Enteral Iron Supplementation in Extremely Preterm Infants and its Positive Correlation with Neurodevelopment; Post Hoc Analysis of the PENUT Randomized Controlled Trial

Abstract

Objectives:

To test whether an increased iron dose is associated with improved neurodevelopment as assessed by the Bayley Scales of Infant Development (BSID-III) among infants enrolled in the Preterm Erythropoietin (Epo) Neuroprotection Trial (PENUT).

Study design:

This is a post hoc analysis of a randomized trial which enrolled infants born at 24 to 28 completed weeks of gestation. All PENUT infants who were assessed with BSID-III at 2 years were included in this study. The associations between enteral iron dose at 60 and 90 days and BSID-III component scores were evaluated using generalized estimating equations models adjusted for potential confounders.

Results:

692 infants were analyzed (355 placebo, 337 Epo). Enteral iron supplementation ranged 0–14.7 mg/kg/day (IQR 2.1–5.8 mg/kg/day) at day 60, with a mean of 3.6 mg/kg/day in placebo-treated infants and 4.8 mg/kg/day in Epo-treated infants. A significant positive association was seen between BSID-III cognitive scores and iron dose at 60 days, with an effect size of 0.77 BSID points per 50 mg/kg increase in cumulative iron dose (P = .03). Higher iron doses were associated with higher motor and language scores, but did not reach statistical significance. Results at 90 days were not significant. The effect size in the Epo-treated infants compared with placebo was consistently higher.

Conclusion:

A positive association was seen between iron dose at 60 days and cognitive outcomes. Our results suggest that increased iron supplementation in preterm infants, at the doses administered in the PENUT Trial, may have positive neurodevelopmental effects, particularly in infants treated with Epo.

Trial Registration:

Clinicaltrials.gov: NCT01378273.

Because the majority of iron transfer occurs late in pregnancy(1), extremely low gestational age neonates (ELGANs) are at risk of iron deficiency if not supplemented appropriately. Rapid growth and phlebotomy further increases iron demand(2). Iron plays key roles in brain development, metabolism and function: it is required for mitochondrial respiration, nucleic acid replication and repair, immune function, cell signaling, as an enzyme co-factor and is required for myelination, dendritogenesis and neurotransmitter production(3–6). Thus, iron deficiency during periods of active brain development may result in maldevelopment of those brain regions with the highest energy needs during the period of deficiency and is associated with developmental deficits including impaired recognition memory(7), motor delays(8) and myelin-dependent speed of processing(9). Iron sufficiency is therefore vital to optimize neurodevelopmental outcomes(5, 9–12). In preterm infants, iron supplementation is frequently necessary to maintain iron sufficiency. However, free iron is a pro-oxidant, and iron-dependent pathways have been implicated in neonatal brain injury(13–15), so excessive iron supplementation may be detrimental.

Iron is increasingly recognized as a critical nutrient for brain development in preterm infants(16), but the optimal supplementation dose to balance the risks of deficiency and potential toxicity is unclear. The American Academy of Pediatrics (AAP) recommends iron supplementation doses of 2 mg/kg/day in preterm infants, starting at 1 month of age(17, 18), however we have previously documented iron deficiency in infants receiving up to 12 mg/kg/day(19, 20). We hypothesized that the AAP recommended doses might be inadequate to meet iron demands in preterm infants particularly those treated with erythropoiesis stimulating agents (ESAs) because iron is prioritized to red cells over the brain(21–24). Whether ESAs exacerbate negative iron balance in ELGANs and worsen neurodevelopment is unknown and is an additional aim of this study.

The Preterm Erythropoietin (Epo) Neuroprotection Trial (PENUT) population represents a well-characterized cohort of extremely preterm infants with standardized neurodevelopmental follow-up data. Although iron guidelines were provided as part of the PENUT recommendations, adherence was not required, leading to a wide range of iron supplementation doses. PENUT subjects are an ideal population in which to evaluate a broad spectrum of iron dosing and associated effects on neurodevelopmental outcomes in Epo- and placebo-treated infants.

We evaluated the association between cumulative iron dose and neurodevelopmental outcomes at 2 years corrected age as assessed by the Bayley Scales of Development, third edition (BSID-III). We hypothesized that there is a positive association between iron dose and BSID-III scores, with higher cumulative iron doses correlating with improved outcomes. Further, we hypothesized that Epo treatment would increase iron demand in Epo-treated infants, leading to a more significant correlation in Epo compared with placebo-treated infants.

Methods

Study Design and Population

We conducted a post hoc analysis of all infants from the PENUT population who underwent BSID-III assessment at two years corrected age. PENUT (NCT01378273) was a multisite, prospective, randomized, placebo-controlled trial conducted to evaluate the neuroprotective effects of Epo in infants born between 24 0/7 weeks and 27 6/7 weeks’ of gestation. The PENUT study design and primary outcome have been published previously(25, 26). Briefly, infants from 30 hospitals in the United States were enrolled and randomized to Epo or placebo. Neonates received Epo 1000 U/kg/dose or placebo intravenously every 48 hours for 6 doses, starting by 24 hours of life, followed by 400 U/kg/dose Epo subcutaneously or sham injection three times per week through 32 weeks postmenstrual age. Demographic information, including race and ethnicity, was self-reported by mothers and collected as part of the trial to assess potential generalizability of the study results. IRB approval was obtained at all study sites and written parental consent was obtained. In this manuscript, the analysis population included all infants enrolled in PENUT who survived and completed BSID-III assessments at 2-years corrected age.

