Reducing anaemia in low income countries: control of infection is essentialBMJ 2018; 362 doi: https://doi.org/10.1136/bmj.k3165 (Published 01 August 2018) Cite this as: BMJ 2018;362:k3165
- Sant-Rayn Pasricha, laboratory head1 2 3 4,
- Andrew E Armitage, senior postdoctoral scientist1,
- Andrew M Prentice, professor5,
- Hal Drakesmith, associate professor1 6
- 1MRC Human Immunology Unit, MRC Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, UK
- 2Walter and Eliza Hall Institute of Medical Research, Melbourne, Victoria, Australia
- 3Department of Medical Biology, The University of Melbourne, Melbourne, Victoria, Australia
- 4Diagnostic and Clinical Haematology, Royal Melbourne Hospital, Melbourne, Victoria, Australia
- 5MRC Unit The Gambia at London School of Hygiene and Tropical Medicine, Banjul, Gambia
- 6Haematology Theme, Oxford Biomedical Research Centre, University of Oxford, Oxford, UK
- Correspondence to: S-R Pasricha
Anaemia affects 273 million children and 529 million women globally,1 accounting for 8.8% of all years lived with disability.2 The prevalence of anaemia in children under 5 years is highest in sub-Saharan Africa (62.3%) and South East Asia (53.8%),1 where rates of concomitant infection are also high. Between 1993 and 2011, the estimated prevalence of anaemia worldwide fell only from 33% to 29% in non-pregnant women and from 47% to 43% in preschool children.3 The 2016 Global Nutrition Report found that progress towards the World Health Organization target of a 50% reduction in anaemia in women by 2025 is 100 years behind schedule.4
Iron is an essential micronutrient required for many biological processes, including oxygen transport, mitochondrial function, and numerous enzymatic pathways. Iron deficiency can therefore compromise diverse physiological functions and ultimately lead to anaemia. Public health strategies to control anaemia emphasise iron replenishment: fortification of food with iron, universal distribution of iron supplements, and home fortification of complementary foods with iron containing multiple micronutrient powders.567 Failures in resolving the burden of anaemia have been attributed to programmatic limitations: problems with financing, supplying, distributing, or adherence to interventions.48 However, as discussed in this article, new understanding of iron-infection interactions, evidence from clinical trials, and anaemia epidemiology cast doubt on the safety and prominence of universal iron interventions as the mainstay of public health anaemia control, especially among young children in areas of high infectious burden. Although this has resulted in modifications to WHO iron intervention guidelines, a further change of approach to anaemia control is needed.
New insights into iron-infection interactions
Anaemia control policy in children was upended in 2006 by a large trial in Pemba, Tanzania, that was prematurely terminated owing to an increased risk of death and hospitalisation, as well as other serious infection related events, among children randomised to receive iron folic acid (box 1). Although a parallel trial in Nepal (which is non-malaria endemic) did not detect these adverse effects, the Pemba study provoked urgent policy changes,10 reinvigorated clinical, epidemiological, and experimental research investigating iron-infection interactions, and coincided with rapid advances in fundamental understanding of iron biology. This work, which we describe below, highlights the importance of infection control in public health prevention and control of anaemia.
Iron-malaria policy since Pemba
The Pemba trial (Tanzania) was powered to show a survival benefit in children receiving iron, randomising over 24 000 children aged 6-36 months to receive placebo, iron folic acid, and iron folic acid with zinc. The trial was stopped early because of an increase in death or hospital admission among children randomised to receive iron folic acid (with or without zinc) (relative risk 1.12, 95% CI 1.02 to 1.23). Iron folic acid also increased hospital admissions (RR 1.11, 1.01 to 1.23), cerebral malaria (1.22, 1.02 to 1.46), serious adverse events (1.32, 1.10 to 1.59), deaths (1.61, 1.03 to 2.52), and admissions to hospital (1.28, 1.05 to 1.55) with non-malarial infections. A substudy of 2413 children found that infection in children receiving iron supplementation was increased only in children without elevated baseline zinc protoporphyrin.
A similar trial evaluating iron folic acid in children in a non-malaria endemic setting (Nepal) was published simultaneously. This trial was also designed to show benefit from iron on child survival, but it was terminated prematurely once there was no possibility of a beneficial effect being detected. It found no effects from iron on infections including diarrhoea.9
As a result of the Pemba trial, WHO suspended its policy of universal iron for children in malaria endemic areas, instead recommending iron folic acid only in anaemic children.10 In 2011, based on a Cochrane review, WHO published new recommendations for micronutrient powder implementation to children where the prevalence of anaemia is high, stating that in malaria endemic areas micronutrient powders should be provided only to children with access to malaria preventive and treatment services.5 Over the past decade there has been rapid scaling up of micronutrient powder programmes, up to 50 million children received the supplements in 2015.
