A critical analysis of iron status indicators in three independent studies of South African primary school children

Abstract

Background The potential dire consequences of iron deficiency (ID) and iron deficiency anaemia (IDA) on childhood development are of major public health concern. Many factors contribute to anaemia, ID being only one progressive factor. The prevalence of ID and IDA must be accurately determined before iron intervention strategies can be safely prescribed. There is continued uncertainty regarding the optimal approach to identifying and measuring ID, as indicators have different roles, explore different aspects of iron metabolism and cannot be directly compared. Furthermore, inflammation and infection have a confounding effect on the commonly applied indicator and acute phase reactant, serum ferritin (SF). In the public health setting, a suitable method to assess iron status in developing countries has to be inexpensive, standardised, established, easy to measure and its applications specific to identifying ID. Aim We conducted secondary analysis of screening data from three independent iron intervention studies to critically evaluate the indicators used to determine iron status in 6-11-year-old primary school children from three South African provinces. Study design and methods A cross-sectional descriptive analysis was performed on the screening data collected in 2009 and 2010 during iron intervention studies in KwaZulu-Natal (n=736), Northern Cape (n= 1045), and North West (n=546). The three distinct study sites were analysed independently and collectively. Children’s haemoglobin (Hb), SF, transferrin receptor (TfR), zinc protoporphyrin (ZPP), and C-reactive protein (CRP) concentrations were measured and body iron calculated. ID prevalence was compared using different methods (namely the single indicators SF, TfR and ZPP, body iron and the multiple criteria model), and the influence of inflammation on SF was considered. Literature suggests that the multiple criteria model provides a more complete assessment of iron status. The performance of single and body iron indicators were compared to the multiple criteria model (by assessing sensitivity, specificity and predictive values). Results Significant positive correlations between CRP (indicator of inflammation) and SF existed in all study sites and the combined sample (p < 0.01). The mean SF concentration was substantially higher in subjects with inflammation than those without. A different SF cut-off to identify ID was applied to subjects with inflammation. The percentage of ID subjects varied using different indicators (4.2 – 26.5% in KwaZulu-Natal; 4.1 – 13.4% in Northern Cape; 7.0 – 24.4% in North West; and 5.4 – 15.2% in the combined sample). The sensitivity, specificity and predictive values of alternate ID indicators varied within and between study sites, compared to the multiple criteria model. Conclusion Simply using Hb as an ID indicator is inaccurate. The vast differences between percentages identified as ID by different indicators is reason for concern. No consistent agreement appeared between single ID indicators, body iron and the multiple criteria model for ID identification after correcting for inflammation in primary school children. The global view of the multiple criteria model as the gold standard for estimating ID is debatable and potentially impractical at a public health level. Current evidence cautions against overestimating the prevalence of ID, as there is more associated harm than deficiency underestimation. This critical analysis has confirmed a need for research to identify a suitable, accurate and precise alternative to Hb as a tool in the South African public health setting. Furthermore, the impact of inflammation on iron status indicators, in particular SF, should be assessed in context to clearly set parameters for its use in nationally-representative nutrition surveys, the cornerstone of iron intervention strategies.