Iron Supplementation

Iron guidelines were developed for PENUT (provided in online material). The guidelines recommended that all infants receive iron supplementation, starting at 3 mg/kg/day enteral iron (feeds plus supplement) when infants were 7 days old and on a minimum of 60mL/kg/day of enteral feeds. Iron dose was recommended to be increased to 6 mg/kg/day when infants reached 100 mL/kg/day of enteral feeds. If infants were not taking 60 mL/kg/day of enteral feeds by day of life 8, intravenous (IV) iron was recommended to be administered until enteral iron was tolerated. Iron status was assessed by serum ferritin or zinc protoporphyrin-to-heme ratio (ZnPP/H) and iron dosing was recommended to be adjusted based on these results up to a maximum of 12 mg/kg/day.

Outcomes

The primary outcome for PENUT was death or severe neurodevelopmental impairment based on standardized neuromotor examination and BSID-III assessment at 22–26 months corrected age. In this post hoc study, BSID-III assessment was the primary outcome.

Statistical Analyses:

All statistical analyses accounted for potential sibship correlation using Generalized Estimating Equations (GEE)(27) with robust standard errors and independence working correlation structure, consistent with prior studies(26).

The primary goal of this study was to evaluate the associations between mean BSID-III component scores and cumulative enteral iron intake at day 60. First, relationships between iron and outcomes were evaluated descriptively using scatterplots. Second, linear associations were examined among all participants using GEE models adjusted for treatment group, gestational age at birth (grouped as 24/25 week or 26/27 weeks), recruitment site, and potential confounding variables (Apgar score at 5 minutes, stage 2b or 3 necrotizing enterocolitis(NEC), intracranial hemorrhage (grade 3 or 4), severe sepsis (defined as culture-positive infection requiring blood pressure support and/or significant escalation in respiratory support), cumulative volume of packed red blood cell (pRBC) transfusion per kg to day 60) and maternal education(28). We fitted similar GEE models but with the addition of interactions between cumulative enteral iron intake and treatment arm, to examine the potential Epo modification of the iron-outcome relationships of interest. The associations between BSID-III scores and cumulative iron intake were assessed within each treatment arm using GEE models adjusted for gestational age group, recruitment site, and similar potential confounding variables.

As a sensitivity analysis, the impacts of IV iron and red blood cell transfusions, were evaluated. Additionally, correlations between iron dose at 90 days of life and BSID-III scores were evaluated as well as the impact of maternal education as a potential confounder affecting neurodevelopment.

All statistical tests were two-sided and statistical significance was set at level 0.05. All analyses were post-hoc and not adjusted for multiple testing. Statistical analyses were performed using the R statistical software package (version 3.5.1, Vienna, Austria)(29).

Results

There were 692 infants, 355 placebo-treated and 337 Epo-treated, included in this analysis. Screening, inclusion, and available analysis data are described in Figure 1 (available at www.jpeds.com). Demographic characteristics are summarized in Table I. There were no significant differences in maternal and neonatal characteristics at baseline between treatment groups. Infants with missing data, on either BSID-III scores or any of the potential confounding variables used in GEE models were excluded from later statistical analyses (see footnote in Table 1). BSID-III scores are summarized in Table 2 (available at www.jpeds.com). BSID-III scores for all infants showed a mean cognitive score of 91.0 (SD 15.7), motor 90.1 (SD 16.8), and language 88.4 (SD 17.6).

Figure 1; Online: CONSORT diagram.

The CONSORT diagram shows the number screened, eligible, randomized, and included in this post-hoc analysis from the parent PENUT Trial.

Table 1:

Maternal demographics and neonatal data at enrollment

	Placebo (N = 355)	Epo (N = 337)
Maternal demographics
Age, mean (SD)	28.6 (6.2)	29.5 (6.1)
Hispanic(a), n (%)	80 (23%)	73 (22%)
Race, n (%)
White	238 (67%)	230 (68%)
Black	74 (21%)	79 (23%)
Others, unknown, or unreported	43 (12%)	28 (8%)
Education(b), n (%)
High School or less	123 (35%)	98 (29%)
Some college	99 (28%)	113 (34%)
College degree or greater	96 (27%)	87 (26%)
Neonatal data at enrollment
Pregnancy induced hypertension, n (%)	27 (8%)	26 (8%)
Prenatal steroids(d), n (%)	326 (92%)	306 (91%)
Prenatal magnesium sulfate(e), n (%)	294 (83%)	268 (80%)
Cesarean delivery, n (%)	241 (68%)	227 (67%)
Delayed cord clamping(g), n (%)	122 (34%)	134 (40%)
Sex - Male, n (%)	179 (50%)	175 (52%)
Birth gestational age, n (%)
24w	83 (23%)	75 (22%)
25w	97 (27%)	81 (24%)
26w	95 (27%)	75 (22%)
27w	80 (23%)	106 (31%)
Mean (SD)	25.9 (1.1)	26.0 (1.2)
Multiple gestation, n (%)	96 (27%)	85 (25%)
Birth weight in grams, mean (SD)	804 (183)	827 (192)
Apgar score at 5 minutes < 5(h)	62 (17%)	59 (18%)

^(a)Ethnicity was unknown or unreported for 1 (<1%) and 1 (<1%) subject in the placebo and Epo groups, respectively.

^(b)Maternal education was unknown or unreported for 37 (10%) and 39 (12%) subjects in the placebo and Epo groups, respectively.

^(c)Risk of chorioamnionitis included prolonged ruptured of membranes, suspected or confirmed chorioamnionitis, pre-term labor, pyrexia 38.1°C in labor, and antibiotic administration

^(d)Prenatal steroids were unknown for 6 (2%) and 3 (1%) subjects in the placebo and Epo groups, respectively.

^(e)Prenatal magnesium sulfate was unknown for 18 (5%) and 8 (2%) subjects in the placebo and Epo groups, respectively.