Updated 2016 WHO guidelines reintroduced daily iron supplementation as a second option (alongside micronutrient powders) in children, and based on an updated Cochrane review,11 reiterated that in malaria endemic areas iron provision to children “should be done in conjunction with public health measures to prevent, diagnose and treat malaria.” Guidelines further state that “provision of iron … should not be made to children who do not have access to malaria-prevention strategies (e.g. provision of insecticide-treated bednets and vector-control programmes), prompt diagnosis of malaria illness, and treatment with effective antimalarial drug therapy.”7RETURN TO TEXT
Humans have evolved to withhold iron from plasma during infection
The discovery that hepcidin, the master iron control hormone, is an inflammation regulated acute phase peptide highlights the direct interaction between iron homeostasis and innate immunity.12 Hepcidin is secreted mainly by hepatocytes, causing internalisation and degradation of the sole iron exporter ferroportin on macrophages and duodenal enterocytes, thereby preventing iron egress.13 Hepcidin upregulation during infection inhibits dietary iron absorption and macrophage release of recycled iron, resulting in acute reductions in plasma iron levels; this impairs bacterial growth in plasma14 but simultaneously starves developing red cells of iron.
At the high doses typical in supplements, oral iron can overwhelm natural hepcidin mediated constraints on iron absorption.15 Treating patients with iron supplements raises transferrin saturation, providing a medium which enhances microbial growth (for example, Eschericia coli, Yersinia, Salmonella) in ex vivo assays.16 Furthermore, iron deficient individuals absorb iron supplements more efficiently, resulting in a spike in plasma iron that can temporarily overcome transferrin’s binding capacity, causing appearance of non-transferrin bound iron,17 which enhances growth of virulent pathogens such as Yersinia and Vibrio spp.18 Thus, at high doses, iron supplementation may promote pathogen growth by overcoming the hepcidin mediated iron starvation response to inflammation.
Iron deficiency protects children against malaria, and iron administration poses a transitory risk for malaria
Cohort studies in Malawi and Tanzania showed that preschool children with baseline iron deficiency living where malaria transmission is intense and where no prophylaxis was distributed have a substantially reduced risk of subsequent malaria parasitaemia, clinical malaria, severe malaria, and all cause and malaria specific mortality.1920
Correspondingly, two iron intervention studies indicate that treating iron deficiency with iron increases malaria susceptibility. After receiving micronutrient powders containing 18 mg iron (with or without zinc), iron deficient Tanzanian children had a 41% increased risk of malaria, while no increased risk was found among iron replete children.21 Likewise, an increased incidence of malaria was found in HIV infected anaemic children in Malawi in the first three months after iron supplementation.22
Although a 2016 Cochrane review found no overall effect of iron interventions on clinical risk of malaria in children, subgroup analysis stratifying by malaria prevention or treatment access revealed that clinical risk increased by 16% (95% confidence interval 2-31%) when iron was administered without co-provision of prevention or treatment, but declined by 9% (16-3%) when administered with control measures. Furthermore, iron increased Plasmodium parasitaemia risk by 11% (0-23%) during supplementation and by 23% (9-40%) in the post-supplementation period,11 supporting the cohort study findings described above. The consequent 2016 WHO guidelines therefore require that iron interventions be given only where strategies to prevent, diagnose, and treat malaria are operational; however, malaria control in settings where iron interventions may be recommended is often suboptimal (see below).
The above data are supported by experiments showing preferential P falciparum in vitro infection of erythrocytes from iron replete and recently iron supplemented donors, over those from iron deficient donors (with donors comprising healthy US adults23 and Gambian children receiving micronutrient powders containing iron).24 Merozoites preferentially invade young erythrocytes (reticulocytes), which increase during recovery from anaemia after iron treatment, providing a mechanistic explanation for these observations.23 Another recent study noted protection from clinical malaria and parasitaemia in pregnant Ghanaian women and Zambian children carrying the ferroportin Q248H mutation, proposed to be mediated via a decrease in red cell iron.25 Correspondingly, mice with erythroblast specific ferroportin knockout (in which red cells have increased iron), were rendered susceptible to murine malaria. Together, these data suggest a cruel paradox: while children with iron deficiency anaemia are most likely to benefit from iron interventions, effective erythropoietic responses after iron treatment place these individuals at greatest risk through enhancing susceptibility to malaria. Any solution that delivers iron for erythropoiesis and successfully restitutes anaemia by definition induces reticulocytosis, casting doubt that in malarial settings any effective iron intervention (supplement, fortificant, or food) will be safe unless accompanied by reliable malaria prevention.
