22).

There is strong evidence for a deficit in the ability ofneutrophils to kill bacteria. A marked decrease was ob-served in killing ofStaph aureus (14, 17, 23), Staph a/bus(16), and E co/i (21, 28). Reversal ofthe abnormality wasreported within 4-7 d by Chandra (14) and within < 15

d by Walter et al (28). It was much slower in the study ofYetgin et al (23) where bacterial killing was improved butstill very abnormal after 1.5 mo ofiron therapy and com-plete after 3 mo oftherapy. Only in one report was thereno abnormality in bacterial killing when E co/i was usedas the test organism (15). When the data on bacterial kill-ing are examined in their entirety, a deficit in this functionis among the abnormalities ofimmune function that havebeen most convincingly demonstrated in iron deficiencyand that would be expected to lead to an increased mci-dence of infection.

Iron deficiency and infection

Because the first major study dealing with iron de-ficiency and frequency of infections was conducted�-60 y ago (33), it may seem surprising that the evidencefor a cause and effect relationship is still considered in-conclusive. It is important to appreciate that ethical issuesand constraints were often a basis for the selection of ex-perimental designs that subsequently yielded ambiguousresults. In addition, the confounding role of acute infec-tions in the diagnosis of iron deficiency has only been

appreciated very recently.Most ofthe studies ofiron deficiency and infection (33-

46; G Heresi et al, unpublished observations) fall into oneof two categories, each involving difficulties in interpre-tation. In the first category an association between irondeficiency and frequency ofinfection was looked for, usu-ally among poor populations. Both iron deficiency andinfections are known to be associated with poverty andlack ofeducation (and probably also with crowded housingand contaminated water supplies). Consequently if aniron-deficient population is found to have a high rate of

infection, it is difficult to prove that iron-deficiency playedan independent role from that ofpoverty, poor education,and associated adverse socioeconomic factors. In the see-ond category prevalence of infections was reported in apopulation that received additional iron compared witha control group that did not. In most instances, the dietor management ofthe two groups also differed in respectsother than iron administration. Even when such studieswere well designed in terms of randomized case selectionand use ofa placebo group, ethical considerations requirediron treatment in the more severe cases ofiron deficiency.Consequently only those milder cases that would be lesslikely to have demonstrable deficits in immune functionwere available for investigation.

In both categories, the criteria used for diagnosing irondeficiency and the manner in which infections were doe-

umented were often inadequate by current standards.

There is now general agreement that the combination ofanemia, low serum ferritin, and an additional laboratoryabnormality (mean corpuscular volume, transferrin sat-

uration, or protoporphyrin) constitutes strong evidenceofiron deficiency. Anemia and low transfernin saturationalone are ambiguous since they may also be present notonly during severe and mild infections but also for sometime after the clinical manifestations ofan infection havebeen resolved. A low serum ferritin by itself merely in-dicates low iron reserves and would not be expected tobe associated with any physiological handicaps. Docu-mentation of infections is most reliably done at frequentintervals by a health aide who visits the home to observeand to obtain a history; a history obtained infrequentlyor on a single occasion is far less conclusive. With theseconsiderations as a background, one can understand whyproof of a relationship between iron deficiency and in-fection rate is difficult to obtain.

The earliest large study of the effect of iron deficiencyon the frequency of infections was that of Mackay (33),who pioneered the use of iron-fortified cow’s milk in theEast End of London in 1928. The use of iron-fortifiedmilk powder or iron medication resulted in a higher he-moglobin concentration and a rate ofrespiratory and gas-trointestinal disease that was about half that of the un-treated control group. The study group of54l infants waslarge and the patients were carefully followed. However,

the infants were not randomly assigned to iron-treatedand control groups, the two groups were studied sequen-tially rather than concurrently, and different forms of ironwere administered in the two clinics that participated inthe study: one used iron-fortified powdered milk and the

other used an iron medication with which there was aproblem ofcompliance. Mackay recognized all these flawsin experimental design but she and others were neverthe-less impressed by the dramatically lower morbidity in theiron-treated group. The paper was very influential andstill is quoted widely.