^(f)Delivery complications were defined as the presence of one or more of the following complications during delivery: prolapsed cord, true knot, tear or rupture of cord, placental abruption, twin-twin transfusion, feto-maternal bleeding, ruptured uterus, or traumatic instrument delivery.

^(g)Delayed cord clamping was unknown for 94 (26%) placebo and 89 (26%) Epo subjects.

^(h)Apgar score at 5 minutes was missing for 3 (1%) subjects in the Epo group.

Table 2, online:

Summary of BSID-III component scores in all participants and by treatment groups

BSID-III Component	Placebo (N=355)	Epo (N=337)	All Participants (N=692)
Cognitive – N; mean (SD)	355; 90.8 (15.8)	337; 91.4 (15.7)	692; 91.0 (15.7)
Motor – N; mean (SD)	348; 90.2 (17.3)	332; 90.0 (16.4)	680; 90.1 (16.8)
Language – N; mean (SD)	345; 88.7 (17.9)	332; 87.4 (17.2)	677; 88.4 (17.6)

Daily enteral iron supplementation ranged 0–14.7 mg/kg/day (IQR 2.1–5.8 mg/kg/day) at day 60 and 0–19.7 mg/kg/day (IQR 0–5.4 mg/kg/day) at day 90. Epo-treated infants received higher daily iron doses, with a mean (SD) of 4.8 (3.3) mg/kg/day and 3.5 (3.9) mg/kg/day at day 60 and 90 days respectively, compared with 3.6 (2.5) and 2.9 (3.2) mg/kg/day in the placebo-treated group. Overall, infants received a cumulative enteral iron mean (SD) of 179.5 (114.3) mg/kg and 289.9 (159.9) mg/kg at day 60 and 90 respectively. Epo-treated infants received higher cumulative iron supplementation, with a mean (SD) of 211.7 (116.7) mg/kg and 333.9 (168.5) mg/kg to day 60 and 90, compared with 149.0 (103.2) mg/kg and 248.1 (139.0) mg/kg in the placebo-treated group. Placebo-treated infants received higher cumulative pRBC transfusion volumes, with a mean (SD) volume of 71.5 (57.7) mL/kg and 77.9 (65.4) mL/kg by 60 and 90 days, compared with 41.9 (46.1) mL/kg and 45.1 (51.2) mL/kg by 60 and 90 days in the Epo-treated group. The mean differences in cumulative iron supplementation, as well as cumulative transfusion volume, at day 60 and day 90 between treatment groups were statistically significant (all p<0.001).

Figure 2 illustrates consistently positive linear relationships between cumulative iron intake at day 60 and BSID-III scores for both treatment arms. Regression model results estimating the associations between cumulative iron intake at day 60 and BSID-III scores in all infants are shown in Figure 3, reported as the estimated effect size for each 50 mg/kg increase in cumulative enteral iron intake at day 60, adjusted for treatment arm and potential confounding variables. There was a significant association between cumulative iron dose and cognitive outcomes: an increase of 50 mg/kg of cumulative enteral iron at day 60 was associated with a higher mean cognitive score of 0.77 points (95% confidence interval (CI) of [0.06, 1.48]; p=0.03). When refitting GEE models with interactions between cumulative enteral iron intake and treatment, the estimated interactions were not statistically significant (all p>0.05).

Figure 2: Unadjusted linear trends in cumulative enteral iron supplementation at day 60 and BSID-III component scores at 2-years of age.

Raw data on cumulative enteral iron supplementation at day 60 and BSID-III scores are shown, with unadjusted linear trends fitted by treatment group for cognitive component (panel A), motor component (panel B), and language component (panel C).

Figure 3: Associations between cumulative iron intake at day 60 and BSID-III component scores at 2-years of age for all infants.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 60 in all infants were examined using GEE models clustering on same-birth siblings and adjusted for fixed effects of treatment group, gestational age group, recruitment site, Apgar score at 5 minutes, necrotizing enterocolitis (stage 2b or 3), severe intracranial hemorrhage, severe sepsis, and cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 60. Cumulative IV iron at day 60 was adjusted in sensitivity analyses.

[1] Results are reported as effect size of 50 mg/kg of cumulative enteral iron at day 60 on BSID-III scores

[2] Positive values indicate that higher iron intake was associated with higher mean BSID-III score, adjusted for potential sibship correlations and confounding variables

Figure 4 shows similar results estimating the associations between cumulative iron intake at day 60 and BSID-III scores in each treatment group. Despite a loss of power when examining treatment arms individually, a significant association between cumulative iron dose and cognitive score was identified in the Epo-treated group, with an effect size of 1.02 (95% CI of [0.11, 1.94]; p=0.03). A similar relationship was seen in the placebo-treated group but was not significant.

Figure 4: Associations between cumulative iron intake at day 60 and BSID-III component scores at 2-years of age by treatment arms.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 60 for infants in each treatment arm were examined using GEE models clustering on same-birth siblings and adjusted for fixed effects of gestational age group, recruitment site, Apgar score at 5 minutes, necrotizing enterocolitis (stage 2b or 3), severe intracranial hemorrhage, severe sepsis, and cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 60. Cumulative IV iron at day 60 was adjusted in sensitivity analyses.

[1] Results are reported as effect size of 50 mg/kg of cumulative enteral iron at day 60 on BSID-III scores

[2] Positive values indicate that higher iron intake was associated with higher mean BSID-III score, adjusted for potential sibship correlations and confounding variables

When adjusting for cumulative IV iron, the positive impact of cumulative enteral iron on BSID-III scores became stronger in magnitude (though did not reach statistical significance in all domains). In the sensitivity analysis evaluating the impact of higher cumulative transfusion volume, we found that higher transfusion volume was associated with lower BSID-III scores (Figure 5; available at www.jpeds.com). When we reran the analyses adjusted for maternal education in the 616 participants for whom this data was available, the results did not change significantly (Figure 9 and Figure 10; available at www.jpeds.com).

