Intervention Review

You have free access to this content

Iron therapy for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anaemia

  1. Bo Wang1,
  2. Siyan Zhan2,
  3. Ting Gong3,
  4. Liming Lee3,*

Editorial Group: Cochrane Developmental, Psychosocial and Learning Problems Group

Published Online: 6 JUN 2013

Assessed as up-to-date: 31 MAY 2013

DOI: 10.1002/14651858.CD001444.pub2


How to Cite

Wang B, Zhan S, Gong T, Lee L. Iron therapy for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anaemia. Cochrane Database of Systematic Reviews 2013, Issue 6. Art. No.: CD001444. DOI: 10.1002/14651858.CD001444.pub2.

Author Information

  1. 1

    Chinese Academy of Medical Sciences, Health Science Popularization Research Center, Beijing, Beijing, China

  2. 2

    School of Public Health, Peking University, Centre for Evidence Based Medicine and Clinical Research, Department of Epidemiology and Biostatistics, Beijing, China

  3. 3

    School of Public Health, Peking University, Department of Epidemiology and Biostatistics, Beijing, Beijing, China

*Liming Lee, Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 36 Xueyuan Road, Haidian District, Beijing, Beijing, 100191, China. lmlee@vip.163.com.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 6 JUN 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Aukett 1986

MethodsRCT. Allocation by investigators not connected with study. Investigators, parents and assessors blind to allocation


Participants110 children aged 17-19 months with IDA (Hb 8-11 g/dL). Community sample


InterventionsFerrous sulphate 24 mg and vitamin C 10 g or vitamin C 10 mg as identical appearing placebo. Compliance checked twice weekly


OutcomesDenver Developmental Screening Tests before and 8-9 weeks after commencement of treatment. Also measured weight change and measures of iron status


Notes-


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskComment: no sufficient information provided

Allocation concealment (selection bias)Low riskQuote (from report): "Allocation was double blind using consecutively numbered bottles"

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Incomplete outcome data (attrition bias)
All outcomes
Low risk6/54 missing from intervention group; 7/56 missing from control group. Characteristics of these 2 groups were similar to each other and to the group who reattended

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Driva 1985

MethodsRCT. No details of allocation process available. Nursery nurse and assessors blind to allocation


Participants40 infants with IDA (Hb < 11.0 g/dL and all suffering from iron deficiency) aged 3-25 months recruited from an orphanage in Greece


InterventionsIM injection of iron 50 mg in the intervention group


OutcomesBayley Scales of infant Development administered pre and 10 days post commencement of intervention


Notes-


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskComment: no sufficient information provided

Allocation concealment (selection bias)Unclear riskComment: no sufficient information provided

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskQuote (from the report): "Neither the tester nor the nursery nurse knew the type of therapeutic intervention"

Comment: Insufficient information on blinding of investigators

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from the report): "Neither the tester nor the nursery nurse knew the type of therapeutic intervention"

Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome measures are available for all children. Complete follow-up in both intervention group and control group

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Honig 1978

MethodsRCT. No details of allocation process available. Investigators, parents and assessors blind to allocation


Participants24 infants with IDA (Hb < 10.5 g/dL and at least 2 abnormal measures of iron
status), aged 9-26 months, recruited from paediatric clinic in New York


InterventionsIron dextran complex at a dose calculated to raise Hb to 12 g/dL. Placebo of IM sterile saline


OutcomesBayley Scales of infant Development and Bayley Infant Behaviour Record administered pre and 5-8 days post commencement of intervention. However, only some subscales of the Behaviour Record were reported


NotesAuthors provided unpublished details of allocation concealment


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskComment: no sufficient information provided

Allocation concealment (selection bias)Low riskComment: Authors provided unpublished details of adequate allocation concealment

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors blind to allocation"

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from the report): "Neither the examiner nor the nurse was informed as to which group any infant belonged until Bayley protocols had been scored"

Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome measures are available for all children. Complete follow-up in both intervention group and control group

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Idjradinata 1993

MethodsRCT. Allocation sequence derived from table of random
numbers, no details provided of who carried this out. Infants grouped by iron status class and randomised separately. Investigators, parents and assessors blind to allocation


ParticipantsChildren aged 12-18 months: 50 with IDA (Hb ≤ 10.5 g/dL, transferrin saturation < 10% and serum ferritin < 12 mg/dL) recruited from a paediatric clinic in Indonesia. 29 non-anaemic, iron deficient and 47 iron sufficient infants also randomised to treatment or control


InterventionsFerrous sulphate 3 mg/kg body weight by mouth for 4 months or placebo syrup of same appearance and flavour. Compliance checked weekly


OutcomesBayley Scales of Infant Development administered pre and 4 months post commencement of intervention. Measures of weight, length and Hb status


