Anda di halaman 1dari 13

See discussions, stats, and author profiles for this publication at: https://www.researchgate.


Metabolic features of the reproductive phenotypes of polycystic ovary


Article  in  Human Reproduction Update · April 2009

DOI: 10.1093/humupd/dmp008 · Source: PubMed


220 60

2 authors:

Lisa Moran Helena Teede

University of Adelaide Monash University and Monash Partners Academic Health Sciences Centre


Some of the authors of this publication are also working on these related projects:

Insulin resistance in women with PCOS and the role of exercise tehrapy View project

Implementation and evaluation of a perinatal mental health screening program in refugee and migrant women View project

All content following this page was uploaded by Helena Teede on 04 January 2016.

The user has requested enhancement of the downloaded file.

Human Reproduction Update, Vol.15, No.4 pp. 477– 488, 2009
Advanced Access publication on March 11, 2009 doi:10.1093/humupd/dmp008

Metabolic features of the reproductive

phenotypes of polycystic ovary
Lisa Moran 1,3,† and Helena Teede 1,2,†
The Jean Hailes Foundation for Women’s Health Research Unit, Monash Institute of Health Services Research, Monash University, Locked
Bag 29, Clayton, VIC 3168, Australia 2Diabetes Unit, Southern Health, Clayton, VIC 3168, Australia
Correspondence address. Tel: þ61-3-9594-7592; Fax: þ61-3-9594-7550; E-mail:

Downloaded from by guest on December 31, 2015

table of contents
† Overview
† Methods
Diagnostic features of PCOS
† Assessment of specific phenotypes for metabolic features
† Non-NIH PCOS: hyperandrogenic ovulatory PCOS (phenotype C)
Non-NIH PCOS: non-hyperandrogenic anovulatory PCOS (phenotype D)
† Discussion

background: Polycystic ovary syndrome (PCOS) is a common condition in women of reproductive age with well established metabolic
abnormalities. There are numerous diagnostic criteria generating several reproductive diagnostic phenotypes [National Institute of Health
(NIH) hyperandrogenic anovulatory PCOS and non-NIH PCOS including hyperandrogenic ovulatory or non-hyperandrogenic anovulatory
PCOS]. There is ongoing debate regarding the optimal diagnostic criteria for PCOS and on the metabolic implications of newer non-NIH
PCOS phenotypes.
methods: We reviewed the literature on the presence of risk factors for type 2 diabetes (DM2) and cardiovascular disease (CVD) across
the reproductive diagnostic phenotypes of PCOS with the aims of comparing the metabolic features of the NIH and non-NIH groups and
identifying potential high metabolic risk phenotypes of PCOS.
results: NIH PCOS patients present with greater obesity, abdominal obesity, insulin resistance (IR) and risk factors for DM2 and CVD
compared with non-NIH ovulatory and non-hyperandrogenic PCOS patients. Where differences in metabolic features exist between the
phenotypes, they are generally related to the degree of total and abdominal obesity. There is emerging evidence suggesting ovulatory
and non-hyperandrogenic PCOS have greater metabolic abnormalities than controls primarily linked to abdominal adiposity. There is cur-
rently no evidence that non-hyperandrogenic PCOS is associated with a less adverse metabolic profile than ovulatory PCOS.
conclusions: Current metabolic evidence appears to justify the inclusion of both non-NIH PCOS groups (ovulatory and non-
hyperandrogenic) as PCOS subgroups. NIH PCOS is associated with a more adverse metabolic profile including greater total and abdominal
obesity, IR and risk factors for CVD and DM2 than non-NIH phenotypes.

Key words: polycystic ovary syndrome / diagnostic criteria / insulin resistance / hyperandrogenism

et al., 2000; Azziz et al., 2004). It is associated with a range of reproduc-

Overview tive, obstetric, metabolic and psychological features. Reproductive and
Polycystic ovary syndrome (PCOS) is a common endocrine condition obstetric manifestations include hyperandrogenism, menstrual dysfunc-
in women of reproductive age with prevalence estimated at 4 –8% tion, infertility and pregnancy complications. These include early preg-
(Knochenhauer et al., 1998; Diamanti-Kandarakis et al., 1999; Asuncion nancy loss, gestational diabetes, pregnancy-induced hypertensive
†L.M. and H.T. both contributed to article conception, design, drafting, manuscript writing, critical revision and final approval of version to be published.