Figure 5; online: Effect sizes of cumulative pRBC transfusion volume at 60 days in statistical models examining the relationships between cumulative enteral iron intake at day 60 and BSID-III component scores.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 60 in all infants overall and by treatment group were examined using GEE models clustering on same-birth siblings and adjusted for potential confounding variables, which included cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 60, as shown in Figures 3 and 4. This figure shows the effect size of cumulative pRBC transfusions in all of these statistical models, indicated as change in mean BSID-III scores corresponding to an increase of 15mL/kg in cumulative pRBC volume at day 60.

[1] Results are reported as effect size of 15 mL/kg of cumulative transfusion volume at day 60 on BSID-III scores

[2] Negative values indicate that higher cumulative transfusion volume was associated with lower mean BSID-III score, adjusted for potential sibship correlations, cumulative enteral iron, and confounding variables

Figure 9; online: Associations between cumulative iron intake at day 60 and BSID-III component scores at 2-years of age for all infants, accounting for maternal education, stratified by treatment group.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 60 in all infants were examined using GEE models clustering on same-birth siblings and adjusted for fixed effects of treatment group, gestational age group, recruitment site, Apgar score at 5 minutes, necrotizing enterocolitis (stage 2b or 3), severe intracranial hemorrhage, severe sepsis, cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 60, and maternal education. Cumulative IV iron at day 60 was adjusted in sensitivity analyses.

[1] Results are reported as effect size of 50 mg/kg of cumulative enteral iron at day 60 on BSID-III scores

[2] Positive values indicate that higher iron intake was associated with higher mean BSID-III score, adjusted for potential sibship correlations and confounding variables, including maternal education

Figure 10; online: Associations between cumulative iron intake at day 60 and BSID-III component scores at 2-years of age by treatment arm, accounting for maternal education.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 60 for infants in each treatment arm were examined using GEE models clustering on same-birth siblings and adjusted for fixed effects of gestational age group, recruitment site, Apgar score at 5 minutes, necrotizing enterocolitis (stage 2b or 3), severe intracranial hemorrhage, severe sepsis, cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 60, and maternal education. Cumulative IV iron at day 60 was adjusted in sensitivity analyses.

[1] Results are reported as effect size of 50 mg/kg of cumulative enteral iron at day 60 on BSID-III scores

[2] Positive values indicate that higher iron intake was associated with higher mean BSID-III score, adjusted for potential sibship correlations and confounding variables, including maternal education

To evaluate the persistence of these findings over time, we evaluated the correlation between cumulative iron dose at 90 days and BSID-III scores, both overall and stratified by treatment group (Figures 6–8; available at www.jpeds.com). Although similar trends were seen, they did not reach statistical significance at this time point.

Figure 6; online: Unadjusted linear trends in cumulative enteral iron supplementation at 90 days and BSID-III component scores at 2-years of age.

Raw data on cumulative enteral iron supplementation at day 90 and BSID-III scores are shown, with unadjusted linear trends fitted by treatment group for cognitive component (panel A), motor component (panel B), and language component (panel C).

Figure 8; online: Effect sizes of cumulative pRBC transfusion volume at 90 days in statistical models examining the relationships between cumulative enteral iron intake at day 90 and BSID-III component scores at 2-years of age.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 90 in all infants overall and by treatment group were examined using GEE models clustering on same-birth siblings and adjusted for potential confounding variables, which included cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 90, as shown in eFigure 3. This figure shows the effect size of cumulative pRBC transfusions in all of these statistical models, indicated as change in mean BSID-III scores corresponding to an increase of 15mL/kg in cumulative pRBC volume at day 90.

[1] Results are reported as effect size of 15 mL/kg of cumulative transfusion volume at day 90 on BSID-III scores

[2] Negative values indicate that higher cumulative transfusion volume was associated with lower mean BSID-III score, adjusted for potential sibship correlations, cumulative enteral iron, and confounding variables

Discussion:

In this post hoc analysis of ELGANs, a greater cumulative iron dose at 60 days was associated with improved cognitive scores, with similar, but not significant, associations seen in motor and language scores among all study infants. These associations were of greater magnitude among Epo-treated infants, and were statistically significant in all domains when adjusted for IV iron. This pattern was seen at both 60 and 90 days but with attenuated strength of association at 90-days. These findings are consistent with our hypothesis that increased iron dose within the PENUT dosing range is associated with improved neurodevelopmental outcomes at 2 years corrected age.

A consistently positive linear association between iron dose and neurodevelopmental outcomes is observed throughout all iron doses assessed (as seen in Figure 2). The absence of a “fall-off” in the curve at higher iron doses suggests that no iron overload or toxicity occurred in the study population. Also, no flattening of the curve in Figure 2 is seen, suggesting that throughout the range of iron doses assessed, no saturation point is reached. This persistently positive association suggests that throughout the range of iron doses administered, iron supplementation helped to replenish negative iron balance without evidence of saturation or toxicity. Evaluating for iron overload is critical as iron supplementation of iron sufficient children has been shown to be detrimental. In a cohort of infants in Chile, Lozoff et al evaluated neurodevelopmental outcomes in infants with high hemoglobin levels who were supplemented with high-iron formula versus those who received the high-iron formula who had low hemoglobin levels. They found that those infants who had higher hemoglobin levels had worse neurodevelopmental outcomes if they received high iron-fortified formula. This suggests that iron supplementation of iron-replete infants may have detrimental effects. In our study of PENUT infants, the persistently positive association between iron dose and outcome suggests that a saturation or overload point was not reached. We speculate that preterm infants have a significant iron deficit due to lack of placental iron transfer, repeat phlebotomy, and the use of ESAs which increase the incorporation of iron for erythropoiesis. Iron supplementation with doses up to 14.7 mg/kg/day showed a positive association with neurodevelopmental outcome, suggesting a significant iron debt. Friel et al calculated that very low birth weight neonates require 1 mg/kg/day of IV iron in order to maintain iron sufficiency(30). Based on enteral iron absorption rates in neonates of approximately 26%−35%(31, 32), this would suggest a need for 3–4 mg/kg/day of iron merely to sustain iron balance, with additional iron needed for losses through phlebotomy.

