Iron supplemented formula milk related to reduction in psychomotor decline in infants from inner city areas: randomised studyCommentary: Iron deficiency and developmental deficit–the jury is still out

BMJ 1999; 318 doi: http://dx.doi.org/10.1136/bmj.318.7185.693 (Published 13 March 1999)
Cite this as: BMJ 1999;318:693

Abstract

Objective: To compare the effect of unmodified cows‘ milk and iron supplemented formula milk on psychomotor development in infants from inner city areas when used as the main milk source.

Design: Double blind, randomised intervention trial.

Setting: Birmingham health centre.

Subjects: 100 infants, mean age 7.8 months (range 5.7 to 8.6 months), whose mothers had already elected to use unmodified cows‘ milk as their infant's milk source.

Intervention: Changing to an iron supplemented formula milk from enrolment to 18 months of age, or continuing with unmodified cows‘ milk.

Main outcome measures: Developmental assessments using Griffiths scales at enrolment and at 18 and 24 months.

Results: 85 participants completed the trial. There were no significant differences in haemoglobin concentration between the two groups at enrolment, but by 18 months of age 33% of the unmodified cows‘ milk group, but only 2% of the iron supplemented group, were anaemic (P<0.001). The experimental groups had Griffiths general quotient scores that were not significantly different at enrolment, but the scores in both groups declined during the study. By 24 months the decrease in the mean scores in the unmodified cows‘ milk group was 14.7 whereas the decrease in the mean scores in the iron supplemented group was 9.3 (P<0.02, 95% confidence interval 0.4 to 10.4). Mean subquotient scores were considerably lower in the unmodified cows‘ milk group at 24 months; significantly so for personal and social scores (P<0.02, –5.4 to 17.2).

Conclusion: Replacing unmodified cows‘ milk with an iron supplemented formula milk up to 18 months of age in infants from inner city areas prevents iron deficiency anaemia and reduces the decline in psychomotor development seen in such infants from the second half of the first year.

Iron supplemented formula milk related to reduction in psychomotor decline in infants from inner city areas: randomised study

  1. J Williams, consultant community paediatriciana,
  2. A Wolff, consultant community paediatriciana,
  3. A Daly, senior paediatric dietitiana,
  4. A MacDonald, head of paediatric dietetic servicesa,
  5. A Aukett, consultant community paediatricianb,
  6. I W Booth (i.w.booth{at}bham.ac.uk), professorc
  1. a Birmingham Children's Hospital NHS Trust, Birmingham B4 6NH
  2. b Northern Birmingham Community Trust, Carnegie Institute, Handsworth, Birmingham
  3. c University of Birmingham, Institute of Child Health, Birmingham B4 6NH
  4. Systematic Reviews Training Unit, Institute of Child Health, University College London Medical School, London WC1N 1EH
  1. Correspondence to: Professor Booth
  • Accepted 10 December 1998

Abstract

Objective: To compare the effect of unmodified cows‘ milk and iron supplemented formula milk on psychomotor development in infants from inner city areas when used as the main milk source.

Design: Double blind, randomised intervention trial.

Setting: Birmingham health centre.

Subjects: 100 infants, mean age 7.8 months (range 5.7 to 8.6 months), whose mothers had already elected to use unmodified cows‘ milk as their infant's milk source.

Intervention: Changing to an iron supplemented formula milk from enrolment to 18 months of age, or continuing with unmodified cows‘ milk.

Main outcome measures: Developmental assessments using Griffiths scales at enrolment and at 18 and 24 months.

Results: 85 participants completed the trial. There were no significant differences in haemoglobin concentration between the two groups at enrolment, but by 18 months of age 33% of the unmodified cows‘ milk group, but only 2% of the iron supplemented group, were anaemic (P<0.001). The experimental groups had Griffiths general quotient scores that were not significantly different at enrolment, but the scores in both groups declined during the study. By 24 months the decrease in the mean scores in the unmodified cows‘ milk group was 14.7 whereas the decrease in the mean scores in the iron supplemented group was 9.3 (P<0.02, 95% confidence interval 0.4 to 10.4). Mean subquotient scores were considerably lower in the unmodified cows‘ milk group at 24 months; significantly so for personal and social scores (P<0.02, –5.4 to 17.2).

Conclusion: Replacing unmodified cows‘ milk with an iron supplemented formula milk up to 18 months of age in infants from inner city areas prevents iron deficiency anaemia and reduces the decline in psychomotor development seen in such infants from the second half of the first year.