& The Author 2009. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email:
478 Moran and Teede

disorders and neonatal complications (Boomsma et al., 2006). Meta- In this setting, the aim of this study was to review the literature on
bolic complications include an elevated risk of impaired glucose toler- the metabolic implications (adiposity, abdominal adiposity and risk
ance (IGT), type 2 diabetes (DM2) (Legro et al., 1999), the metabolic factors for DM2 and CVD) across the reproductive diagnostic pheno-
syndrome (Apridonidze et al., 2005), elevated cardiovascular risk types of PCOS. Specifically, we aimed to compare the metabolic fea-
factors (Meyer et al., 2005a, b) and potential increased risk for cardio- tures of the NIH and non-NIH groups and identify potential high
vascular disease (CVD) (Shaw et al., 2008). Women with PCOS also metabolic risk phenotypes of PCOS to target future screening and
have psychological features with increased depression, poor self-image prevention of metabolic diseases. We reviewed the metabolic features
and reduced quality of life (Himelein and Thatcher, 2006). PCOS thus of PCOS with a particular focus on the differences between the repro-
constitutes a significant health and economic burden estimated at ductive diagnostic phenotypes: namely, NIH versus non-NIH PCOS;
over $4 billion in the USA with 30.1% of costs related to hormonal non-NIH PCOS versus non-PCOS controls and the different
treatment of menstrual dysfunction, 12.2% infertility care, 14.2% treat- non-NIH subgroups compared with each other (hyperandrogenic ovu-
ment of hirsutism and 40.5% of DM2 (Azziz et al., 2005). An economic latory PCOS and non-hyperandrogenic anovulatory PCOS). We also
evaluation of PCOS recently advocated screening, diagnosis and inter- examined whether differences in metabolic features between the
vention, justifiable by ameliorating or preventing serious sequelae. This reproductive phenotypes were due to different anthropometrics
is in keeping with the International Diabetes Federation recommen- (total or abdominal adiposity) or differences in features such as IR
dations to focus on early intervention and avoidance or delay or pro- and hyperandrogenism that vary across the diagnostic groups.
gression to DM2 (Alberti et al., 2007). However, before this is To address these aims, the key clinical questions with regard to the

Downloaded from by guest on December 31, 2015

advocated greater understanding of appropriate screening, long-term metabolic features of the PCOS reproductive phenotypes were shown
risks and effective interventions to minimize metabolic morbidity is in Fig. 1.
urgently needed in women with PCOS (Azziz et al., 2005).
Both the reproductive and metabolic features of PCOS are under-
pinned by insulin resistance (IR). This is a key factor in the aetiology of
PCOS, with insulin stimulating ovarian androgen production and
decreasing hepatic sex hormone-binding globulin (SHBG production)
and increasing total and free androgens (Diamanti-Kandarakis and
Papavassiliou, 2006). The majority of PCOS women have underlying
IR, and there is ongoing debate as to whether this IR is intrinsic to
PCOS, related to obesity alone or related to both factors. There is
additionally potentially an increased prevalence of obesity and abdomi-
nal obesity in PCOS (Escobar-Morreale and San Millan, 2007), which
worsens the IR-associated clinical features (Kiddy et al., 1990; Balen
et al., 1995). It is hypothesized that lean women with PCOS therefore
have PCOS-specific IR or intrinsic IR, which is augmented by the pre-
sence of obesity-specific IR (Teede et al., 2007).
There remains considerable controversy on the optimal diagnostic
criteria for PCOS (Azziz, 2006; Franks, 2006). Although the National
Institute of Health (NIH) criteria (hyperandrogenism and anovulation)
were proposed in 1992 (Zawdaki and Dunaif, 1992), these have now
been expanded to include the non-NIH diagnostic criteria. In 2003,
European Society for Human Reproduction and Embryology/
American Society for Reproductive Medicine (ESHRE/ASRM) diag-
nostic criteria were formulated as two of the three criteria of hyper-
androgenism, polycystic ovaries on ultrasound (PCO) and irregular
anovulatory periods. This introduced two new PCOS phenotypes of
hyperandrogenic ovulatory women with PCO or non-hyperandrogenic
anovulatory women with PCO (2004). An amendment to these cri-
teria was proposed in 2006 by the Androgen Excess Society (AES)
to exclude the phenotype of PCO and irregular cycles without hyper-
androgenism (Azziz et al., 2006, 2009). The majority of literature to
date has focused on the NIH diagnosis of PCOS, and research on
the metabolic implications of non-NIH phenotypes is only just emer-
ging. To date, there is limited understanding of the relative prevalence
of risk factors for metabolic diseases (DM2 and CVD) across the
reproductive diagnostic phenotypes of PCOS. Clarification of this
Figure 1 Summary of relevant clinical questions with regard to the
will aid in determining whether to include non-NIH phenotypes as metabolic implications of the reproductive diagnostic phenotypes of
part of the complex condition of PCOS and in identifying if specific PCOS.
reproductive PCOS phenotypes have elevated metabolic risks.
Metabolic features of PCOS phenotypes 479

