Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11

Contents lists available at ScienceDirect

Best Practice & Research Clinical

Obstetrics and Gynaecology
journal homepage: www.elsevier.com/locate/bpobgyn

Diagnostic criteria for PCOS: Is there a need for a

rethink?
Didier Dewailly, MD, Professor a, b, *
a
Service de Gyn

ecologie Endocrinienne et M
edecine de la Reproduction, Ho^pital Jeanne de Flandre, CHRU, F-
59037 Lille cedex, France
b
Facult

e de M
edecine, Universit

e Lille Nord de France, 59000 Lille, France

The diagnostic criteria for polycystic ovarian syndrome (PCOS)

Keywords:
have been grouped in different classications that have been
polycystic ovary syndrome
androgens conicting for many years. At present, the classication of Rot-
anovulation terdam is the most used, but with varying frequency depending on
diagnosis the country and medical specialties. This classication is now >10
ultrasound years old. Although its fundamental principle (two criteria
anti-Mllerian hormone required out of three) is still valid, each of its three items (oligo-
anovulation (OA), hyperandrogenism (HA), and polycystic ovarian
morphology (PCOM)) needs to be updated. The denition of bio-
logical HA is still unresolved. The criteria used to dene OA are
insufcient. The denition of PCOM proposed in 2003 is now
obsolete when using the latest generation of ultrasound machines.
The serum anti-Mllerian hormone (AMH) assay seems increas-
ingly to be an excellent substitute for follicular count and is likely
to emerge as the ofcial PCOM marker. A new consensus confer-
ence is urgently needed.
2016 Published by Elsevier Ltd.

Introduction

Several studies have focused on the diagnostic criteria for polycystic ovary syndrome (PCOS), which
have been the subject of debate for many years. As the name suggests, PCOS is a syndrome, not a
disease, which means a more or less complete set of phenotypic characteristics, variously associated,

* Corresponding author. Faculte
de Me

decine, Universite

Lille Nord de France, 59000 Lille, France. Tel.: 33 3 20 44 63 09;
Fax: 33 3 20 44 64 07.
E-mail address: didier.dewailly@chru-lille.fr.

http://dx.doi.org/10.1016/j.bpobgyn.2016.03.009
1521-6934/ 2016 Published by Elsevier Ltd.
6 D. Dewailly / Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11

with the lack of a constant element on which the diagnosis could be based (gold standard). According
to the predominance of a particular symptom, patients with PCOS are directed to various specialists
who, therefore, have a partial and oriented view of the syndrome. Because of its prevalence, PCOS is
often associated with other conditions with symptoms that are sometimes mistakenly viewed as being
directly linked to PCOS. A signicant example here is obesity followed by metabolic abnormalities. The
clinical signs and biological disorders related to obesity often pollute the presentation of PCOS
phenotype, thereby creating discrepancies in the literature. Other conditions, less frequent than
obesity, may be associated with PCOS and may be unrecognized if their symptoms overlap with those
of PCOS. Finally, with the advent of new diagnostic markers, we recognize several moderate or latent
forms of PCOS. This makes it especially difcult to set diagnostic thresholds for different markers
allowing separation, with good sensitivity and good specicity, of normal women from patients with
PCOS. Thus, as posed in the title, the diagnostic criteria must be reassessed.

