IgM in samples containing IgM MC and for IgA in samples containing IgA MC, compared with samples

MC, especially problems due to antigen excess (6), we compared the results of assays of non-MC-containing samples and of MC-containing samples separately, as well as the overall results (Fig. 1). The ratios (Beckman/ Behring) of the mean values for samples and the variances within results from a given assay and reference material have remained the same for all samples during the 5-year period evaluated here. The ratios of results for all three immunoglobulins are substantially closer to unity (1.0) with the standardization referenced to RPPHS than was the case previously. The persisting differences may relate to combinations of several factors, including the following: (a) Either or both manufacturers did not use the recommended method of value transfer (1); (b) the reactivities of the different antisera are different, i.e., the antiserum from one manufacturer recognizes different epitopes, or possibly complexes, from those recognized by the other; or (c) matrix differences exist between the samples and the manufacturers’ calibrants. That at least the second factor is likely is supported by the different ratios of concentrations obtained for IgM in samples containing IgM MC and for IgA in samples containing IgA MC, compared with samples with no MC. A similar difference was not seen in samples with IgG MC (Fig. 1). The changes in values secondary to changes in the manufacturers’ assigned values for the three immunoglobulins before and after the introduction of RPPHS are shown in Table 1. Each manufacturer’s change is expressed as the ratio of the value after conversion to RPPHS to the value before conversion, using only survey samples without MC. That the Beckman values changed relatively little is not surprising; previous CAP reference materials (RPSP1–RPSP4) were referenced against the US National Reference Preparation (2), as is RPPHS (1). In addition, values were assigned to RPSP4, the CAP reference material that immediately preceded RPPHS, using the same protocol that was used for RPPHS. In contrast, Behring previously referenced against in-house purified proteins. Although the observed differences in concentration between the manufacturers are small (2–5%), they are statistically significant (two-tailed P values by paired t-test were all ,0.0005). If “universal” reference intervals for plasma proteins are to be used (7), it is critical for manufacturers to assign values to calibrants and controls as accurately as possible, preferably using the recommended method (1).