Iron optimisation in pregnancy: a Haematology in Obstetric and Women's Health Collaborative consensus statement

Abstract

Anaemia is a well‐recognised and widely accepted consequence of iron deficiency (ID); however, the two diagnoses are not synonymous with the effects of ID occurring long before the development of anaemia. In adults, ID can cause physical and neuropsychological symptoms, including lethargy, altered mood and poor concentration, reducing an individual's quality of life. Foetal and neonatal ID has been associated with impaired neurocognitive development with lasting effects despite iron replacement in early life. Obstetric ID is common, affecting up to 70% of Australian pregnancies. The impact, at both an individual and a population level, remains underappreciated and consensus on the identification and management of obstetric ID is lacking. This consensus statement was developed by the Haematology in Obstetrics and Women's Health (HOW) Collaborative and utilised the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system to evaluate evidence and strength of recommendations. Recommendations are as follows: (i) Routine ferritin screening should be performed in all pregnant women (GRADE 1C) at booking and 24–28 weeks. Repeat testing should be performed at 36 weeks if clinically indicated or if the woman is previously unscreened. (ii) ID in pregnancy should be defined as a ferritin level <30 μg/L (GRADE 1D). (iii) An appropriate oral iron formulation should be offered as first‐line therapy for obstetric ID (GRADE 1B). (iv) Alternate‐day oral dosing can be considered to limit side effects in women with obstetric ID (GRADE 2B). (v) Intravenous iron should be offered to women with ID/ID anaemia who are intolerant of or refractory to oral iron or in the third trimester (GRADE 1B).

Introduction

Iron deficiency (ID) is the most common single‐nutrient deficiency worldwide, affecting approximately 2 billion people annually, and is the most prevalent cause of anaemia. ID disproportionately affects women throughout menstruation, pregnancy and childbirth. When anaemia develops, it is the leading cause of years lived with disability worldwide. ¹

Anaemia is a well‐recognised and widely accepted consequence of ID; however, the two diagnoses are not synonymous. The effects of tissue ID occur long before the development of anaemia. In adults, ID is associated with physical and neuropsychological manifestations such as lethargy, altered mood and poor concentration. Such states have been linked to reduced physical capacity and quality of life (QoL). ² , ³ , ⁴ Foetal and neonatal ID has been associated with impaired neurocognitive development with lasting effects despite iron replacement in early life. ⁵ , ⁶

Obstetric ID is common. In an Australian clinical practice improvement (CPI) project, routine ferritin screening identified ID in 60%–70% of pregnant women, as defined by a ferritin level <30 μg/L. ⁷ Similar rates were seen internationally, with 53% of pregnancies in a Canadian cohort and 42% of pregnancies in a US cohort complicated by ID. ⁸

The impact of ID, at both an individual and a population level, is underappreciated. Consensus statements on the optimisation of obstetric ID are lacking. Herein, we describe the pathophysiology of obstetric ID, highlighting the non‐erythropoietic roles of iron and the detrimental effects of tissue ID to both mother and infant and providing consensus recommendations on optimisation of ID in pregnancy.

Methods

This document was conceived by the Haematology in Obstetrics and Women's Health (HOW) Collaborative (www.howcollaborative.org), a multidisciplinary group of haematologists, obstetricians, gynaecologists, midwives, anaesthetists and research scientists from Australia and New Zealand. The Accurate Consensus Reporting Document (ACCORD) ⁹ was used given the lack of high‐quality data available to guide recommendations.

Members of the HOW Collaborative steering committee, with a professional interest in ID, were invited to participate in the consensus statement. Additional leading experts were sought to ensure broad multidisciplinary and geographic representation.