The timing of iron supplementation is also salient in this preterm cohort. Steinmacher et al compared the effects of early (from 2 weeks) versus late (from 2 months) iron supplementation in preterm infants <1301 g, and showed early iron supplementation was associated with better neurocognitive and psychomotor development at 5 years (though p>0.05)(33). Similarly, Christian et al showed that maternal-fetal iron supplementation, versus postnatal, has the most significant impact on neurodevelopment in term infants(34). Given that PENUT infants were of extremely low gestational age, iron supplementation in this age group likely impacted critical processes of brain development, specifically in rapidly developing regions like the hippocampus(35), which is important for cognitive development, with the motor and language areas less affected. The importance of early (versus late) iron supplementation may also potentially explain the difference in significance seen in the 60, as opposed to 90-day iron dosing in this study, with the first 60 days representing a time when the brain is particularly vulnerable to iron deficiency.

Although a positive but statistically insignificant association between iron dose and BSID-III scores was seen in all infants, the observed effect size was higher in the Epo-treated group but not significantly different from the effect among placebo-treated neonates. This is in keeping with our hypothesis that iron needs are greater in infants treated with ESA’s, due to the increased iron requirements for augmented erythropoiesis on top of the already high iron demands of the growing infant at this age. In two studies by our group, preterm infants, many of whom were treated with ESAs, showed downtrending or low serum ferritin values despite receiving up to 12mg/kg/day of iron supplements(19, 20). This highlights the importance of providing adequate iron supplementation to infants treated with ESAs, as their iron debt may be larger. The high iron need during ESA treatment may also potentially explain the difference in significance seen between the association between iron dosing at 60 days versus 90 days and BSID-III score. All PENUT infants in the Epo-treated group received Epo until 32 weeks’ post-menstrual age. Those born at 24 0/7 weeks’ gestation received Epo for 56 days, and those born at 27 6/7 weeks received Epo for only 29 days. Thus, the differential effects of Epo treatment may have diminished by 90 days. Although we may conclude that iron supplementation in Epo-treated infants is particularly critical, infants in the placebo-treated arm also demonstrated improved BSID-III scores with higher iron doses in this analysis, suggesting that all preterm infants benefit from iron supplementation. Further studies are needed to explore the relationship between iron, ESA’s and neurodevelopment.

Based on current literature and these findings, we propose that supplemental iron administration in ELGANs is safe and may improve outcomes. Iron supplementation guidelines were part of PENUT and are in place at the University of Washington and Minnesota. Although no prospective trials have compared iron dosing regimens, based on available evidence, we recommend initiating enteral iron supplementation at 4 mg/kg/day once infants are 7 days old and tolerating 60 mL/kg/day of enteral feeds. Iron doses should be adjusted to achieve ferritin and ZnPP/H values within goal of 76–400 ng/mL and 30–183 micromol/mol, up to 12 mg/kg/day. Further studies to identify iron biomarkers that optimize neurodevelopment are needed.

Figure 3, and Figures 5 through 10 were adjusted for potential confounders that were associated with worse neurodevelopmental outcomes in this cohort, including pRBC transfusion. This suggests that the beneficial effects of increased iron endowment from pRBC transfusions are negated by potential detrimental effects of transfusions on development. Although several studies have compared the neurodevelopmental effects of restrictive and liberal transfusion guidelines and shown no difference in outcome, these studies have not compared the outcome of infants with no transfusions compared with those with one or more transfusion, or the effects of transfusion volume on outcomes(28, 36–38). There is some suggestion that pRBC transfusions may be associated with worse neurodevelopmental outcomes(28, 39, 40). Large studies have evaluated the association between different hematocrit thresholds and outcome, but have not evaluated the impact of transfusion volume on neurodevelopment(36–38, 41). Given potential negative effects of pRBC transfusions, including infection, inflammation(42), transfusion-related NEC(43) and potentially adverse developmental effects, it is ideal to implement strategies to minimize the need for transfusions, including delayed cord clamping, limiting phlebotomy, use of restrictive transfusion guidelines and ESA’s.

A limitation of this study is that it represents a post-hoc analysis of a randomized clinical trial. Although PENUT included a large cohort of ELGANs who received a wide range of iron doses, the study was not designed to evaluate the association between iron dose and outcome. Therefore, as with all post-hoc analyses, results must be interpreted with caution. PENUT participating centers were provided with iron dosing guidelines, however, guideline adherence was variable, leading to a non-homogeneous iron supplementation practice. The variability in treatment practices were exploited here to evaluate the effects of different iron supplementation doses.

Challenges arise when using BSID-III assessment as the primary outcome measure to evaluate the neurodevelopmental effects of iron supplementation. A comprehensive outcome measure such as the BSID-III is advantageous in providing a comparable measure of development, but can be confounded by many other contributing factors. We have tried to control for these in our analysis by controlling for clinical factors known to affect development; however, the post-discharge environment has a significant impact on neurodevelopment and could not be adequately controlled for in this study.

