Introduction: Current clinical practice guidelines recognise that a calculated free testosterone (T) level is the single most-useful, clinically sensitive marker of androgen excess in women, but there is no clear guidance as to the best way to measure free T. Several equations have been proposed to calculate clinically useful estimates of free T including the free androgen index (FAI) and calculated free T (cFT). The FAI is not used in men but it is still commonly used in the investigation of hyperandrogenism in women.
To our knowledge the relationship between FAI and calculated free T, at the lower extremes of SHBG concentration has not been fully investigated in women.
Methods: Serum samples from women for the investigation of hyperandrogenism (n=53) were measured for T by LC-MS/MS, serum albumin and SHBG (SHBG range, 10132 nmol/L) were measured on the Abbott Architect. Calculated free T was determined using the Vermeulen equation and the FAI was calculated.
We also recorded the FAI results from 20,124 women investigated for hyperandrogenism over a ten year period, all samples had T measured using a validated LC-MS method.
Results: The ratio between FAI and cFT was found to increase at lower concentrations of SHBG in women (<30 nmol/L).
From a total of 20,124 results 4223 were found to have a normal T (<1.6 nmol/L) and an SHBG at the lower end of the concentration range (<30 nmol/L). A gradual increase in the FAI was seen in women with a normal T concentarion as the SHBG concentration decreased. The FAI varied from 5 to 40 in these women, showing an 8 fold difference in results, indicating that the FAI is unreliable at low SHBG concentrations.
Conclusion: We would recommend using cFT in women instead of the FAI because of its better agreement across all SHBG concentrations.