Spuriously low concentrations of serum iron measured with generation 14 Kodak Ektachem slides: prevalence, possible causes, and partial improvement with generation 16 slides.

For sera with iron (Fe) concentrations < 4 mumol/L, Kodak Ektachem slides Generation (GEN) 14 (without ascorbic acid) yielded systematically lower results for Fe than did liquid Ferrozine-based reagents from Baker containing ascorbic acid (10 g/L, final concentration) and adapted to Cobas-Bio. During an 8-month comparison period, outliers (defined as [Fe]Cobas - [Fe]Kodak > 4 mumol/L) were seen in 21 of the 8731 sera (0.24%) tested, corresponding to < 5% of the sera with [Fe]Kodak < 4 mumol/L. In vitro addition of ascorbic acid and (or) Fe identified at least two types of outliers: type 1 (approximately 70%), characterized by [Fe]Kodak > 0.4 mumol/L, by (supra)normal Fe recovery in Kodak slides in the presence or absence of ascorbic acid (10 g/L), and by between-method differences in serum Fe (Cobas - Kodak) that were significantly correlated with serum Zn content (P < 0.0004); and type 2 (approximately 30%), tentatively ascribed to contamination by EDTA, with serum Fe by Kodak < 0.4 mumol/L and Fe recovery near 0%, both of which could be significantly and dose-dependently increased by addition of ascorbic acid (5-20 g/L). For both types of outliers, flameless atomic absorption spectrometry (AAS) yielded results that were significantly higher than concentrations by Kodak with GEN 14. Use of GEN 16 slides (containing ascorbic acid) improved concordance of Kodak results with Cobas, and hence with flameless AAS, for both types of outliers; abolished Zn dependency of results; and increased Fe results in sera with type 2 outliers, although these remained substantially lower than by Cobas. However, like other ascorbic acid-containing reagents, GEN 16 slides were more sensitive to interference by dextran-bound Fe, as assessed during in vitro addition experiments and comparisons involving samples from Fe-dextran-treated patients. GEN 16 slides are hence expected to more frequently overestimate the physiologically available protein-bound Fe in hemodialysis patients. In hospital laboratories, this new interference will probably arise more frequently than the spuriously low results with GEN 14, hence warranting further efforts in optimizing Fe slides.