Historical background

In 1935, Stein and Leventhal [1] designated PCOS as a syndrome due to its association with varying
degrees of hirsutism, chronic anovulation, bilateral ovarian enlargement, and obesity. With the advent
of hormonal assays, the diagnostic criteria have been for some time based on the increase in serum
androgen and luteinizing hormone (LH) levels, with an increased LH/follicle-stimulating hormone
(FSH) ratio. In 1990, during an expert conference sponsored in part by the National Institute of Child
Health and Human Disease (NICHD) of the NIH [2], the participants agreed that the major diagnostic
criteria for PCOS should include (in the order of importance) i) hyperandrogenism (HA) and/or
hyperandrogenemia, ii) menstrual dysfunction, and iii) exclusion of other known disorders.
This denition has prevailed for many years, ignoring the advent of ultrasound (U/S) and the
possibility of reverting to a morphological description of polycystic ovaries (PCOM; polycystic ovarian
morphology). It must be recognized that in the 1980e1990s, the U/S criteria had long been uncertain
and variable. Based on a rigorous evaluation of these criteria, the Rotterdam consensus conference held
in 2003 [3] decided to retain the following sonographic description of PCOM: either 12 or more follicles
per whole ovary measuring 2e9 mm in diameter, or increased ovarian volume (>10 cm3). This de-
nition was then added to the NIH criteria, thus introducing a classication based on three items (Table
1). The most signicant advance in this consensus conference was the requirement of at least two
items. This exible association generates the possibility of four phenotypes (Table 2) with various
concerns, which are not accepted universally. Certainly, the phenotype without overt HA (phenotype
D) is the most controversial [4].
Indeed, concerned with the laxity of this denition, some have later proposed to adjust it by
making the criterion HA a mandatory prerequisite (AES classication) [5]. At present, the classi-
cation of Rotterdam is the most used, but with varying frequency depending on the country and
medical specialties. Whether this classication has been well understood remains unclear. For
example, during a large survey of IVF centers in the world [6], >80% of the physicians reported using
this classication; however, almost 40% also considered HA as a mandatory criterion. In addition, this
classication is >10 years old and has not been updated, leading to inconsistencies in the literature.

What were the benets of applying the Rotterdam criteria?

First, this classication now allows recognizing moderate phenotypes lacking either HA or oligo-
anovulation (OA). As these phenotypes are not recognized by the NIH classication, the rst conse-
quence has been the increase in the prevalence of PCOS by about 50% in different populations [7]. This

Table 1
Rotterdam 2003 criteria (two out of three).

1. Oligo- and/or anovulation (OA)

2. Clinical and/or biochemical signs of hyperandrogenism (HA)
3. Polycystic ovarian morphology (PCOM) and exclusion of other etiologies (congenital adrenal hyperplasias, androgen-
secreting tumors, Cushing's syndrome, hyperprolactinemia, and hypothalamic anovulation)
 D. Dewailly / Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11 7

Table 2
The Rotterdam phenotypes.

OA HA PCOM Phenotype

yes yes yes A

yes yes no B
no yes yes C
yes no yes D

increase in prevalence appears to be divided equally between the C and D phenotypes (isolated HA or
OA, respectively). In addition, the Rotterdam classication considers the ethnic variations in the
phenotypic expression of PCOS, which explains why the prevalence of PCOS as dened by the Rot-
terdam criteria was higher in some populations than in others [8,9].
Conversely, the Rotterdam consensus conference has specied that the isolated presence of PCOM
on U/S was not sufcient to make the diagnosis of PCOS. This is extremely important because, although
using different criteria, a large number of studies from different countries reported a prevalence of
20e30% of PCOM in the general population of women exhibiting neither HA or OA (reviewed in
Ref. [10]). Furthermore, using the strict U/S criteria retained in 2003, the prevalence of asymptomatic
PCOM in the general population has increased markedly in recent years, reaching 80% in some series.
This is clearly an articial deviation and the reasons are outlined later.

Should the Rotterdam classication be reviewed?

In answer to this question, the classication should perhaps not be reviewed in its fundamental
principle (two criteria required out of three). Only this exibility allows recognizing the moderate
forms of PCOS. Conversely, all three items used for this classication should be updated.

A e Hyperandrogenism

Its clinical denition is established with three main symptoms: hirsutism, acne, and alopecia.
Regarding hirsutism, most authors agree to recognize the subjectivity of this symptom, despite the
semiquantitative evaluation provided by the score of Ferriman and Gallwey. This problem is still not
resolved. In fact, many authors consider that the sole complaint of unwanted hair growth, in the
absence of frank hirsutism on physical examination, may signal the presence of PCOS. It is the same for
acne, which is far from being a specic sign of PCOS, especially among adolescents. A more precise
classication of acne (type of lesions, topography, etc.) would possibly increase the specicity, but it
does not exist to date. Of course, alopecia must meet the denition of androgenic alopecia so that it can
be distinguished from several other causes of hair loss. When isolated, it is not specic and PCOS does
not seem to be the most common cause.
The denition of biological HA is an issue that is still unresolved to date. It is treated in one of the
following articles in this issue. Most authors agree on the fact that the currently available assays for
plasma androgens have poor sensitivity. The advent of mass spectrometry technique does not seem to
have much improved this problem. The free androgen index (FAI), which requires measuring both total
testosterone and SHBG (sex hormone-binding globulin), is recommended by most authors, as it seems
to be the most sensitive biological index of HA. However, SHBG production at the liver level is
extremely sensitive to the negative effects of insulin. Therefore, the FAI is articially increased in
conditions of hyperinsulinism, thereby creating a major bias in the denition and recognition of HA,
particularly in obese women. For some, the main goal of androgen assessment is to exclude other states
of HA, especially androgen-secreting tumors or hyperthecosis. This assessment must be based on a
gradual strategy of investigations to avoid unnecessary expense. It seems futile to demonstrate a
biological HA when there are no clinical symptoms of HA in women with PCOM, especially if they also
have OA. Finally, follicular excess being most likely secondary to intraovarian HA, markers of this
excess, essentially follicle count at U/S and serum anti-Mllerian hormone (AMH) level, can be
considered as indirect markers of HA [11].
8 D. Dewailly / Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11