Malaria is not the only infection exacerbated by iron interventions
In the colon, beneficial bacteria help prevent colonisation by pathogenic species and may not require iron (for example, lactobacilli). By contrast, iron favours growth of enteric pathogens.26 Unabsorbed iron from supplements, micronutrient powders, and fortificants can exacerbate the risk of intestinal colonisation by pathogenic species.27 Studies in non-malaria endemic settings in Pakistan found increased diarrhoea (including bloody diarrhoea) associated with micronutrient powders containing iron,28 besides increased rapid chest indrawing, raising concern that iron may also exacerbate the risk of respiratory infection. By contrast, the large Nepal trial (box 1) found no effects from supplemental iron on infections, including diarrhoea.9 Nevertheless, systematic reviews published in 2016, 2013, and 2002 (that is, before these large trials) each found that iron interventions increased diarrhoea risk (by 15% (6-26%),11 4% (1-6%),29 and 11% (1-23%)30 respectively).
Studies evaluating how iron fortification influences intestinal microbiota provide a plausible mechanism underlying increased diarrhoea after iron treatment. Ivorian children given iron fortified biscuits had reduced commensal lactobacilli, increased pathogenic enterobacteria, and raised intestinal inflammatory biomarkers.31 Likewise, increased carriage of intestinal pathogens including Escherichia, Shigella, and Clostridium and intestinal inflammation was found in Kenyan infants receiving micronutrient powders containing iron.27 Such effects were not seen in South African children with low baseline prevalence of pathogenic bacteria, which suggests a harmful effect of iron on intestinal microbiota may manifest in settings where carriage of pathogenic species is already prevalent, perhaps relating to underlying water, sanitation, and hygiene conditions. These trials were not powered to discern effects on gastrointestinal illness.
Control of malaria alone may decrease iron deficiency
Symptomatic and asymptomatic malaria parasitaemia inhibit iron absorption via hepcidin upregulation. In Beninese women with asymptomatic P falciparum infection, antimalarial treatment increased mean dietary iron absorption from 10% to 18%.32 Complementing this, the prevalence of iron deficiency and iron deficiency anaemia was significantly higher at the end of malaria season compared with the start in rural west and east African preschool children, suggesting suppression of iron absorption by infection.33 Moreover, interruption of malaria transmission for 12 months in rural Kenya reduced the prevalence of iron deficiency in young children from 35% to 26%, and anaemia from 54% to 32%.34
Cochrane reviews indicate that malaria control can improve haemoglobin or reduce anaemia: for example, intermittent preventive therapy reduced moderate anaemia by about 40% in pregnant women35 and 29% in children36; bednet use increased packed cell volume by 1.5%37; and indoor residual spraying improved haemoglobin by 0.85 g/dL (based on a single trial).38 Thus, malaria control alone may alleviate iron deficiency and anaemia even without iron provision.
Iron supplementation to pregnant women in malaria endemic areas is likely safe and improves child outcomes
Two large trials have recently evaluated the effects of antenatal iron supplementation on malaria risk during pregnancy. Kenyan women showed no evidence of increased clinical malaria, placental malaria, or parasitaemia in those randomised to oral iron, even among those who received limited or no intermittent preventive therapy.39 Importantly, iron yielded clear benefits to both mother and child, increasing birth weight by 150 g, lengthening gestation duration by 3.4 days, and reducing risk of low birth weight and premature birth by 58% and 7% respectively. Benefits were enhanced among women with baseline iron deficiency. Similar benefits of iron have been identified previously,4041 but this trial was one of the first to include women with anaemia, potentially explaining the large impact on birth outcomes. In a second recent placebo controlled trial in non-iron deficient, non-anaemic, pregnant women in Tanzania, iron supplementation did not increase placental malaria, although no benefit on birth weight or gestation duration was found, perhaps reflecting higher baseline iron status in this study population.42 These trials indicate that antenatal iron supplementation is safe even where malaria is highly endemic and malaria control strategies incompletely implemented (perhaps because of more developed host immunity and/or tolerance to pathogen load in adulthood),43 and has important benefits on maternal and neonatal health when iron deficiency is prevalent. Detailed investigations of potential risks from iron in pregnancy on other infections are still warranted.