The next large study was conducted by Andelman andSered (34) almost 40 y later and included > 1000 infants

from a low socioeconomic group in Chicago. Patients wererandomly assigned to iron and control groups. However,the iron group received an iron-fortified milk formula thatwas provided gratis whereas an evaporated milk formulaand a separate vitamin supplement (without iron) wererecommended for the control group. Information aboutillnesses was only obtained at routine clinic visits. Theiron group had significantly fewer respiratory infections(by at least #{189}during each of the 3-mo intervals duringthe first year oflife (p < 0.0005). Because iron deficiencyis unusual before age 6 mo, other factors associated withthe diets and formula preparation in the two groups prob.-ably were responsible for the difference in frequency ofinfections.

In the intervening years there have been a number ofadditional studies, some of which indicated no predis-position to infection in iron-deficient individuals whereasothers did. In some studies of iron-treated vs control in-

at TO

KY

O IK

A S

HIK

A U

NIV

on February 13, 2012

ww

w.ajcn.org

Dow

nloaded from

332 DALLMAN

dividuals, iron deficiency was either mild or nonexistentin the control group as judged, for example, by a lack ofdifference in hemoglobin concentration in the two groups(38, 40). In such circumstances the lack ofa difference inthe infection rate is inconclusive. Other studies report onlyan association of anemia and low serum iron with infec-tion (34, 41, 46). It is now recognized that infection resultsin these laboratory changes irrespective of iron status.Consequently an association of these findings is likely tobe unrelated to iron deficiency.

Two other studies deserve comment because the mor-bidity data were collected frequently and with great care.In a recent field trial in Chile (G Heresi et al, unpublishedobservations), morbidity data were collected on 100 in-fants who received either a milk formula fortified withiron or a similar unfortified formula between ages 3 and15 mo, a period when infections are normally common.Laboratory studies were done at age 3, 9, and 15 mo andmorbidity was closely monitored by a diary and by weeklyhome visits. In this closely monitored study, anemia waspresent in � 30% of the unsupplemented group and 5%the supplemented group but there was no difference infrequency ofrespiratory or diarrheal disease. The negativeresults indicate that mild iron deficiency resulted in nomajor increase in morbidity. However, small differencesin morbidity could have been missed, because the totalnumber of anemic infants was relatively low: only 14 of47 in the nonfortified group vs 2 of 53 in the fortifiedgroup. In addition, the large proportion of infants thathad to be excluded makes one less confident about inter-preting the results.

An analogous study of 383 village preschool childrenin India (45) showed only a small hemoglobin differencebetween an iron-plus-folate-treated group and a placebo-treated control group and no significant difference in fre-

quency or duration of respiratory or enteric infections.Morbidity data also were collected during weekly homevisits. After 12 mo of supplementation, 1 1% of the sup-plemented group was anemic compared with 26% of theunsupplemented group. However, hemoglobin resultswere obtained on only about half of the children and theaccuracy of the results was limited by the use of skinpuncture samples collected on a filter paper disc. Althoughthere were cultural obstacles to obtaining better and morecomplete blood samples, the data suggest that iron defi-ciency could not have resulted in a very large increase inmorbidity but do not exclude the possibility of smallerbut important changes in incidence and duration of in-fection.

The situation may be different among populations withmore severe iron deficiency and/or poorer environmentalsanitation. This possibility was suggested by the study ofCantwell (39) in which Maori newborn infants were ran-domly selected to receive 200-250 mg iron as iron dextranor no injection. Between ages 9 and 18 mo, the meanhemoglobin concentration in the untreated group aver-aged ‘...-90 g/L and was > 20 g/L below that of the iron-treated group. Unfortunately, morbidity was only eval-uated by the number of hospitalizations for 2 y, 42% in

the untreated group of 144 and 32% in the treated group

of94. Clearly, morbidity in this study was less thoroughlymonitored than in the Chilean and Indian studies andthe difference in number of hospitalizations between theiron-treated and untreated groups was relatively smallthough statistically significant.

