The purpose of measurement standardization is to achieve closer comparability of results obtained using different commercial systems. Regarding serum protein immunoassays, a reference preparation (BCR-470) was released in 1993 and adopted by manufacturers across the world to value-assign their assay calibrators for routine methods to reduce method-dependent variation. Moving from nephelometric (Beckman Immage 800) to turbidimetric determination (Roche Cobas c 501) of seven serum proteins, we preliminarily checked the comparability of results between the two systems. The study was performed according to the CLSI EP9-A protocol on 30 fresh sera, tested on each system in duplicate, and subdivided on two different days, without recalibration and using manufacturers' control materials to validate the runs. Both manufacturers' package inserts provide statements that kit calibrators are traceable to BCR-470. Suggested reference intervals are also the same. Although a fairly good correlation was observed (r = 0.955), the comparison of ceruloplasmin methods produced evidence of highly significant proportional (regression slope, 0.572) and constant bias (intercept, 0.05 g/L). Absolute and percentage mean differences were -0.11 g/L (95% confidence interval (CI) -0.13 to -0.10 g/L) and -39.1% (CI -43.1 to -35.2%), respectively. No other evaluated proteins showed similar problems. Lacking a ceruloplasmin reference method, it is impossible to demonstrate that one of the two assays produces true ceruloplasmin values. The problem is, however, that results coming from the two assays are clearly not comparable. This may be either due to a lack of commutability of the reference material with biological samples in the evaluated assays or to calibration problems by manufacturers in one of the stages of the calibration hierarchy.