Notes-


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote (from report): "Random assignment of the infants to either the iron treatment or placebo intervention was done with a table of random numbers"

Allocation concealment (selection bias)Unclear riskComment: no sufficient information provided

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Incomplete outcome data (attrition bias)
All outcomes
Low risk1/25 missing from intervention group; 2/25 missing from control group. Missing outcome data balanced in numbers across groups

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Kimmons [pers comm]

MethodsAllocation sequence generated by random number table prepared by pharmacy and intervention performed by research nurse who was not further involved in study. Assessors blind to allocation


Participants42 children aged 6-24 months with Hb < 10.6 g/dL and MCV < 73 fL were randomised. Children recruited from general paediatric outpatients


InterventionsIron dextran complex calculated to raise Hb to 12 g/dL or dot of India ink at covered injection site


OutcomesBayley Scales of Infant Development and Bayley Infant Behaviour Record
administered pre and 1 week post commencement of intervention


Notes-


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote (from the report): "Allocation sequence generated by random number table prepared by pharmacy"

Allocation concealment (selection bias)Low riskQuote (from the report): "Allocation by nurse not connected with study"

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: no sufficient information provided

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from correspondence): "Assessors were blind to allocation"

Incomplete outcome data (attrition bias)
All outcomes
Low risk4/21 missing from either of both groups. Missing outcome data balanced in numbers across groups

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Lozoff 1982

MethodsRCT. Allocation by investigators not involved in other
aspects of the study. Investigators, parents and assessors blind to allocation


Participants28 children aged 6-24 months with IDA (Hb < 10.5 g/dL and at least 2 abnormal measures of iron status) and 40 children without IDA (Hb > 12 g/dL) randomised. Volunteers invited to have iron screening from a community in Guatemala


InterventionsFerrous ascorbate 5 mg/kg body weight orally twice daily for 7 days or identical
placebo prepared by pharmacy and administered by researcher blind to allocation


OutcomesBayley Scales of Infant Development administered pre and 6-8 days post
commencement of intervention


Notes-


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskComment: no sufficient information provided

Allocation concealment (selection bias)Low riskQuote (from the report): "Both anaemic and nonanaemic infants were randomly assigned to oral iron or placebo treatment groups by investigators who did not participate in screening , testing"

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Incomplete outcome data (attrition bias)
All outcomes
Low risk0/15 (MDI) or 3/15 (PDI) missing from intervention group; 0/13 (PDI) or 1/13 (MDI) missing from control group. Missing outcome data almost balanced in numbers across groups

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Lozoff 1987

MethodsRCT. Allocation by investigators not involved in other
aspects of the study. Investigators, parents and assessors blind to allocation


Participants52 children with IDA (Hb ≤ 10.5 g/dL and at least 2 abnormal measures of iron status)

45 children with Hb > 10.5 to < 12 g/dL and 2 abnormal measures of iron status

21 children with Hb > 12 g/dL and 2 abnormal measures of iron status

38 children with Hb > 12 g/dL and only serum ferritin < 12 μg/dL

35 children with Hb > 12 g/dL and normal measures of iron status

Those with Hb < 12 g/dL were randomised to IM iron followed by placebo, IM placebo followed by oral iron or IM placebo followed by oral placebo. The remaining children were randomised to IM placebo followed by either oral iron or oral placebo


InterventionsIM iron (Imferon) calculated to raise Hb to 12.5 g/dL or IM placebo then ferrous
sulphate 5 mg/kg or similar placebo given twice daily by investigators for 7 days


OutcomesBayley Scales of Infant Development administered pre and 7 days post
commencement of intervention


NotesOutcome assessments performed on all children at 1 week but data not easily interpretable (see analysis section). The authors report no significant differences between intervention and control groups on MDI or PDI scores


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskComment: no sufficient information provided

Allocation concealment (selection bias)Low riskQuote (from correspondence): "Allocation by investigators not involved in other aspects of the study"

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow-up in both intervention group and control group

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

Walter 1989

MethodsRCT. Method of allocation unclear. Children within each
iron-status group were randomly assigned to intervention and control. Investigators, parents and assessors blind to allocation


Participants39 anaemic (Hb < 11.0 g/dL) children aged 12 months allocated to iron (n = 24) or placebo (n = 15). Children recruited from within community-based RCT of iron supplementation


InterventionsFerrous sulphate 15 mg 3 times daily for 10 days or placebo


OutcomesBayley Scales of Infant Development administered pre and 11 days post commencement of intervention


Notes-


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskComment: no sufficient information provided

Allocation concealment (selection bias)Unclear riskComment: no sufficient information provided

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote (from correspondence): "Investigators, parents and assessors were blind to allocation"

Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome assessments available on all children. Complete follow-up in both intervention and control group

Selective reporting (reporting bias)Low riskAll expected outcomes were reported

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Deinard 1986No non-treated controls

