A comparison of eleven commercial kits for the determination of serum ferritin levels.

Eleven commercial kits for serum ferritin have been compared. Seven kits used RIA and four IRMA methodology. Seven kits used human liver ferritin, two human spleen ferritin and two undefined human ferritin as standard material. Incubation times varied between 30 minutes and 36 hours. Only four kits gave normal ranges determined with the kit in question. All values were compared with the kit in routine use. Although correlation of results was good in 10 cases, (r = 0.903-0.993) the numerical values varied greatly, (slope of the regression line b = 0.437-1.94). These values were obtained from at least 30 sera for each kit. Seven kits gave significantly different results from the routine kit (p less than 0.05 -- Wilcoxon matched pairs test -- two tail value). Inter and Intra-assay coefficients of variation were under 12.5% in all cases within the range of interest. Four kits showed a pronounced non-linear correlation with the routine kit, and one of the IRMA kits exhibited a high-dose hook effect within the standard curve. The other three IRMA kits showed no such effect up to 2000 micrograms . liter-1 (under assay conditions).