Article Text

Utility of red cell distribution width in screening for iron deficiency
  1. Roderick F Hinchliffe1,
  2. Graham J Bellamy1,
  3. Adam Finn2,
  4. Frank Bell3,
  5. Ajay J Vora1,
  6. Lynne Lennard4
  1. 1Department of Paediatric Haematology, Sheffield Children's NHS Foundation Trust, Sheffield, UK
  2. 2School of Clinical Sciences, Faculty of Medicine and Dentistry, University of Bristol, Bristol, UK
  3. 3Division of Infectious Disease, University of Washington, Seattle Children's Hospital, Seattle, Washington, USA
  4. 4Clinical Pharmacology Unit, Department of Human Metabolism, University of Sheffield Medical School, Sheffield, UK
  1. Correspondence to Dr Lynne Lennard, Clinical Pharmacology Unit, Department of Human Metabolism, Floor E, University of Sheffield Medical School, Beech Hill Road, Sheffield S10 2RX, UK; l.lennard{at}


Aims To assess the sensitivity of an adult-derived red cell distribution width (RDW) reference limit in the detection of iron deficiency in young children.

Methods Haematological analysis performed on a cohort of 13-month-old healthy term infants of North European ancestry.

Results 21/98 infants were iron-deficient (>2.5% hypochromic red cells). Of the remaining 77, 35 with RDW >13.9% also had evidence of incipient iron deficiency on the basis of significantly lower haemoglobin (11.5 vs 11.8 g/dl, p=0.046), mean cell volume (75.6 vs 77.8 fl, p=0.002) and mean cell haemoglobin (25.4 vs 26.2 pg, p=0.002) values and higher zinc protoporphyrin (55 vs 44 μmol/molhaem, p<0.001) values than those of the 42 with RDW ≤13.9%.

Conclusions An adult-derived RDW reference limit has utility in screening for iron deficiency at the age of 13 months. The incidence of non-anaemic iron deficiency in this group was 52.8%.

Statistics from

What is already known on this topic

  • The red cell distribution width, an automated index of red cell anisocytosis, is a sensitive marker of the early stages of iron deficiency.

  • Previous studies have suggested that reference limits in young children are higher than in adults.

What this study adds

  • This study provides evidence that the use of an adult-derived red cell distribution width reference limit is appropriate at 13 months of age.

  • The use of an adult-derived value indicated an incidence of non-anaemic iron deficiency of 52.8% in North European infants at 13 months of age.


The incidence of iron deficiency remains high in infants and young children, especially among those from socially and economically deprived groups and those in developing countries. There is a strong association between iron deficiency anaemia and impaired psychomotor development in young children and some evidence to suggest that non-anaemic iron deficiency (NAID) may also irreversibly affect behaviour and neurodevelopment.1 These concerns have prompted the assessment of newer measures of iron status arising from the increasing sophistication of automated red cell analysis, such as the percentage of hypochromic red cells (%hypo) and reticulocyte haemoglobin content, in the search for improved detection of NAID. An increase in the red cell distribution width (RDW), a long-established index of variability in red cell size (anisocytosis), is also an early indicator of iron deficiency.2 Previous studies have indicated higher RDW values in healthy young children than in adults.3 In this report, we question these conclusions and present our findings on the sensitivity of RDW to NAID, using an adult-derived upper reference limit, in a cohort of healthy term infants of North European ancestry at 13 months of age.

Subjects and methods

Parents of healthy infants were approached on the postnatal wards of the two obstetric units serving the city of Sheffield for enrolment in a trial of a combination acellular pertussis/diphtheria/tetanus/Hib (DTaP/Hib, Pasteur Merieux) vaccine. Families expressing interest in the study and who spoke English and were registered with a Sheffield GP were seen at home to provide further information and obtain informed consent. Infants born less than 36 weeks gestation or weighing less than 2.5 kg at birth were excluded from the study. Venous blood was taken at 2, 5 and 13 months of age to assess response to the vaccine. With parental and local research ethics committee approval an additional sample was taken at each time point to determine the blood count (Bayer H1, Siemens, Frimley, UK) and red cell zinc protoporphyrin concentration (ZPP; Helena Laboratories, Gateshead, UK). The results of the vaccine study, which took place during 1995–1996, together with an analysis of the leucocyte counts of these infants, have been previously published.4–6 We have recently analysed the red cell data from this study with the aim of producing age-related reference values. As part of this analysis, the 13-month data, presented here, were critically assessed in order to exclude those with iron deficiency anaemia and NAID. Statistical analysis of results was by unpaired t test or, where variables were not normally distributed, by the Mann–Whitney test.


