Background: Though mass spectrometry (MS) assays are increasingly used for routine clinical measurements of serum total testosterone (TT), information about the variability of results is limited. This study assessed the variability of TT measurement results from routine MS assays.
Methods: Twenty serum samples (12 females, 8 males) were analyzed on 2 days by seven high performance liquid chromatography (HPLC), and one gas chromatography (GC)-tandem mass spectrometry (HPLC-MS/MS, GC-MS/MS) assays. Two samples (male and female) were provided in five replicates to assess the within-run variability. Results were compared against those obtained at National Institute of Standards and Technology (NIST). The within- and between-laboratory variability was assessed for each sample. Comparisons to the NIST results were performed using bias plot and Deming regression analysis.
Results: The overall coefficient of variation of the results obtained with MS assays was <15%CV at >1.53 nmol/L and <34%CV at 0.3 nmol/L. The between-assay variability was the major contributor to the overall variability. The assay precision was the highest (<3%CV) with assays using liquid-liquid extraction for sample preparation or GC-MS/MS. The mean percent difference to the reference assay was 11%. The slopes of Deming regression analysis of the MS assays were between 0.903 and 1.138 (correlation coefficient: >0.996). TT concentrations for one assay were above the measurement range.
Conclusions: The variability of TT measurement results among MS assays is substantially smaller than that reported for immunoassays. The type of sample preparation may affect assay precision. Standardizing assays can further reduce the variability of measurement results.