A positive association was identified between iron dose at 60 days and cognitive outcomes, with non-significant but positive associations seen in motor and language developmental domains. When adjusted for IV iron, these relationships were more pronounced, with a significant association between cumulative enteral iron dose and neurodevelopmental scores in all domains in the Epo-treated subjects. Findings at 90 days did not reach significance. This suggests that the first 60 days are a critical period for iron and brain development in preterm infants and that increased iron supplementation in ELGANs even at the relatively high doses administered in the PENUT population, may have positive neurodevelopmental effects, particularly in infants treated with ESAs. We speculate that greater iron supplementation addresses the negative iron balance induced by prematurity, phlebotomy and ESAs. Our results add to the mounting body of evidence suggesting that current AAP iron supplementation guidelines may be inadequate to support optimal brain development and to address the iron debt in preterm infants, particularly in the context of increased use of ESAs.

Figure 7; online: Associations between cumulative iron intake at day 90 and BSID-III component scores at 2-years of age for all infants and by treatment groups.

The associations between mean BSID-III component scores (cognitive, motor, and language) and cumulative enteral iron intake at day 60 in all infants were examined using GEE models clustering on same-birth siblings and adjusted for fixed effects of treatment group, gestational age group, recruitment site, Apgar score at 5 minutes, necrotizing enterocolitis (stage 2b or 3), severe intracranial hemorrhage, severe sepsis, and cumulative volume of pack red blood cell (pRBC) transfusion per kg at day 60. Cumulative IV iron at day 60 was adjusted in sensitivity analyses. Similar analyses were performed for infants in each treatment group.

[1] Results are reported as effect size of 50 mg/kg of cumulative enteral iron at day 90 on BSID-III scores

[2] Positive values indicate that higher iron intake was associated with higher mean BSID-III score, adjusted for potential sibship correlations and confounding variables

Abbreviations:

ELGANs

extremely low gestational age neonates

AAP

American Academy of Pediatrics

Epo

Erythropoietin

Epo

Preterm Erythropoietin

PENUT

Neuroprotection Trial

BSID-III

Bayley Scales of Development, third edition

IV

intravenous

ZnPP/H

zinc protoporphyrin-to-heme ratio

GEE

Generalized Estimating Equations

pRBC

packed red blood cell

CI

confidence interval

ESA

erythropoietic stimulating agent

NEC

necrotizing enterocolitis

IQR

Inter-quartile range

SD

Standard Deviation

APPENDIX. Additional members of the PENUT Consortium

PENUT Site PI’s

Rajan Wadhawan², Sherry E. Courtney³, Tonya Robinson⁴, Kaashif A. Ahmad⁵, Ellen BendelStenzel⁶, Mariana Baserga⁷, Edmund F. LaGamma⁸, L. Corbin Downey⁹, Nancy Fahim¹⁰, Andrea Lampland¹¹, Ivan D. Frantz, III¹², Janine Khan¹³, Michael Weiss¹⁴, Maureen M. Gilmore¹⁵, Jean Lowe ^16, Nishant Srinivasan¹⁷, Jorge E. Perez¹⁸, Victor McKay¹⁹

PENUT Consortium Co-investigators

Billy Thomas MD, MPH³, Nahed Elhassan MD, MPH³, Sarah Mulkey MD, PhD³, Vivek K. Vijayamadhavan MD⁵, Bradley Yoder MD⁷, Jordan S. Kase MD⁸, Jennifer Check MD, MS⁹, Semsa Gogcu MD, MPH⁹, Erin Osterholm MD¹⁰, Thomas George MD¹⁰, Camilia R. Martin MD, MS¹², Deirdre O’Reilly MD, MPH¹², Nicolas Porta MD¹³, Raye-Ann de Regnier MD¹³, Catalina Bazacliu MD¹⁴, Frances Northington MD¹⁵, Raul Chavez Valdez MD¹⁵, Patel Saurabhkumar MD, MPH¹⁷, Magaly Diaz-Barbosa MD¹⁸.

Affiliations

1. University of Washington (Seattle, Washington)

2. AdventHealth for Children, (Orlando, Florida)

3. University of Arkansas for Medical Sciences (Little Rock, Arkansas)

4. University of Louisville, (Louisville, Kentucky)

5. Methodist Children’s Hospital (San Antonio, Texas)

6. Children’s Minnesota (Minneapolis, Minnesota)

7. University of Utah (Salt Lake City, Utah)

8. Maria Fareri Children’s Hospital at Westchester Medical Center (Valhalla, New York)

9. Wake Forest School of Medicine (Winston-Salem, North Carolina)

10. University of Minnesota Masonic Children’s Hospital (Minneapolis, Minnesota)

11. Children’s Minnesota (St. Paul, Minnesota)

12. Beth Israel Deaconess Medical Center (Boston, Massachusetts)

13. Prentice Women’s Hospital (Chicago, Illinois)

14. University of Florida (Gainesville, Florida)

15. Johns Hopkins University (Baltimore, Maryland)

16. University of New Mexico (Albuquerque, New Mexico)

17. Children’s Hospital of the University of Illinois (Chicago, Illinois)

18. South Miami Hospital (South Miami, Florida)

19. Johns Hopkins All Children’s Hospital (St. Petersburg, Florida)

20. University of North Carolina School of Medicine (Chapel Hill, North Carolina)

21. University of California San Francisco School of Medicine (San Francisco, California)

22. National Institute of Neurological Disorders and Stroke (Bethesda, Maryland)

23. University of Iowa (Iowa City, Iowa)

Appendix 7. Clinical Guidelines for PENUT Trial – Iron Supplementation

Goal: To maintain iron sufficiency in all study subjects.

