Endocrine Abstracts (2011) 25 P285

Measurement of salivary testosterone in female samples using a highly sensitive LC-MS/MS assay

Brian Keevil1, Philip McDonald1, Wendy McDowell2, Alan Wallace3 & Fred Wu4


1University Hospital of South Manchester, Manchester, UK; 2London School of Hygiene and Tropical Medicine, London, UK; 3Glasgow Royal Infirmary, Glasgow, UK; 4Manchester Royal Infirmary, Manchester, UK.


Introduction: We have developed a highly sensitive LC–MS assay in an attempt to improve the measurement of salivary testosterone in female samples.

Methods: A 200 μl saliva sample, calibrators or QC were mixed with 10 μl working internal standard (0.1 μg/l) and 1 ml of methyl tert-butyl ether (MTBE). Vortex mixed for 4 min and frozen at −80 °C (1 h). Unfrozen organic layer was transferred to a glass tube and evaporated. The residue was reconstituted with 100 μl of 50:50 mobile, vortex mixed and transferred to a 96-well microtitre plate.

Liquid chromatography was performed using a Waters Acquity UPLC system. Extract (30 μl) was injected directly from the microtitre plate onto a C18 Acquity 1.8 μm HSS T3 analytical column (2.1×50 mm). A solvent gradient was used to elute testosterone and D5Testosterone from the column. TMS was performed on a Waters Quattro Xevo TQ-S in positive ionization mode. Multiple reaction monitoring transitions for testosterone were m/z 289.2>108.8 (quantifier) and m/z 289.2>96.8 (qualifier) and for D5-testosterone internal standard was m/z 294.1>99.8.

Results: Lower limit of quantitation was 2 pmol/l, recovery was 98–109%. Interbatch precision (CV) was 4.4, 6.0 and 2.8% at 17, 100 and 154 pmol/l respectively.

The median female range (n=96) was 10–15 pmol/l and the median male range (n=96) was 100–150 pmol/l. Although cross validation of calibrators was performed to eliminate any calibration issues/bias we still found the LC–MS method gave 25% lower values than the RIA method in males and 400% lower in females

Conclusions: Using a highly sensitive mass spectrometer we have developed an assay for salivary testosterone with greatly improved detection limits and confirmed that RIA was unsuitable for measuring salivary testosterone especially in female samples. We established reference ranges in saliva samples and found good demarcation between the male and female reference ranges.

Article tools

My recent searches

No recent searches.