Anemia, Iron Supplementation, and the Brain

Introduction

The human brain develops rapidly in the late fetal and early neonatal period. The brain is not a homogenous organ with a single developmental trajectory. Instead, brain development can be conceptualized as a regional process with each brain region having a different developmental trajectory with different timing of onset and completion.¹ A brain region is particularly vulnerable to extrinsic environmental events such as anemia and iron deficiency during periods of its rapid development, which are often termed critical or sensitive periods.² Primary regions that support fundamental brain functions such as learning and memory and speed of processing begin their period of rapid development shortly before term. For example, the hippocampus has a rapid onset of differentiation starting at 28 weeks post-conceptional age, while myelination increases rapidly starting at 32 weeks gestation.¹ Risk factors that disrupt these processes not only cause acute dysfunction, but also jeopardize brain regions that are later developing, eg, prefrontal cortex, because they rely on the integrity of connections from these more primary regions.³ These late effects are only appreciated in long-term neurodevelopmental follow-up and are the true cost to society of early life disruption of brain development.³

Anemia and the Developing Brain

Early life anemia presents a significant risk to the developing brain because it compromises the delivery of oxygen and iron. The newborn brain is highly metabolic, utilizing 60% of the body’s total oxygen consumption rate, compared with 20% in the adult.⁴ The total body oxygen consumption rate of the newborn is 2–2.5x greater than in adulthood. The high metabolic rate of the newborn and particularly the newborn brain derives from high energy cost of anatomic growth and development. Thus, the brain is highly reliant on an adequate supply of substrates that support its high oxygen consumption rate including oxygen, iron, glucose, amino acids, copper and iodine. In addition to the brain risk from reduced oxygen and iron delivery, anemia can induce a pro-inflammatory state in the brain.^5,6 Inflammation disrupts white matter development and neuroinflammation may contribute to developmental psychopathologies such as autism and schizophrenia.^7–9 Premature infants are at a higher risk of autism.¹⁰ Whether inflammation due to either anemia or packed red blood cell (pRBC) transfusions contribute to this risk is unknown, but the role of inflammation is suggested by pre-clinical models.^6,11

Studies of infants with postnatal iron deficiency and iron deficiency anemia clearly demonstrate negative effects on short and long-term brain development and function.¹² Iron is a critical nutrient for the developing brain because it has a direct synthetic role in myelination,¹³ energy metabolism for dendritogenesis,¹⁴ monoamine neurotransmitter synthesis,¹⁵ and regulation of synaptic plasticity genes through chromatin modification.¹⁶ Infants born at term are rarely anemic although multiple gestational conditions such as moderate maternal iron deficiency,¹⁷ intrauterine growth restriction,^18,19 maternal smoking,²⁰ and gestational glucose intolerance²¹ compromise fetal iron stores and tissue levels.^22,23 Non-anemic term infants with low iron stores, defined as a serum ferritin < 76 mcg/L in the neonatal period, have compromised brain function and neurodevelopment.^19,24–27

Preterm infants are far more likely to be anemic during the time period between 28 and 40 weeks gestation because they are born with lower total body iron stores and are frequently phlebotomized.^28,29 Phlebotomy removes 3.46 mg of elemental iron per gram of hemoglobin (Hgb) lost. Because iron is prioritized to red cells over the brain, dietary and storage iron are largely utilized to support erythropoiesis at the expense of all other tissues including the brain.³⁰ Iron deficiency, irrespective of anemia, in the preterm neonate compromises neurodevelopment; adequate and timely iron supplementation is effective in promoting neurodevelopment.^31–37

Anemia and Neurodevelopment in Preclinical Models

Anemia has been induced by phlebotomy in developmentally appropriately aged mice to study its effects on regional brain metabolism and gene expression and behavior.^6,11,38,39 Daily phlebotomy of mice from postnatal day 3, which is the neurodevelopmental equivalent of a 26 week gestational human infant, to postnatal day 14, which is equivalent to just past term in the human, induces a state of anemia commensurate with the Hgb/hematocrit levels seen in preterm infants. Phlebotomy to a hematocrit of 25% reduces brain iron by 40%. In the hippocampus, it increases lactate levels two fold, reduces expression of synaptic plasticity genes such as Brain Derived Neurotrophic Factor (BDNF) and increases Vascular Endothelial Growth Factor (VEGF) and Transferrin Receptor-1 expression.^38,39 The latter indicates that the brain is both hypoxic and iron deficient.

