Iron in Infancy and Long-term Development

In this issue of the Archives, Lozoff et al¹ report the findings of a long-term follow-up of children at 10 years of age who were participants in a double-masked, randomized controlled trial conducted from 1991 through 1994 in Santiago, Chile, testing iron-fortified (12.7 mg/L) vs low-iron (2.3 mg/L) infant formula use from 6 to 12 months of age.² All infants enrolled in the original trial were screened to have no iron deficiency anemia at the outset. Among 57% of the original sample of children who were reached, an intent-to-treat analysis showed that the mean scores on tests of spatial memory and visual-motor integration at 10 years of age were lower in the iron-fortified group compared with the control.¹ The mean effect sizes for both these outcomes were small (−0.21), whereas effect sizes for other measures, such as overall IQ, arithmetic, visual perception, and motor coordination, showing a similar negative trend were even smaller, ranging from −0.08 to −0.16, and not significant. Stratified analyses revealed large negative effect sizes (−0.85 to −1.36) of the iron formula among children who had high hemoglobin (Hb) levels (Hb level >12.7-13.0 g/dL, 5.0%-5.5% of the sample) at the start of the trial, whereas those with low Hb level (<10.4-10.7 g/dL, 9%-24% of the sample) benefitted somewhat with the intervention. For a large proportion of children with Hb concentrations in the middle of the distribution (10.8-12.7 g/dL, 70%-90% of the sample) the iron intervention did not have any impact on the measured outcomes. (To convert Hb to grams per liter, multiply by 10.0.)

The importance of this study lies in its evaluation of the long-term developmental outcomes of an early-infancy iron intervention. Few such studies exist in the literature. To interpret the negative results of the study, it is important to briefly examine the existing knowledge on the topic of iron deficiency in infancy and cognition, compare its findings with other study results not considered, examine some of its limitations, and finally briefly consider the broader implications of the study findings.

The relationship between iron deficiency and iron deficiency anemia and developmental outcomes in children has been controversial, and clear causal evidence for a link found lacking when existing trials have been carefully evaluated.³ Lozoff et al¹ have made significant contributions to establishing this link and demonstrated in longitudinal observational studies that iron deficiency in early life can result in cognitive and motor impairments that may be irreversible.^{4 - 5} They have previously also documented longer-term associations in 5-year-olds and 10-year-olds showing poorer developmental outcomes in children who were iron deficient and anemic in the first year of life.^{6 - 7} Although longitudinal, these studies have not allowed for making causal inferences. When examined with respect to response to treatment, studies in infants and young children have yielded mixed results. In short-term studies of treatment with iron in deficient infants, 2 studies showed an improvement in a subset of scales,^{8 - 9} while 2 others found no improvement in any measured scale.^{10 - 11}

With regard to prevention, there are few randomized supplementation trials of iron among infants examining short-term cognitive and motor outcomes. Some have shown positive effects on varying aspects of infants' behavior and function including exploration, psychomotor development, and visual perceptual skills.^{8 ,12 - 13} Among 4 trials that used iron-fortified formula, 2 found reductions in psychomotor decline^{14 - 15} whereas 1 found no effect on cognitive performance.¹⁶ The fourth study is by Lozoff et al¹⁷ of an early follow-up of developmental outcomes at 12 months of age in this same study cohort, which showed positive behavioral differences among those who had received the iron-fortified vs the low-iron formula. In 2 recent supplementation trials, age at first unassisted walking with daily supplementation with iron (12.5 mg) and folic acid was examined.^{18 - 19} In one of the studies done in Pemba, Zanzibar, supplementation from 5 to 11 months of age decreased age at first unassisted walking by a mean of 15 days (P < .04).¹⁸ In contrast, in Nepal, iron-folic acid was associated with a mean delay of 28 days (95% CI, 11.3-44.7) in age at walking unassisted; the delay was even more pronounced (mean of 60 days later) in the smaller infants (mid–upper arm circumference <9.5 cm).¹⁹ Thus, short-term effects of iron fortification or supplementation interventions in infancy seem to be inconsistent.

Limitations of the study, as also highlighted by Lozoff et al,¹ including (1) a high loss to follow-up, (2) use of a nonspecific measure of Hb with no information regarding true iron status of infants, and (3) confounding factors such as exposure to maternal smoking and female sex in the high Hb level group, should indicate caution in the interpretation of the study results. The latter 2 concerns make it hard to discern who the “high Hb level” group of children represent. While the authors dismiss the higher proportion of females in the high Hb stratum as not a potential explanatory factor, in Nepal we have shown development test scores of school-aged girls to be lower than those of boys,²⁰ with differences in type of schooling, sex-related cultural differences in child rearing and parenting practices, and other factors perhaps explaining these sex differences in this environment. The high loss to follow-up can result in an unrepresentative sample, and examining within this attenuated sample treatment differences in the lower and upper tail of the distributions of Hb can yield spurious findings.

In addition, findings of a recent follow-up of a trial cohort that are relevant need to be considered. This study was a follow-up among Thai children at 9 years of age who had participated in a placebo-controlled trial of daily iron (10 mg), zinc (10 mg), and iron plus zinc supplementation starting at age 4 to 6 months for 6 months.²¹ The follow-up study revealed no differences in full-scale, verbal, and performance IQ on the WISC-III (Wechsler Intelligence Scale of Children, Third Edition), school performance scores, and Raven's colored progressive matrices by supplement group relative to the placebo. Of note, there was no evidence of any adverse impact with iron supplementation with or without added zinc. The prevalence of anemia (Hb level, <10.0 g/dL) in infancy was 28.8% but iron deficiency (serum ferritin level, <12 ng/mL [26.96 pmol/L]) was virtually nonexistent, ranging from 0% to 3.4% across groups, indicating that the anemia was unlikely due to iron deficiency. It seems that the 2 studies may have comparable contexts, tested about the same amount of iron in infancy, and followed the children at a similar age, and both did not have demonstrable iron deficiency in the infant population when the intervention was tested. Yet disparate findings of a potential adverse impact in Chile vs no impact in Thailand were observed. It is not clear at present how to reconcile these 2 sets of findings.

It is well recognized that iron stores needed in infancy are accumulated in utero, and breastfed infants of deficient mothers and those born preterm or with low birth weight are likely to have depleted iron stores by 3 to 4 months of age. Recently our group has shown significant improvements in aspects of intelligence, executive function, including working memory and inhibitory control, and fine motor function with antenatal iron-folic acid supplementation relative to a control in prospective study of children at ages 7 to 9 years whose mothers participated in an antenatal micronutrient supplementation trial in rural Nepal.²⁰ Our findings reveal that in utero iron is critical for the development of the central nervous system, with implications for strategies such as antenatal iron supplementation for impacting long-term developmental outcomes among children in iron-deprived settings.

Caution is needed in generalizing the results of the follow-up study by Lozoff et al,¹ which stands, as yet, alone in showing small-sized negative consequences on developmental outcomes among iron-sufficient children exposed to iron-fortified vs low-iron formula during infancy. The study from Thailand, also in a setting where iron deficiency was not prevalent, showing no effect of iron supplementation on cognitive function, when considered simultaneously would suggest that iron deficiency may be a prerequisite for finding a benefit. Whether iron deficiency in infancy, manifest largely due to deficiency in utero, can be overcome with supplementation during infancy for improving central nervous system development and function needs to be further examined in rigorous studies of short and long duration.