(measured by equilibrium dialysis) or calculated as a function of total tes-

Methods tosterone and SHBG levels (free androgen index) or estimated from equi-
A review was conducted of the medical literature to identify studies eval- librium binding equations (Vermeulen et al., 1999). Although there is
uating the implications of reproductive diagnostic criteria or reproductive considerable variation between hormone assays and a lack of accurate
phenotypes on the metabolic features of PCOS. Supplementary references ranges for well-defined non-PCOS populations (Hart et al., 2004), elev-
were obtained from initial citations. We searched the database MEDLINE. ated circulating androgens are nevertheless present in 60 – 80% of
Search terms as keywords in article text or as subject headings included women with PCOS (Azziz et al., 2006).
polycystic ovarian syndrome, polycystic ovary syndrome, polycystic As a syndrome, PCOS is by definition a constellation of features with no
ovaries, PCOS or PCO, and diagnostic criteria, diagnosis, presentation specific diagnostic test. There is considerable variability in the diagnosis of
or phenotype with searches limited to females and humans. This search biochemical or clinical hyperandrogenism, menstrual features (anovulation,
resulted in 4512 papers. On screening for title or abstract, we selected amenorrhoea or oligomenorrhoea) and PCO leading to great diversity in
23 for our review that compared metabolic features of PCOS [impaired the clinical presentation of PCOS (Geisthovel and Rabe, 2007). The limit-
fasting glucose (IFG), IGT, DM2, metabolic syndrome, IR, glucose toler- ations and challenges with regard to the need for accurate and consistent
ance, lipid levels, blood pressure and adipokines and family history of cor- definition of each diagnostic parameter (Geisthovel and Rabe, 2007) have
onary artery disease, CVD, DM2 or hypertension] between different been extensively discussed elsewhere. Indeed, the diagnostic criteria for
diagnostic phenotypes (NIH and non-NIH) of PCOS. In addition, we hand- PCOS remain controversial with several proposed criteria still in use, as
searched references of relevant reviews and systematic reviews and detailed in Table I and below.
included studies to locate other potentially eligible studies. The results
were presented as a comparison of (i) NIH PCOS with the non-NIH Diagnostic criteria for PCOS

