Objectives: To assess the effects of sample dilution, peroxidase concentration, and chloride ion (Cl(-)) on plasma unbound bilirubin (B(f)) measurements made using a commercial peroxidase methodology (UB Analyzer) in a study population of ill, premature newborns.
Design and methods: B(f) was measured with a UB Analyzer in 74 samples at the standard 42-fold sample dilution and compared with B(f) measured at a 2-fold sample dilution using a FloPro Analyzer. B(f) was measured at two peroxidase concentrations to determine whether the peroxidase steady state B(f) (B(fss)) measurements were significantly less than the equilibrium B(f) (B(feq)), in which case it was necessary to calculate B(feq) from the two B(fss) measurements. B(f) was also measured before and after adding 100 mmol/L Cl(-) to the UB Analyzer assay buffer.
Results: B(feq) at the 42-fold dilution was nearly 10-fold less than but it correlated significantly with B(feq) at the 2-fold dilution (mean 8.2+/-5.2 nmol/L versus 73.5+/-70 nmol/L, respectively, p<0.0001; correlation r=0.6). The two UB Analyzer B(fss) measurements were significantly less than B(feq) in 42 of 74 (57%) samples, and Cl(-) increased B(feq) in 66 of 74 (89%) samples by a mean of 82+/-67%.
Conclusions: B(fss) measured by the UB Analyzer at the standard 42-fold sample dilution using assay buffer without Cl(-) and a single peroxidase concentration is significantly less than the B(feq) in undiluted plasma. Accurate B(f) measurements can be made only in minimally diluted serum or plasma.