Abstract

Reference values for two ferritin assays (Tina-quant^®a Ferritin, Enzymun^®, both Roche Diagnostics, Mannheim, Germany) were established (136 males and 139 females). To rule out inflammation as well as iron deficiency in the reference population, subjects with the C-reactive protein concentration < 5 mg/l, and zinc protoporphyrin < 40 µmol/mol heme and the soluble transferrin receptor < 3 mg/l were selected. Taking into account latent iron deficiency as well as hereditary hemochromatosis the 5–95 percentile range was as follows: male, 27–365 μg/l; female, 13–148 μg/l for Tina-quant^®a and 12–151 μg/l for Enzymun^®. The Tinaquant^® a Ferritin assay showed a very good correlation (r≧0.990) to Enzymun^® ferritin, Ferritin Abbott^® (Abbott Diagnostics, Delkenheim, Germany), N Latex Ferritin^® (Dade Behring, Marburg, Germany) and the Ferritin Chiron^® (Chiron Diagnostics, Fernwald, Germany). However, the slopes of the standard principal component method analysis were calculated to be between 1.03 (Enzymun^®) and 1.41 (N Latex Ferritin^®). For four assays the median recovery of the 3rd International Recombinant Ferritin Standard (NIBSC 94/572) measured by serial dilution was 89–109 %. The N Latex Ferritin^® assay recovered half of the target values. Because of the good correlation with other assays, a matrix effect is likely. The question arises whether the manufacturers' agreement on the recombinant