Iron deficiency and immune responsein experimental animals

Well-designed experiments with animals have the ad-vantages ofdietary iron as the major experimental variableand of uniform conditions of age, duration, and severity(32, 47-54). The animal experiments support the human

data by demonstrating convincing impairments in the de-layed-hypersensitivity skin-test response (cell-mediatedimmunity) (49). Phagocytic function is subnormal asjudged by decreased bacterial killing and NBT reduction(54), low activities of myeloperoxidase, and a decrease inthe oxidative burst (53).

The animal studies differ from human studies in thatthey show an impairment in humoral or antibody-me-diated immunity. Nalder et al (32) found a substantialdecrease in the antibody response to tetanus toxoid inrats. More recently Kuvibidila et al (50) and Kochanowskiand Sherman (55, 56) used the sensitive Jerne plaque assayto assess the production of IgG and 1gM in spleen cellsfrom iron-deficient rodents. There was a marked decreasein both compared with control (50, 55, 56) and pair-fed

animals (50). The abnormality could be reversed in micemade iron-deficient after weaning (50) but not in rats thatwere deficient during the nursing period (55). These find-ings in experimental animals may make it worthwhile toreexamine this issue in man, where no deficit in antibody-mediated immunity has been demonstrable.

Nutritional immunity

Recently it was proposed that iron administration mightactually be harmful by predisposing to infection. Thereis a great deal ofexperimental evidence that iron-bindingproteins protect animals from infection by withholdingiron from the invading organisms that require it for theirgrowth (12). This phenomenon, sometimes referred to asnutritional immunity, explains why excessive adminis-tration ofiron in large doses by injection could predisposeto infection. In most studies an increased risk of infectionwas demonstrated only when transferrin was nearly sat-

urated with iron as it would be in diseases associated withexcessive storage of iron. Under ordinary circumstances,transferrin is < 35% saturated and only minimal changesin saturation occur with consumption of iron-fortifiedfoods or treatment with oral iron (3).

Iron-deficient mice had an increased survival after in-fection with Salmonella (57) and mice with depleted iron

stores were more resistant to Trypanosoma cruzi infectionthan were controls (58). These studies indicate that mildiron deficiency may actually protect against infection un-der very specific laboratory conditions.

at TO

KY

O IK

A S

HIK

A U

NIV

on February 13, 2012

ww

w.ajcn.org

Dow

nloaded from

IRON DEFICIENCY AND INFECTION 333

Optimal intake of iron

The possibility that iron deficiency could both impairand enhance the immune response has been described aspresenting a paradox or a dilemma (9, 10). We are ac-customed to the idea that the deficiency of a nutrient isharmful in all respects. However, in the case ofiron (andprobably other nutrients such as unsaturated fatty acids),it is becoming difficult to avoid the conclusion that theoptimal range ofintake differs according to the conditionsunder which a specific body or cell function is examined.For example, in a protected environment an iron intakelow enough to impair oxygen delivery and ATP produc-tion may actually prove to decrease mortality and mor-bidity with certain types ofinfections (57, 58). It is doubt-ful that the same would be true under most real life cir-cumstances that require the body to respond to very manydifferent environmental stimuli and hazards. For this rca-

son, what we consider the optimum intake for a nutrientincreasingly can be seen to represent a composite or com-promise among often conificting values for individualbody functions and environmental conditions.

Conclusion

The literature dealing with the effects ofiron deficiencyon the immune response has become quite large. Theevidence for defects in cell-mediated immunity and bac-terial killing is impressive but it is more difficult to de-termine from existing studies whether morbidity from in-fections is increased. Consequently there is still a need todetermine whether iron deficiency predisposes to infec-tion.