Phlebotomy-induced anemia also alters the mouse hippocampal transcriptome in a dose and sex dependent manner.⁶ Male mouse pups show dysregulation of gene expression in synaptic plasticity and pro-inflammatory pathways when bled to target hematocrits of 25% or 18%. A similar number and type of genes are differentially expressed compared to non-bled controls at both hematocrit levels, suggesting a threshold response at a relative modest degree of anemia. In contrast, females have 91% fewer differentially expressed genes at the 25% hematocrit level compared to the males, but 57% more at the 18% level. This anemia-dose dependent response in females was characterized by more pro-inflammatory and less synaptic plasticity pathway abnormalities compared to males.⁶ The findings are of interest because of the dose- and sex-specific differences in cognitive performance and pro-inflammatory cytokines reported by Benavides et al in preterm infants transfused for varying degrees of anemia.⁴⁰

Neonatal anemia also alters adult cognitive and social-cognitive behavior in a sex and dose dependent manner, with males with more severe anemia being most affected.¹¹ The abnormality in social-cognitive behavior is characterized by a desire for social isolation and avoidance of novel environments, which can be interpreted as an anxiety or autism-like phenotype. The effect of red cell transfusion on these outcomes has not been assessed, but would need to be in order to directly translate the findings in humans.40,41

Neonatal Anemia and its Treatment to Promote Brain Development

Iron Supplementation

The detrimental effect of iron deficiency on the developing brain is well established. The literature on iron supplementation in the developmental window starting between 24 and 40 weeks post-conceptional age is more limited. Studies in the last decade have identified this time period as perhaps the most important time to maintain iron sufficiency as it relates to the developing brain. Iron supplementation of pregnant women improves developmental outcomes at school age for offspring born at term.⁴² Interestingly, postnatal iron supplementation of infants born to mothers who received placebo in the Christian et al study did not improve their neurodevelopment,⁴³ emphasizing the critical nature of the last trimester time window for iron-dependent neurodevelopment. Infants born prematurely are by definition in the same developmental time window, which begs the question whether iron supplementation for preterm infants improves Hgb concentrations and neurodevelopment.

The literature confirms the concept that iron supplementation supports hematology and neurodevelopment. Earlier iron supplementation in preterm infants maintains better iron status and improves mental processing composite scores at 5 years of age.³³ Later gestational age preterm infants also have improved hematology outcomes and better Griffiths Mental Development Scales at 1 year of age if they are supplemented with 2 mg/kg body weight of iron vs placebo.³⁵ Similarly, breastfed preterm infants with birthweights between 2000 and 2500 grams, whether due to prematurity or intrauterine growth restriction, have lower rates of iron deficiency and iron deficiency anemia at 6 months of age if supplemented with 1–2 mg/kg body of iron vs placebo.³⁶ The supplemented infants have less behavioral issues at 3 and 7 years.^36,37 Younger gestational age preterm infants also benefit neurodevelopmentally at 2 years of age by having greater cumulative iron dosing in the first 60 or 90 days of the NICU stay.³⁴ Preclinical studies of iron repletion of iron deficient animals support the concept that earlier repletion spares long-term neurodevelopment.¹²

Erythropoietin Stimulating Agents

Erythropoiesis stimulating agents (ESAs) such as erythropoietin and darbepoietin have been shown repeatedly to improve hematocrit and reduce the number of transfusions for preterm infants.^44,45 Evidence for the neurodevelopmental benefits of ESAs has been mixed.

