A report on two interlaboratory surveys of serum ferritin estimation

Two surveys of the proficiency of performance of serum ferritin estimation in Australia have been conducted, one in 1978 and one in 1979. The majority of participants in both surveys used assay kits supplied by Clinical Assays, Roche Diagnostics and Ramco. The first survey showed a median intrabatch variation 9% and median interbatch variation of 17%. The overall assay variation had a median level of 22%. Assays by the various kits gave values which were not significantly different for each specimen (including purified liver and spleen ferritin solutions). The second survey showed an improvement in the overall precision of the participants with a median coefficient of variation of 15%. Contrary to the first survey, significant differences were evident between the results in 4 of the 5 specimens assayed. Some participants returned results for the high ferritin serum that differed substantially depending on the dilution of the sample used. The surveys showed a need for improvement in the assay performance of some participants, the need for ferritin levels to be assayed at least in two dilutions, and the need for standardization of assays, reference ranges and reporting units. Recommendation are made on all these points.