Obviously, little progress has been made since 2003 in dening HA. The question as to whether HA
should be a mandatory criterion is still debated. Obviously, PCOS is a hyperandrogenic disorder [12],
but oligo-anovulatory women with PCOM and without overt HA are part of the syndrome [4,7].

B e Oligo-anovulation

The criteria used by the Rotterdam consensus conference to dene OA are insufcient (primary or
secondary amenorrhea, oligomenorrhea dened by less than eight episodes of menses a year or cycle
length >35 days). This denition does not include the subclinical dysovulation, which is found with
some prevalence in the general population, with or without symptoms of HA. Some authors propose
to consider moderate PCOS cases as those women with regular cycles, some actually anovulatory,
even if they have no HA. Indeed, the higher the frequency of anovulation, the higher the testosterone
and serum AMH levels, which tend to overlap with the values of genuine cases of PCOS [13]. It is
therefore recommended, particularly in conditions of infertility to check the serum progesterone
level in two consecutive cycles, between days 20 and 24 of the cycle in women with regular cycles
and PCOM. Conversely, in the adolescent, stricter criteria must be required to dene menstrual ir-
regularity. The WHO (World Health Organization) classication is very sensitive (yielding a preva-
lence of OA of 51%) and creates many false-positive cases [14]. A stricter denition is needed, such as
that of Van Hooff et al. [15], that is, <8 menstruations and/or 2 cycles for <22 or >42 days per year).
With this denition, the prevalence of OA drops to 14% and OA becomes highly predictive of PCOS
later in life.

C e Polycystic ovarian morphology

This is, without question, the most contentious issue concerning the Rotterdam classication.
The denition of PCOM is the subject of a subsequent article in this issue. It is clear that the
denition proposed in 2003 is now obsolete when using the latest generation of devices with
maximum probe frequencies >8 MHz. This applies mainly to follicle count, while the threshold for
ovarian volume does not appear to be inuenced by the technical improvement of U/S. With the
new appliances, the threshold of 12 follicles per ovary is surpassed, and its application leads to an
articial overestimation of the PCOM prevalence in the general population. A panel of experts has
recently addressed the problem and proposes to use a threshold to 25 with new U/S machines [16].
It must be emphasized, however, that this threshold of 25 corresponds to the 95th percentile of the
populations of normal women. As 20e30% of these women have asymptomatic PCOM, the
question of their exclusion from the control groups is currently debated. This can only be done by a
biomathematical method avoiding the need for preestablished thresholds, which obviously would
bias the results. By cluster analysis, we previously proposed a threshold of 19 follicles per ovary for
selecting control women without PCOM [17]. Finally, the serum AMH assay seems increasingly to
be an excellent substitute for follicular count [18]. This is discussed in a following article in this
issue. After a few aberrations related to the technical problems [19,20], this assay is likely to
emerge as the approved PCOM marker. The better control of AMH assay and especially its auto-
mation [21] will allow its dissemination and it now becomes imperative to dene a universally
agreed threshold.

Is the Rotterdam classication fully reliable?

