Comparison of four clinically validated testosterone LC-MS/MS assays: Harmonization is an attainable goal.

INTRODUCTION: Immunoassays and liquid chromatography-tandem mass spectrometry assays are commonly employed in clinical laboratories for measurement of total testosterone in serum. Results obtained from either of these methodologies compare poorly due to differences in calibration and/or inadvertent detection of interfering substances by the immunoassays. Standardization efforts are underway, but recent studies indicate that accuracy remains an issue. METHODS: This study compares the results from four independently developed and validated LC-MS/MS assays for total testosterone. The calibration for each assay was verified using National Institute of Standards and Technology Standard Reference Material 971. RESULTS: Initially, one of the four assays had a mean percent difference of +11.44%, compared to the All Method Mean, but following re-verification of all five non-zero calibrator concentrations with the NIST SRM 971, the mean percent difference decreased to -4.88%. Subsequently, the agreement between all four assays showed a mean bias of <5% across the range of all testosterone concentrations (0.13-38.10 nmol/L; 3.7-1098 ng/dL), including at low concentrations of <1 nmol/L (<29 ng/dL). CONCLUSIONS: Excellent agreement between four independently developed LC-MS/MS assays demonstrates that harmonization using standard reference material is attainable. However, as we found in this study, to ensure accurate calibration it is critical to validate the concentrations of new lots of calibrators.