Maternal gestational iron status and infant haematological and neurodevelopmental outcomes

Prevention of iron deficiency (ID), the most common micronutrient deficiency in infants and children, begins prenatally by ensuring adequate fetal loading. Adequate intrauterine iron status is crucial for normal fetal brain development, postnatal brain performance and prevention of early postnatal iron deficiency, particularly in infants fed exclusively human milk. Adequate fetal loading may be achieved in some cases through adequate maternal iron levels prior to pregnancy and oral iron supplementation during pregnancy. However, because so many women are iron‐deficient leading up to pregnancy, coupled with the negative iron balance induced by pregnancy, a large number of women remain iron‐deficient during pregnancy. More consistent iron‐specific early screening and more effective iron delivery approaches are needed to solve this global problem.


| I N TRODUC T ION: ROL E OF IRON I N BR A I N DEV E LOPM E N T A N D F U NC TION
Iron is a fundamental nutrient for optimal brain development and function from conception until senescence.][3][4][5][6][7][8][9][10][11][12][13][14][15][16] Iron deficiency (ID) is the most common micronutrient deficiency in the world, affecting close to 2 billion people, mostly women and their offspring. 17he risk for ID varies across the lifespan of humans and depends on factors such as iron availability and rates of utilisation.Many women enter pregnancy with compromised iron status due to a prior history of low dietary iron intake and excessive blood loss, especially from menstruation. 18regnancy increases iron demand over the non-gravid state because of increased maternal, placental and fetal iron demands. 19Postnatally, iron requirements are greatest in the first years of life when the brain is most rapidly growing. 20ron deficiency during the first 1000 days from conception has been associated with multiple neurodevelopmental abnormalities, most of which are consistent with the known role of iron in the brain (Table 1).

| FETA L IRON ACCR ETION A N D R ISK FAC TOR S FOR R EDUCED FETA L IRON L OA DI NG
The fetus accretes the majority of its iron during the third trimester, maintaining an average of 75 mg of elemental iron per kilogram bodyweight. 21The vast majority of fetal/ neonatal iron, 55 mg/kg bodyweight, is found in the form of haemoglobin.Unless an acute fetal haemorrhage occurs, newborns are rarely anaemic at birth because of the prioritisation of available fetal iron to the red blood cells. 22The term infant has 12 mg/kg bodyweight of storage iron, mostly stored in the liver as ferritin. 21The mean serum ferritin concentration in the term infant is 135 mcg/l, considerably higher than later in childhood and adulthood. 23These liver stores are utilised in the first months of postnatal life to support blood production and tissue iron requirements while the infant is on breast milk, which has a low iron content.Serum ferritin concentrations decrease during the first 18 months as these stores are catabolised. 24In gestational conditions where iron transport to the fetus is compromised, nonhaematological tissues (brain, heart, skeletal muscle) and storage tissue (predominantly liver) are affected before the red cells, resulting in anaemia being the last stage of ID. 25,26 Therefore, newborns with ID are identified by primarily by having a low serum ferritin concentration, indicating that they have consumed their iron stores for the purposes of maintaining tissue and red cell iron status.When iron stores are depleted, as indexed by a low serum ferritin, the brain is at risk for tissue ID. 27,28 Maternal ID is the most common risk condition.Using sensitive tests that detect pre-anaemic ID, Auerbach et al. 29 documented a 42% rate of ID in a random, relatively low-risk cohort of US women.The figure in lower resource countries may be as high as 85%.Pregnancy contributes to negative iron balance because of expansion of the maternal blood volume (3.46 mg of elemental iron are required for each gram of haemoglobin) as well as feto-placental iron demands. 19lthough maternal fetal iron transport favours the fetus at the expense of the mother, fetal iron status is ultimately compromised when the entire maternal-placental-fetal triad is sufficiently iron-deficient. 30A maternal haemoglobin concentration <100 g/l or a serum ferritin concentration <13.4 mcg/l defines thresholds where fetal iron status begins to be compromised. 30,31Preterm birth in the absence of any other maternal-fetal iron pathology compromises fetal iron accretion, as iron is accreted largely in the third trimester. 23etal iron status can also be compromised in ironsufficient women by two common gestational conditions: maternal hypertension and maternal glucose intolerance. 23,25,26Half (50%) of intrauterine growth-restricted fetuses born to mothers with hypertension have low ferritin concentrations. 25Maternal glucose intolerance, whether present prior to gestation or limited to gestation, causes low fetal iron stores via a complex fetal pathophysiology. 32Maternal hyperglycaemia results in fetal hyperglycaemia and fetal hyperinsulinaemia.Both cause increased fetal oxygen consumption beyond the placental capacity to transport oxygen, thereby resulting in fetal hypoxaemia.The fetus mounts a compensatory erythropoietic response resulting in polycythaemia, a process that requires 3.46 mg/g of haemoglobin synthesised.The fetal iron requirement in this condition exceeds the placental capacity to transport iron, 33 thereby forcing the fetus to utilise iron stores and divert any available iron from developing tissues such as the brain for erythropoiesis. 26This diversion is consistent with the principle that fetal iron is prioritised over red blood cells, as well as the corollary principle that anaemia is a late and therefore poor biomarker of tissue, including brain, ID. 20,22,28 T A B L E 1 Neonatal and early infancy neurodevelopmental abnormalities induced by fetal iron deficiency as related to known iron functions in the developing fetal/neonatal brain.

