Iron status in early infancy is associated with trajectories of cognitive development up to pre-school age in rural Gambia

Abstract

Introduction: Iron deficiency is among the leading risk factors for poor cognitive development. However, interventions targeting iron deficiency have had mixed results on cognitive outcomes. This may be due to previous interventions focusing on the correction of iron deficiency anaemia in late infancy and early childhood, at which point long lasting neural impacts may already be established. We hypothesise that the relationship between iron status and cognitive development will be observable in the first months of life and will not be recovered by 5 years of age.

Methods: Using data from the Brain Imaging for Global Health (BRIGHT) Study in Gambia (n = 179), we conducted mixed effects modelling to assess the relationship between iron status at 5 months of age and trajectories of cognitive development from 5 months- 5 years using (i) a standardised measure of cognitive development (Mullen Scales of Early Learning) and (ii) an eye-tracking assessment of attention processing (visual disengagement time).

Results: All infants were iron sufficient at 1 month of age. At 5 and 12 months of age 30% and 55% of infants were iron deficient respectively. In fully adjusted analyses, infants in the lowest tercile of soluble transferrin receptor (sTfR) (best iron status) achieved MSEL Cognitive Scores on average 1.9 points higher than infants in the highest sTfR tercile (p = 0.009, effect size = 0.48). There was no evidence that this group difference was recovered by 5 years of age. Infants in the lowest sTfR tercile had visual disengagement time 57ms faster than the highest tercile (p = 0.001, effect size = 0.59). However, this difference diminished by early childhood (p = 0.024).

Conclusion: Infants are at risk of iron deficiency in early infancy. A relationship between iron status and cognitive development is apparent from 5 months of age and remains observable at 5 years of age. One mechanism by which iron availability in early infancy impacts brain development may be through effects on early attentional processing, which is rapidly developing and has substantial nutritional requirements during this period. To support neurocognitive development, prevention of iron deficiency in pre- and early postnatal life may be more effective than correcting iron deficiency once already established.

Fig 1. Flowchart of participant retention and data collection.

GA incompatible: gestational age of participant incompatible with study recruitment schedule. The study was set up so that a maximum of 15 expected deliveries were accepted each month, in order to ensure the study team could fulfil the requirements of the study. Data of withdrawn participants was retained in analysis with the exception of those diagnosed with a developmental delay. In these cases, all participant data was excluded from the dataset.

Fig 2. Incidence of iron deficiency onset.

N = 109 (Infants met the criteria for iron deficiency at 1, 5, 8 or 12 months) iron deficiency defined as ferritin <12μg/L if CRP < 5mg/L or ferritin <30μg/L if CRP≥ 5mg/L.

Fig 3. MSEL cognitive score trajectories by sTfR tercile at 5 months of age.

The Figure shows the modelled mean trajectories of MSEL Cognitive score from 5 months to 5 years of age, for each tercile of sTfR at 5months of age. The large oval shows a magnified version of the section of the graph highlighted in the small oval. Infants in the lowest tercile of sTfR scored on average 2 points higher than those on the highest tercile at 5 months of age. This group difference was still observable at 5 years of age. Model Specification; MSEL Cognitive score = Bl*Age + B2*Age³ + B3*Med*sTfR + B4*Lowest_sTfR + Constant Age interactions were not included in the equation as they did not significantly improve the fit of the model.

Fig 4. Visual disengagement time trajectories by sTfR tercile at 5months of age.

The Figure shows the modelled mean trajectories of Visual Disengagement Time from 5 months to 5 years of age, for each tercile of sTfR at 5months of age. Infants in the lowest tercile of sTfR responded on average 47ms faster than those in the highest tercile of sTfR at 5 months of age. However, this group difference diminished from around 20 months of age, with negligible group differences by 40 years of age. Model Specification; Visual Disengagement time = Bl*Age + B2*ln(Age) + B3*(ln(Age)² + B4*Med*sTfR + B5*Lowest_sTfR B6*Med_sTfR_Age + B7*Lowest sTfR_Age + Constant.