We thank Dr Puppala for her letter and for her interest in our article.
We agree that calculation of the free androgen index (FAI) can have a role in the identification of women with hyperandrogenism. However, there are a number of limitations in the use of the FAI. First, as the FAI is calculated using total testosterone and sex hormone-binding globulin (SHBG) measurements, the FAI will reflect any inaccuracies in the measurement of these analytes. If the total testosterone is measured using an immunoassay method prone to interference, the result can be double the true total testosterone level and can lead to a misleading result for FAI. We studied the difference that analytical methodology can make in our own laboratory. Using data from our study, a hypothetical patient who has testosterone levels at the 95 percentile and an average SHBG (median 32 nmol/L), the FAI can vary by around 3% depending upon which testosterone method is used.
Second, Dr Puppala mentions that FAI is considered a more sensitive marker of androgen excess in women. Many studies would support this, although the benefits are relatively modest compared with either calculated bioavailable testosterone or total testosterone.[2, 3] The sensitivity of a test is determined by the cut-off value used. Many studies have used a cut-off value of 5% irrespective of how the testosterone has been measured. Other studies have method-dependent cut-off points of 3.36% (mass spectrometry) and 8% (immunoassay). The use of the FAI gives increased sensitivity for the diagnosis of hyperandrogenism only when appropriate method-dependent cut-off points are used that take into account local methods for SHBG and testosterone.
Third, as SHBG is negatively correlated with body mass index and insulin, the FAI may be misleadingly high in healthy obese women or women with insulin resistance.
We agree that obesity-induced hyperandrogenism, hypothyroidism, hyperprolactinaemia and iatrogenic hyperandrogenism are all possible causes of virilisation in women. The aim of our article was not to provide an exhaustive list of potential causes, but rather to discuss the role of biochemical testing in the differential diagnosis of hyperandrogenism and some important limitations in clinical practice.