Controversy around the definition of PCOS continues

Additional research on individual phenotypes created by the 2003 Rotterdam criteria may help lead to consensus.

In early 2008, the European Society of Human Reproduction and Embryology and the American Society for Reproductive Medicine published a consensus on treating infertility in women with PCOS. Because of the controversy surrounding the topic, the two societies set the following recommendations: Provide preconceptional counseling; use anti-estrogen clomiphene citrate as a first-line intervention, exogenous gonadotrophins or laparascopic ovarian surgery as a second-line intervention and in vitro fertilization as a third-line intervention; and limit the use of metformin to women with glucose intolerance. In fact, the researchers recommended against routine use of metformin for infertility treatment in these women.

Despite this consensus, a larger disagreement still continues; that is, the controversy over the proper diagnosis of the disease, stemming from the existence of two major sets of consensus guidelines, as well as differing international and specialty preferences.

Richard S. Legro

“The controversy reflects both political and territorial beliefs of the parties involved,” Richard S. Legro, MD, professor of obstetrics and gynecology at Penn State University College of Medicine, Hershey, Pa., told Endocrine Today. “The political battle can be drawn between the United States and Europe and the territorial battle can be drawn between medical endocrinologists and gynecologists.”

Physicians in the United States seem to prefer to diagnose the syndrome based on the criteria set during the 1990 National Institutes of Health international conference on PCOS, while physicians in Europe are more apt to follow the newer 2003 Rotterdam criteria. This is best demonstrated by the 2006 position statements from Ricardo Azziz, MD, MPH, chairman of the department of obstetrics and gynecology and director of the center for androgen related disorders at Cedars-Sinai Medical Center, and Stephen Franks, MD, FMedSci, professor of Reproductive Endocrinology at the Imperial College in London.

“The common ground is that all parties agree that PCOS is a condition of ovarian dysfunction, they just have different diagnoses,” said Legro. “They’re really both consensus guidelines by expert opinion, which is the lowest level of medical evidence. So we’re really splitting hairs to argue which definition is better.”

Endocrine Today spoke with Legro, Azziz, Franks and other physicians close to the consensus agreements to gather the most current opinions on the controversy and to find out how near physicians are to a true consensus.

The evolution of the guidelines

Ricardo Azziz, MD, MPH, wrote a position statement on PCOS in 2006. Photo by Bill Pollard

In 1990, the NIH held an international conference on PCOS, during which participants set the criteria for diagnosing the syndrome based on a survey of the audience and speakers. According to the NIH criteria, to be diagnosed with PCOS, a woman had to present with chronic oligoanovulation and either biochemical or clinical signs of androgen excess, according to John E. Nestler, MD, one of the conference attendees and chair of the division of endocrinology and metabolism at Virginia Commonwealth University in Richmond.

Recognizing that the diagnosis may be broader than these two features, the European Society of Human Reproduction and Embryology and the American Society for Reproductive Medicine jointly sponsored a PCOS consensus workshop in 2003, which essentially created the revised 2003 consensus on diagnostic criteria, better known as the Rotterdam criteria.

The Rotterdam criteria expanded the definition to diagnose women with PCOS if they have two of the following characteristics: chronic oligoanovulation or anovulation, androgen excess or polycystic ovaries on ultrasound.

Those who attended the Rotterdam consensus meeting decided that polycystic ovaries should be considered one of the possible criteria for PCOS based on available studies by Pache et al in 1992, van Santbrink et al in 1997 and Jonard et al in 2003.

Essentially, the Rotterdam criteria added two phenotypes of women with PCOS: Women with normal menstrual periods and normal fertility, but who have androgen excess and polycystic ovaries on ultrasound; and women with oligomennorhea and polycystic ovaries on ultrasound, but normal androgen excess.

“This basically says that if you only have two of three, you could have a woman who has evidence of hirsutism and has polycystic ovaries, but ovulates every month. And for what it’s worth, in our old definition we never said that anybody who ovulates every month could have polycystic ovary syndrome,” Robert W. Rebar, MD, executive director of the American Society for Reproductive Medicine, told Endocrine Today.

Why expand the criteria?

