Precision and accuracy of serum ferritin measurements.

For the epidemiological evaluation of iron stores the precision of serum ferritin measurement in the low range are important for the accurate determination of the prevalence of iron deficiency and the detection of subtle changes in serum ferritin levels after food iron manipulation. In addition the recognition of iron overload and the evaluation of its severity are important. In this study samples with low serum ferritin values were repetitively measured in a classic radioimmunoassay (RIA) and in two "2-site" immunoradiometric assays (IRMA), one of which used a polystyrene bead and the other a polystyrene tube fas the solid phase. Variability was significantly less with the IRMA using a bead than that using a tube. Optimum precision was noted when samples were run at the lowest possible dilution (1:10) with relatively long reaction times. The bead IRMA was also more precise that the RIA which had a standard curve with 50% radioactive binding (maximum precision) of 40 ng/ml. Within the normal serum ferritin range (12 to 300 ng/ml) extremely similar results were obtained with an RIA and IRMA. However, when iron overload samples (serum ferritin values greater than 2000 ng/ml) were examined the RIA gave values significantly lower than those obtained by IRMA. The lower values by RIA may be related to the immunological heterogeneity of serum ferritin which is maximized by an assay performed in antigen excess (RIA) and minimized by one performed in antibody excess (IRMA). These observations indicate a need for the development of specific serum ferritin assays for epidemiological studies. By manipulating the components of the standard curve the RIA and IRMA can be optimized to provide maximum precision when low serum ferritin values are being measured. in choosing an assay the ability of the method to determine the severity of iron overload must also be taken into consideration.