Darbepoietin has an extended half-life that allows for less frequent dosing than erythropoietin. A randomized controlled trial of darbepoietin was completed in preterm infants with a birth weight between 500–1250 g to evaluate whether transfusion need would be decreased, similar to erythropoietin, while requiring less frequent injections.⁴⁵ They compared weekly darbepoietin dosing from 48 hours of age to 35 weeks to three times weekly erythropoietin dosing or placebo. The primary findings confirmed decreased transfusions and donor exposures and increased hematocrit in the darbepoietin and erythropoietin groups compared to placebo. Importantly for brain development, there were no differences in mortality, retinopathy of prematurity, or intracranial hemorrhage among treatment groups. A follow-up analysis of these groups performed at 18–22 months suggested improved cognitive outcomes after darbepoietin and erythropoietin therapy.⁴⁶ Despite low patient numbers (27 darbepoietin, 29 erythropoietin, 24 placebo), they found significantly higher cognitive scores in the darbepoietin and erythropoietin groups compared to placebo. There were also higher object permanence scores in both treatment groups compared to placebo and a trend towards improved language scores.

A larger randomized controlled trial of preterm infants treated with high-dose erythropoietin was undertaken to explore its potential cognitive benefits therapy. In the Preterm Erythropoietin Neuroprotection Trial (PENUT), infants between 24 and 28 weeks were treated with erythropoietin starting within 48 hours of age through 32 weeks with the primary outcome of death or severe neurodevelopmental impairment that included severe cerebral palsy, Bayley Scales of Infant Development (BSID)-III cognitive score <70 or BSID-III motor score <70.⁴⁷ However, contrary to Ohls et al,⁴⁶ no differences between groups were found in the primary outcome between treatment groups.

Why differing neurodevelopmental outcomes have been found between studies of ESAs is unclear. Brain iron deficiency could be a variable affecting outcomes. As discussed above, iron utilization is prioritized to red cells over other tissues including the brain.^30,48 A concern with erythropoietin therapy is that the resulting expansion of red cells could result in a relative brain iron deficiency.⁴⁹ Despite an iron supplementation protocol, infants in the PENUT study received varied amounts of iron. In a post-hoc analysis, German et al showed that cumulative iron dosing at 60 days of age in preterm infants from either randomized group was positively correlated with mental, motor and language composite scores on the BSID-III.³⁴ Importantly, the effect was greater in infants that received erythropoietin, suggesting that augmented erythropoiesis in the erythropoietin group prioritized iron to the red cells over the brain and appropriate (higher) dosing of iron improved neurodevelopmental outcomes, consistent with preclinical literature.³⁰ Inability to give iron, however, did not result in worse outcomes for erythropoietin treated infants, but rather a lack of neurodevelopmental benefit. Importantly, there was not a decline of BSID-III scores at the highest doses of iron and therefore no appearance of iron toxicity.³⁴

Preclinical studies have provided insight behind the mechanisms by which erythropoietin may improve neurodevelopmental outcomes. Erythropoietin is necessary for proper brain development; an absence of erythropoietin results in embryonic neurogenesis defects.⁵⁰ Erythropoietin also stimulates production of BDNF,^51,52 a hormone necessary for neuronal growth and development. Erythropoietin protects the brain during injury. It promotes neurogenesis and oligodendrocyte development following hypoxic injury,^53,54 while a lack of erythropoietin results in reduced cell proliferation following stroke injury.⁵⁰ Following hypoxic injury, erythropoietin has anti-inflammatory effects in the brain, reducing cytokine load and infiltrating leukocytes, and prevents brain atrophy.⁵⁵ Erythropoietin’s anti-inflammatory properties may be particularly effective for neuroprotection in the treatment of neonatal anemia, since we have shown increased inflammation in the neonatal hippocampus following phlebotomy-induced anemia.⁶

Erythropoietin improves brain metabolic dysfunction. Rescue treatment of PIA with erythropoietin restores hematocrit and thereby reduces the hypoxic burden as indicated by normalization of VEGF expression; however, Transferrin Receptor-1 expression remains elevated, indicating worsening iron deficiency.³² The finding is consistent with shunting of dietary and storage iron preferentially into the red cells over the brain.^30,48 PIA suppresses phosphorylation of mammalian target of rapamycin (mTOR) pathway proteins in the hippocampus, a pathway important for monitoring critical brain metabolites such as oxygen and iron.³⁹ This pathway was partially rescued by erythropoietin treatment, likely improving hippocampal growth and development.