Rationale:

Iron status may influence neurodevelopmental outcomes. Iron is an essential nutrient for the normal functioning of all cells, because iron-containing enzymes are involved in critical pathways such as oxidative metabolism, neurotransmitter production, cell replication, and energy metabolism.¹⁻³ Its deficiency is associated with symptoms ranging from reversible hematologic abnormalities to potentially irreversible neurodevelopmental abnormalities.^{4, 5} The third trimester is an especially important time for brain development, and inadequate nutrition, specifically insufficient iron accretion, during this time can have permanent consequences for brain function.^{6, 7} In utero iron accretion rate is 1 mg/kg/day. Too much iron can also be harmful in preterm infants since iron has powerful oxidant properties and can be toxic when present in excess.⁸

Iron balance is the net result of stores present at birth, iron utilization, and iron loss. The prenatal transfer of iron is affected by placental function, maternal factors such as smoking, diabetes and her iron status, and fetal factors such as intrauterine hypoxia.⁹ Infants with intrauterine growth retardation and infants of diabetic mothers are known to be at high risk for iron deficiency.⁹ Due to the timing of their birth (end of second trimester), ELGANs are denied the normal placental transfer of iron that occurs during the third trimester. Thus if they are inadequately supplemented after birth, they can become iron deficient over time, even if they were iron sufficient for gestational age at the time of birth.¹⁰ Postnatally, iron stores are influenced by iron loss (phlebotomy), iron utilization (Epo treatment, erythropoiesis) and iron intake (parenteral or enteral).

The AAP recommends enteral iron supplementation of preterm infants < 1500 gm at 2 to 4 mg/kg/day.¹¹ Epo treated infants require higher dosing due to the prioritization of iron to erythropoiesis.¹²⁻¹⁷ When infants are unable to tolerate enteral feedings, iron is still required for normal growth and development. Because there is no loss due to compromised absorption, parenteral dosing of iron is lower than enteral. Our goal is to avoid contributing to poor neurologic outcomes due to iron deficiency,^{6, 18, 19} while avoiding any potential increase in oxidative injury.⁸ Iron is prioritized to erythropoiesis under normal circumstances, and in the presence of Epo, this effect will be intensified, thus we feel it is imperative to provide adequate iron to avoid iron deficiency at the tissue level, particularly in brain.²⁰ Because preterm newborns have insufficient radical scavenging systems in the first week of life, no supplemental iron (above that contained in formula) will be given until after 7 days of age.²¹ In the context of iron deficiency, preterm infants given oral iron doses of up to 12 mg/kg/day tolerate this well and show no increase in free iron or in oxidative stress as measured by blood and urine isoprostanes.^{22, 23} Administering iron to all subjects will help preserve the blinded condition.

Assessing Iron Balance.

Serum ferritin reflects the iron stores in the body. Iron stores increase over the last trimester, with ferritin increasing from a mean concentration of 63 ng/mL at 23 weeks to 171 ng/mL at term.^{9, 24} Low serum ferritin concentrations accurately reflect iron deficiency, but an elevated serum ferritin may reflect inflammation, iron overload, or may be transiently increased after red blood cell transfusion. Serum ferritin concentration of 80 ng/mL is at the 25th percentile for preterm infants; <40 ng/mL has been correlated with abnormal ERP in IDMs and iron deficiency in brain on neonatal autopsy specimens, and concentrations of <15 ng/mL reflect severe iron deficiency.^{3, 9, 25, 26} Other studies show that serum ferritin <75 ng/mL is associated with abnormal neurological reflexes and ABR in preterm infants.^{27, 28} Ferritin can be measured on 0.4 mL of blood.

Measuring zinc protoporphyrin to heme ratios (ZnPP/H) is another way to assess iron sufficiency: this ratio reflects the availability of iron to be incorporated into the protoporphyrin molecule to form heme.²⁹ If iron is unavailable, zinc is incorporated instead, thus the ratio increases with iron deficiency. ZnPP/H deceases over the last trimester as iron stores increase in the fetus.^{30, 31} Elevated ZnPP/H occurs with iron deficiency, but may also occur with enhanced iron utilization as occurs with Epo-stimulated erythropoiesis.³¹ ZnPP/H is a more sensitive test than serum Ferritin. This test is performed on the red blood cell pellet.

Treatment Plan

When enteral feedings are started, a standard iron containing formula will be used if breast milk is unavailable. After the first week of life, and once infants (all subjects) have an enteral intake of 60 mL/kg/d, they will be started on enteral iron at a dose of 3 mg/kg/d total (feeds + supplement). Enteral iron will be increased to 6 mg/kg/d total when infants achieve an enteral intake of 100 mL/kg/d.³² This is a well tolerated dose that helps to prevent iron deficiency in preterm infants.³² Serum ferritin or ZnPP/H ratios should be checked at 14 days and again at 42 days. If adjustments in iron supplementation are required, consider following iron studies every two to three weeks, adjusting iron accordingly.

• If the serum ferritin is <40 ng/mL, increase the supplementation by 2 mg/kg/day.

• If the serum ferritin is 40 ng/mL to 100 ng/mL, increase the supplementation by 1 mg/kg/day.

• If the serum ferritin is >100 ng/mL but less than 400 ng/mL, continue same dose.

• If the serum ferritin is >400 ng/mL, hold the iron and check ferritin in 2 weeks.

• If ZnPP/H > 184, increase the supplementation by 2 mg/kg/day.

• If ZnPP/H values are in the normal range (ZnPP/H 30 to 183), continue 6 mg/kg/day.

The iron dose may be increased up to 12 mg/kg/day to maintain iron sufficiency.

Figure 1.