Linos 1990No non-treated controls

Lozoff 1996No non-treated controls

Lozoff 2003Children were free of iron-deficient anaemia

Moffatt 1994Some children were free of iron-deficient anaemia

Oski 1983No non-treated controls

Stoltzfus 2001Some children were free of iron-deficient anaemia

Walter 1983No non-treated controls

 
Comparison 1. Iron treatment versus placebo in children with iron deficiency anaemia

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Tests of psychomotor development performed 5-30 days after study entry: change in Bayley Scale Psychomotor Development Index (PDI)5162Mean Difference (IV, Random, 95% CI)-1.25 [-4.56, 2.06]

 2 Tests of mental development performed 5-30 days after study entry: change in Bayley Scale Mental Development Index (MDI)5164Mean Difference (IV, Random, 95% CI)1.04 [-1.30, 3.39]

 3 Tests of psychomotor development performed 5-30 days after study entry: post treatment Bayley Scale PDI4123Mean Difference (IV, Random, 95% CI)-1.62 [-9.79, 6.55]

 4 Tests of mental development performed 5-30 days after study entry: post treatment Bayley Scale MDI4125Mean Difference (IV, Random, 95% CI)-0.01 [-9.17, 9.15]

 5 Tests of psychomotor development performed more than 30 days after study entry: change in Bayley Scale PDI147Mean Difference (IV, Random, 95% CI)18.4 [10.16, 26.64]

 6 Tests of mental development performed more than 30 days after study entry: change in Bayley Scale MDI147Mean Difference (IV, Random, 95% CI)18.8 [10.17, 27.43]

 7 Tests of psychomotor development and mental development performed more than 30 days after study entry: change in Denver Developmental Screening Test scores197Mean Difference (IV, Random, 95% CI)0.80 [-0.18, 1.78]

 8 Tests of psychomotor development performed more than 30 days after study entry: post treatment Bayley Scale PDI147Mean Difference (IV, Random, 95% CI)14.5 [8.82, 20.18]

 9 Tests of mental development performed more than 30 days after study entry: post treatment Bayley Scale MDI147Mean Difference (IV, Random, 95% CI)15.20 [8.96, 21.44]

 
Summary of findings for the main comparison. Iron therapy compared with placebo for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anaemia

Iron therapy compared with placebo for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anaemia

Patient or population: children less than 3 years of age with iron deficiency anaemia

Settings: communities or hospitals

Intervention: iron therapy

Comparison: placebo

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

PlaceboIron therapy

Bayley Scale PDI

within 30 days of commencement of therapy
The mean change in pre- to post-treatment test scores ranged across control groups from 0.2 to 11The mean change in pre- to post-treatment test scores in the intervention groups was 1.25 lower (95% CI -4.56 to 2.06)-162

(5 studies)
++OO
low1
-

Bayley Scale MDI

within 30 days of commencement of therapy
The mean change in pre- to post -treatment test scores ranged across control groups from 4.1 to 13.58The mean change in pre- to post-treatment test scores in the intervention groups was 1.04 higher (95% CI -1.30 to 3.39)-164

(5 studies)
++OO
low1
-

Bayley Scale PDI

more than 30 days of commencement of therapy
The mean change in pre- to post-treatment test scores in the control group was 5.1The mean change in pre- to post-treatment test scores in the intervention group was 18.4 higher (95% CI 10.16 to 26.64)-47

(1 study)
+++O2

moderate
-

Bayley Scale MDI

more than 30 days of commencement of therapy
The mean change in pre- to post-treatment test scores in the control group was 0.5The mean change in pre- to post-treatment test scores in the intervention group was 18.8 higher (95% CI 10.17 to 27.43)-47

(1 study)
+++O2

moderate
-

Denver Developmental Screening Test

more than 30 days of commencement of therapy
The mean change in pre- to post-treatment test scores in the control group was 3.2The mean change in pre- to post-treatment test scores in the intervention group was 0.8 higher (95% CI -0.18 to 1.78)-97

(1 study)
+++O2

moderate
-

CI: confidence interval; MDI: Mental Development Index; PDI: Psychomotor Development Index.

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 The funnel plot indicated small study effects. Number of children was less than 400.
2 Number of children was less than 400.
 
Table 1. Methods to be used in future updates, where applicable

Unit of analysis issues

 
Cluster-randomised trials

Effect estimates and their standard errors from correct analyses of cluster-randomized trials will be meta-analysed using the generic inverse-variance method in RevMan (RevMan 2011). For cluster-randomised trials analysed by incorrect statistical methods (not taking the clustering into account), we plan to adjust their sample sizes using the methods, described in the Cochrane Handbook for Systematic Reviews of Intervention (Higgins 2011), using an estimate of the intracluster correlation coefficient (ICC) derived from the trial or other sources. If we use ICC from other sources, sensitivity analyses will be conducted to investigate the effect of variation in the ICC.