Role of iron Peak timing risk in humans
Acute brain effect in preclinical models

Likely related neurobehavioral effect in humans
Adequate myelination depends on spectrum of fatty acids synthesised by ironcontaining desaturases 1 32 weeks post-conceptional age -2 years postnatal age 2 Hypomyelination 3 Slow speed of brain processing 4 ATP production through oxidative phosphorylation depends on iron-containing cytochromes 5,6 Brain growth is regional.Each region has different timing of peak growth.Peak hippocampus growth from 28 weeks post-conceptional age to 18 months postnatal age 2 Reduced oxygen consumption rate and ATP generation results in simplified dendrite, axon and synapse production in hippocampus 7,8 Reduced learning and memory capacity 9,10 Monoamine neurotransmitters are synthesised by hydroxylases that depend on iron for their enzymatic activity 11 Mid-gestation to 3 years of age 12 Altered dopamine, serotonin and norepinephrine metabolism 13 Poor maternal-infant bonding 14 Iron regulates brain gene expression via epigenetic modification of chromatin through iron-containing histone demethylases 15 Unknown Gestational/lactational ID reduces Brain-Derived Neurotrophic Factor expression and induces autism and schizophrenia pathways in hippocampus 16 Reduced learning and memory capacity 9,10 Each of these fetal is associated with poorer neurodevelopmental outcome.

