Polycystic Ovarian Syndrome: Long-Term Health Consequences

Temeka Zore, MD; Nikhil V. Joshi, MD; Daria Lizneva, MD, PhD; Ricardo Azziz, MD, MPH, MBA


Semin Reprod Med. 2017;35(3):271-281. 

In This Article

Understanding the Phenotypes of PCOS

Over the past 25 years, there have been three sets of diagnostic criteria set forth by different societies and conferences. The first was developed as the result of a survey of participants at a meeting on the subject at the National Institutes of Health (NIH) in April 1990. The resulting criteria proposed a diagnosis of PCOS that included, in order of importance: (1) hyperandrogenism/hyperandrogenemia; (2) oligo-ovulation (ovulatory dysfunction or OD); (3) exclusion of other known disorders such as Cushing's syndrome, hyperprolactinemia, or congenital adrenal hyperplasia; and possibly (4) polycystic ovarian morphology (PCOM) on ultrasound. The last criterion was considered, however, particularly controversial.[4] These criteria resulted in two phenotypes for PCOS, phenotypes A and B (Fig. 1), often called "classic PCOS."

Figure 1.

Comparing the diagnostic classification proposed in the proceeding of the National Institutes of Health conference held in 1990 (NIH 1990), an expert conference cosponsored by the European Society for Human Reproduction and Human Embryology/American Society for Reproductive Medicine held in Rotterdam in 2003 (Rotterdam 2003), and the Androgen Excess and PCOS Society Task Force recommendations published in 2006 (AE-PCOS 2006). Phenotypes A and B are often considered to represent "classic" polycystic ovarian syndrome (PCOS), while phenotype C is often termed "ovulatory" PCOS, and phenotype D is "non-hyperandrogenic" PCOS. HA, hyperandrogenemia.

The second set of criteria was based on the results of an expert gathering held in Rotterdam, the Netherlands, in May 2003, cosponsored by the European Society for Human Reproduction and Embryology (ESHRE) and the American Society for Reproductive Medicine (ASRM). The ESHRE/ASRM guidelines proposed that PCOS be diagnosed, after exclusion of other etiologies, if the subject met two of the following three features, including: (1) oligo- or anovulation, (2) clinical and/or biochemical signs of hyperandrogenism, and (3) polycystic ovaries.[5,6] The use of the Rotterdam criteria effectively broadened the diagnostic criteria and introduced two additional phenotypes, that is, phenotype C often called "ovulatory PCOS," and phenotype D or "non-hyperandrogenic PCOS," compared with the NIH criteria (Fig. 1). The Rotterdam 2003 criteria also increased the number of patients diagnosed with the disorder.

Subsequently, an expert task force assembled by the Androgen Excess and PCOS Society was tasked with systemically reviewing the published literature and using an evidence-based approach to determining the best diagnosis for the disorder. Using metabolic dysfunction as the remote complication of PCOS, the guidelines published in 2006 recommended that PCOS be defined, first and foremost, a diagnosis of androgen excess, accompanied either by OD and/or PCOM, after exclusion of related or mimicking disorders. These recommendations thus excluded the non-hyperandrogenic phenotype (i.e., phenotype D) from the diagnosis of PCOS as had been endorsed by the 2003 Rotterdam criteria (Fig. 1).[7]

Owing to the varying diagnostic criteria set forth by the recommendations and differences in research and clinical practice globally, a November 2012 NIH-sponsored consensus conference reviewed existing evidence, and recommended the use of the broader 2003 Rotterdam criteria albeit with the caveat that the phenotype of the subject be specified. Critically, consideration of PCOS phenotype, rather than aggregating all phenotypes into one disorder, is critical to understanding which patients are at risk for long-term complications.[8]

In the clinical setting, phenotypes A and B, the "classic" forms, are typically the most common, occurring in up to two-thirds of women diagnosed with PCOS. The other two phenotypes appear to be equally distributed.[2] However, it is important to note that typically estimates on the prevalence of the PCOS phenotypes arise from medically biased data based on patients examined in the clinical setting. A recent meta-analysis of 43 populations, including a total of 13,796 patients with PCOS, indicated that the prevalence of the PCOS phenotype in referral versus unselected populations was higher for phenotype A (50 vs. 19%, respectively) but lower for phenotype B (13 vs. 25%, respectively) and phenotype C (14 vs. 34%, respectively).[9] There were little differences in the prevalence of phenotype D (17 vs. 19%, respectively) between the two groups. Furthermore, referral PCOS subjects had a greater mean body mass index (BMI) than local controls, a difference that was not apparent in unselected PCOS.

Of note, a longitudinal study following a referred PCOS population over 20 years demonstrated a reduction in the more severe phenotypes, improved ovulation, and amelioration in the degree of hyperandrogenemia in these women by their fourth decade of life.[10] Overall, it is critical that when considering the risk of long-term consequences in PCOS, a clear understanding of the various phenotypes present, and their impact by referral status and by age, be taken into account.