According to Franks, physicians recognized the need to expand the PCOS definition in 2003 because of research showing different phenotypes could exist. Studies had shown that within families, one sister could have regular menses while another has amenorrhea or oligomenorrhea — but both have androgen excess and polycystic ovaries. Another finding revealed that simply gaining or losing weight could move a woman up or down the spectrum. For example, irregular cycles may become regular as a woman loses weight.

In fact, in 2006, Welt et al found that women with PCOS defined by irregular menses and hyperandrogenism had higher BMI, suggesting that weight may aggravate PCOS symptoms.

According to Nestler, who is also an Endocrine Today Editorial Board member, much of the Rotterdam criteria had to do with the Europeans preference to view polycystic ovaries on ultrasound as a means to diagnose PCOS. On the American side, investigators traditionally have thought that an ultrasound of the ovaries is not necessary.

“A lot of PCOS is treated by endocrinologists, and they don’t do ultrasounds in the office. And if you send a patient to radiology, you don’t always get reports that tell you enough information,” Nestler told Endocrine Today. “Also, it’s known that if you take 100 women who are perfectly normal — normal menstrual periods, no androgen excess — somewhere around 20% to 25% of them will have polycystic ovaries on ultrasound. It’s a common finding even in normal women.”

“The truth of the matter is that these two phenotypes are not that frequent,” Nestler added. “Even if you use the Rotterdam criteria, that makes up a relatively small percentage of women who have PCOS.”

How the phenotypes match up

Franks admitted that the phenotype without androgen excess was essentially ill defined at first. But his group of investigators and others began to look at the subgroups in more detail and found similarities with those patients who are diagnosed by the NIH criteria.

In 2006, Dewailly et al looked at 66 patients with oligoanovulation and polycystic ovaries without hirsutism or elevated serum androstenedione and testosterone levels. Compared with normal ovulating women, these women had higher mean waist circumference and higher mean levels of serum testosterone, androstenedione and luteinizing hormone. However, compared with women diagnosed with PCOS based on the presence of hyperandrogenism, oligoanovulation and polycystic ovaries, these patients demonstrated milder endocrine and metabolic abnormalities.

In Barber et al, Franks and his colleagues found that women with PCOS by oligomenorrhea and polycystic ovaries had higher luteinizing hormone levels and lower sex hormone-binding levels than controls, which the researchers said supports the view that these women are part of the spectrum of PCOS.

“Obviously androgen levels were different, because by definition they didn’t have androgen excess,” Franks said. “But luteinizing hormone levels were significantly higher in that phenotype group than in controls, and that’s consistent with PCOS. And in our series, we found that sex hormone-binding globulin levels were significantly lower in this group, even after adjusting for BMI. So that, from a biochemical perspective, put them in part of the same spectrum.”

Androgen Excess Society Guidelines

In 2005, the Androgen Excess Society commissioned a seven-member task force to look into the controversy surrounding PCOS and to recommend a definition to its members based on a review of published evidence, according to Azziz.

Published in The Journal of Clinical Endocrinology and Metabolism in 2006, the investigators created a more evidence-based definition of PCOS, but it was again based on expert review.

The task force concluded that physicians should accept the original 1990 NIH criteria with modifications, taking into account the 2003 Rotterdam criteria. A principal conclusion, according to the study, was that PCOS should be first considered a disorder of androgen excess or hyperandrogenism; however, some members considered that there may be forms of PCOS without overt evidence of hyperandrogenism.

Variation among the phenotypes

Legro said he saw the Rotterdam criteria as simply adding another category and another way to diagnose women with this disease. “Studies have shown that the Rotterdam criteria may increase the prevalence of PCOS by as much as 50%, but we have to acknowledge that the increased proportion is less severely affected, at least in terms of insulin resistance and metabolic abnormalities,” Legro said.

Women with PCOS are known to have insulin resistance, putting them at higher risk for type 2 diabetes and metabolic syndrome, Nestler said. “What we don’t know is, do those new phenotypes also carry that same risk? Or is there no increase in risk, or is it intermediate — somewhere between a normal woman and a classic woman with PCOS? We don’t know exactly what those phenotypes represent in terms of metabolic abnormalities.”

Robert W. Rebar

Researchers have already begun to investigate the differences between the phenotypes. Welt et al found that women with PCOS characterized by irregular menses and hyperandrogenism are the most severely affected of the phenotypes when it comes to androgen levels, ovarian volumes and insulin levels.