Iron Management (Ferritin or ZnPP/H may be used to assess iron status)

If the subject is not taking oral/enteral feeding volumes totaling 60 mL/kg/day by study day 8, iron should be supplemented intravenously. To maintain in utero accretion rates, 1 mg/kg/day is needed, and doses of up to 2 mg/kg/day have been demonstrated to maintain ferritin levels with good safety profiles. Patient needs will vary depending on their iron status at birth, phlebotomy losses, and transfusion rates. To be conservative, all subjects will receive 1.5 mg/kg/dose twice a week of parenteral iron (iron dextran or iron sucrose)³³ by intravenous infusion until they achieve an enteral intake of 60 mL/kg/d.³⁴ Iron dextran can be added to the parenteral nutrition solution and administered over 24 h, or diluted in 2 mL of D10W or normal saline and administered over 4 h. Iron dextran is not compatible with lipid emulsions, so must go through a separate line, or lipid emulsion must be turned off during the iron dextran infusion. Iron sucrose is an alternative option. Unlike iron dextran, iron sucrose can be administered over 20 to 30 minutes, so this may be more convenient. As with iron dextran, we stop the TPN and lipid emulsion, flush the line, give the iron sucrose, and flush the line. This dose is within the range of normal requirements for growing premature infants, and is meant to ensure that iron deficiency is not a confounding factor for neurodevelopmental outcomes in this study. If subjects are made NPO at a later point in their course, parenteral iron should be used if no feedings are given for a period of 7 days or greater.

• If the serum ferritin is <40 ng/mL, supplement with 2 mg/kg three times a week.

• If the serum ferritin is > 40 ng/mL to 100 ng/mL, supplement with 2 mg/kg twice times a week.

• If the serum ferritin is >100 ng/mL but less than 400 ng/mL, continue same dose (1.5 mg/kg twice a week).

• If the serum ferritin is >400 ng/mL, hold the iron and check ferritin in 2 weeks.

If serum ferritin remains low, continue to increase iron supplements up to 2 mg/kg/day IV. Similar incremental increases in iron dosing can be made using ZnPP/H.

• If ZnPP/H > 184, increase the supplementation incrementally to a maximum of 2 mg/kg/day.

• If ZnPP/H values are in the normal range (ZnPP/H 30 to 183), continue current dosing.

After study drug dosing is completed (32–6/7 weeks PMA).

Once study drug has been completed at 32–6/7 weeks, iron supplementation may need to be adjusted. If the ferritin level done at 42 days remains low, continue or increase supplements and recheck in 2 weeks. If the ferritin is in the normal range, consider decreasing supplements to 2–4 mg/kg/day as shown below. If the ferritin is above 400, hold the iron and recheck the ferritin in 2 weeks. Keep in mind that if the baby has an infection (sepsis, NEC, etc.), the ferritin may be falsely elevated.

• If the serum ferritin is <100 ng/mL, increase the supplementation by 1 mg/kg/day.

• If the serum ferritin is >100 ng/mL to 200 ng/mL, consider using 4 mg/kg/day.

• If the serum ferritin is >200 ng/mL to 400 ng/mL, consider using 2 mg/kg/day.

• If the serum ferritin is >400 ng/mL, hold the iron and check ferritin in 2 weeks.

• If ZnPP/H > 184, increase the supplementation by 1 mg/kg/day.

• If ZnPP/H values are in the normal range (ZnPP/H 30 to 183), consider using 2 to 4 mg/kg/day.

Measuring serum Ferritin

Principle:

The DxI Ferritin assay is a two-site immunoenzymatic (“sandwich”) assay. A sample is added to a reaction vessel with goat anti-ferritin-alkaline phosphatase conjugate, and paramagnetic particles coated with goat anti-mouse:mouse anti-ferritin complexes. Ferritin in the sample binds to the immobilized monoclonal anti-ferritin on the solid phase, while the goat anti-ferritin enzyme conjugate reacts with different antigenic sites on the ferritin molecules. Separation in a magnetic field and washing removes materials not bound to the solid phase. A chemiluminescent substrate, Lumi-Phos* 530, is added to the reaction vessel and light generated by the reaction is measured with a luminometer. The photon production is proportional to the amount of ferritin in the sample. The amount of analyte is determined by means of a stored, multi-point calibration curve.

Specimen:

Absolute minimum volume is 170 mcL of serum or plasma.

This includes the reaction vessel dead volume (60 mcL), sample pipettor overdraw (20 mcL), dead volume of 0.5 mL sample cup (80 mcL) and test volume (10 mcL). Blood may be collected in a plasma separator tube, a sodium heparin tube, an SST tube, but NOT EDTA, citrate or oxalate tubes. Specimens can be stored at room temperature (15–30 °C) for up to 8 hours and at 2–8 °C if testing will be done within 48 hours. If not, freeze specimens at −20 °C. Avoid multiple thaw-freeze cycles.

Measuring ZnPP/H

Principle:

The zinc protoporphyrin/heme ratio (ZnPP/H) is measured by hematofluorometry. The hematofluorometer detects the ratio of ZnPP fluorescence to hemoglobin absorption. A drop of whole blood is added to 2 drops of “ProtoFluor” Reagent, which stabilizes the hemoglobin in its oxygenated form. As a 420 nm light beam strikes a film of blood cells, hemoglobin will absorb a portion of this light and the remaining light excites zinc protoporphyrin (ZnPP). The excited ZnPP fluoresces at 594 nm, and the fluorescence is measured by front-surface fluorometry. The hematofluorometer uses these data to calculate the μmol ZnPP/mol heme ratio.

Specimen:

0.1 mL whole blood, anticoagulated with EDTA or heparin, in blood collection tubes or capillary tubes. Samples are stable for 1–2 months refrigerated.