BACKGROUND: The reliability of bilirubin measurements is not satisfactory due to the marked variability of methodologies since discovery. METHODS: Imprecision of the three analytical systems (diazo method, vanadate oxidase method, and dry chemistry method) were calculated in accordance with CLSI guidelines EP15-A2. According to Clinical and Laboratory Standards Institute (CLSI) guideline (EP9-A2), 40 analytes were detected to evaluate method performance and comparability among three routine measurements for serum bilirubin in a China laboratory. RESULTS: The within-run precisions and the intermediate imprecision of the three methods all met the CLIA'88 requirement of bilirubin measurement. All three evaluated methods were closely correlated with correlation coefficients ranging from 0.995 to 0.999. The 95% predicted bias confidence intervals between diazo method and vanadate oxidase method for total bilirubin were acceptable. However, for direct bilirubin, the 95% predicted bias confidence intervals was only acceptable at the medium medical decision point. The 95% predicted bias confidence intervals between diazo method and dry medical method for total bilirubin was all acceptable. CONCLUSIONS: All three evaluated routine methods for serum bilirubin in our lab were closely correlated. For measurements of total and direct bilirubin, both dry medical method and vanadate oxidase method had significant differences compared with the diazo method. Further studies are required to reduce the biases among these methods.
BACKGROUND: The reliability of bilirubin measurements is not satisfactory due to the marked variability of methodologies since discovery. METHODS: Imprecision of the three analytical systems (diazo method, vanadate oxidase method, and dry chemistry method) were calculated in accordance with CLSI guidelines EP15-A2. According to Clinical and Laboratory Standards Institute (CLSI) guideline (EP9-A2), 40 analytes were detected to evaluate method performance and comparability among three routine measurements for serum bilirubin in a China laboratory. RESULTS: The within-run precisions and the intermediate imprecision of the three methods all met the CLIA'88 requirement of bilirubin measurement. All three evaluated methods were closely correlated with correlation coefficients ranging from 0.995 to 0.999. The 95% predicted bias confidence intervals between diazo method and vanadate oxidase method for total bilirubin were acceptable. However, for direct bilirubin, the 95% predicted bias confidence intervals was only acceptable at the medium medical decision point. The 95% predicted bias confidence intervals between diazo method and dry medical method for total bilirubin was all acceptable. CONCLUSIONS: All three evaluated routine methods for serum bilirubin in our lab were closely correlated. For measurements of total and direct bilirubin, both dry medical method and vanadate oxidase method had significant differences compared with the diazo method. Further studies are required to reduce the biases among these methods.