Downloaded from by guest on December 31, 2015

PCOS diagnostic phenotypes (hyperandrogenic ovulatory PCOS or non-
NIH diagnostic criteria
hyperandrogenic anovulatory PCOS), (ii) non-NIH hyperandrogenic ovula-
NIH diagnostic criteria have been used for the past 15 years based on bio-
tory PCOS or non-hyperandrogenic anovulatory PCOS with non-PCOS
chemical or clinical hyperandrogenism and anovulation (excluding other
controls and (iii) non-NIH hyperandrogenic ovulatory PCOS with non-
secondary causes including thyroid dysfunction, non-congenital adrenal
hyperandrogenic anovulatory PCOS (Fig. 1).
hyperplasia or hyperprolactinaemia) (Zawdaki and Dunaif, 1992). Accord-
ing to these criteria, 4 – 8% of women in a general population have PCOS
Diagnostic features of PCOS (Knochenhauer et al., 1998; Diamanti-Kandarakis et al., 1999; Asuncion
PCOS is a heterogeneous condition, and diagnostic criteria are based on et al., 2000; Azziz et al., 2004). The majority of research has used NIH cri-
reproductive features: anovulation or menstrual disturbance, polycystic teria for diagnosis of PCOS.
ovaries on ultrasound (PCO) and hyperandrogenism. Menstrual disturb-
ances comprise anovulation, amenorrhoea (lack of menstruation for .3 Non-NIH diagnostic criteria: ESHRE/ASRM and AES diagnostic criteria
months) and oligomenorrhoea (irregular menstruation). A recent review In 2003, the ESHRE/ASRM (2004a, b) criteria were developed. These cri-
reported that 78.4% of PCOS patients present with oligomenorrhoea teria diagnosed PCOS based on two of the three criteria of hyperandro-
and 18.1% present with eumenorrhoea, amenorrhoea and anovulation genism, irregular anovulatory periods and PCO, thus introducing two
(Goldzieher, 1981; Franks, 1989; Balen et al., 1995). PCO are defined additional phenotypic subsets (PCO and anovulation/menstrual dysfunc-
as the presence of 12 or more follicles measuring 2 – 9 mm in each tion or hyperandrogenism) (2004). There is currently no data on the
ovary and/or increased ovarian volume (.10 ml or cm3) (Balen et al., prevalence of PCOS as defined by the ESHRE/ASRM criteria in the
2003) and are present in up to 22% of women in the general population general population, although these would be expected to include more
(Farquhar et al., 1994) and 75% of women with PCOS (Azziz et al., 2006). women than NIH criteria. There is some prevalence data in infertile or
Assessment of clinical hyperandrogenism involves subjective clinical anovulatory oligomenorrheic populations; Broekmans et al. (2006) esti-
scoring of degrees of hirsutism (excessive growth of terminal hair in mated ESHRE/ASRM diagnosed PCOS to account for 91% of World
women in a male-like pattern) such as the Ferriman – Gallwey score or Health Organisation II anovulatory infertility compared with 55% for NIH-
later modifications (Ferriman and Gallwey, 1961; Hatch et al., 1981). diagnosed PCOS. Belosi et al. (2006) found that the ESHRE/ASRM criteria
Other hyperandrogenic features include acne (Oberemok and Shalita, expanded the population of women diagnosed with PCOS by over 20%.
2002; Witkowski and Parish, 2004), seborrhoea and androgenic alopecia.
The recommended assessment of biochemical hyperandrogenism is pri- AES diagnostic criteria
marily through measurement of the ovarian androgens testosterone as In 2006, the AES published a position statement which suggested that andro-
total testosterone (free and bound to SHBG), free testosterone gen excess is the key component of PCOS related to clinical presentation and

Table I The different diagnostic criteria for PCOS

NIH criteria Consensus European Society for Human Reproduction and AES Position Statement (Azziz et al., 2006)
Statement (Zawdaki and Dunaif, Embryology/American Society for Reproductive
1992) Medicine Consensus Statement (2004a, b)
Oligo-ovulation and clinical and/or Two out of three oligo- and/or anovulation or clinical Hyperandrogenism (hirsutism and/or
biochemical signs of hyperandrogenism and/or biochemical signs of hyperandrogenism or hyperandrogeniaemia) and ovarian dysfunction
and exclusion of other aetiologies* polycystic ovaries and exclusion of other aetiologies* (oligo-anovulation and/or polycystic ovaries) and
exclusion of other androgen excess related disorders*