Cross-over trials

For cross-over trials, we will extract and analyse data from the first period only.

Studies with multiple treatment groups

For factorial studies, we will include all comparisons that differ only in the presence or absence of iron. For other studies, multiple eligible intervention groups will be combined.

Assessment of reporting biases

 
Funnel plots can investigate the relationship between effect size and standard error. Asymmetry may be attributable to publication bias, but might also reflect true heterogeneity. If we find such a relationship between trial size and effect size, we will examine clinical variation (such as intensity of intervention, underlying risk) of the studies. In current version of this review, no such analysis was undertaken due to the small number of included studies.

Data synthesis

 
If a high level of heterogeneity is detected (I2 ≥ 75%), we will combine studies by narrative summary rather than in a meta-analysis, and explore the sources of heterogeneity by conducting predefined subgroup analyses.

Subgroup analysis and investigation of heterogeneity

 
If significant heterogeneity is identified within a meta-analysis and there is a sufficient number of included studies (> 10), 3 possible sources will be investigated separately:

  • baseline haemoglobin level, i.e. ≤ 90 g/L versus >90 g/L
  • intervention dose, i.e. ≤ 10 mg/day versus > 10 mg/day
  • duration of iron therapy, i.e. ≤ 10 days versus >10 days

 
Table 2. Methods used to impute standard deviations for changes values

StudyGroupOutcome measureP valueImputationImputed SDsImplied correlationPlausibility of imputation

Driva 19851InterventionPDI> 0.05Impossible10.940.69Plausible

MDI< 0.01Possible10.940.78Plausible

ControlPDI> 0.05Impossible10.940.82Plausible

MDI> 0.05Impossible10.940.79Plausible

Honig 19782InterventionPDI0.10Possible21.22-0.26Plausible

MDI0.01Possible15.14-0.27Plausible

ControlPDI> 0.05Impossible21.220.20Plausible

MDI> 0.05Impossible15.140.57Plausible

Kimmons [pers comm]3InterventionPDI0.36Possible9.730.79Plausible

MDI0.27Possible3.900.97Plausible

ControlPDI0.36Possible9.730.79Plausible

MDI0.27Possible3.900.97Plausible

Lozoff 19824InterventionPDI-Impossible7.40.92Plausible

MDI-possible7.40.93Plausible

ControlPDI-Impossible7.40.94Plausible

MDI-possible7.40.998Plausible

Lozoff 19875InterventionPDI> 0.80Impossible---

MDI> 0.69Impossible---

ControlPDI> 0.80Impossible---

MDI> 0.69Impossible---

Idjradinata 19936InterventionPDI< 0.001Possible17.91-0.59Plausible

MDI< 0.001Possible25.09-1.84Implausible

ControlPDI< 0.001Possible17.91-0.51Plausible

MDI-Impossible25.09-1.70Implausible

 MDI: Mental Development Index; PDI: Psychomotor Development Index; SD: standard deviation.
1 P values resulted from paired t-tests (for within group changes). The SD for the change MDI in the intervention group (10.94) was used for the SDs of the change in PDI in both groups and the change MDI in the control group in the analyses, which were impossible to be imputed.
2 P values resulted from paired t-tests (for within group changes). The SD for the change PDI in the intervention group (21.22) and for the change in MDI (15.14) were used for the SDs of the change in PDI and MDI in the control group, respectively, which were impossible to be imputed.
3 P values derived from a repeated measures analysis of variance were treated as if resulting from unpaired t-tests (for the between group differences in changes).
4 Here the SDs for the change in MDI were imputed from the reported power of the study to detect differences of specified sizes in MDI given the observed SDs of the changes (although not giving the values of these SDs), rather than from P values. The SDs for the change MDI were used for the SDs of the change in PDI in the analyses, which were impossible to be imputed.
5 P values derived from a repeated measures analysis of variance were treated as if resulting from unpaired t-tests (for the between group differences in changes).
6 P values were derived from a repeated measures analysis of variance. For PDI, P values were treated as if resulting from unpaired t-tests (for the between group differences in changes). For MDI, P values were treated as if resulting from paired t-tests (for within group changes). The SD for the change MDI in the intervention group (25.09) was used for the SDs of the change in MDI in the control group, which was impossible to be imputed. However, the estimated SDs for change in MDI would imply correlations between pre- and post-test scores substantially in excess of -1, which is impossible. It was, therefore, decided instead to estimate SDs from these reported scores and standard errors, following methods described by Follmann (Follmann 1992), assuming a correlation of 0 between pre- and post-treatment test scores. This provided an estimate of the SD for the change in MDI of 15.26 for the control group and 14.90 for the intervention group and of the SD for the change in PDI of 14.21 for the intervention group and 14.58 for the control group.