The authorship group met to establish the priority areas for the literature review. Search engines included PubMed, MEDLINE and Cochrane Central Register of Controlled Trials. References of included articles were checked for further literature. Search terms included: pregnancy, obstetric, neonatal, foetal, infant, iron and outcomes. The search was restricted to articles in English and published over the past 20 years until 31 December 2022. The first and last author independently reviewed the abstracts of the 161 identified articles and excluded 117 due to duplication or a focus only on ID anaemia (IDA) in pregnancy. The group discussed the results of the review and collaboratively developed recommendations iteratively through multiple meetings. Consensus was achieved using the modified Delphi process. Recommendations where consensus was met were accepted. There were no major discrepancies and minor discrepancies were discussed and reconciled. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) ¹⁰ system was used to evaluate levels of evidence and strength of recommendations. There were no financial incentives for committee members and all conflicts of interest were declared at the commencement of the project.

The consensus paper has been endorsed by the Royal College of Pathologists Australasia, Royal Australian College of General Practitioners, Australian and New Zealand College of Anaesthetists, Royal Australasian College of Pathologists, Society of Obstetric Medicine of Australia and New Zealand, Australian and New Zealand Society of Blood Transfusion, Haematology Society of Australia and New Zealand and Society of Hospital Pharmacists of Australia and consumer feedback was obtained.

Discussion and Reccomendations

Diagnosis

Serum ferritin is the most reliable marker of ID, except in the context of inflammation. ¹¹ While chronic disease is uncommon in pregnancy, when present, a higher ferritin threshold should be considered, and transferrin saturation is often useful in this setting.

Debate persists regarding the optimal diagnostic threshold of ID. ⁸ The World Health Organisation (WHO) defines ID as a serum ferritin value <15 μg/L based on the analysis of 12 studies (541 healthy individuals) where bone marrow iron stores were absent at a mean ferritin concentration of 15.1 μg/L and present at 70.4 μg/L. ¹¹ Additional studies have shown a ferritin level <15 μg/L has a specificity of 99% but sensitivity of only 75% in identifying ID. When this threshold is increased to 30 μg/L, specificity increases to 92% but the sensitivity improves to 99%. ¹² A recent Australian cohort study of women aged 18–35 years found nearly 25% of diagnoses would be missed when utilising the lower diagnostic threshold of 15 μg/L. The Royal College of Pathologists Australasia defines ID as a serum ferritin level <30 μg/L. ¹³

It is unclear how many of the aforementioned studies included pregnant women. A study of 93 pregnant women comparing serum ferritin with bone marrow iron stores found a serum ferritin level of 30 μg/L was the optimal lower limit of normal. ¹⁴

Internationally there is a varied approach to the investigation of obstetric ID. In low‐ and middle‐income countries, routine iron replacement is recommended. Recent UK guidelines recommend ferritin screening for ‘at‐risk’ individuals otherwise empiric iron therapy when anaemia is identified. ¹⁵ The Royal Australian and New Zealand College of Obstetricians and Gynaecologists ¹⁶ recommend screening for ID once anaemia develops; however, other groups have advocated for routine ferritin screening in pregnancy. ⁷

Foetal iron stores were previously thought to be preserved even in light of maternal ID; however, recent studies have shown that neonates can develop ID, as defined as a cord ferritin level <76 μg/L (lowest quartile of cord ferritin concentrations), with or without anaemia. ¹⁷ , ¹⁸ Additionally, in cohort studies, a mean ferritin concentration of infants born to ID mothers is significantly lower than those born to iron‐replete mothers (88.2–98.5 μg/L compared to 135.6–147.2 μg/L) with a maternal ferritin level <10–14 μg/L the critical threshold where foetal iron stores are compromised. ¹⁷ , ¹⁸ , ¹⁹

Recommendations are as follows:

• Routine ferritin screening should be performed in all pregnant women (GRADE 1C) at booking and at 24–28 weeks. Repeat testing should be performed at 36 weeks if clinically indicated or if the woman is previously unscreened.

• ID in pregnancy should be defined as a ferritin level <30 μg/L (GRADE 1D).