Key messages

  • Iron deficiency anaemia is common in infants from inner cities who are given unmodified cows‘ milk in the first year of life

  • Giving an infant iron supplemented formula milk instead of cows‘ milk not only prevents anaemia but reduces the decline in developmental performance observed in those given only cows‘ milk

  • An iron supplemented formula milk rather than cows‘ milk should be provided free of charge for infants up to the age of 18 months who are living in inner cities and who are not receiving breast milk

Introduction

Iron deficiency anaemia–that is, a haemoglobin concentration <110.0 g/l–still occurs in 10 to 30% of preschool children living in inner cities in the United Kingdom. 1 2 There is a well established association between iron deficiency anaemia and developmental delay, and randomised studies providing oral iron supplements suggest that this may be causal.38 We have previously shown that iron deficiency anaemia in infants and toddlers receiving unmodified cows‘ milk as their main milk source is eliminated by changing to an iron supplemented formula milk between 6 and 18 months of age.9 Our study aimed to address an additional and pragmatic question: does randomisation to receive an iron supplemented formula milk between 6 and 18 months of age lead to an additional developmental advantage compared with continuing receipt of unmodified cows‘ milk? Detailed haematological and nutritional data from the study have already been published.9 We now present the developmental outcomes.

Subjects and methods

Recruitment

Our keyworker (AD) received the names of all infants aged 6–8 months (567 identified) living in an inner city area of Birmingham from health visitors dealing with that area. AD visited the families, and the parents of only those infants whose mothers had already changed their infant's diet to unmodified cows‘ milk (n=116) were asked to consider including their infant in the study. All mothers were given both verbal and written explanations of the study.

The mean age of infants at recruitment (47 boys and 53 girls) was 7.8 months (range 5.7 to 8.6 months). The population was 75% Caucasian, 24% AfroCaribbean, and 2% Asian (Indians). We excluded all preterm infants. Figure 1 shows the withdrawals and losses of participants from the study.

Fig 1.

Withdrawals and losses from study

The participants lived in a socially deprived area with poor housing, high unemployment, and poor public amenities–locally there was only one bank and no large supermarkets. The small local shops were expensive and had limited stocks of food, particularly fresh fruit and vegetables.

Power calculation– We performed a power calculation, which showed that if 47 participants were allocated to each dietary group this would provide a study power of 95% at a significance level of 5% for a difference in haemoglobin concentration of 7.5 g/l between groups.

Study design

After recruitment we randomised the infants in the pharmacy department at Birmingham Children's Hospital by random numbers in blocks of four to receive either an iron supplemented formula milk or to continue on unmodified cows‘ milk. We gave the results of randomisation to AD who was therefore unblinded. At 18 months, those infants randomised to change to an iron supplemented formula milk were transferred back to cows‘ milk, and both groups continued on the cows‘ milk until 24 months of age. Serial haematological, anthropometric, and developmental assessments using the Griffiths scales were made at enrolment and at 18 and 24 months of age.10 We excluded those participants whose haemoglobin concentration decreased to <90 g/l and referred them to their general practitioner.

We supplied the iron supplemented formula milk free of charge, and we gave those mothers whose infants remained on the cows‘ milk a monthly payment equivalent to the cost of 500 ml cows‘ milk daily. Mothers from both groups on income support were still entitled to claim free cows‘ milk with milk tokens. However, as not all parents were in receipt of income support, and therefore not entitled to the cows‘ milk, the cows‘ milk group received funding to purchase 500 ml cows‘ milk per day. Table 1 lists the nutrient content of the cows‘ milk and iron supplemented formula milk.

View this table:
Table 1.

Nutrient composition of unmodified cows‘ milk and iron supplemented formula (0.28 MJ of energy per 100 ml each)

Developmental assessments

The Griffiths scale calculates an overall developmental score (general quotient), which is the mean of five subscales: locomotor, personal and social, hearing and speech, eye and hand coordination, and performance (manipulation and precision).10

Five trained and experienced observers performed the Griffiths scales. The observers were blinded to the group randomisations.

Statistical analyses– We performed statistical analyses with either paired and unpaired Student's t tests, >>2 tests, or Fisher's exact tests, and analysis of variance.

Ethical approval– We obtained ethical approval from the South Birmingham Health Authority's ethics committee. We obtained informed written consent from caregivers.

Results

Withdrawals and losses

Some data points were missing due to intercurrent illness in a participant, transiently being unable to locate children, or insufficient volume of blood for assay. Out of 269 contacts, a developmental score was unavailable on 11 occasions (3%).

During the course of the study some children were found to have conditions that led their caregivers to withdraw them from the study. This was either because recommended treatment (for example, diet for an infant with renal disease) interfered with the study or because the diagnosis made continuing participation impractical (fig 1).

Sociodemographic characteristics of the study groups– After randomisation we found no significant difference between the two groups (table 2).

View this table:
Table 2.

Characteristics of two study groups

Haematology

At enrolment there were no statistically significant differences in mean haemoglobin concentration between the two groups; 16% of the cows‘ milk group and 13% of the iron supplemented formula milk group were already anaemic.

