Exp Clin Endocrinol Diabetes 2006; 114(4): 182-187
DOI: 10.1055/s-2006-924062

J. A. Barth Verlag in Georg Thieme Verlag KG Stuttgart · New York

Is it Necessary to Measure Free Testosterone to Assess Hyperandrogenemia in Women? The Role of Calculated Free and Bioavailable Testosterone

A. Mueller[*] 1 , R. Dittrich[*] 1 , S. Cupisti1 , M. W. Beckmann1 , H. Binder1
  • 1Department of Obstetrics and Gynecology, Erlangen University Hospital, Erlangen, Germany
Further Information

Publication History

Received: September 20, 2005 First decision: December 29, 2005

Accepted: January 11, 2006

Publication Date:
17 May 2006 (online)


Hirsutism in women is defined as excessive facial and/or body terminal hairs showing a masculine distribution; the condition affects approximately 7 % of women of reproductive age, and mild forms of hirsutism in particular are the most common ([Souter et al., 2004]). Various scoring systems have been evaluated for quantifying hirsutism; the most popular is a modified version of the Ferriman-Gallwey score (mF‐G score), originally proposed by Ferriman and Gallwey in 1961 ([Souter et al., 2004]; [Ferriman and Gallwey, 1961]). Women with unwanted hair growth should be evaluated endocrinologically, as 50 % are found to have an androgen excess (AE) disorder - e.g., polycystic ovary syndrome (PCOS), hyperandrogenic insulin-resistant acanthosis nigricans syndrome, nonclassic adrenal hyperplasia, and isolated chronic oligo- or anovulation ([Souter et al., 2004]).

Chronic anovulation is a common problem for infertile couples, with a rate of 20 - 25 % ([van Santbrink et al., 1997]). As recommended by the World Health Organization (WHO), women with anovulation are classified on the basis of two endocrine parameters - the endogenous gonadotropins and estrogens ([WHO, 1993]; [ESHRE Capri Workshop Group, 1995]; [Laven et al., 2001]). Up to 80 % of women with chronic anovulation have serum levels of follicle-stimulating hormone (FSH) and estradiol (E) within the normal range. These women are classified as having normogonadotropic normo-estrogenic anovulatory infertility, more commonly referred to as WHO group 2, and constitute a notoriously heterogeneous population ([Hart et al., 2004]). In polycystic ovarian syndrome (PCOS), characterized by chronic anovulation, hyperandrogenemia is also the most common cause of normogonadotropic normo-estrogenic anovulatory infertility, and such patients certainly belong to WHO group 2 ([Laven et al., 2001]; [Hart et al., 2004]).

PCOS is a syndrome of ovarian dysfunction that is frequently associated with the systemic condition of insulin resistance, with the cardinal features of hyperandrogenism and/or polycystic ovarian morphology ([Laven et al., 2002]). The definition of PCOS has been a matter of controversy, and aspects of the pathophysiology and natural history of the condition remain unclear. Previously established diagnostic criteria were not universally accepted, as there were markedly divergent views of the etiology, pathogenesis, and clinical appearance ([Hart et al., 2004]). The diagnostic criteria for PCOS were recently revised by an expert conference sponsored by the European Society for Human Reproduction and Embryology and the American Society for Reproductive Medicine. The criteria are characterized by clinical and biochemical signs, as well as by ovarian morphology (Table [1]). PCOS is diagnosed if two of three criteria are met ([Rotterdam ESHRE/ASRM-sponsored PCOS Consensus Workshop Group, 2004 a],[b]). In the consensus view, polycystic ovaries (PCO) were considered to be one of the possible criteria for PCOS. These criteria again recognize that PCOS is also still a diagnosis based on the exclusion of other related disorders. It was considered that free testosterone (FT) or the free androgen index (FAI) were more sensitive methods of assessing hyperandrogenemia in women with possible PCOS ([Rotterdam ESHRE/ASRM-sponsored PCOS Consensus Workshop Group, 2004 a],[b]).

