Iron Overload and Outcomes in Very Low Birth Weight Infants

Iron is essential for growth and haematopoiesis but iron overload can have detrimental effects. The aim of this study is to correlate iron indices at 1 month of age in very low birth weight infants undergoing intensive care prior to additional iron supplementation and correlate results with clinical measures and 2 year corrected neurodevelopmental outcomes. A prospective observational cohort study was undertaken of very low birth weight infants at a tertiary neonatal intensive care unit. In 45 eligible infants, iron overload existed in 14 (31%) who had signicantly decreased: mean (SD) gestational age at birth [26(2) v 28(2) weeks], birth weight [814(231)g v 1011(238)g], weight on assessment [1180(369)g v 1502(423)g], haemoglobin level at birth [13.8(3.1)g/dL v 16.1(2.5)g/dL] and mean corpuscular volume at assessment [85(4)fL v 90(6)fL] with increased number of red cell transfusions [4(2.5-5) v 1(0-2)] and volume of transfusions [65 (38)mL v 27 (23)mL]. Regression analysis determined the number of transfusions had the strongest correlation with iron status on testing. Neurodevelopmental outcomes at 2 years were not signicantly different. Conclusion: Very low birth weight infants at risk of iron overload ( ≥ 2 red cell transfusions) may benet from the measurement of iron indices prior to oral iron supplementation.


Introduction
Iron essential for growth and haematopoiesis but excessive free iron has been associated with brain injury in models of hypoxic-ischemic brain injury and infection [1][2][3][4]. Preterm neonates require iron supplementation in the management of anaemia of prematurity and receive red cell transfusions contributing to iron load. Low circulating levels of transferrin and other iron-binding proteins in preterm infants allow increased circulation of non-protein-bound iron. Neonatal animal models have shown a predisposition to iron overload due to an inability to down-regulate their iron absorption [5].
There is no de nitive test to measure iron overload in preterm infants, and elevated serum ferritin is a nonspeci c indicator of severe hepatocyte injury and in ammation. Iron plays a vital role in cellular function in all organ systems and is essential for optimal development and function [3]. It is routine to supplement iron in preterm infants to prevent iron de ciency anaemia. The optimum dose, duration and point of intervention with iron supplementation in preterm and low birth weight infants has yet to be determined, though the recommended enteral iron intake for very low birth weight (VLBW) infants is 2-3mg/kg/day [6]. Serum ferritin (SF) concentration, a measure of tissue iron stores, is currently considered the best surrogate biochemical measure of iron overload in the absence of active infection or elevated Creactive protein [7]. Excess exogenous iron may contribute to oxidative injury in preterm babies exacerbating conditions such as necrotising enterocolitis, retinopathy of prematurity, bronchopulmonary dysplasia and inducing neuronal injury [3,4,8,9]. We aimed to correlate iron indices with clinical measures and 2 years corrected neurodevelopmental outcomes in VLBW infants.

Method
Ethical committee approval was received from a tertiary referral, university-a liated maternity hospital for this prospective study of infants birth weight <1500 g admitted to the neonatal intensive care unit (NICU). Iron indices [ferritin, transferrin, transferrin saturation, total iron binding capacity (TIBC), iron], full blood count, biochemistry pro le, liver pro le (transaminases, albumin, bilirubin), C-reactive peptide (CRP) were measured prior to the initiation of routine oral iron supplements. Infants were excluded if infection was suspected or con rmed acute illness, recipients of a red cell concentrate transfusion (RCC Tx) in the previous 48 hours, chromosomal disorders, discharged or died prior to achievement of full enteral feeds.
Serum ferritin (SF) is a clinically available proxy measure for iron status in the absence of confounders. The upper limit of normal of serum ferritin (300 mg/L) was derived from Obladen et al, where on days 24-26 (n=128) the 90th percentile was 300 and the 97th percentile was 383 mg/L (383 ng/mL) [9,10]. Serum ferritin less than 12 ng/L was suggestive of iron de ciency in preterm infants [3]. Neurodevelopmental assessment was conducted by a developmental psychologist (MS) at 2 years of age corrected using the Bayley Scales of Infant and Toddler Development (Bayley-III). Serial cranial ultrasounds at 0-24h, 24-72h, day 7, one month and day of discharge were performed by paediatric radiologist (VD) with an interest in neonatal brain injury. Magnetic resonance imaging (MRI) was conducted at 2 years corrected.