Proportion of cases of anaemia that are iron responsive is lower than previously estimated
Previously, at least half of cases of anaemia were attributed to iron deficiency and therefore potentially iron responsive. Anaemia prevalence in preschool children is highest in Africa, the eastern Mediterranean, and Asia. In these settings, WHO estimates that in preschool children only 32%, 38%, and 41% of cases of anaemia, respectively (42% globally), are iron responsive.1 These figures are higher among women.1 Thus, non-iron responsive anaemia is prevalent in young children in low income, high infection settings—both malaria endemic and non-endemic. The population response to iron is likely determined by the proportion of individuals able to absorb and utilise iron for erythropoiesis. Physiologically, iron absorption and utilisation is determined by hepcidin status: in a study of 1313 Gambian and Kenyan infants, we found that only 27% had hepcidin concentrations below the threshold permitting effective iron absorption and utilisation.44
Causes of anaemia other than iron deficiency in these settings may include anaemia of inflammation, malaria, and carrier states for or clinically evident inherited red cell disorders (for example, thalassaemia, sickle cell disease, and G6PD deficiency). Diagnosing the cause(s) of anaemia remains challenging in resource poor settings: biomarkers for iron deficiency are expensive and difficult to interpret with concurrent inflammation, while sophisticated red cell testing is difficult to implement routinely in the field. Nevertheless, the epidemiology of conditions that cause anaemia should be considered when planning anaemia control programmes because treatment of non-iron deficiency anaemia with iron is at best ineffective and may increase the risk of harm in settings with high infection burden. Coexistent multiple micronutrient deficiencies (for example, of other haematinic micronutrients or vitamin A) could impair the haemoglobin response to iron. Vitamin A deficiency may impair immune function, exacerbating infection risk, promoting inflammation, and impairing iron utilisation. Studies testing effects of vitamin A supplementation (with or without iron) on anaemia have yielded mixed results.45 Data comparing micronutrient powders with iron supplements alone in children are lacking, although a meta-analysis of four trials in pregnant women found no additional benefit from micronutrient powders on anaemia compared with iron alone.46
Environmental enteropathy is a subclinical malabsorptive condition highly prevalent among children in low income countries, attributed to recurrent infection and chronic malnutrition and characterised by intestinal villous atrophy, impaired barrier function, and intestinal inflammation.47 The role of intestinal functional impairment on iron absorption remains unclear, and causal links between environmental enteropathy and iron deficiency or anaemia have not yet been defined. However, it remains plausible that failure of luminal absorptive function may render anaemia non-iron responsive.
Implementation of malaria prevention is often inadequate for safe delivery of universal iron interventions
WHO guidelines currently recommend universal iron or micronutrient powders where anaemia prevalence exceeds 20-40%, even in malaria endemic areas, as long as interventions to “prevent, diagnose and treat” malaria are also provided. The appropriate forms of malaria control to accompany iron interventions are not defined. For example, insecticide treated bednets reduce the incidence of uncomplicated clinical malaria, but only by about 50%.37 A trial in Ghana where all participants used insecticide treated bednets did not find an increase in malaria from micronutrient powders containing iron, but there was a 23% (2-49%) increase in hospitalisation during (but not after) the intervention period, suggesting bednets do not prevent iron induced increases in serious adverse events.48
The coverage of malaria prevention, diagnosis, and treatment measures remains inadequate in high risk settings. The 2016 World Malaria Report indicated that only 57% of at-risk sub-Saharan African populations slept under a bednet (prevention), 56% of febrile children are taken to trained healthcare providers (diagnosis), and 30% of children with evidence of P falciparum infection and a history of fever receive antimalarial drugs (treatment).49 Thus, under routine conditions, it is likely that fewer than half of children in malaria endemic settings would be covered such that iron interventions could be deployed safely; unselected, “universal” distribution is likely to be unsafe. Although the malaria burden has declined over the past two decades, the 2017 World Malaria Report suggests that gains have plateaued or are even being reversed.50
Risk-benefit of iron interventions in infection endemic regions must be considered
An important consideration before iron distribution is whether reductions in the prevalence of anaemia or benefits on functional health outcomes, such as development and growth in young children (not yet supported by high quality evidence from randomised controlled trials5152), justify the increased risk of infections.
