Effect of aggregates on albumin standardization.

In order to investigate the consequences of presence of aggregates in human albumin standards, pools of monomer, dimer and polymer albumin were prepared and quantitated by three total-protein methods (biuret, Folin-Lowry and spectrophotometry at 279 nm) and by four different albumin methods (dye-binding by bromcresol green, electroimmunoassay, radial immunodiffusion and automated immunoprecipitation). Biuret was chosen as the reference method and the monomer was used as the standard in all methods. Both the total-protein and albumin methods gave values for aggregated albumin different from the biuret values. The maximal bias occurred in radial immunodiffusion where quantitation of the dimer and polymer pools gave only 67% and 48% of the biuret values, respectively. In five commercial albumin preparations investigated, the content of di- and polymer varied from 3 to 34%. Uncritical use of albumin preparations in standardization may thus introduce bias in the measurements.