Several lessons drawn from the difficulties in inter-preting existing reports can be applied to the design offuture studies. A first prerequisite is a sufficiently largepopulation to assure that small differences in morbiditycan be detected. A revealing and authoritative survey ofrandomized clinical trials showed that many therapieshave been labeled ineffective because sample size was in-adequate to provide a fair test (59). The technical aspectsof selecting the proper sample size for a clinical trial de-serve attention and were discussed in the paper. An ap-

propriate population for study should also have a largepercentage of individuals who meet strict diagnostic cri-teria for iron deficiency including, at the minimum, ane-mia and low serum ferritin. The diagnosis should be madewhen the individual is entirely well and when there hasbeen no infection within the previous 2-4 wk. Once anappropriate population has been identified, iron-treatedand placebo groups must be randomly selected and com-pliance should be verified. Morbidity should be recordedat least once a month, ideally at more frequent intervals,and should include minor illnesses. Lastly, the durationof the study should be at least 6 mo.

The design ofsuch a study certainly represents a majorchallenge. However, it will remain difficult to evaluate theimplications of laboratory abnormalities in immune

function in the iron-deficient patient until there is con-elusive evidence regarding morbidity. I anticipate that thesubstantial laboratory abnormalities in immune functionin iron deficiency will eventually be shown to be associatedwith a decreased resistance to infection. U

References

1. Dailman PR. Biochemical basis for manifestations ofiron deficiency.

Ann Rev Nutr l986;6: 13-40.

2. Luckens iN. Iron deficiency and infection. Am J Dis Child l975;l29:

160-2.3. Committee on Nutrition, Academy of Pediatrics. Relationship be-

tween iron status and incidence of infection in infancy. Pediatrics

l978;62:246-S0.

4. Strauss RG. Iron deficiency, infectious, and immune function: areassessment. Am J Gin Nutr l978;3l:660-6.

5. Chandra RK. Interactions of nutrition, infection and immune re-sponse. Acts Paediatr Scand l979;68: 137-44.

6. Gross RL, Newberne PM. Role ofnutrition in immunologic function.Physiol Rev l980;60: 188-304.

7. Stockman JA. Infections and iron. Am J Dis Child l981;135:18-20.

8. Beisel WR. Single nutrients and immunity. Am J Gin Nutr1982;35(suppl):4 17-68.

9. Humbert JR. Mcore LL. Iron deficiency and infection: a dilemma.J Pediatr Gastroenterol Nutr l983;2:403-6.

10. Keusch GT, Farthing MJG. Nutrition and infection. Ann Rev Nutrl986;6: 131-54.

1 1. Vyas D, Chandra RK. Functional implications of iron deficiency.In: Stekel A, ed. Nutrition in infancy and childhood. New York,

NY: Raven Press, 1984:45-59.12. Weinberg ED. Iron withholding: a defense against infection and

neoplasia. Physiol Rev 1984;64:65-l02.13. Joynson DHM, Walker M, Jacobs A, Dolby AE. Defect of cell-

mediated immunity in patients with iron-deficiency anaemia Lancet1972;II: 1058-9.

14. Chandra RK. Reduced bactericidal capacity ofpolymorphs in irondeficiency. Arch Dis Child 1973;48:864-6.

15. Kulapongs P, Suskind R, Vithayasi V, Olson RE. Call-mediated

immunity and phagocytosis and killing function in children withsevere iron-deficiency anaemia. Lancet l974;II:680-91.

