The Value of Androgen Measures for Diagnosing Polycystic Ovary Syndrome (PCOS) in an Unselected Population

Polycystic Ovary Syndrome (PCOS) is diagnosed by a combination of three features: hyperandrogenism (biochemical and/or clinical), ovulatory dysfunction, and polycystic ovarian morphology, usually detected by ultrasonography. Our study aimed to determine the need for androgen measurements by using hirsutism to establish hyperandrogenism for diagnosing PCOS in a medically unbiased population.

We utilized a pre-existing cohort of unselected (medically unbiased) females aged 18–45 years. All underwent a history and physical, including a modified Ferriman-Gallwey (mFG) hirsutism score. Subjects were categorized clinically as eumenorrheic non-hirsute (CONTROLS), menstrual dysfunction only (OLIGO-ONLY), hirsutism only (HIRSUTE-ONLY), or menstrual dysfunction and hirsutism (OLIGO + HIRSUTE). All subjects underwent measurements of androgens using high-quality assays. CONTROLS established the upper normal limit for androgen levels. We defined PCOS using the NIH 1990 criteria.

Of 462 individuals with complete evaluations, 311 (67.3%) were CONTROLS, 71 (15.4%) were OLIGO-ONLY, 64 (13.9%) were HIRSUTE-ONLY, and 16 (3.5%) were OLIGO + HIRSUTE. Neither HIRSUTE-ONLY nor OLIGO-HIRSUTE women required androgen measures to demonstrate hyperandrogenism. Among OLIGO-ONLY, 19 (26.8%) demonstrated hyperandrogenemia without hirsutism, with White women significantly more likely than Black women to demonstrate this.

In our study of medically unbiased reproductive-aged women using the NIH 1990 criteria for PCOS, only 15.4% of women evaluated (those with menstrual dysfunction only) required androgen measurements. In these women only one-quarter demonstrated hyperandrogenemia. These data provide a strategy to minimize the need for androgen assays, including firstly categorizing subjects by clinical presentation and then assessing circulating androgens in the subgroup with menstrual dysfunction only.