According to Legro, other studies have already shown that women with PCOS who have polycystic ovaries as one of the two required criteria are more likely to have a more mild metabolic risk than those women who have irregular periods and androgen excess per the 1990 NIH criteria.

Franks and his colleagues found that women with PCOS characterized by oligomenorrhea and polycystic ovaries, but who had normal androgen levels, were almost indistinguishable from healthy controls in terms of prevalence metabolic syndrome was concerned. These women also had fewer metabolic features than women with PCOS who had hyperandrogenism.

Franks said that he and his group have also found that patients with the “classic” syndrome of anovulation and androgen excess were more insulin resistant, and even when correcting for BMI, that phenotype of patients had metabolic abnormalities that the other phenotypes did not. In another study, Broekmans et al, women with PCOS according to the Rotterdam criteria had a lower frequency of obesity, hyperglycemia and insulin resistance, compared with women diagnosed according to the NIH criteria.

“The interesting thing from the point of view of this new category — those with anovulation without androgen excess — is that they don’t have evidence of insulin resistance,” Franks said. “That raises a few questions, but is important from a prognostic point of view. Physicians can look at the different phenotypes and differentiate between those that are at long-term risk for diabetes and those with less risk.”

Future studies necessary

Presenting complaints may be the key to future clinical study, according to Legro. “We need trials that focus on complaints of our patients and outcomes of importance to the individuals and to the health of the community,” Legro said.

For example, Legro suggested focusing studies on infertility, hirsutism, endometrial hyperplasia and cancer — all presenting complaints of women with PCOS. “Admittedly every study cannot look at this big outcome, but some studies need to,” Legro said. “There has to be a greater movement and greater collaborative approach from the National Institutes of Health and from industry to perform the multicenter trials that will answer these questions.”

Longitudinal studies are also necessary, specifically those that perform follow-up on the women with PCOS who have impaired glucose tolerance, and to evaluate conversion to type 2 diabetes, Franks said.

What may lead to true consensus?

Stephen Franks

A large problem with PCOS is that investigators still do not know the underlying cause. “We know that there is an important genetic component. We think there’s probably a primary ovarian abnormality and that the insulin resistance and metabolic abnormality interact with that and act as a second hit,” Franks said. “That is a theory that a lot of us subscribe to, but it is still a theory. Until we actually have a specific metabolic or genetic marker for PCOS, there is going to be a lot of controversy.”

Azziz said that PCOS may likely be caused by a number of genetic traits, and the risk for developing PCOS is then further increased by environmental factors, including obesity and nutrition. “The true definition of PCOS will be arrived at by a couple of different strategies, one of which is to do fairly extensive genetic analyses of the population in an attempt to determine the genetic/pathophysiologic phenotypes,” Azziz said. “This controversy will not be resolved until we get much more extensive, large studies of genetic and family risks of the disorder.” – by Tina DiMarcantonio

Does PCOS increase risk for CVD?

For More Information:

• Azziz R, Carmina E, Dewailly D, et al. Criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: An Androgen Excess Society guideline. J Clin Endocrinol Metab.2006;91:4237-4245.
• Barber TM, Wass JAH, McCarthy MI, Franks S. Metabolic characteristics of women with polycystic ovaries and oligo-amenorrhea but normal androgen levels: implications for the management of polycystic ovary syndrome. Clin Endocrinol. 2007;66:513-517.
• Broekmans FJ, Knauff EAH, Valkenburg O, et al. PCOS according to the Rotterdam consensus criteria: change in prevalence among WHO-II anovulation and association with metabolic factors. BJOG. 2006;113:1210-1217.
• Dewailly D, Catteau-Jonard S, Reyss AC, et al. Oligoanovulation with polycystic ovaries but not overt hyperandrogenism. J Clin Endocrinol Metab. 2006;91:3922-3927.
• The Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Hum Reprod.
• The Thessaloniki ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. Consensus on infertility treatment related to polycystic ovary syndrome. Hum Reprod. 2008;23:462-477.
• Welt CK, Gudmundsson JA, Arason G, et al. Characterizing discrete subsets of polycystic ovarian syndrome as defined by the Rotterdam criteria: the impact of weight on phenotype and metabolic fractures. J Clin Endocrinol Metab. 2006;91:4842-4848.