| Neurodevelopmental consequences of fetal iron underloading
Studies over the past 20 years have delineated risks of fetal iron underloading to the developing brain.Neurodevelopmental risks are closely related to the timing of iron insufficiency during gestation, ranging from the peri-conceptional time-period through the end of the third trimester.Low peri-conceptional maternal iron intake is correlated to a greater risk of autism in the offspring. 34First and second trimester iron deficiency increases the risk for low birthweight (LBW), due to either fetal growth restriction (FGR) or preterm delivery; 35 both conditions increase the risk of poorer developmental outcome.Also, second trimester ID is associated with an increased risk of schizophrenia. 36][39][40][41][42][43] Recent studies continue to document the negative effects of fetal iron underloading on long-term neurodevelopment 39,40,43 and the potential positive effect of maternal supplementation on haematological health and neurodevelopment. 41McCarthy et al. 39 assessed cord serum ferritin concentrations and 5-year developmental outcome in 697 newborns.Infants with cord ferritin concentrations <76 mcg/l had more internalizing and total behavioural problems at 5 years of age. 39Geng et al. 43 reported on neurodevelopmental outcomes of infants enrolled in the preand postnatal ID study by Shao et al. 31 Using event-related potentials to assess cognitive processing, they showed that infants with neonatal ID, defined as a serum ferritin concentration <76 mcg/l) had more cognitive processing problems at 9 months than did infants who had strictly postnatal ID. 43 El Alfy et al. 40 compared 50 infants born to mothers with ID anaemia with 50 infants born to iron-sufficient mothers.Infants born to ID mothers had lower cord blood haemoglobin and ferritin concentrations and longer latencies on auditory brainstem responses indexing slower neural processing speed.Arija et al. 41 showed that children born to mothers with serum ferritin concentrations between 12 and 60 mcg/l and daily iron intakes of 14.5-30 mg had better executive function and working memory at 7 years of age.
Studies by Christian's group are particularly salient in demonstrating the neurodevelopmental importance of maintaining iron sufficiency in pregnancy.This group randomised pregnant women at high risk for ID to prenatal iron and micromineral supplements or placebo and assessed subsequent child neurodevelopment. 38,42Children born to mothers who were supplemented had better neurodevelopmental performance at school age.Postnatal supplementation of the children whose mothers received placebo during pregnancy did not show neurodevelopmental benefits, clearly demonstrating the importance of fetal iron on brain development. 42

| Recent discoveries causing a paradigm shift for screening and treatment regimens of pregnant women and their children
Studies in the past two decades have clarified that ID in infants and toddlers is not solely, or perhaps even primarily, a function of postnatal dietary intake and intestinal blood loss as previously thought. 44Historically, from the 1960s to the 1990s, it was thought that neonates could not be irondeficient because infants born to iron-deficient mothers were not anaemic at birth.Therefore, the major studies of the effect of postnatal iron deficiency on the developing brain of infants and toddlers did not factor fetal iron endowment into the equation and study design.Postnatal strategies to prevent ID were aimed at fortifying or supplementing infant foods with iron.Practitioners, researchers and public health policymakers questioned whether the breastfed infant could maintain iron sufficiency, given the low iron content of human milk.The risks of enteral iron supplementation on young infants, including slow growth and altered microbiome, were also not appreciated at that time. 45,46he two research fields of fetal/neonatal and postnatal iron deficiency in humans were quite separate despite evidence from rodent models that induced postnatal iron deficiency by feeding pregnant dams iron deficient diets.The merging of the two research fields combined with research on the value of delayed cord clamping on subsequent postnatal iron status resulted in an elegant, unified conceptualisation of early life iron metabolism in humans.
The first important paradigm shift occurred with the discovery of the importance of delaying umbilical cord clamping for up to 2 minutes after delivery.'Delayed' cord clamping is more physiological than immediate umbilical cord clamping and results in the transfer of sufficient iron to increase the infant's haemoglobin concentration by 4 g/l at 2 months of age. 47sing a factorial approach, Dewey calculated that an appropriate weight-for-gestational-age term infant who receives delayed cord clamping, is exclusively breastfed, and grows at the average rate of the WHO standard curve, has sufficient iron stored from gestation until at least 4 months without the need for dietary iron supplements.The calculation underscored that the amount of iron in human milk is adequate to maintain iron sufficiency in newborn infants, consistent with WHO and AAP guidelines that advocated exclusive breastfeeding for the first six postnatal months for term infants. 48ubsequent research by the Zimmermann and Krebs groups independently found that supplementary dietary iron given to young infants induced intestinal dysbiosis, a risk factor for diarrhoeal disease. 45,46The altered microbiome was hypothesised to be due to unabsorbed dietary iron-stimulating growth of potentially pathogenic siderophilic bacteria in the colon including Escherichia coli and Salmonella.Beneficial bacteria, e.g.lactobacillus, found characteristically in human milk-fed infants, have low iron requirements.The medical concern was that human newborns, who are relatively immunecompromised in their first weeks of life, would be at greater risk for serious disease from this dysbiosis.At 4-6 months of infants are far more immune-competent than at birth and would likely tolerate a source of enteral iron better, through either supplements or the introduction of complementary foods.
A second paradigm shift occurred in the 1990s and 2000s with the identification of common maternal gestational conditions beyond severe maternal iron deficiency that compromise fetal and neonatal iron status (see above).These findings drove the design of a transformational study by Lozoff's group in the 2010s that demonstrated the major contribution of prenatal factors to the offspring's iron and neurodevelopmental status at 9 months of age. 31,44They documented maternal iron status prior to delivery, infant iron status at delivery and infant iron and neurodevelopmental status postnatally in a longitudinal manner.They found that infants with inadequate fetal iron loading had a greater risk of ID at 9 months of age accompanied by neurodevelopmental deficiencies, many of which had been ascribed to postnatal ID in previous studies where newborn iron status was not measured.Christian's randomised controlled trials (RCTs) of maternal compared with postnatal iron supplementation cited earlier in this review were consistent with the Lozoff findings. 38,42verall, these studies caused a shift to consider maternal iron supplementation to ensure adequate fetal iron loading as an important tool to prevent postnatal iron deficiency.

