Abstract
The analysis of serum albumin by bromocresol purple (BCP), by bromocresol green (BCG), and by electroimmunoassay (EIA) has been compared in a mixed group of 100 hospital patients. Good agreement between the immunochemical and BCP methods, but a positive bias with the BCG method led to routine use of the BCP method. In a group of 19 paediatric patients on chronic haemodialysis the mean BCP-albumin was found to be lower than the true albumin, as determined by EIA, whereas the positive bias using the BCG method remained. The difference between BCP-albumin and EIA-albumin in the haemodialysis group varied from 0 to −;13 g/L (mean −7·1 g/L) but was effectively constant for most patients over a 3-month period. Possible mechanisms and necessary further investigations are discussed.