Endocrine Abstracts

Background: While investigating hyperandrogenism in women, local laboratory policy dictates adult females with a serum testosterone of >3.5 nmol/l should be confirmed with an alternative method to exclude interference by cross-reacting substances. Patients who were on testosterone replacement or had a previous result in the same year were excluded. We explore the biochemical investigation of discrepant testosterone in females.

Method: Retrospective audit over 1 year of testosterone requesting commencing April 2022. Testosterone was measured using Roche Cobas electrochemiluminescence immunoassay. Elevated female testosterone levels were confirmed with liquid chromatography tandem mass spectroscopy (LCMS). Results were reviewed independently by two individuals and discrepant results were further evaluated.

Results: An elevated testosterone was found in 1515 adult females (24.4% of total), with 89 (5.9%) sent for confirmation testing. Of these, 14 female patients with discrepant testosterone results were identified with a mean testosterone of 4.5 nmol/l (range 3.6-5.3) when measured with immunoassay, and 1.12 nmol/l (0.2-2.8) with an LCMS method (paired t-test P<0.05). Mean testosterone for females with non-discrepant results (and not on replacement therapy) was 4.4 nmol/l (immunoassay) and 4.2 nmol/l (LCMS) (P=0.35). Follow-up investigations consisted of repeat testosterone (50% of cases), SHBG (75%), DHEA-S (25%), Androstenedione (13%), and 17-OHP (6%). A review of the drug history showed Progesterone-only Pill was prescribed in 25%, Noresthisterone (13%), Depo-Provera (6%), HRT (6%), pregnancy (6%), and no findings (31%).

Conclusion: Elevated testosterone in females is not an unusually finding, however, this audit demonstrates the necessity of confirming elevated testosterone in females where the cause is unclear. Results show a wide variation in the follow up of patients with discordant testosterone results. Interference by Norethisterone in immunoassay testosterone is not a newly described phenomenon, however, clinicians should be aware of interferences in immunoassay to prevent unnecessary investigations into an analytical problem.