Comparison of fresh frozen serum to traditional proficiency testing material in a College of American Pathologists survey for ferritin, folate, and vitamin B12.

CONTEXT: Comparison of different analytical methods in proficiency surveys may be affected by the artificial nature of the survey material. OBJECTIVE: To compare intermethod differences in proficiency survey results between 2 types of survey material, conventional proficiency testing material (PTM) and fresh frozen human serum (FFS), for 3 markers of anemia: ferritin, folate, and vitamin B12. DESIGN: Data were gathered from a 2003 survey event in the College of American Pathologists Ligand ("K") Series, in which the specimens to be tested by each participating laboratory included 1 vial of FFS and 2 vials of PTM with different analyte concentrations. The more than 1600 laboratories subscribing to the survey were not advised as to the nature of the specimens. MAIN OUTCOME MEASURES: The bias of each method relative to the median of method means for each analyte and each type of survey material, and the interlaboratory coefficient of variation for each method. RESULTS: For each of the 3 analytes, moderate to large method biases were observed. For ferritin, method biases correlated strongly between comparable PTM and FFS specimens (Spearman r = 0.863, P < .001), whereas virtually no correlation was found for folate (r = -0.224, P = .48), and a marginally significant correlation existed for B12 (r = 0.55, P = .049). CONCLUSIONS: With ferritin, proficiency survey performance of PTM is similar to that of FFS, implying that method biases relate mainly to calibration. With folate and to a lesser extent with B12, PTM and FFS exhibit different method biases, implying that the biases reflect analyte heterogeneity and/or matrix effects.