*Congenital adrenal hyperplasia, androgen-secreting tumours, Cushing’s syndrome, 21-hydroxylase-deficient non-classic adrenal hyperplasia, androgenic/anabolic drug use or abuse,
syndromes of severe IR, thyroid dysfunction, hyperprolactinaemia.
480 Moran and Teede

For simplification, PCOS can be subdivided into four reproductive pheno-

Table II Comparison of the different reproductive types: NIH- diagnosed PCOS either with (phenotype A) or without PCO
diagnostic criteria for PCOS resulting in potentially (phenotype B); biochemical/clinical hyperandrogenism with PCO but no
different phenotypes oligo/anovulation (phenotype C), or no biochemical/clinical hyperandro-
genism with PCO and oligo/anovulation (phenotype D) (Fig. 1, Table II).
Features Phenotypes Although it has been suggested that these newer non-NIH phenotypes (C
A B C D E F G and D) generally demonstrate a milder reproductive presentation than NIH-
........................................................................................ diagnosed PCOS (A and B) (Belosi et al., 2006), there is conflicting data on
Hyperandrogenism þ þ þ 2 þ 2 2 the metabolic implications of the different phenotypes.
(biochemical or clinical)
Reproductive severity reflected by the number of clinical features of
Oligo- or anovulation þ þ 2 þ 2 2 þ PCOS (menstrual irregularity, acne, hirsutism, elevated testosterone and
Polycystic ovaries þ 2 þ þ 2 þ 2 elevated luteinising hormone) is associated with metabolic severity with
p p increased body fat, waist-to-hip ratio (WHR) and IR (fasting insulin and
NIH criteria
p p p p HOMA-IR) (Michelmore et al., 2001). Hence, with the introduction of
ESHRE/ASRM criteria
p p p the newer diagnostic phenotypes with milder reproductive features of
AES criteria
PCOS, there is the potential for the inclusion of women with milder meta-
bolic presentations and lower long-term reproductive and metabolic risks.
Furthermore, PCOS is associated with impaired psychosocial health
(Himelein and Thatcher, 2006), and the relative contribution of the differ-