Non‐erythropoietic roles of iron

Approximately 1 g of additional iron is required throughout pregnancy to meet the competing needs of the maternal–placental–foetal unit, with the highest requirements seen in the third trimester. Iron is essential for increased maternal erythropoiesis, placental development and metabolic activity, foetal erythropoiesis and establishment of adequate iron stores. ²⁰ Maternal and foetal erythropoiesis is prioritised to maintain foetal oxygen delivery, with tissue ID developing before anaemia. ¹⁷ Once ID is established, there is a sequential loss of tissue iron, with cardiac and skeletal muscles affected prior to the brain. ¹⁷

Perinatally, iron is essential for early brain development through its supporting role in neuronal and glial energy metabolism, neurotransmitter synthesis and myelination. ¹⁷ Structures with high nutritional requirements in the latter stages of pregnancy are most vulnerable including the hippocampus, which is important for learning and memory, and neuronal myelination. ²¹ In pre‐clinical models, foetal ID compromises hippocampal synaptic plasticity and metabolism ²² and can result in a hypomyelinated corticospinal tract and basal ganglia structures, which are involved in motor activity, learning, execution of sequential movements and coordination. ²¹

Clinical consequences of foetal‐neonatal ID

The neurocognitive effects of foetal‐neonatal ID can be seen as early as 2 months of age with iron‐replete, not ID, infants able to distinguish their mother's voice from a stranger's, as demonstrated by hippocampal electrophysical activity. ¹⁹ Social and emotional behaviour of a cohort of 9‐ to 10‐month‐olds were adversely affected by ID, with ID infants (with or without anaemia) shyer, more difficult to soothe and less likely to engage. ²³ In the same cohort at 9–12 months, ID with or without anaemia was associated with inferior motor performance on testing of gross motor development, performance, coordination and sequencing. ²¹ Children at 5 years of age who had foetal‐neonatal ID as defined by cord ferritin in the lowest quartile demonstrated poorer performance in the domains of language ability, fine‐motor skills and tractability. ²⁴ It must be noted that the sample sizes of all studies were small.

Foetal‐neonatal ID has been found to impact long‐term motor and mental functioning with the largest cohort of iron‐deficient infants followed out to 19 years of age. At each follow‐up time point, those with severe ID at recruitment as defined by either moderate IDA or uncorrected ID after 3 months of oral iron, performed worse on testing of motor and mental functioning. At 10–14 years, more of these children had repeated a grade and parental/teacher concerns regarding anxiety and depression were increased. ²⁵ While recruitment was not based on cord ferritin results, the study participants were healthy infants free of traditional risk factors for ID such as prematurity, low birthweight, malnutrition or chronic illness, making foetal‐neonatal ID a major risk factor for infant ID (with or without anaemia). ²⁴ The validity of this assumption is supported in a prospective study of nearly 200 infants where cord ferritin was found to be a predictor of iron stores at 6, 12 and 24 months of age. ²⁶

The effects of iron supplementation on motor and mental development in early childhood are inconsistent and appear dependent on the timepoint when ID developed. A large systematic review and meta‐analysis of 2827 children (1412 who had received iron replacement and 1415 placebo) showed no benefit in children younger than 3 years. These findings support the notion that perinatal ID has long‐lasting effects, with correction in infancy unable to compensate for deficits that occur due to maternal ID in pregnancy. In contrast, modest improvement in intelligence test scores was seen in children diagnosed and treated over 7 years of age, akin to adult ID. ²⁷

Maternal ID

Iron is essential for synthesis of cofactors and enzymes necessary for cellular respiration and oxygen delivery, including the generation of cellular adenosine triphosphate. ⁵ The more metabolically active tissues including skeletal muscle, cardiac myocytes and brain are most sensitive to ID. ⁵ Symptoms include myalgias, increased fatigue, reduced exercise tolerance, shortness of breath, poor concentration, reduced cognitive function and headache. ²⁸