The Rotterdam classication is not fully reliable. Firstly as mentioned earlier, this is due to the lack of
sensitivity and/or specicity of each item taken one by one. Secondly, experience and common clinical
sense alone allow wise application of the Rotterdam classication. This classication can fail in chal-
lenging clinical situations, particularly, in women with hypothalamic anovulation and associated
PCOM. Whether this association is fortuitous [22] or causally related [23], it is a pitfall for the clinician
as there is no consensual diagnostic criterion for hypothalamic anovulation. Therefore, OA might be
erroneously attributed to PCOS, especially when wrong thresholds are used to dene PCOM [24]. Only
the vigilance of the clinician and his interest in the psycho-nutritional status of his patients may avoid
 D. Dewailly / Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11 9

misdiagnosing PCOS and ignoring the hypothalamic problem. Likewise, PCOS might be overdiagnosed
in adolescents, for the reasons mentioned earlier, or conversely unrecognized because it is difcult to
assess the U/S criteria. Hence, the diagnosis of PCOS by the classication of Rotterdam must always be a
diagnosis of exclusion. It is for the clinician to systematically search any other condition that may give
HA and/or OA. This was clearly formulated during the Rotterdam consensus conference in 2003, but
clinicians should bear this in mind.
Conversely, as discussed earlier, the Rotterdam classication excludes asymptomatic women with
PCOM. However, should these women be excluded from the denition? They are not subfertile [25], but
at baseline, the serum AMH and FSH level is increased or decreased, respectively [26]. When stimu-
lated, they have PCO-like features in terms of androgen secretion [27] or follicle response to controlled
ovarian hyperstimulation (COH) [28]. In a recent genetic study, a specic single-nucleotide poly-
morphism (SNP) in the C9orf3 gene was associated with isolated PCOM [29]. If including these women
in the PCOS spectrum is premature, the question remains whether they should be maintained in the
control groups used to dene the thresholds for PCOM.

Is the classication in different Rotterdam phenotypes clinically and scientically relevant?

Obviously, PCOS phenotypes at metabolic risk should be better identied. Whether the Rotterdam
classication addresses this issue properly is still unclear. In one study, the classic phenotype and, to a
lesser extent, the ovulatory phenotype were independently associated with insulin resistance, whereas
the normoandrogenic phenotype was not [30]. Likewise, in another study, obese women without
androgen excess had a different metabolic prole compared with those in the NIHePCOS group, with
lower rates of hyperglycemia and hyperinsulinism, despite comparable distributions of body weight
[7]. Lastly, women with hyperandrogenic PCOS have a worse cardiometabolic prole and a higher
prevalence of the cardiovascular (CV) risk factors compared with those with nonhyperandrogenic PCOS
[31]. Hence, some authors proposed that the name PCOS be retained for the reproductive phenotypes
and that a new name be created for the phenotypes at high metabolic risk [32]. It is not clear, however,
that every patient would t such a clear-cut classication.
The scientic use of the Rotterdam classication is also a major issue, particularly for genetic
studies. The phenotypic variability of PCOS, as dened by the Rotterdam classication, may be partly
genetically determined, as indicated by the study where independent associations were found be-
tween some SNPs and each of the Rotterdam criteria separately [29]. However, as seen in daily practice,
some patients with PCOS may mutate from one phenotype to another, depending on the extrinsic
factors of the disease such as age, weight, or environment. Thus, OA and to a lesser degree HA regress
with age [33], and conversely overweight women with PCOS are more prone to ovulation disorder and
HA than lean women [34]. PCOM also changes with age, while the inuence of the weight seems
minimal or negligible. It is the same for the plasma level of AMH. It, nevertheless, remains that some
women with PCOS appear to be protected from OA or HA, for reasons that remain unresolved to date
and that genetics may help elucidate.
Finally, the application of the Rotterdam classication for scientic purposes should be more
rigorous. In many articles, the authors claim to have applied this classication but do not give enough
details, particularly regarding the thresholds used to dene HA, follicle excess at U/S, and/or excessive
serum AMH level. The techniques used to collect this information are not detailed enough. We urgently
need ofcial guidelines for the selection not only of patients but also of control women to improve the
scientic rigor of studies about PCOS.

Conclusion

This brief review does not claim to be exhaustive. Only a new consensus conference, followed by a
comprehensive report with rigorous and strict recommendations, will be so. Hopefully, this new
consensus conference will be held as soon as possible.
10 D. Dewailly / Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11

Practice points

1) The definition of HA should be primarily clinical. The main goal of androgen assays is to
exclude other states of HA.
2) In the adolescent, strict criteria must be required to define menstrual irregularity suggestive
of PCOS.
3) The threshold used to define follicle excess at ultrasound (U/S) should be adapted to the
machine used; follicle count is likely to be replaced by the AMH assay in future.
4) The isolated presence of PCOM on U/S is not sufficient to make the diagnosis of PCOS.
5) The diagnosis of PCOS by the classification of Rotterdam must always be a diagnosis of
exclusion.