Downloaded from by guest on December 31, 2015

long-term morbidity. The AES suggested that the diagnostic criteria be modi-
fied to include only those with hyperandrogenism (biochemical or clinical) ent features of PCOS to this (metabolic, hirsutism, impaired reproductive
and PCO and/or ovarian dysfunction (oligo-anovulation and/or polycystic function, obesity) across the phenotypes is not clear. Thus, potentially the
ovaries) (with the exclusion of other related disorders) (Azziz et al., overall severity of reproductive features may track with metabolic and
2006). This definition excluded the phenotypic subset of PCO and ovarian psychological features, rendering the phenotypes A and B more severe
dysfunction but no hyperandrogenism (Table II). than C and D.
Concerns with the current diagnostic criteria include the fact that the The relative prevalence of the PCOS subgroups are also difficult to
ultrasound criteria are non-specific enough to be observed in other determine from the current literature (Supplementary Appendix I) with
common conditions, including functional hypothalamic anovulation sec- bias introduced by the variable recruitment strategies from endocrinology
ondary to situations such as stress, weight loss or exercise. As noted by or fertility clinics rather than the general population. Depending on the
Dewailly et al. (2006), this may lead to artificial PCOS phenotypes being population recruited from, up to 18% of women with PCOS by ESHRE/
created with medical and psychological implications. However, this classi- ASRM criteria can have non-hyperandrogenic PCOS (D) and up to 25%
fication (World Health Organisation Class I anovulation or hypogonado- of women can have ovulatory PCOS (C). This indicates an increasing
trophic hypogonadal ovulation) is not normogonadotrophic and number of women with PCOS who may experience different reproductive
therefore not World Health Organisation Class II and should not be con- and metabolic risks, when compared with those who have NIH PCOS
sidered as PCOS. Where there is potentially uncertainty, it is important to with potential implications for research, screening and clinical practice.
exclude World Health Organisation I or hypothalamic anovulation. This
can be excluded either with vaginal ultrasound for endometrial thickness
or progestin-induced withdrawal bleed to test for estrogen status. Assessment of specific
Women with relatively mild endocrinological disturbances may also,
potentially, be inappropriately diagnosed with PCOS. For example, a
phenotypes for metabolic
case of acne and PCO in the context of normal biochemical hyperandro- features
genism and ovulation may still be theoretically defined as PCOS. However,
this is controversial and many would consider this group to not have Within these four main phenotypes of PCOS (Fig. 1), it is clinically rel-
PCOS (Norman et al., 2007). evant to understand how the metabolic features track with the repro-
Hyperandrogenism itself, in the absence of ovulatory disturbances, may ductive phenotypes in order to provide insights into how to target
also contribute to an adverse metabolic profile and IR. Emerging evidence metabolic screening and prevention strategies. There are currently
suggests that hyperandrogenism increases abdominal obesity, which in turn few studies that specifically compare the different reproductive diag-
increases IR (Escobar-Morreale and San Millan, 2007). Pre-adipocytes are nostic phenotypes or presentations of PCOS and their metabolic
known to have androgen receptors, and adipose cell function is regulated implications.
by androgens at a mechanistic level. Increased androgens have also been
shown to induce selective IR in cultured adipocytes (Corbould, 2007).
This is further supported by the effects of anti-androgens like spironolac- NIH PCOS: phenotypes A and B
tone, which appear to reduce IR in some (Ganie et al., 2004) but not all IR is a key pathophysiological feature of PCOS and a significant contri-
studies (Sahin et al., 2004; Vrbikova et al., 2004; Gambineri et al., butor to both reproductive and metabolic complications. The link
2006). Furthermore, the metabolic syndrome in women without PCOS
between IR, IGT, DM2 and CVD in the general population is well
has been linked to increasing androgen status at menopause in the
established (Haffner et al., 1990; Lundgren et al., 1990; Lillioja et al.,
setting of increased abdominal obesity (Janssen et al., 2008).
1993; Folsom et al., 1997; Ruige et al., 1998; Lehto et al., 2000;
Rutter et al., 2005), suggesting women with PCOS and IR have an
Implications of different diagnostic criteria elevated risk of these metabolic conditions. Consistent with the
The development of the new ESHRE/ASRM and AES criteria has intro- increased IR, women with PCOS display an increased prevalence of
duced greater heterogeneity into PCOS from a reproductive and possibly IGT and DM2. A recent study reported a 5-fold risk of developing
from a metabolic perspective. DM2 over a period of 8 years in women with PCOS, compared
Metabolic features of PCOS phenotypes 481

with the risk calculated using age- and weight-matched controls Orbetzova, 2007) and IR on euglycaemic hyperinsulinaemic clamp
(Boudreaux et al., 2006). There is also emerging evidence that women (Belosi et al., 2006) for NIH PCOS when compared with ovulatory
with PCOS have a greater chance of developing gestational diabetes, with PCOS or non-hyperandrogenic PCOS. Where NIH and non-NIH
a recent meta-analysis reporting an OR of 2.94 (Boomsma et al., 2006). PCOS women were of similar BMI (Cenk Sayin et al., 2003; Diamanti-
There is currently no good data to establish whether clinical cardio- Kandarakis and Panidis, 2007) and WHR (Diamanti-Kandarakis and
vascular events are increased in PCOS and there are limited long-term Panidis, 2007), similar metabolic features such as surrogate markers of
studies appropriately addressing this question (Pierpoint et al., 1998; IR were generally reported (Cenk Sayin et al., 2003; Diamanti-Kandarakis
Wild et al., 2000; Shaw et al., 2008). However, women with PCOS and Panidis, 2007), although one study reported a less adverse lipid profile
have an increased prevalence of traditional (hyperlipidaemia) and for non-NIH PCOS (Cenk Sayin et al., 2003). Differences in total and
novel (increased homocysteine, inflammation, oxidative stress and leu- abdominal adiposity between NIH PCOS and non-NIH PCOS may there-
kocyte counts and impaired fibrinolysis and metabolic syndrome) risk fore account for differences in IR and metabolic risk. However, these
factors for CVD mostly compared with weight-matched controls. studies did not assess metabolic and anthropometric comparisons
They also display increased clinical and subclinical markers of premature within the non-NIH diagnostic phenotypes to determine the relative
atherosclerosis [endothelial dysfunction, impaired pulse wave velocity degree of metabolic impairment for phenotypes C and D (Fig. 1).
(PWV)]; increased carotid intima-media wall thickness (IMT); presence
of carotid plaque and increased coronary artery calcification, as demon-
strated by our group and others (Apridonidze et al., 2005; Meyer et al., Non-NIH PCOS:

Downloaded from by guest on December 31, 2015

2005a, b; Bhattacharya, 2008; Cussons et al., 2008). PCOS is thus hyperandrogenic ovulatory
defined as a population with a high risk of developing IGT and DM2
and a potentially high risk for CVD. These elevated risks occur indepen- PCOS (phenotype C)
dent of obesity and are worsened by the presence of obesity (Legro
et al., 1999, 2001; Boudreaux et al., 2006; Ehrmann et al., 2006).
Non-NIH PCOS (hyperandrogenic ovulatory
However, the majority of the aforementioned literature on the meta- PCOS) compared with NIH PCOS
bolic implications of PCOS was conducted prior to the introduction of A number of studies have compared NIH PCOS with non-NIH hyper-
the newer non-NIH diagnostic phenotypes. As such, it is not yet androgenic ovulatory PCOS (Fig. 1). The majority of studies report less
known whether these recognized metabolic implications of PCOS are adverse metabolic profiles for ovulatory PCOS. Elevated BMI (Carmina
applicable to all the reproductive diagnostic phenotypes of PCOS. et al., 2005, 2006a, b; Welt et al., 2006b; Barber et al., 2007) and waist
or hip circumference (Welt et al., 2006b; Barber et al., 2007) were
NIH PCOS with or without PCO: phenotypes observed for NIH PCOS compared with hyperandrogenic ovulatory
A and B PCOS in association with significantly increased IR [fasting insulin,
HOMA-IR or quantitative insulin sensitivity check index (QUICKI)]
The majority of women with NIH PCOS also have PCO (phenotype A).
(Carmina et al., 2005; Welt et al., 2006b; Barber et al., 2007) and meta-
Of those that do not, the few studies specifically comparing NIH PCOS
bolic syndrome prevalence (Welt et al., 2006b; Barber et al., 2007) and
with or without PCO (phenotype A versus B) have generally shown
worsened cardiovascular risk factors (elevated triglycerides, hsCRP and
similar glucose tolerance, IR [fasting or oral glucose tolerance test
homocysteine and lower HDL-C) (Carmina et al., 2005).
(OGTT) insulin or insulin sensitivity index] or lipid profiles (Loucks
Not all metabolic parameters differed between phenotypes and no
et al., 2000; Legro et al., 2005; Dewailly et al., 2006; Hahn et al., 2006;
differences in DM2, IFG, glycosylated haemoglobin A1c (HBA1c) and
Welt et al., 2006b; Diamanti-Kandarakis and Panidis, 2007; Hsu et al.,
fasting glucose were observed between ovulatory and NIH PCOS
2007; Shroff et al., 2007). Conversely, in one study increased IR
(Welt et al., 2006b). Furthermore, on adjustment for age and BMI,
[assessed by insulin tolerance test (ITT)] was reported for PCOS with
no differences in lipids, systolic blood pressure (SBP) and diastolic
PCO (phenotype A) compared with PCOS without PCO (phenotype
blood pressure (DBP) (Welt et al., 2006b) or HDL-C and triglycerides
B) (Najmabadi et al., 1997). In addition, higher low-density lipoprotein
(Barber et al., 2007) were observed between subjects. This indicates a
cholesterol (LDL-C), total cholesterol, family history of DM2 (Shi
contribution of elevated adiposity to the worsened metabolic risk
et al., 2008) and triglycerides (Chae et al., 2008) were reported for
factors present in classic NIH PCOS. Age differences also existed
women with NIH PCOS without PCO, when compared with NIH
across the PCOS subsets. This could account for differences in meta-
PCOS with PCO despite similar weight, BMI, waist or hip circumference.
bolic parameters between groups, although specific statistical details
Overall, the evidence suggests a similar risk of metabolic disease for the
were not provided (Barber et al., 2007).
two phenotypes of NIH PCOS.
However, in these studies due to the lack of weight matching, it is
not possible to determine whether the worsened cardiometabolic
Non-NIH PCOS: phenotypes C and D profile in NIH compared with non-NIH ovulatory PCOS is due to
The metabolic literature examining the newer phenotypes of PCOS intro- differences in adiposity or other factors. It is possible that weight
duced by the ESHRE/ASRM is contradictory. There is emerging evidence may be the modulating factor that primarily worsens IR and reproduc-
that these phenotypes have a less adverse metabolic profile than NIH tive and metabolic function in PCOS (Carmina et al., 2005). In support
PCOS. A number of studies have demonstrated elevated BMI, WHR, of this, a small number of studies comparing weight-matched NIH
waist circumference (Belosi et al., 2006; Hsu et al., 2007; Pehlivanov PCOS and non-NIH hyperandrogenic ovulatory PCOS reported simi-
and Orbetzova, 2007), fasting insulin, homeostasis assessment of IR larities in the groups including: similar fasting glucose (Diamanti-
(HOMA-IR), metabolic syndrome prevalence (Pehlivanov and Kandarakis and Panidis, 2007), IR [fasting insulin (Carmina and
482 Moran and Teede

