Three commercial immunoradiometric "kit" assays for serum ferritin evaluated.

We evaluated three commercially available immunoradiometric assays for serum ferritin, with particular emphasis on the statistical validity of the results. The data show that it was unusual (a) for dose/response relationships of the standards to be linear over the whole concentration range suggested by the manufacturers and (b) for dose/response relationships of test sera to parallel those of the standards. These findings cast doubt on the ability of any of the assays reliably to discriminate small differences in ferritin concentrations. Nevertheless, all three methods gave reproducible results, but this reflects technical expertise rather than accuracy of the results per se. The method based on liver ferritin as tracer detected less of the iron-binding protein than those based on spleen ferritin, despite significant cross reactivity of the two antibodies.