16. Macdougall LG, Anderson R, MacNab GM, Katz J. The immune

response in iron-deficient children: impaired cellular defense mech-

anisms with altered humoral components. J Pediatr 1975;6:833-43.17. Chandra RK. Impaired immunocompetence associated with iron

deficiency. J Pediatr l975;86:899-902.18. Fletcher J, Mather J, Lewis Mi, Whiting G. Mouth lesions in iron-

deficient anemia: relationship to Candida albicans in saliva and toimpairment oflymphocyte transformation. J Infect Dis 1975;l31:

44-50.19. Gross RI, Reid JVO, Newberne PM, Burgess B, Marston BAR, Hift

W. Depressed cell-mediated immunity in megaloblastic anemia dueto folic acid deficiency. Am J Gin Nutr 1975;28:225-32.

20. Sawitsky B, Kanter R, SaWItSky A. Lymphocyte response to phy-tomitogens in iron deficiency. Am J Med Sci l976272:l53-60.

21. Srikantia SG, Bhaskaram C, Prasad JS, Krishnamachari KAVR.Anaemia and immune response. Lancet 1976;II:l307-9.

22. Prasad JS. Leucocyte function in iron-deficiency anemia. Am J GinNutr 1979;32:550-2.

23. Yetgin 5, Altay C, Ciliv G, Laleli Y. Myeloperoxidasc activity andbactericidal function of PMN in iron deficiency. Acta Haematol(Basal) 1979;6l:lO-4.

24. Bagchi K, Mohanram M, Reddy V. Humoral immune response in

at TO

KY

O IK

A S

HIK

A U

NIV

on February 13, 2012

ww

w.ajcn.org

Dow

nloaded from

334 DALLMAN

children with iron-deficiency anaemia. Br Med J l980;280:l249-51.

25. Gupta K�K, Dhatt PS, Singh H. Cell-mediated immunity in childrenwith iron-deficiency anaemia. Indian J Pediatr l982;49:507-l0.

26. Prema K, Ramalakshmi BA, Madhavapeddi R, Babu S. Immunestatus of anaemic pregnant women. Br J Obstet Gynaecol l982;89:

222-5.27. Grosch-W#{246}rner I, Grosse-Wilde H, Bender-G#{246}tze Ch, Schafer

KH. Lymphozytenfunktionen bei Kindern mit Eisenmangel. KilnWochenschr l984;62: 109 1-3.

28. Walter T, Arredondo 5, Ar#{233}valoM, Stekel A. Effect ofiron therapyon phagocytosis and bactericidal activity in neutrophils of iron-dc-

ficient infants. Am J Gin Nutr l986;44:877-82.29. Thelander L, Graslund A, Thelander M. Continual presence of ox-

ygen and iron required for mammalian ribonucleotide reduction:possible regulation mechanism. Biochem Biophys Res Communl983;l 10:859-65.

30. Lederman HM, Cohen A, Lee JW, Freedman MH, Gelfand EW.

Deferoxamine: a reversible S-phase inhibitor ofhuman lymphocyteproliferation. Blood l984;64:748-53.

31. Mainou-FowlerT, BrockiH. Effect of iron deficiency on the response

of mouse lymphocytes to conconavalin A: the importance of trans-ferrin-bound iron. Immunology l985;54:325-32.

32. Nalder BN, Mahoney AW, Ramakrishnan R, Hendricks DO. Sen-

sitivity of the immunological response to the nutritional status of

rats.J Nutr l971;l02:535-42.33. Mackay HMM. Anaemia in infancy: its prevalence and prevention.

Am J Dis Child l928;3:l 17-46.34. Andelman MB, Sered BR. Utilization ofdietary iron by term infants�

Am J Dis Child l966;l 11:45-55.35. Giles C, Brown JAH. Urinary infection and anaemia in pregnancy.

BrMedJ 1962;2:l0-3.36. Little PJ. The incidence ofurinary infection in 5000 pregnant women.

Lancet l966;2:925-8.37. Lovric VA. Normal haematologic values in children aged 6 to 36

months and socio-mcdical implications. Med J Aust 1970;2:366-

77.38. Burman 0. Haemoglobin levels in normal infants aged 3 to 24

months and the effect ofiron. Arch Dis Child 1972;47:26l-7l.

