Assessment of iron deficiency in Kenyan children from capillary blood

Abstract

In absence of a non-invasive gold standard, iron deficiency anemia (IDA) stages are best characterized by multi-indices. However, choice of an appropriate single iron biomarker to replace multi-criteria model for ID screening in resource poor settings is of critical need. We compared ID defined as ¡Ý2 of 3 abnormal values in ferritin (SF), transferrin receptor (sTfR), or zinc protoporphyrin (ZP) concentrations (i.e. multi-criteria model) to ID defined by abnormal levels in any of iron biomarkers (SF alone, or sTfR alone, or ZP alone). This was a cross-sectional survey. Capillary blood was obtained from 680 children (aged 6C35 mo) for measurement of iron status using SF, sTfR, and ZP. Based on multi-criteria model, ID prevalence (¡ÀSE) was 62.1¡À1.9%, whilst those based on abnormal values in SF, sTfR, and ZP, were 40.0¡À1.9%, 57.7¡À1.9%, and 77.3¡À1.6%, respectively. The қ statistics for agreement between multi-criteria model and other iron indicators were 0.5 to 0.8; with sTfR having the best agreement (қ=0.84). Receiver-operating characteristic (ROC) curve analysis showed sTfR (area-under-the-curve (AUC) = 0.69) to be superior to both SF (AUC=0.61) and ZP (AUC=0.63) in predicting IDA (p<0.001). sTfR, compared to SF and ZP, accurately estimates the prevalence of ID in children based on multi-indices and least affected by inflammation. Funded by CDC and Nevin Scrimshaw International Nutrition Foundation. Grant Funding Source: NUTRITION (ASN)