By 18 months, 2% of the iron supplemented formula milk group and 33% of the cows‘ milk group were anaemic (P<0.0001). At 24 months, when the iron supplemented formula milk group had been returned to cows‘ milk for 6 months, 26% of the cows‘ milk group were anaemic but none of the iron supplemented formula milk group were anaemic (P=0.0017). Similar changes occurred in mean corpuscular volume and ferritin concentration: there was no difference at enrolment, but at 18 and 24 months there were significantly lower values in the cows‘ milk group.9

Developmental outcomes

Interobserver differences– We found no statistically significant differences when the developmental scores for the different observers were compared by analysis of variance for both iron supplemented formula milk and cows‘ milk groups.

Griffiths general quotient scores– At enrolment there were no significant differences between the two groups. Scores in each group declined during the study (table 3, fig 2). By 18 months the mean general quotient score had decreased by 8.3 (P=0.002) and 6.7 (P=0.02) points in the cows‘ milk and iron supplemented formula milk groups respectively.

Fig 2.

Changes in Griffiths general quotient score between enrolment and 24 months of age in infants receiving unmodified cows‘ milk or iron supplemented formula milk (P=0.02)

View this table:
Table 3.

Analysis of developmental scores for unmodified cows‘ milk and iron supplemented formula milk groups from enrolment to 18 months (46 and 46 infants respectively) and from enrolment to 24 months (44 and 42 infants respectively)

By 24 months there had been a further decrease of 6.4 points in the mean score of the cows‘ milk group (P<0.001), whereas a decrease of only 2.6 points had occurred in the iron supplemented formula milk group. The decrease in general quotient score between enrolment and 24 months was significantly greater in the cows‘ milk group than in the iron supplemented formula milk group (14.7 versus 9.3 respectively, P<0.02; 95% confidence interval 0.4 to 10.4) (fig 2).

Subquotient scores

At enrolment there were no significant differences in mean subquotient scores between the two study groups, but they declined in both groups throughout the study (table 3, fig 3).

The decrease in subquotient scores from enrolment was consistently greater in the cows‘ milk group than in the iron supplemented formula milk group, both at 18 and 24 months, and in all subscales. However, only the decrease in personal and social skills between enrolment and 24 months was significantly greater in the cows‘ milk group than the iron supplemented formula milk group (P=0.02).

Developmental scores and haematological status– There was no significant linear correlation between haemoglobin concentration and general quotient scores at 24 months. However, those participants allocated cows‘ milk were significantly clustered towards both a lower haemoglobin concentration and lower general quotient score than those receiving iron supplemented formula milk, who were clustered significantly towards both a higher haemoglobin concentration and general quotient score. Thus, of the 24 participants with both a haemoglobin concentration <120 g/l and general quotient score <100 at 24 months, 20 had received cows‘ milk. In contrast, 13 of the 16 with a haemoglobin concentration >120 g/l and a general quotient score >100 had received iron supplemented formula milk (P<0.0001).

Fig 3.

Changes in Griffiths subscales between enrolment and 24 months of age

Growth and nutrient intakes– Both groups grew satisfactorily on both the iron supplemented formula milk and the cows‘ milk.9

Discussion

Our study shows that in a population of socioeconomically deprived infants, changing from unmodified cows‘ milk to an iron supplemented formula milk from 7 to 18 months of age prevented iron deficiency anaemia at 24 months, and significantly reduced the decline in psychomotor performance seen in those infants randomised to continue on cows‘ milk.

Nutritional basis of the observed effects– In contrast to other randomised studies, we chose to look at a realistic and practical dietary intervention. 6 7

We have previously shown that this intervention prevents the development of anaemia, but the precise nutritional basis of the developmental advantage in the group receiving iron supplemented formula milk is uncertain. The intakes of the two groups also differed substantially in nutrients other than iron.9 However, the strength of the recognised association between iron deficiency anaemia and developmental delay, and the scale of the difference in iron status between the two groups, lead us to suggest that it is the disparity in iron status between the two groups that is the most plausible explanation for the observed difference in developmental performance.

Comparison with previous studies– Our findings support previous studies of supplementation with oral iron in children with iron deficiency anaemia, where an improvement in developmental performance was noted. 6 7 12

Moffatt and colleagues conducted a similar longitudinal cohort study to our own and showed a developmental advantage at 9 and 12 months, which was no longer detectable at 15 months of age.8 The transient nature of the effect may have been due to differences in the timing, duration, and mildness of the iron deficiency compared with our study group.

Basis of the developmental advantage– The Griffiths scale has been well validated, and the subscales provide useful insights into the basis of the differences in developmental scores. 10 13 In our study the major difference was in the personal and social subscale. This supports the view that iron deficiency anaemia may exert its effects on developmental performance by alterations in affect, thereby making a child clingy, lethargic, irritable, and listless,5 and leading to impaired learning skills.

