Background A dry-slide bromocresol green method for serum albumin
consistently demonstrated negative bias in a proficiency testing scheme, particularly
at low albumin concentrations. Methods and Results Albumin was measured
using both a dry-slide method and an immunonephelometric method in patient serum
samples and external quality assessment samples. Results obtained using the dry-slide
method were lower than those obtained using immunonephelometry (mean bias - 1·0 g/L,
P = 0·0039), but this difference was exaggerated in the quality
assessment samples. In three other albumin preparations with assigned concentrations,
recovery with the dry-slide method decreased upon dilution of the matrix.
Conclusions Our results suggest that apparent under-recovery of
albumin using this technology is probably due to the diffusive properties of the
quality assessment sample matrix.