Red Blood Cell Transfusion

The standard treatment for anemia in premature infants is pRBC transfusion. More than 70% of preterm infants <1000 grams are transfused at least once.⁵⁶ Some infants have their entire blood volume phlebotomized and replaced over as short as a two week period in the newborn intensive care unit (NICU).²⁹ Injuries associated with transfusion are well-known in adult literature, but more recent observational studies raised concerns about severe preterm injuries, including intraventricular hemorrhage and resulting neurodevelopmental deficits.⁵⁷ While the studies of the effects of neonatal ID, with or without anemia, on the developing preterm brain are persuasive, the neurodevelopmental risks of anemia (at different hematocrit levels) vs transfusion to preterm infants are not well understood. Therefore, multiple randomized controlled trials of preterm infants sought to determine outcomes based on transfusion randomized to a higher (liberal) or lower (restrictive) threshold of anemia.^58–60

The single-site trial of 100 newborns at the University of Iowa (Iowa study)⁵⁸ found more frequent major adverse neurological effects (increased severe intraventricular hemorrhage and periventricular leukomalacia) during the neonatal period in the restrictive group compared to the liberal group, leading them to conclude that restrictive transfusion practices may be harmful. A multicenter randomized controlled trial (RCT) in Canada named the Premature Infants in Need of Transfusion (PINT) trial studied 451 infants with similar anemia thresholds for red cell transfusions.⁵⁹ This trial showed no differences in major morbidities although there was a trend toward better neurodevelopmental outcomes at 2 years of age in the infants randomized to a liberal transfusion threshold similar to the Iowa study, including the primary composite of death or neurodevelopmental impairment, cognitive delay, and neurosensory impairment.⁶¹ The recent multicenter Transfusion of Prematures (TOP) RCT studied 1692 infants with similar anemia threshold randomization to PINT and the Iowa trials.⁶⁰ No differences in death or morbidities including neurodevelopment at 2 years were seen between the two groups, which differed in average threshold Hgb by ~ 20 g/L.

Long-term evaluation from the Iowa study provided a nuanced assessment of outcomes in a select group of 56 children followed up to 13 years. Their studies suggest differential neurodevelopmental outcomes related to transfusion are influenced by sex. An MRI at 12 years of age revealed no differences in brain volumes between liberal and restricted transfusion groups, although comparison to non-preterm controls show significantly lower volumes in the liberal group but not restricted group.⁶² Comparison between sexes showed that females in the liberal group had more structural abnormalities than males in the liberal group: lower volumes of multiple brain regions and total brain tissue. Comparison between liberal and restricted females showed lower intracranial volumes (ICV) in the liberal group and a negative correlation between cerebral white matter and mean hematocrit level, ie. higher hematocrit equaled less white matter in females.⁶³ In contrast, a lower pre-transfusion Hgb for males correlated with lower ICV, the opposite finding from females.⁴¹ When neurocognitive outcomes were measured at 8–15 years of age, the liberally transfused group had worse outcomes in intelligence and neuropsychological functioning compared to the restricted group.⁶⁴ However, there were also sex-specific differences. Higher pre-transfusion Hgb correlated with opposite outcomes between males and females; females had a lower pooled BSID-III score with higher Hgb.⁴¹ Overall, these studies suggest that females are more harmed by transfusions, but males are more harmed by severe anemia.

The causative mechanisms differentiating male and female outcomes to level of anemia and transfusion are not well understood, but inflammatory response may be involved. As described above, preclinical models of anemia demonstrate different inflammatory responses between males and females.⁶ In neonates, multiple circulating pro-inflammatory cytokines increased following transfusions.⁶⁵ The cytokine responses appear to be sex-specific. In the Iowa study patients, monocyte chemotactant protein-1 (MCP-1) increased in female neonates with cumulative transfusions, but not in males.⁴⁰ Importantly, MCP-1 levels negatively correlated with BSID-III cognitive scores.⁴⁰ The authors hypothesized that red cell transfusions may have caused damage to developing inflammation-sensitive structures, particularly myelin.