The neurocognitive sequelae of maternal ID are best demonstrated in an interventional cohort study of iron‐ replete ID and IDA women of childbearing potential (113 women). All women received 4 months of iron replacement and underwent testing in the domains of attention, learning and memory pre‐ and post‐treatment. Correction of ID improved the quality of the individual's performance, whereas correction of anaemia improved performance speed. ³

Iron is a co‐factor for dopamine and serotonin synthesis and is vital for glutamate and γ‐aminobutyric acid homeostasis. These neurotransmitters are involved in mood regulation, thus linking ID with behavioural changes, including depression. ⁵ , ²⁸ , ²⁹

A systematic review found that ID with or without anaemia was associated with postnatal depression (PND) with a ferritin value <12 μg/L associated with a 2.3 times increased risk of PND. Iron replacement in the postpartum period was associated with the improvement of PND, with resolution of symptoms in 43% of women who received iron replacement for 6 weeks compared to 20% in the untreated group (P = 0.03). Additionally, in those women where PND persisted, ID was more common (27.1% compared to 4.5% (P = 0.02)). No differences in outcomes were seen between oral or intravenous iron (IVI) replacement. ³⁰

In a randomised controlled trial (RCT) of IDA mothers recruited 10 weeks post‐partum, those who received iron replacement at recruitment were more engaged with and responsive to their infant at 9 months compared to those who did not receive iron replacement based on a parent/caregiver scale. At follow‐up, all women had normalised haemoglobin (Hb); however, only those who received iron replacement were iron replete. ³¹

Anaemia, the end consequence of ID, is associated with adverse obstetric outcomes including low birthweight, ³² pre‐term delivery ³³ and risk of blood transfusion. ³⁴ An Australian CPI project found that routine ferritin screening reduced anaemia at birth from 12.2% to 3.6%. ⁷

Appropriate treatment

Two large systematic reviews and meta‐analyses concluded that oral iron improves Hb concentrations in pregnant women, with increases of 4.59 g/L (95% confidence interval (CI) 3.72–5.46) to 13.4 g/L (95% CI 2.7–24.2) and reduces the risk of ID (0.59, 95% CI 0.46–0.79) and IDA (0.40, 95% CI 0.26–0.60) at birth compared to placebo, but is associated with gastrointestinal side effects. ³² , ³⁵

Consensus is lacking regarding the optimal dose of oral iron in those with ID without anaemia. ³⁶ , ³⁷ A recent study assessing iron absorption in women with ID found that the maximal fractional absorption of iron occurred at 60 mg of elemental iron. ³⁸ Hepcidin, a key regulator of oral iron, increases post‐oral iron, leading to reduced absorption for 24 h. Intermittent oral iron dosing overcomes this suppression. A systematic review of intermittent compared to daily iron (as defined as any iron compound administered orally) in pregnancy found similar maternal and infant outcomes, although the risk of anaemia near term was increased in the intermittent dosing group but side effects, namely gastrointestinal, were reduced. ³⁹ Several small RCTs in both ID and IDA women have since been conducted with similar results. ⁴⁰ , ⁴¹ While alternate‐day dosing increases the fractional iron absorption, ⁴⁰ it may lead to inferior cumulative absorbed doses in IDA women. Further research continues in this area. ⁴²

The pivotal study of IVI in pregnancy, FER‐ASAP, an industry‐sponsored trial, ⁴³ randomised 252 pregnant women with IDA to receive ferric carboxymaltose (1000–1500 mg in divided doses) or ferrous sulphate (200 mg/day) for 12 weeks and showed that IVI was superior to oral iron in the degree and speed of Hb recovery, with 84% of women in the IVI group having normalised Hb compared to 70% in the oral iron group (odds ratio 2.06, P = 0.031) and within a shorter time frame (median 3.4 vs. 4.3 weeks). Additionally, vitality (P = 0.025) and social functioning (P = 0.049) were significantly improved prior to delivery in the IVI group. Treatment‐related adverse events were similar between the groups and none were severe.

