Endocrine Abstracts

The measurement of serum testosterone in women is difficult with most automated immunoassays because of poor sensitivity, accuracy and precision at low concentrations. Mass spectrometry is the gold standard for these measurements but it is slower and requires special expertise and equipment. We investigated whether the modification of a commercial RIA allowed improved testosterone measurements in laboratories without access to mass spectrometry.

A commercial RIA (Siemens coat-a-count) was modified by increasing the volume of serum and diluting the low calibrator to extend the lower limit of the standard curve from 0.70 to 0.35 nmol/l. The new method had a functional sensitivity of 0.22 nmol/l and showed recoveries of 93–122% when measuring specimens made up by adding testosterone to blank serum (range 0.27–4.3 nmol/l). The between batch imprecision was 12.7% at a testosterone concentration of 0.39 nmol/l (n=21).

The results of the new assay were compared with those of an automated immunoassay (Siemens Centaur) using a variety of serum specimens referred to the laboratory for testing. The RIA gave lower results than the Centaur by ~0.5 nmol/l. In women with raised serum sex hormone binding globulin as a result of oestrogen therapy the difference was increased to ~1.0 nmol/l suggesting that the automated assay was adversely affected by changes in the hormone binding environment. In 56 women referred because of low libido, serum testosterone by RIA was reduced in 34% although it was typically normal in this group when measured using the automated immunoassay.

In conclusion, a simple modification of a commercial RIA enabled more reliable measurements of serum testosterone in women. The changed results from the modified assay could have important therapeutic implications in different groups of patients.