| Achieving the goal of adequate fetal iron status
Achieving the goal of an adequately iron-nourished fetus is not easy because the rate of ID prior to pregnancy is so high, 29 screening guidelines for ID during pregnancy around the globe are inconsistent, prescription of iron supplements during pregnancy is highly variable, and sources of iron are either scarce (e.g.meat) or unpalatable (e.g.oral medicinal iron supplements).Compliance with oral iron therapy during pregnancy is low. 49

| Screening for ID: when and how
Screening for ID should ideally begin prior to pregnancy as part of a general nutritional assessment because maternal pre-conceptional iron status is associated with risks to offspring neurodevelopmental outcome. 34As first and second trimester ID is a risk for low birthweight of the infant due to preterm birth, fetal growth restriction or both, 35 and each is an independent risk factor to the developing brain, screening for ID should be performed early in pregnancy.While screening for anaemia may be the most practical laboratory approach, emerging evidence suggests that preanaemic ID in pregnant women is also a risk for suboptimal gestational outcome. 35Non-haematologic iron assessments, e.g.serum ferritin or %total iron binding capacity saturation could therefore be considered in addition to routine haemoglobin concentration screening.Ideally, a determination of iron status should be in place by the beginning of the second trimester to leverage optimal intervention strategies for those with ID, either with or without anaemia.While universal iron supplementation of pregnant women remains controversial, 50 little question exists regarding intervention for those with iron deficiency. 51Failure to address low iron status in pregnant woman is highly associated with fetal iron underloading.A recent analysis of 408 maternal-fetal dyads without intentional iron intervention showed that 78% of babies born to mothers with serum ferritin concentrations <10 mcg/l had cord blood ferritin concentrations <75 mcg/l, 52 the level at which brain development is compromised. 4,10,37,39,43he inflection point of maternal iron stores at which fetal iron stores are compromised is similar to the level defined by Shao et al. 31 2.3.2 | Maternal intervention options and their relations to infant haematological and neurodevelopmental outcomes: the RAPIDIRON and RAPID-IRONKIDS trials Besides promoting better maternal health outcomes during pregnancy, iron intervention for gestational ID can also benefit the haematological and, potentially, neurobehavioural health of the offspring.The standard of care treatment route is oral iron therapy.A recent randomised controlled trial of 66 mg of daily ferrous fumarate versus placebo showed higher cord serum ferritin concentrations in the treatment group. 53In certain populations, iron delivered orally as part of a multimineral supplementation packet as opposed to only iron plus folic acid may be more efficacious. 53ral iron is notoriously poorly tolerated by pregnant women and thus compliance is often an issue. 49Moreover, even if compliance is good, there is concern about whether oral iron treatment during pregnancy is efficacious with respect to fetal iron loading.For example, a study of daily oral iron, initiated at or prior to 20 weeks' gestation and continued until delivery, showed an improvement of iron parameters.Nevertheless, 45% of the newborns in the study had evidence of ID at birth. 54The findings suggest that oral iron supplementation alone may not reach the developing fetus, either because of poor compliance or for other physiological reasons. 54An alternative, more effective approach is necessary to protect the developing fetal and postnatal brain.
