Assessment of Potential Bias in a Reference Method for Progesterone in Human Serum

摘要

In clinical practice, the most common methods used for the quantification of steroids in biological fluids are immunoassays. Although immunoassays are simple to perform and inexpensive, they have the potential to produce falsely elevated results if the antibodies used cross-react with other steroids in the sample. To estimate the degree of possible bias associated with their test results, laboratories are encouraged to participate in proficiency testing schemes. In such schemes, typically the median of the results reported by all participating laboratories is used as the "target value" for assessment of accuracy. However, the ideal situation is where reference values characterised with low measurement uncertainties are provided by a reference laboratory using reference methodologies. To address the need for reference methods in clinical chemistry, the National Measurement Institute of Australia (NMIA) has developed a high-accuracy reference method for the quantification of progesterone in human serum. The method was based on the primary ratio method of isotope dilution mass spectrometry (IDMS). In this method, samples undergo solvent extraction with quantification by ultra-performance liquid chromatography tandem mass spectrometry. A secondary analytical method was also developed for confirmatory analyses of samples. The complementary method involves the immuno-affinity extraction of samples and determination by gas chromatography-mass spectrometry. The availability of two orthogonal analytical methods for sample analysis facilitated the rigorous investigation of bias in the reference method resulting in the very low measurement uncertainty of less than 3percent (at the 95percent confidence level). The validity of the reference method was successfully demonstrated in an international key comparison study organised by the Consultative Committee for the Amount of Substance (CCQM). The method was also used to provide reference values for a proficiency testing study coordinated by the Royal College of Pathologists Australia (RCPA).