Research agenda

1) Decide whether women with asymptomatic PCOM should be excluded from the control
groups.
2) Define a universally agreed threshold of serum AMH that defines PCOM.
3) Improve identification of PCOS phenotypes that are at metabolic risk.
4) Elucidate why some women with PCOS appear to be protected from OA or HA.

Conict of interest

The author declares having no conict of interest.

References

[1] Stein IF, Leventhal ML. Amenorrhea associated with bilateral polycystic ovaries. Am J Obstet Gynecol 1935;29:181e91.
[2] Zawadzki JK, Dunaif A. Diagnostic criteria for polycystic ovary syndrome: towards a rational approach. In: Dunaif A,
Givens JR, Haseltine FP, et al., editors. Polycystic ovary syndrome. Boston: Blackwell Scientic Publications; 1992. p.
377e84.
*[3] Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group. Revised 2003 consensus on diagnostic criteria and
long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 2004;19(1):41e7.
[4] Dewailly D, Catteau-Jonard S, Reyss AC, et al. Oligoanovulation with polycystic ovaries but not overt hyperandrogenism.
J Clin Endocrinol Metab 2006;91(10):3922e7.
*[5] Azziz R, Carmina E, Dewailly D, et al. The Androgen Excess and PCOS Society criteria for the polycystic ovary syndrome:
the complete task force report. Fertil Steril 2009;91(2):456e88.
[6] Ning N, Balen A, Brezina PR, et al. How to recognize PCOS: results of a web-based survey at IVF-worldwide.com Reprod
Biomed Online 2013;26(5):500e5.
*[7] Broekmans FJ, Knauff EA, 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(10):1210e7.
[8] Wang S, Alvero R. Racial and ethnic differences in physiology and clinical symptoms of polycystic ovary syndrome. Semin
Reprod Med 2013;31(5):365e9.
[9] Zhao Y, Qiao J. Ethnic differences in the phenotypic expression of polycystic ovary syndrome. Steroids 2013;78(8):
755e60.
[10] Hart R, Hickey M, Franks S. Denitions, prevalence and symptoms of polycystic ovaries and polycystic ovary syndrome.
Best Pract Res Clin Obstet Gynaecol 2004;18(5):671e83.
[11] Dewailly D, Pigny P, Soudan B, et al. Reconciling the denitions of polycystic ovary syndrome: the ovarian follicle number
and serum anti-Mullerian hormone concentrations aggregate with the markers of hyperandrogenism. J Clin Endocrinol
Metab 2010;95(9):4399e405.
*[12] Azziz R, Carmina E, Dewailly D, et al. Positions statement: criteria for dening polycystic ovary syndrome as a pre-
dominantly hyperandrogenic syndrome: an Androgen Excess Society guideline. J Clin Endocrinol Metab 2006;91(11):
4237e45.
[13] Sjaarda LA, Mumford SL, Kissell K, et al. Increased androgen, anti-Mullerian hormone, and sporadic anovulation in
healthy, eumenorrheic women: a mild PCOS-like phenotype? J Clin Endocrinol Metab 2014;99(6):2208e16.
 D. Dewailly / Best Practice & Research Clinical Obstetrics and Gynaecology 37 (2016) 5e11 11