Implications of the study– We acknowledge that it is difficult to quantify precisely the developmental advantage in the infants receiving iron supplemented formula milk, but neverthless believe that this study has a number of important implications. Firstly, it confirms the well recognised observation that socioeconomic deprivation places infants at increased risk of adverse developmental outcomes. 14 15

Secondly, this developmental deficit seems, in part, to be nutritionally mediated. Thirdly, iron deficiency anaemia is common in high risk populations,14 and both this and the developmental disadvantage are susceptible to a simple intervention: the provision of an iron supplemented formula milk in place of cows‘ milk.

Breast milk is clearly the milk of choice for the developing infant.16 Our study suggests that in those mothers who find breast feeding impractical, iron supplemented formula milk seems to be effective and acceptable, and benefits high risk infants and children up to the age of at least 18 months.15

Acknowledgments

We thank the health visitors at Nechells Health Centre for their help, Dr P Davies for his statistical advice, and Dr M Huntley for advice on the Griffiths scales.

Footnotes

  • Contributors IWB, AMacD, and AA had the original idea for the study, initiated it, and supervised its conduct. AD carried out the dietary intervention and data collection. TW took part in the developmental assessment and the preliminary data analyses. JW took part in the developmental assessment, collated and analysed the data and, with IWB, participated in writing the paper. IWB and JW will act as guarantors for the paper.

  • Funding Farley Health Products.

  • Conflict of interest None.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.

Commentary: Iron deficiency and developmental deficit–the jury is still out

  1. Stuart Logan, director
  1. a Birmingham Children's Hospital NHS Trust, Birmingham B4 6NH
  2. b Northern Birmingham Community Trust, Carnegie Institute, Handsworth, Birmingham
  3. c University of Birmingham, Institute of Child Health, Birmingham B4 6NH
  4. Systematic Reviews Training Unit, Institute of Child Health, University College London Medical School, London WC1N 1EH

    It has been estimated that around 10% of young children in the developed world and perhaps 50% in poor countries are iron deficient, with the burden falling disproportionately on less privileged children.1 In their randomised controlled trial, Williams and colleagues report a 5.4 point smaller decline between 6–8 and 24 months in Griffiths developmental quotient in infants given iron supplemented feeds than in those given unmodified cows‘ milk (95% confidence interval 0.4 to 10.4). If, as this finding suggests, iron deficiency in infancy causes developmental deficit its prevention should be a public health priority.

    This trial was small but generally well conducted: central randomisation ensured appropriate allocation concealment and assessment of outcome was blind to treatment group. It is unfortunate that a small number of children were excluded from analysis for what seem to be inappropriate reasons, including “failed protocol,” “autism,” thalassaemia, and anaemia. In small trials, exclusions after randomisation may be important in the interpretation of results, particularly where confidence intervals barely exclude 0. Interpretation is further complicated by the finding of virtually no difference between groups at 18 months.

    The evidence from earlier studies is conflicting. A Canadian randomised controlled trial of iron supplementation in infants suggested a beneficial effect of supplementation at 9 and 12 months but not at 6 or 15 months and was weakened by substantial losses to follow up.2 A randomised controlled trial of iron supplementation between 6 and 12 months in 944 previously breastfed infants found no differences between groups in Bayley scale scores at 12 months although iron deficiency anaemia was common in the unsupplemented group.3 Two trials of iron treatment in toddlers with iron deficiency anaemia have reported on developmental findings 2–4 months after starting treatment. In a placebo controlled trial including 50 infants with iron deficiency anaemia, Idjradinata and Pollitt reported a significantly greater improvement in Bayley scale scores in the iron treated group.4 However, Aukett and colleagues in a similar study found no significant differences between treated and control groups in changes in Denver developmental screening test scores although, as a result of inappropriate dichotomisation of continuous data, the study is frequently quoted as suggesting a positive effect of treatment.5

    Although this trial tilts the balance of probability towards belief in a causal link between iron deficiency and developmental deficit, the evidence remains unclear. A call for more research may be interpreted as an attempt to avoid decision making, but, particularly where interventions are costly or may have adverse consequences, clarity of evidence is essential. Prevention of iron deficiency is difficult. In spite of the high bioavailablity of iron in breast milk, both breastfed and bottle fed infants are at risk of developing iron deficiency unless sufficient iron is provided by the weaning diet. Dietary advice may not be effective in preventing iron deficiency anaemia,6 and the administration of elemental iron may have side effects.7 Large trials of both iron supplementation in infants and iron treatment in children with iron deficiency anaemia are urgently needed.

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.

    THIS WEEK'S POLL