Taken at face value, these relatively large RCTs suggest that within the Hgb range tested, all in the anemic range, no discernable effect on gross neurodevelopmental measurements can be detected. Data from underpowered subanalyses of the clinical studies and strong corroborative support from pre-clinical models suggest that the effect of anemia and its treatment with pRBC transfusion may be far more complex, particularly regarding sex, than were assessed in the RCTs. Multiple reasons exist for negative neurodevelopmental outcomes in large RCTs (Box 1).³ Several of these factors exist in the PINT and TOP trials, most notably the use of an insensitive neurodevelopmental battery relative to the expected neurobiological impact and the fact that all infants were anemic and transfused, regardless of which randomization group. If the neurobiologic effects of anemia or transfusion are driven by threshold, rather than dose-response kinetics, no differentiation of outcomes would be expected between the groups.

Box 1: Common causes for null neurodevelopmental outcome studies in RCTs.

• Wrong timing of intervention with regards to brain development

• Target population must not already be sufficient

• Corollary: Population must not have remained deficient after intervention

• Wrong (or insensitive) neurodevelopmental test battery

• Wrong timing of assessment with regards to affected brain region

• Failure to identify subpopulations with opposite effects (e.g., M vs F)

Taken together, isolating causality of either anemia or pRBC transfusion on neurodevelopment is not possible from these three RCTs. The effect of anemia without transfusion on preterm neonate neurodevelopment is not possible to detect because all infants in the trials were transfused. Isolating the effect of pRBC transfusions is impossible to pick out as the root cause because of the ubiquitous existence of anemia in the population. Similarly, in trials of ESAs vs placebo, it is also hard to isolate the role of the ESAs when most if not all the infants in those trials were transfused.

To date, preclinical studies of the neurodevelopmental effects of anemia and pRBC transfusion are absent. Clues can be gathered from the single published study of anemia and neonatal transfusion where they assessed gut outcomes. MohanKumar et al showed that anemia alone caused increased infiltration of macrophages into the gut.⁶⁶ Only after transfusion, however, was there a significant increase in gut inflammatory markers and susceptibility to severe gut injury. In this model, anemia appears to precondition gut tissue to inflammatory injury caused by the transfusion. A similar pathology may occur in the brain; upregulation of inflammatory pathways occur in the hippocampus following phlebotomy-induced anemia.⁶ However, neurologic factors such as the developmental stage of each brain region and the blood-brain barrier and blood product factors such as storage time and donor characteristics may change how transfusion affects brain injury and neurodevelopmental outcomes. Further preclinical studies in appropriate developmental models are much needed to address these questions.

Delayed Cord Clamping

Delayed clamping of the umbilical cord after birth is recommended for both preterm and full-term births. Although there is no universally accepted definition, clamping of cord 2 to 3 minutes after delivery of the infant or when cord pulsations cease is typically considered delayed cord clamping (DCC).⁶⁷ DCC allows transfer of an additional 25–35 mL/kg of placental blood to the infant.⁶⁷ DCC is associated with better hemodynamic stability and decreased need for transfusions in the neonatal period. A meta-analysis of 20 randomized trials involving >3,500 infants found that compared with early cord clamping, DCC is associated with higher Hgb and serum ferritin at 6–10 weeks in preterm infants.⁶⁸ In full term infants, DCC reduces the incidence of iron deficiency and iron deficiency anemia up to one year of age.^68–70 Additional studies have reported better myelin content in the brain regions associated with motor, visual, and sensory function⁶⁹ and better neurodevelopment, especially in boys, with DCC.⁷¹ In regions where maternal iron deficiency in pregnancy is common, DCC is important for improving the iron status of infants.⁷² Both DCC and cord milking are equally effective in this respect.^72,73

Biomarkers and Bioindicators of Risk to the Brain

A Hgb-based diagnosis and treatment strategy may not achieve neuroprotection in perinatal iron deficiency. Hgb has poor predictive sensitivity and specificity for diagnosing brain iron deficiency. Due to prioritization of available iron to the RBCs over all other organs, tissue iron deficiency, including in the developing brain, predates anemia.^23,74 The risk is highest in the perinatal period when iron needs of brain development compete with those of erythropoiesis.⁴⁸ Similarly, anemia is corrected before tissue iron deficiency during iron treatment,⁷⁵ leaving the developing brain iron deficient for a longer duration. Data in older infants show that treatment after the onset of anemia is inadequate for correcting the neurological deficits,⁷⁶ likely due to persistent brain iron deficiency. A management strategy based on biomarkers of brain iron status and health in the pre-anemic period is therefore desirable. Some commonly used biomarkers are reviewed below:

Serum Ferritin

Serum ferritin reflects iron storage. Low serum ferritin indicates iron deficiency. A cord blood ferritin <35 μg/L predicts brain iron deficiency and impaired recognition memory at birth and lower psychomotor development at 1 year of age in full-term infants of diabetic mothers with iron deficiency.⁷⁷ Other studies have demonstrated that a cord blood ferritin ≤75 μg/L correlates with slower auditory brainstem evoked responses, suggestive of reduced auditory tract myelination, in both full term and preterm infants.^32,78,79 A higher cord blood serum ferritin is also associated with impaired mental and psychomotor development at 5 years of age in full term infants.⁸⁰ A problem with serum ferritin is that levels could be falsely increased in inflammation. Iron supplementation is typically held when serum ferritin is >350–400 μg/L.^47,49 Recent studies have demonstrated feasibility of measuring urinary ferritin in both full term and preterm infants.^81,82 Urinary ferritin correlates with serum ferritin^81,82 and offers a noninvasive screening method. In a recent study, urine ferritin <12 ng/mL corrected for urine creatinine and specific gravity had 82% sensitivity and 100% specificity for detecting iron-limited erythropoiesis, with a positive predictive value of 100%.⁸² However, the method requires a relatively large volume of urine and may not be sensitive in severe iron deficiency.⁸¹

Zinc Protoporphyrin to Heme Ratio

Zinc protoporphyrin to heme ratio (ZnPP/H) is an indicator of iron-limited erythropoiesis. When iron is not available, zinc is incorporated into the protoporphyrin molecule, leading to increased ZnPP/H. Higher ZnPP/H indicates iron deficiency and has greater sensitivity than Hgb and serum ferritin for detecting iron deficiency. ZnPP/H from the immature erythrocyte fraction has greater sensitivity for detecting mild iron deficiency than whole blood ZnPP/H.⁸³ ZnPP/H is not affected by inflammation but is affected by pRBC transfusions and erythropoietin treatment. ZnPP/H deceases over the last trimester of pregnancy.⁸⁴ Cord blood ZnPP/H is higher in infants with perinatal iron deficiency.⁸⁴ A cord blood ZnPP/H >118 μM/M predicts worse recognition memory at 2 months in infants with perinatal iron deficiency.²⁵ A serum ferritin <75 μg/L was not sensitive for such prediction in this study.²⁵ Similarly, a retrospective analysis demonstrated that BSID at 2 years correlate better with ZnPP/H values than serum ferritin values in extremely low gestational age neonates (ELGAN).⁸⁵

Reticulocyte Hemoglobin

Reticulocyte hemoglobin, typically defined as reticulocyte Hgb content (CHr) or reticulocyte Hgb equivalent (RET-He), depending on the hematology analyzer used,⁸⁶ is an indicator of the hemoglobinization of reticulocytes and correlates with bone marrow iron level. Since reticulocytes are in the circulation only for 1–2 days, Ret-He provides a more real time information on bone marrow iron status than Hgb, which is an average of the entire RBC population, each with a lifespan of 90 to 120 days. As in other age groups, a low CHr or RET-He is an indicator of iron deficiency in newborn infants.^87,88 Animal data show that a low RET-He also predates the onset of brain iron deficiency in the postnatal period,⁸⁹ and thus could be a biomarker of brain iron status. Our recent data in a nonhuman primate model show that Ret-He has comparable predictive accuracy for early detection of iron deficiency and anemia as the conventional iron indices, and is more sensitive than ZnPP/H.⁹⁰ Moreover, Ret-He is a component of the complete blood count in some hematology analyzers and does not require additional blood samples. Ret-He is affected less by inflammation than iron indices.⁹¹ The small coefficient of variation also makes RET-He useful for monitoring temporal trends in the iron status of individual patients (e.g., during iron or erythropoietin treatment). Reference Ret-He values for the first 90 days after birth are available from 22 to 42 weeks of gestation.⁹² A value 25 pg corresponds to the 2.5 percentile in human newborn infants.⁹³ A Ret-He < 29 pg had 85% sensitivity and 73% specificity for detecting iron deficiency at 3–4 months corrected age in one study.⁹⁴ A low RET-He can be seen in certain hemoglobinopathies, such as α and β thalassemia. Additionally, not all hematology analyzers can generate RET-He results.