[14] Hickey M, Doherty DA, Atkinson H, et al. Clinical, ultrasound and biochemical features of polycystic ovary syndrome in
adolescents: implications for diagnosis. Hum Reprod 2011;26(6):1469e77.
*[15] van Hooff MH, Voorhorst FJ, Kaptein MB, et al. Predictive value of menstrual cycle pattern, body mass index, hormone
levels and polycystic ovaries at age 15 years for oligo-amenorrhoea at age 18 years. Hum Reprod 2004;19(2):383e92.
*[16] Dewailly D, Lujan ME, Carmina E, et al. Denition and signicance of polycystic ovarian morphology: a task force report
from the Androgen Excess and Polycystic Ovary Syndrome Society. Hum Reprod Update 2014;20(3):334e52.
[17] Dewailly D, Gronier H, Poncelet E, et al. Diagnosis of polycystic ovary syndrome (PCOS): revisiting the threshold values of
follicle count on ultrasound and of the serum AMH level for the denition of polycystic ovaries. Hum Reprod 2011;26(11):
3123e9.
[18] Pigny P, Jonard S, Robert Y, et al. Serum anti-Mullerian hormone as a surrogate for antral follicle count for denition of the
polycystic ovary syndrome. J Clin Endocrinol Metab 2006;91(3):941e5.
[19] Iliodromiti S, Kelsey TW, Anderson RA, et al. Can anti-Mullerian hormone predict the diagnosis of polycystic ovary
syndrome? A systematic review and meta-analysis of extracted data. J Clin Endocrinol Metab 2013;98(8):3332e40.
*[20] Dewailly D, Andersen CY, Balen A, et al. The physiology and clinical utility of anti-Mullerian hormone in women. Hum
Reprod Update 2014;20(3):370e85.
[21] Pigny P, Gorisse E, Ghulam A, et al. Comparative assessment of ve serum antimllerian hormone assays for the diagnosis
of polycystic ovary syndrome. Fertil Steril 2016;105(4):1063e9.
[22] Robin G, Gallo C, Catteau-Jonard S, et al. Polycystic Ovary-Like Abnormalities (PCO-L) in women with functional hypo-
thalamic amenorrhea. J Clin Endocrinol Metab 2012;97(11):4236e43.
[23] Carmina E, Fruzzetti F, Lobo RA. Increased anti Mullerian hormone levels and ovarian size in a subgroup of women with
functional hypothalamic amenorrhea: further identication of the link between polycystic ovary syndrome and func-
tional hypothalamic amenorrhea. Am J Obstet Gynecol 2016 Jan 6. http://dx.doi.org/10.1016/j.ajog.2015.12.055. pii:
S0002-9378(15)02659-9. [Epub ahead of print].
[24] Lauritsen MP, Pinborg A, Loft A, et al. Revised criteria for PCOS in WHO Group II anovulatory infertility e a revival of
hypothalamic amenorrhoea? Clin Endocrinol (Oxf) 2015;82(4):584e91.
[25] Hassan MA, Killick SR. Ultrasound diagnosis of polycystic ovaries in women who have no symptoms of polycystic ovary
syndrome is not associated with subfecundity or subfertility. Fertil Steril 2003;80(4):966e75.
[26] Catteau-Jonard S, Bancquart J, Poncelet E, et al. Polycystic ovaries at ultrasound: normal variant or silent polycystic ovary
syndrome? Ultrasound Obstet Gynecol 2012;40(2):223e9.
[27] Chang PL, Lindheim SR, Lowre C, et al. Normal ovulatory women with polycystic ovaries have hyperandrogenic pituitary-
ovarian responses to gonadotropin-releasing hormone-agonist testing. J Clin Endocrinol Metab 2000;85(3):995e1000.
[28] Sigala J, Sifer C, Dewailly D, et al. Is polycystic ovarian morphology related to a poor oocyte quality after controlled
ovarian hyperstimulation for intracytoplasmic sperm injection? Results from a prospective, comparative study. Fertil
Steril 2015;103(1):112e8.
*[29] Cui L, Li G, Zhong W, et al. Polycystic ovary syndrome susceptibility single nucleotide polymorphisms in women with a
single PCOS clinical feature. Hum Reprod 2015;30(3):732e6.
*[30] Moghetti P, Tosi F, Bonin C, et al. Divergences in insulin resistance between the different phenotypes of the polycystic
ovary syndrome. J Clin Endocrinol Metab 2013;98(4):E628e37.
*[31] Daan NM, Louwers YV, Koster MP, et al. Cardiovascular and metabolic proles amongst different polycystic ovary syn-
drome phenotypes: who is really at risk? Fertil Steril 2014;102(5):1444e51. e3.
[32] Dunaif A, Fauser BC. Renaming PCOSea two-state solution. J Clin Endocrinol Metab 2013;98(11):4325e8.
[33] Elting MW, Korsen TJ, Rekers-Mombarg LT, et al. Women with polycystic ovary syndrome gain regular menstrual cycles
when ageing. Hum Reprod 2000;15(1):24e8.
[34] Kiddy DS, Hamilton-Fairley D, Bush A, et al. Improvement in endocrine and ovarian function during dietary treatment of
obese women with polycystic ovary syndrome. Clin Endocrinol (Oxf) 1992;36(1):105e11.