Endocrine Abstracts

Testosterone levels in normally cycling women are generally assumed to be elevated during ovulation. The clinical relevance of changing testosterone levels during the menstrual cycle, however, is unclear, due to poor performance of current direct immunoassays for testosterone at low concentrations.

An isotope dilution-liquid chromatography–tandem mass spectrometry (ID-LC–MS/MS) method with sufficient reliability and sensitivity¹ was used to measure testosterone in serum samples obtained daily during the menstrual cycles of 25 healthy women, characterized by biochemical and physical examination. Since ID-LC–MS/MS is not accessible to all clinical laboratories, we also evaluated the performance of the recently available direct immunoassay Architect 2nd Generation Testosterone.

Performance of the ID-LC–MS/MS method was comparable to the reference method² (weighted Deming regression: y=1.007×−0.056 nmol/l; r=0.9998). Comparison of the 2nd generation immunoassay results to ID-LC–MS/MS yielded y=1.095×+0.104 nmol/l (r=0.90). Testosterone levels were significantly higher mid-cycle, although a peak was not discernable in all individuals. Intra-individual variation exceeded the group average in the menstrual cycle; the ratio of extremes (max/min) found in individuals ranged from 1.6–3.7. Apart from a persistent positive bias, the immunoassay measured the same testosterone profiles. The reference interval was 0.30–1.69 nmol/l for ID-LC–MS/MS and 0.50–2.00 nmol/l for the immunoassay.

Our ID-LC–MS/MS method measured low testosterone levels accurately. The Architect 2nd generation testosterone immunoassay had acceptable performance across the entire range measured in women. On average, the elevation of mid-cycle testosterone levels is statistically significant, although not relevant for clinical practice since the day-to-day variation is higher and independent of menstrual cycle. In light of this, we recommend measuring testosterone on at least two independent occasions for diagnostic purposes.

1. Bui 2010 AnnClinBiochem.

2. Thienpont 2008 ClinChem.

Declaration of interest: I fully declare a conflict of interest. Details below.

Funding: This work was supported, however funding details unavailable.