Importantly, similar transferrin saturation levels were documented between the two groups. This is relevant as transplacental iron transfer is transferrin‐dependent. From these results, it can be inferred that IVI is not associated with excessive transplacental iron transfer. ²⁰ No newborns born to women in the IVI‐treated group had a serum ferritin level above the normal range. ⁴³

Different IV formulations

In Australia, iron polymaltose, ferric carboxymaltose and ferric derisomaltose are licensed for use in the obstetric setting. Several systematic reviews and RCTs have compared iron polymaltose and ferric carboxymaltose to a multitude of comparators in pregnancy, with a lack of strong evidence to support IVI over oral, ³⁵ , ⁴⁴ , ⁴⁵ and no statistically significant differences noted in outcomes or adverse events between the preparations. QoL measures were reported in only two studies, with improvements in general health (P = 0.021) and physical energy (P = 0.016), reduced psychological down heartedness (P = 0.005) clinical depression (P = 0.003) and overall improved mental health status (P < 0.001) with IVI compared to oral in one of the two studies. Limited included studies looked at the role of IVI in women with ID without anaemia. The largest, a retrospective review with a subset of 234 ID pregnant women, ⁴⁶ showed that Hb increased by a mean of 5.1 g/L (±1.3 g/L) at 3 weeks post‐infusion. No further increases in Hb or evidence of haemoconcentration were seen. More recently, a comparative RCT on ferric derisomaltose in pregnant women with refractory ID was published. The authors concluded that IVI was superior in preventing anaemia compared to oral iron (91% of women in the IVI group did not develop anaemia compared to 73% in the oral iron group, P < 0.001) and was associated with improved fatigue and QoL. Similar rates of adverse events were noted between the two groups, with two allergic reactions in the IVI group requiring antihistamines and corticosteroids but not adrenaline. ⁴⁷ There is emerging evidence that ferric derisomaltose may be associated with a lower incidence of hypophosphataemia compared to ferric carboxymaltose. ⁴⁸

Recommendations are as follows:

• An appropriate oral iron formulation should be offered as first‐line therapy for obstetric ID (GRADE 1B).

• Alternate‐day oral dosing can be considered to limit side effects in women with obstetric ID (GRADE 2B).

• IVI should be offered to women with ID/IDA who are intolerant of or refractory to oral iron or in the third trimester (GRADE 1B).

The authors note that oral iron tolerability is challenging particularly as pregnancy progresses and compliance is variable. These factors need to be considered when prescribing oral iron. There is insufficient evidence to support alternate‐day dosing in IDA to enhance tolerability. IVI is therefore the authors' preferred mode of replacement in women with IDA in the late second and third trimesters of pregnancy (Fig. 1).

Figure 1.

Obstetric iron optimisation flowchart.

An intersectional approach to iron deficiency in pregnancy

Obstetric ID is an example of intersectional biases in medicine where gender, socio‐economic and/or geographic factors compound, resulting in inequitable health outcomes. ¹ The symptoms of ID can be normalised by physicians and underappreciated by women especially during pregnancy. Patients with any barriers to expressing or articulating these symptoms due to language, literacy or disability are at an even greater disadvantage. Clinicians must advocate for their patients and unborn babies by being cognisant of the impacts of ID with proactive optimisation.

Call to action

While the management of ID in pregnancy is safe and effective, iron optimisation prior to pregnancy would ensure the availability of adequate iron stores to meet the increased needs of the maternal, foetal placental unit during pregnancy, thus avoiding the development of ID and the associated maternal and infant adverse outcomes.

Conclusion

The prevalence of ID in pregnancy has been estimated to be as high as 70%, affecting both mother and child with physical, psychological and neurocognitive consequences. Optimisation of ID in pregnancy improves maternal well‐being, prevents the development of anaemia and the associated morbidity and mortality and ensures the best opportunity for the health and welfare of the neonate. The lack of public health interventions despite the known prevalence and impact of ID has evolved into inequitable health outcomes and unnecessary suffering on the lives of women and their children.