RESUMEN
BACKGROUND: Usually serum albumin is measured with dye-binding assay as bromocresol green (BCG) and bromocresol purple (BCP) methods. The aim of this paper was to examine the differences in albumin measurements between the Advia2400 BCG method (AlbBCG), Dimension RxL BCP (AlbBCP) and capillary zone electrophoresis (CZE). METHODS: Albumin concentrations from 165 serum samples were analysed using AlbBCG, AlbBCP and CZE. CZE was employed to estimate different serum protein fractions. Influence of globulins on albumin concentration discrepancies between methods was estimated as well as the impact of the albumin method on aCa concentrations. Medcalc was employed for statistical analysis, setting a value of P < 0.05 as significant. RESULTS: Correlation of AlbBCG and AlbBCP was r = 0.948 (p < 0.0001), but mean difference was large. Bland-Altman plots showed greater bias at lower albumin concentrations. AlbBCG were positively biased versus CZE (3.54 g/L). There was good agreement between CZE and ALbBCP (< 1 g/L). The AlbBCG assay bias shows a good correlation with alpha-1-globulin concentrations (r = 0.758); moderate and weak correlations were observed with CRP (r = 0.729) and alpha-2-globulin (r = 0.585); we found no correlation with beta-globulin (r = 0.120) or gamma-globulin (r = -0.303). Mean aCa based on AlbBCG and AlbBCP methods were 2.34 ± 0.15 mmol/L and 2.46 ± 0.16 mmol/L (p < 0.01), with a mean BCG-BCP difference of -0.12. CONCLUSION: Albumin results from the BCP and BCG methods may result in unacceptable differences and clinical confusion, especially at lower albumin concentrations. Serum acute phase proteins contribute to overestimating the albumin concentration using AlbBCG.