Comparison of eight routine unpublished LC-MS/MS methods for the simultaneous measurement of testosterone and androstenedione in serum

Liquid-chromatography tandem mass spectrometry (LC-MS/MS) has become the method of choice in steroid hormone measurement. However, little information on the mutual agreement of LC-MS/MS methods is available. We compared eight routine unpublished LC-MS/MS methods for the simultaneous measurement of testosterone and androstenedione. Sixty random serum samples from male and female volunteers were analysed in duplicate by eight routine LC-MS/MS methods. We performed Passing-Bablok regression analyses and calculated Pearson's correlation coefficients to assess the agreement of the methods investigated with one published method known to be accurate. Intra-assay CV of each method was calculated from duplicate results, recoveries for each method were calculated from six spiked samples. Furthermore, a CV between the investigated methods was calculated. The concentrations ranged from 0.05-1.26 nmol/L, 6.15-24.44 nmol/L and 0.15-4.78 nmol/L for testosterone in females, testosterone in males and androstenedione, respectively. The intra-assay CVs were between 3.7-16.0%, 0.9-5.2% and 1.2-9.5% for testosterone in females, testosterone in males and androstenedione, respectively. The slopes of the regression lines ranged between 0.90-1.25, 0.87-1.24 and 0.94-1.31 for testosterone concentrations in females, all testosterone values and androstenedione, respectively. Inter-method CVs were 24%, 14% and 29% for testosterone for concentrations in females and males and androstenedione, respectively. These compare unfavourably to the variation found earlier in published methods. Although most routine LC-MS/MS methods investigated here showed a reasonable agreement, some of the assays showed a high variation. The observed differences in standardization should be taken into account when applying reference values, or should, preferably, be solved