Recently, several randomised controlled trials of intravenous iron versus oral iron therapy have been registered and their protocols published.Among them, the RAPIDIRON trial has the advantage of planned assessment not only of maternal but also infant haematological outcomes. 55In this trial, 4000 mothers with moderate iron deficiency at 12-14 weeks of pregnancy are being randomised to one of three iron treatment arms; one of two arms using different i.v.iron formulations (ferric carboxymaltose or iron isomaltoside) or one oral iron arm.The infants will have their iron status assessed at birth via cord blood assessment of haemoglobin, transferrin saturation (TSAT) and ferritin concentrations, and again at weeks post-delivery for general health.The hypothesis is that infants in the i.v.iron arms of the trial will have better newborn iron status as indexed by their cord ferritin, haemoglobin and TSAT concentrations compared with infants in the oral iron.
Importantly, 600 infant participants from each arm of the maternal RAPIDIRON trial will be enrolled in a follow-up study called RAPIDIRON-KIDS and an associated neuroimaging sub-study for haematological and neurodevelopmental outcome over a 3-year period.Whereas the infant outcome rationale for the RAPIDIRON trial is to reduce maternal complications of ID during pregnancy that affect offspring neurodevelopment (e.g.low birthweight), the two primary hypotheses of the RAPIDIRON-KIDS trial are that infants of mothers randomised to i.v.iron will have better iron status at 4 and 12 months and better neurodevelopmental outcomes at 24 months compared with infants whose mothers were randomised to oral iron therapy.Cochrane reviews have noted the lack of studies assessing relevant infant outcomes following maternal oral iron supplementation in general and specifically i.v.iron supplementation. 56he RAPIDIRON-KIDS trial is unique because it not only assesses general child development using the Bayley Scales of Infant Development but also focuses on behaviours dependent on specific brain regions targeted by fetal/neonatal ID such as recognition memory processing, attention and social-cognitive development (Table 1).
The neuroimaging sub-study of RAPIDIRON-KIDS has the potential to address a critical gap in iron neurodevelopmental research -documentation of brain ID effects in the newborn.Currently, inferences about brain iron status are based on iron biomarkers in peripheral blood (e.g.haemoglobin, serum ferritin).The association of these biomarkers with brain iron concentrations is unknown because of the difficulty in measuring low brain iron concentrations by magnetic resonance imaging.In lieu of measuring brain iron concentrations, the neuroimaging sub-study will measure consequences of fetal/neonatal iron deficiency based on known (and imageable) biology (Table 1).Sequential magnetic resonance imaging (MRI) will be performed in the offspring to assess white and grey matter to index irondependent myelin and neuronal growth status, respectively.Prior studies have determined that better maternal oral iron intake is related to greater grey matter complexity 57 and delayed cord clamping improves white matter integrity. 58The current longitudinal study on a large sample size has the opportunity to provide much needed detail with respect to long-term brain changes induced by fetal iron deficiency.

| CONCLUSION
A major shift in approach to prevention of iron deficiency in children has occurred through improved awareness of the role of fetal iron loading on postnatal iron status.Trials that assess the best strategies to maintain optimal maternal-placental-fetal iron status are crucial to protect the offspring's brain and mental health.

AU T HOR C ON T R I BU T ION S
MKG wrote the manuscript.

R E F E R E NC E S
L IC T OF I N T E R E S T S TAT E M E N T None declared.DATA AVA I L A BI L I T Y S TAT E M E N T Not applicable.E T H IC S A PPROVA LNot applicable.