A total of 123 infants of North European ancestry were studied and the 13-month samples, collected at 395–425 (median 400) days, were used in this study. Because of illness at the time of sampling, analysis failure or withdrawal from the study, results from 98 children were available at this age. In all, 21 (21.4%) had >2.5% hypo and were considered to be iron-deficient.7 This group contained the 9 (9.2%) anaemic children (Hb <10.5 g/dl) in the cohort. There were statistically significant differences between the values of most red cell variables of these children and the 77 with ≤2.5% hypo (table 1A,B). The data from these 77 were then divided into two further groups based on an RDW cut-off of 13.9%, this being the reported upper limit in adults using Siemens instruments.8 Haemoglobin, mean cell volume and mean cell haemoglobin values of the 35 with RDW values >13.9% were significantly lower and ZPP values significantly higher than those of the 42 with RDW values >13.9% (table 1C,D), indicating the existence of incipient iron deficiency in this group. All 21 children with iron deficiency indicated by >2.5% hypo had RDW values >13.9%. In conclusion, of the 98 children studied nine were anaemic, but of the remaining 89 non-anaemic children, 47 (52.8%) had laboratory evidence of NAID (12 with %hypo >2.5 and RDW >13.9% and 35 with RDW >13.9%).

Table 1

A summary of analyses performed on data from 98 children at 13 months of age


Iron deficiency is rare before 4–6 months of age in term infants but the incidence increases rapidly thereafter before falling between 3 and 4 years. This coincides with the age range at which higher RDW values have been reported3 and suggests the control groups in these studies contained subjects with NAID. Prompted by this possibility, we further analysed the 77 children with ≤2.5% hypo using an upper reference limit for RDW of 13.9%, drawn from a study of healthy adults using the identical (Siemens) technology.8 This resulted in a further 35 children being found to have laboratory evidence of NAID over and above those with more advanced haematological changes detected by the use of a %hypo value of >2.5, supporting the contention that the use of higher RDW reference values in young children is misleading and masks the true incidence of NAID at this time. The 21 children found to be iron-deficient based on raised %hypo values all had RDW values >13.9% and would therefore have been detected had RDW been used alone to search for deficiency.

This study has shown that, using an adult-derived upper reference value, RDW is a highly sensitive index of NAID at 13 months of age, indicating a higher incidence than expected in a cohort of healthy term infants of North European ancestry. A problem with RDW measurements however is their lack of specificity, as raised values are found in most forms of anaemia. However, in young children, almost all cases showing mild red cell changes, with or without anaemia, are secondary to iron deficiency or infection, with thalassaemia traits an important cause in some ethnic groups. Reticulocyte haemoglobin content, %hypo and RDW values are all affected by thalassaemia trait while infectious illness, secondary to its effects on iron metabolism, is also likely to affect these measures. Thus, among the common red cell disorders of early childhood, RDW is probably no less specific than %hypo and reticulocyte haemoglobin content, while having the important advantages of being available on all current automated blood count analysers at no extra cost. Reference limits for RDW vary between analysers depending upon the technology used for red cell analysis, and wider application of our findings would require the use of the appropriate instrument-specific reference limit. We conclude that RDW has a role to play in the detection of NAID, although in routine clinical practice it would be prudent to confirm the diagnosis before introducing iron therapy.


View Abstract
  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors RFH performed laboratory analysis, data analysis, data interpretation and co-wrote the manuscript. GJB performed laboratory analysis and collated results. AF designed and ran the clinical study and edited the manuscript. FB clinically evaluated each child prior to blood sampling, collected blood samples for the study and reviewed the manuscript. AJV critically revised the paper and edited the manuscript. LL performed data analysis and interpretation and co-wrote the manuscript. All authors approved the final version of the paper.

  • Funding This work was supported by a grant from The Parents’ Association for Children with Tumours and Leukaemia, The Children's Hospital, Sheffield. LL was supported by Leukaemia and Lymphoma Research, London (grant numbers 08048, 09027 and 10051).

  • Competing interests None.

  • Ethics approval South Sheffield Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The haematological parameters from this study represent unpublished data. We aim to publish all these data. Any data not included in the short report submitted to the Archives of Disease in Childhood we will publish at a later date as reference ranges.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.