Endocrine Abstracts

Introduction: 11-ketotestosterone (11KT) and 11β-hydroxyandrostenedione (11OHA4) are new biomarkers for hyperandrogenic disorders. Steroids can be measured in saliva, allowing non-invasive sampling by patients. We modified a published LC-MS/MS method¹ for quantification of 11-oxygenated androgens in saliva with respect to sample volume, extraction procedure and equipment, and assessed the potential impact of different sampling devices on results.

Method: Calibrators were prepared from primary standards (Merck KGaA, Steraloids Inc.) in artificial saliva. Samples were extracted by supported-liquid-extraction with dichloromethane without online solid-phase-extraction. 11-oxygenated androgens were quantified using a 1290 Infinity II HPLC (Agilent Technologies; injection volume 20 µL) coupled to a 6500+ QTRAP mass spectrometer (AB Sciex) via gradient elution and multi-reaction monitoring in positive ESI mode. After method validation, 30 samples from volunteers and patients with adrenal diseases covering a broad concentration range were measured in parallel at our laboratory and the laboratory which originally described the method. To test a potential impact of the sampling device, spiked artificial, and individual or pooled native saliva was applied to three sampling tools (Salivette® and Salivette® Cortisol (Sarstedt), SalivaBio Infant’s Swab (Salimetrics)), or used directly. Sample volumes of 20-300 µL were compared.

Results: Performance characteristics were largely comparable to the original method, although our limits of quantification were higher (11KT: 30 vs. 5 pmol/l; 11OHA4: 60 vs. 50 mol/l). Measured concentrations correlated well (intercept 3.18/10.19, slope 1.14/1.54 for 11KT/11OHA4, respectively). Bias between the methods was 17% for 11KT, but 59% for 11OHA4. At very high cortisol and low 11OHA4 concentrations we observed a weak interference from cortisol on 11OHA4 that was not described for the original method. Results from measurement by our method did not differ for saliva volumes between 50 and 300 µL. Measured concentrations were significantly affected by the choice of the sampling tool. Samples collected by the SalivaBio Infant’s Swab exhibited the least deviation (-/+6%) from sampling without tool.

Conclusion: We transferred and modified an established method for quantification of 11-oxygenated androgens to allow application if no online SPE is available. While results for 11KT agreed, a significant bias in results for 11OHA4 was observed, perhaps relating to differences in calibrators. To avoid interference from extreme cortisol concentrations on 11OHA4 signals we recommend sampling before hydrocortisone intake. Required volume can be reduced to 50  µl, facilitating analysis of samples from newborns. The impact of sampling devices on measured concentrations must be considered for future studies.

Reference: 1 Schiffer L et al. Ann Clin Biochem. 2019;56(5):564-573.