Hepcidin

Similar to adults, newborn infants are capable of iron regulation through hepcidin.^70,81,95,96 Downregulation of hepcidin in iron deficiency promotes iron absorption in the gastrointestinal tract. Serum hepcidin levels correlate positively with serum ferritin, Ret-He and Hgb, and negatively with transferrin and soluble transferrin receptor in full term and preterm infants.^70,95–97 Reference ranges for cord blood hepcidin from 24 to 42 weeks of gestation are available.⁹⁸ Serum hepcidin levels double during the first month after birth in full-term infants.⁹⁹ A level <16 ng/mL at 4 months of age indicates iron deficiency.⁷⁰ As with ferritin, it is possible to determine urine hepcidin. Urine hepcidin levels correlate with serum hepcidin in preterm infants.⁹⁷ A recent study showed that urine hepcidin/creatinine ratio correlates positively with serum ferritin and negatively with ZnPP/H in ELGAN.⁹⁵ A problem with hepcidin is that it is affected by erythropoietin treatment, pRBC transfusion, iron treatment and inflammation.^87,95–98

Biomarkers of Iron-Dependent Brain Health

The above-mentioned biomarkers primarily index iron deficiency in the heme compartment and not brain iron deficiency and brain health. Molecular biomarkers of iron-dependent brain health in the blood will be better for ensuring neuroprotection during iron deficiency, especially if they can be identified in the pre-anemic period when the chances of preventing iron deficiency-induced adverse neurological effects are better. Exosomes, which are small, cell-derived vesicles found in all the biofluids, carry the same classes of molecules as the parent cell and function as a snapshot of their cell of origin. Cord blood levels of exosomal contactin 2, a neural-specific glycoprotein important for brain development, are lower in newborn infants at risk for iron deficiency, especially males.¹⁰⁰ Conversely, exosomal BDNF levels negatively correlate with serum ferritin in females, suggesting their biomarker potential for early detection of iron deficiency-induced brain dysfunction in females.¹⁰⁰

A series of proteomic and metabolomic experiments in a nonhuman primate model of infantile iron deficiency anemia from our lab have demonstrated the presence of neurologically important metabolites (dopamine, serotonin and N-acetylaspartylglutamate) in the serum of infants with iron deficiency anemia.^101–103 Iron treatment led to additional changes in these metabolites.^101,102 Our recent study of paired serum and cerebrospinal fluid (CSF) in this animal model demonstrated concurrent changes in several neurologically important proteins and metabolites, such as phospholipid transfer protein, serpin family G member 1, transthyretin, and acute phase proteins in the heme and CSF compartments in the pre-anemic period,¹⁰³ suggesting their potential as biomarkers of brain health in early-life iron deficiency.

Biomarker-based Iron Supplementation and Neurodevelopment

A biomarker-based standardized iron supplementation strategy has been used to minimize the risk of brain iron deficiency during erythropoietin therapy in ELGAN.^47,49 Such a strategy results in iron supplementation at a dose higher than the currently recommended dose¹⁰⁴ and potentially improves neurodevelopment by allowing greater cumulative iron supplementation during the first 60 days after birth.³⁴ However, a single biomarker may not be sensitive for both monitoring for iron deficiency and response to treatment. For example, while a serum ferritin-based strategy is useful for monitoring the risk of brain iron deficiency, it is not useful as an early biomarker of response to therapy since storage iron is the last compartment to get corrected during iron treatment, especially during active erythropoiesis.⁴⁹ An RBC based biomarker (e.g., Ret-He or reticulocyte ZnPP/H) is probably better for monitoring response to iron therapy.⁸⁴ An RBC based biomarker also appears better than serum ferritin for monitoring brain health during iron deficiency.¹⁰⁵ However, additional data are needed before a biomarker-based monitoring and treatment strategy can be recommended for clinical practice. An ongoing randomized trial comparing the effects of standard iron dose (4 mg/kg/d from 2 weeks of age) with early, high iron dose (from 1 week of age and dose adjustments to maintain serum ferritin 70–400 ng/mL until 36 weeks postmenstrual age) on 2-year neurodevelopment in preterm infants (NCT04691843) may provide additional data.