39. Cantwell Ri. Iron deficiency anaemia of infancy. Gin Pediatr1972;l 1:443-9.

40. Fuerth JH. Iron supplementation of the diet in full term infants: acontrolled study. J Pediatr l972;80:974-9.

41. Higgs JM, Wells RS. Chronic muco-cutancous candidiasis: associatedabnormalities of iron metabolism. Br J Dermatol 1972;86(suppl):88-102.

42. Walker DM, Dolby AE, Joynson DHM, Jacobs A. Candida and theimmune defects in iron deficiency. J Dent Res l973;52:938-9.

43. Fletcher J, Mather J, Lewis Mi, Whiting G. Mouth lesion in iron-

deficient anemia: relationship to Candida a/bicans in saliva and toimpairment oflymphocyte transformation. J Infect Dis l975;13l:44-50.

44. Davidson F, Hayes JP, Hussein S. Recurrent genital candidosis andiron metabolism. Br J Vener Dis l977;53: 123-5.

45. Damsdaran M, Naidu AN, Sarma KVR. Anemia and morbidity in

rural preschool children. Indian J Med Res l979;69:448-56.

46. Oppenheimer Si. Anaemia of infancy and bacterial infections inPapua, New Guinea. Ann Trop Med Parasitol l980;74:69-72.

47. Baggs RB, Miller SA. Nutritional iron deficiency as a determinant

ofhost resistance in the rat. J Nutr 1973;l03:l554-60.48. Duncombe VM, Bolin TD, Davis A. The effect ofiron and protein

deficiency on the development ofacquired resistance to reinfectionwith Nippostrongy/us brasiliensis in rats. Am J Gin Nutr l979;32:553-8.

49. Kuvibidila SR. Baliga BS, Suskind RM. Effects of iron deficiencyanemia on delayed cutaneous hypersensitivity in mice. Am J GinNutr 198 l;34:2635-40.

50. Kuvibidila SR. Baliga BS, Suskind RM. Generation ofplaque form-

ing cells in iron deficiency anemic mice. Nutr Rep Int l982;26:861-71.

51. Soyano A, Candellet D, Layrisse M. Effect ofiron deficiency on themitogen-induced proliferative response ofratlymphocytes. Int ArchAllergy Appl Immunol l982;69:353-7.

52. Kuvibidila 5, Nauss KM. Baglia BS, Suskind RM. Impairment of

blastogenic response ofsplenic lymphocytes from iron-deficient mice:in vivo repletion. Am J Gin Nutr 1983;37:15-25.

53. MacMer B, Person R, Ochs H, Finch CA. Iron deficiency in the rat:effects on neutrophil activation and metabolism. Pediatr Res l984;l8:549-51.

54. Moore LL, Humbert JR. Neutrophil bactericidal dysfunction towardsoxidant radical-sensitive microorganisms during experimental irondeficiency. Pediatr Res 1984;l8:684-9.

55. Kochanowski BA, Sherman AR. Decreased antibody formation iniron-deficient rat pups-effect of iron repletion. Am J Gin Nutrl985;4l:278-84.

56. Kochanowski BA, Sherman AR. Cellular growth in iron deficientrats: effect ofpre- and postweaning iron repletion. J Nutr l985;llS:

279-87.57. Puschmann M, Ganzoni AM. Increased resistance of iron-deficient

mice to Salmonella infection. Infect Immun l977;l7:663-4.58. Lalonde RG, Holbein BE. Role of iron in Trypanosoma cruzi in-

fection in mice. J Gin Invest 1984;73:470-6.59. Freiman JA, Chalmers TC, Smith H, Kuebler RR. The importance

of beta, the type II error and sample size in the design and inter-pretation of the randomized control trial. Survey of 7 1 “negative”

trials. N Engl J Med 1978;299:690-4.

at TO

KY

O IK

A S

HIK

A U

NIV

on February 13, 2012

ww

w.ajcn.org

Dow

nloaded from