Table 1 Revised 2003 diagnostic criteria for polycystic ovary syndrome (Rotterdam ESHRE/ASRM-sponsored PCOS Consensus Workshop Group, 2004 a,b). Two out of three symptoms must be present 1 Oligo-ovulation or/and anovulation 2 Clinical and/or biochemical signs of hyperandrogenism 3 Polycystic ovaries Exclusion of other etiologies (congenital adrenal hyperplasia, androgen-secreting tumors, Cushing's syndrome)

The majority of women with hirsutism or PCOS have an androgen excess

However, it is unclear which androgen fraction reflects the clinical situation most accurately and correlates best with clinical symptoms of hyperandrogenemia. Testosterone circulates in plasma nonspecifically bound to albumin and specifically bound to sex hormone-binding globulin (SHBG); a small fraction is unbound as free testosterone (FT). Although immunoassay of blood total testosterone (TT) has long been the standard measurement, the TT concentration is not a reliable index of bioavailable T, as it depends on variations in the concentration of the binding proteins. There is good evidence that the FT and the hormone fraction not specifically bound to albumin in plasma - referred to as the bioavailable testosterone (BT) fraction - more accurately reflects the clinical situation than total hormone levels in plasma ([Vermeulen et al., 1999]). Several methods of estimating FT and BT in plasma have been established. FT measurement by equilibrium dialysis (apparent FT concentration, AFTC) ([Vermeulen et al., 1971]; [Vermeulen et al., 1999]) and centrifugal ultrafiltration ([McCann et al., 1985]; [Hammond et al., 1980]; [Vlahos et al., 1982]) are the reference measurement procedures (RMPs) for correct measurement of the free fraction of T in vivo. However, this is a time-consuming and complex manual procedure that is not routinely practicable in large laboratories, which rely increasingly on automated multiplex assay platforms.

Alternative methods of estimating FT have therefore been developed, using additional assay steps, including the free testosterone analogue assay. FT can be measured by an analogue ligand immunoassay method (aFT), which is the easiest and fastest but shows substantially lower values than those obtained with the RMPs ([Ooi et al., 1998]; [Wilke et al., 1987]; [Rosner, 1997]). Models have also been developed for calculating FT (cFT) and BT (cBT) from total T, SHBG, and albumin in one sample ([Vermeulen et al., 1999]; [Sodergard et al., 1982]). Another simple calculation model used by many researchers is an indirect parameter of FT, known as the free androgen index (FAI), which is obtained as the quotient 100 T/SHBG ([Mathur et al., 1981]).

Vermeulen et al. presented an excellent evaluation of a calculation method for estimating the free testosterone fraction in serum ([Vermeulen et al., 1999]). They compared AFTC values with cFT and aFT levels as well as with the FAI. Neither aFT nor FAI was a reliable parameter for the FT fraction, but cFT was a reliable index of the FT fraction, the calculated nonspecifically bound T (cBT) reliably reflected non-SHBG‐T (BT), and immunoassayable SHBG was a reliable measure of SHBG-binding sites.

Is it necessary to use RMPs to estimate FT when calculation is equivalent?

The question remains of whether there is a need to measure FT using RMPs when cFT or cBT are equally appropriate markers for assessing hyperandrogenemia in clinical situations such as hirsutism or PCOS in women. The aim of the present study was to evaluate the diagnostic importance of calculated values (cFT, cBT, and FAI) in hirsute women (the hirsutism group) and in women meeting the diagnostic criteria for PCOS (the PCOS group) and to compare these parameters with the values in women with neither hirsutism nor PCOS (the control group).


1 Andreas Mueller and Ralf Dittrich contributed equally to the authorship

Dr. Andreas Mueller

Department of Obstetrics and Gynecology
Erlangen University Hospital

Universitätsstraße 21 - 23

91054 Erlangen


Phone: + 49-9131-8533553

Fax: + 49-9131-8533552

Email: andreas.mueller@gyn.imed.uni-erlangen.de