Statistics
Statistical analysis was performed using SPSS version 18. Analysis of Gaussian and non-Gaussian data used unpaired t-test and Mann-Whitney test, respectively. Paired data was analysed by paired t-test or Wilcoxon test. Categorical variables were reported as percentages and binomial measurements were compared using χ2 test. Quanti cation of the relationships was determined using Spearman and Pearson correlations. Stepwise logistic regression modelling was performed to identify variables associated with the iron status. In a stepwise fashion covariates that did not make a statistically signi cant contribution (p>0.1) to the model were removed from the model. All tests were two sided and a p < .05 was considered statistically signi cant.
The mean (SD) gestational age and birth weight of the entire cohort was 27(2) weeks and 949 (251) grams respectively. There were 20 female and 25 male infants. Fourteen infants were iron overloaded (Table 1). Infants with iron overload had a signi cantly lower gestational age at birth, birth weight, weight on assessment, haemoglobin at birth, and mean corpuscular volume at assessment with signi cantly increased number of transfusions and volume of erythrocyte transfusions (Table 1). No signi cant relationship or correlation between maternal preeclampsia, number of days to establish full feeds, type of feed (breast milk with or without non-iron rich forti er or formula milk), presence of intraventricular haemorrhage, incidence of high frequency oscillatory ventilation, usage of postnatal steroids, necrotising enterocolitis, presence of patent ductus surgery, liver transaminases, total bilirubin and C-reactive peptide was established between groups. The relationship between duration of intermittent positive pressure ventilation and retinopathy of prematurity was not signi cant between comparison groups.
Serial cranial ultrasound evaluation and neurodevelopmental outcomes (cognition, language and motor outcomes) at two years were not signi cantly different between the groups. Nine infants had 2-year follow-up MRIs and there was no predominance of abnormal MRI in either group. Three of the ve MRIs revealed periventricular white matter abnormalities, individual cases separately revealed deep white matter abnormalities, cortical dysplasia and abnormal myelination.
Correlation analysis of the entire cohort revealed a signi cant negative association between ferritin levels and gestation, birth weight, haemoglobin at birth, and weight on testing. A signi cant positive association was identi ed between ferritin levels and the number of erythrocyte transfusions, total transfused blood volume, daily iron intake, the duration of invasive ventilation and mean corpuscular volume on testing. All patients (n=8) who received 3 or more blood transfusions had iron overload. Four of nine patients (44%) who received 2 transfusions were overloaded and 1 of 7 (14%) who received 1 transfusion. Total iron intake including iron from feeds and RCC transfusions was signi cantly increased in those infants with iron overload. The factors identi ed to be associated with iron status with a p < 0.1 (Table I) were included in the model building for linear regression analysis. Stepwise logistic regression determined that the relationship (B) between every erythrocyte transfusion received, the ferritin increased by 62 ng/mL (B= 62ng/ml, 95% CI 28 to 96, p<0.001). Regression determined that the relationship (B) for every increasing gram of weight on testing the ferritin decreased by 0.169 ng/mL (B=-0.169 ng/ml, 95% CI -0.317 to -0.021, p=0.027). No infant had a ferritin less than 12 ng/mL.

Discussion
Nutritional iron supplementation is routine in most neonatal units without a baseline clinical or laboratory assessment of iron status. The immaturity of the regulation of gastrointestinal iron absorption and potential for increased cerebral uptake and neuronal injury with excess amounts of iron supplementation in preterm infants is unknown [3, 5 , 9, 11].
This study uniquely assessed iron status prior to additional oral iron supplementation and discovered 14 (31%) infants with iron overload which is higher than previously reported at just 19% at 35 weeks corrected [12]. All 8 infants who received 3 or more RCC Tx met criteria for iron overload which is also higher than previously reported at just 50% at 35 weeks corrected [12], although the cohort in our study were of lower gestational age with no signi cant difference between the total iron from feeds and iron status. Stepwise linear logistic regression modelling established a signi cant relationship between number of erythrocyte transfusions and weight at assessment with ferritin level. There was no signi cant impact on neurodevelopmental outcomes at 2 years associated with iron overload, which supports the neurodevelopmental ndings of Amin et al at 8 to 12 months, extending it to 2 year outcomes [13].
The strengths of our study are that its prospective nature, using a well-accepted biological test that is readily available in clinical practice. We performed the analysis upon a uniform population which was de ned prior to testing and prior to intervention with additional enteral iron supplements. However, the study is limited by relatively small numbers.
Our ndings suggest that iron status at 1 month is extremely variable among preterm VLBW infants and is predicted by the number of blood transfusions prior to additional oral iron supplementation. We suggest that infants at risk of iron overload (i.e. ≥2 transfusions received) should have iron indices measured prior to starting iron supplementation. For infants with iron overload, iron supplementation may be potentially harmful and future studies are required to study this and to determine the optimal dose, timing and contraindications to iron supplementation for preterm infants.

Declarations ACKNOWLEDGEMENTS:
We would like to acknowledge the support and assistance of the parents and staff in the Neonatal Intensive Care unit of the National Maternity Hospital. We would like to acknowledge Dr. Ranawaya for the contribution to data collection. Availability of data and material: All data collected was anonymised and stored in electronic format on locked hospital property and was destroyed thereafter as per hospital ethical guidelines.
Code availability: Not applicable.
Authors' contributions: MJB wrote the paper, and gave substantial contributions to the conception and design of the work, the acquisition, analysis and interpretation of data for the work. RMC co-designed and conceived the project and revised the manuscript. AMG revised and reviewed the manuscript for publication. MS conducted the neurodevelopmental assessments and revised the manuscript. VD performed and scored neuroimaging and revised the manuscript. AM co-designed the project and contributed to data collection and manuscript review and revision. BF gave substantial contributions to the conception and design of the work and the acquisition, analysis, and interpretation of data for the work. GH gave substantial contributions to the conception and design of the work, the acquisition, analysis and interpretation of data for the work. EI gave substantial contributions to the conception and design of the work, the acquisition, analysis and interpretation of data for the work. JQ analysed blood samples and gave substantial contributions to the conception and design of the work, the acquisition, analysis and interpretation of data for the work. EJM designed, supervised the project and co-wrote the paper. All authors were involved in drafting the work or revising it critically for important intellectual content, nal approval of the version to be published and agreement to be accountable for all aspects of the work in