Global burden of disease data indicate that in 2016, the disability adjusted life years attributable to diarrhoea, malaria, and iron deficiency anaemia in 1-4 year old children were 11.1 million, 27.3 million, and 1.8 million respectively in sub-Saharan Africa, and 3.5 million, 1.0 million, and 2.4 million respectively in South Asia.53 If iron interventions exacerbate the risk of infectious diseases (even while reducing anaemia) they may raise the overall burden of disease in these settings. By contrast, controlling malaria and diarrhoea decreases the burden of these infections, and by suppressing inflammation and hepcidin, could simultaneously improve iron absorption and utilisation and contribute to reducing the burden of anaemia.
Alleviating the worldwide burden of anaemia remains a critical global health priority. However, settings of greatest prevalence of anaemia commonly coincide with those where infections are also highly endemic. In children, iron alone will not solve the problem of anaemia and could be harmful unless great care is taken to minimise risk of infection. Development and deployment of strategies that control infection will help make iron safer and more effective. As well as improving health outcomes overall, solutions to tackle infection may help alleviate the burden of anaemia independently of iron. To begin to deal with these challenges, we make policy recommendations in box 2 and propose critical research priorities in box 3.
Reducing the burden of anaemia should be considered an important additional rationale for, and outcome of, programmes to control infectious diseases
As recommended by WHO, iron interventions should be withheld from children in malaria endemic settings unless malaria prevention is provided. We propose that this implies delivery of iron and effective malaria control interventions should have identical coverage—malaria prevention and iron interventions should converge at the level of the individual child. In settings of high infection intensity, more aggressive approaches (for example, more than just bednets) such as combining iron interventions with use of malaria chemoprevention, should be explored
When effective in raising haemoglobin, the risk of iron exacerbating malaria is likely to be highest in initially anaemic children. Therefore, when treating anaemic children with iron, they must also receive malaria prevention during recovery
Iron supplementation seems unambiguously beneficial in pregnancy, and hence delivery of, and adherence to, iron should be prioritised in pregnant women in low income countries
Areas for future research
The optimal type of malaria prevention measures to be co-implemented with iron that render iron interventions safe need to be defined in adequately powered prospective studies. The burden of malaria (for example, clinical incidence, parasite prevalence, entomological inoculation rate) and coverage of effective malaria prevention strategies at which universal iron interventions can be considered safe also need to be defined
What are the optimal strategies (technologies and clinical protocols) to improve diagnosis and management of non-iron responsive anaemia in low resource settings?
What is the impact of non-malarial infections (for example, diarrhoea and respiratory and skin infections, all common in low income countries) on iron absorption?
What is the effect of water, sanitation, and hygiene interventions on iron absorption and utilisation, iron deficiency, and anaemia?
What are the lowest doses of iron in pregnancy that achieve desirable clinical outcomes for mother and baby?
Functional health benefits from iron provision in young children (for example, on cognitive and physical development, wellbeing) require precise definition in high quality trials so that evidence based decisions can be made and risk-benefit analysis can be undertaken.
Universal iron interventions designed to alleviate anaemia in young children living in low income countries may exacerbate the risk of infection, especially malaria and diarrhoea
Iron interventions alone will not resolve most of the burden of anaemia in young children living in sub-Saharan Africa and Asia. To achieve anaemia control in high infection settings, tackling infection is at least as important as iron interventions
Strategies to control and prevent malaria (and other infections) may improve the burden of anaemia independently of iron interventions, and if combined with iron, may make iron interventions safer and more effective
Iron intervention should be prioritised in pregnant women, where safety has been shown and benefits to maternal and neonatal outcomes are clear
Contributors and sources: S-RP conceived the manuscript through extensive consultation with other experts in the field at many technical meetings and academic conferences; he wrote the first draft. S-RP has extensive experience in public health, clinical and scientific aspects of anaemia control and biology, has advised several international organisations on aspects of anaemia control, and leads several large randomised controlled trials testing efficacy and safety of anaemia control interventions. AEA and HD are experts in iron-infection interactions through experimental biology. AMP leads several trials aimed at addressing anaemia burden in children and advises several international governmental and non-governmental agencies on nutrition policy. All authors wrote, edited, and approved the final manuscript. S-RP is guarantor for this manuscript.
Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following: S-RP reports grants from Medical Research Council UK, the Bill & Melinda Gates Foundation, the National Health and Medical Research Council Australia, and Cooley’s Anemia Foundation during the study. AEA, AMP, and HD report grants from Medical Research Council UK, the Bill & Melinda Gates Foundation. HD reports support from BRC Blood Theme, NIHR Oxford Biomedical Research Centre, Oxford, United Kingdom.
Provenance and peer review: Not commissioned; externally peer reviewed.