Future Directions

The most urgent need is to generate biomarkers of brain risk when anemia is present in neonates and to clarify which individual characteristics, including patient sex, influence the risk to the brain by anemia and by therapy with either an ESA or pRBC transfusion. Another underexplored area is to define the risk to the brain from the heterogeneity of the transfused products. The negative effects of red cells used for transfusion can be driven by controllable factors such as length of storage, conditions of storage and donor sex.

Key Points:

1. It is unclear what level of anemia affects brain development in human neonates.

2. Phlebotomy induced anemia likely causes brain iron deficiency because iron is prioritized to red cells over the brain.

3. Erythropoiesis stimulating agents improve hematocrit and reduce the number of transfusions for preterm infants; evidence for their neurodevelopmental benefits is mixed.

4. The neurodevelopmental risks of anemia at different hematocrit thresholds versus the risks and benefits of red cell transfusion (the standard treatment) are not well established.

5. While there are biomarkers of anemia and iron status in the heme compartment, reliable biomarkers of brain iron status and brain health are not available for clinical practice.

Synopsis:

The developing brain is particularly vulnerable to extrinsic environmental events such as anemia and iron deficiency during periods of rapid development. Studies of infants with postnatal iron deficiency and iron deficiency anemia clearly demonstrated negative effects on short and long-term brain development and function. Randomized interventional trials studied erythropoiesis stimulating agents and hemoglobin-based red blood cell transfusion thresholds to determine how they affect preterm infant neurodevelopment. Studies of red blood cell transfusion components are limited in preterm neonates. A biomarker strategy measuring brain iron status and health in the pre-anemic period is desirable to evaluate treatment options and brain response.

Best Practices.

What is the current practice for the treatment of neonatal anemia?

There remains no standard practice, ie, no hemoglobin concentration target, for the treatment of neonatal anemia that protects neurologic outcomes. Although red blood cell transfusion is the standard therapy, iron supplementation and erythropoiesis stimulating agents (ESAs) are used variably among neonatal intensive care units. The primary objective of all current interventions is limited to achieve a goal hemoglobin concentration since optimal neurodevelopment targeting has not been established.

What changes in current practice are likely to improve outcomes?

Preclinical and Preclinical and clinical evidence suggest that the treatment of anemia should be sex-specific. Therapy practices that account for patient sex will likely improve outcomes. Numerous studies have demonstrated the neuroprotective benefits of ESAs. Adequate iron supplementation is likely a necessary component. Therefore, the combined treatment will likely improve outcomes.

Major Recommendations:

• Further investigation into hemoglobin transfusion threshold with sex-specific treatment as a primary outcome.

• Further investigation into ESAs with adequate iron supplementation as neuroprotectants.

Bibliographic Source(s):

• Bell EF, et al. Randomized trial of liberal versus restrictive guidelines for red blood cell transfusion in preterm infants. Pediatrics. 2005;115:1685–91.

• Kirpalani H, et al. The Premature Infants in Need of Transfusion (PINT) study: a randomized, controlled trial of a restrictive (low) versus liberal (high) transfusion threshold for extremely low birth weight infants. J Pediatr. 2006;149:301–307.

• Kirpalani H, et al. Higher or Lower Hemoglobin Transfusion Thresholds for Preterm Infants. N Engl J Med. 2020;383:2639–2651.

• Ohlsson A, Aher SM. Early erythropoiesis-stimulating agents in preterm or low birth weight infants. Cochrane Database Syst Rev. 2020;2:CD004863.

• McCarthy EK, Dempsey EM, Kiely ME. Iron supplementation in preterm and low-birth-weight infants: a systematic review of intervention studies. Nutr Rev. 2019;77:865–877.