Lobo, 1999; Diamanti-Kandarakis and Panidis, 2007; Shroff et al., differences in IFG, DM2, fasting glucose, lipid profile or family history
2007; Kauffman et al., 2008), OGTT insulin and minimal model of coronary artery disease or DM2 (Shroff et al., 2007). Similarly,
insulin (Kauffman et al., 2008)], QUICKI, glucose-to-insulin ratio (G/ where abdominal adiposity (waist circumference or WHR) was elevated
I) and HOMA-IR) (Shroff et al., 2007), lipid profile (Shroff et al., for ovulatory PCOS compared with age- and weight-matched controls,
2007; Kauffman et al., 2008) IFG, DM2, metabolic syndrome, family greater IR (fasting insulin, G/I or QUICKI) (Carmina and Lobo, 2001;
history of coronary artery disease and DM2 (Shroff et al., 2007). Carmina et al., 2003, 2005; Dewailly et al., 2006) and lipid profile
In contrast to this, NIH PCOS (phenotypes A/B) had significantly (increased hsCRP, total cholesterol and LDL-C or decreased HDL-C)
higher IR [ITT (Robinson et al., 1993), fasting insulin (Sharp et al., were observed (Carmina and Lobo, 2001; Carmina et al., 2005).
1991; Carmina et al., 2003) and QUICKI (Carmina et al., 2003)] Where controls and hyperandrogenic ovulatory women with PCOS
than ovulatory PCOS (phenotype C) despite similar age and weight were closely age- and weight-matched, no differences in lipids, fasting
(Sharp et al., 1991; Robinson et al., 1993; Carmina et al., 2003). or post-OGTT insulin and glucose or minimal model glucose testing
Abdominal obesity was not measured in these studies which could were observed (Kauffman et al., 2008). In another study, higher OGTT
account for the worsened metabolic risk reported in NIH PCOS. glucose and fasting insulin (Robinson et al., 1993) but equivalent fasting
This is supported by the finding that waist circumference is one of insulin and glucose (Robinson et al., 1993; Diamanti-Kandarakis and
the most sensitive markers of the metabolic syndrome in women Panidis, 2007) and surrogate markers of IR (OGTT insulin and ITT)
with NIH PCOS (Ehrmann et al., 2006). In support of this, similar (Robinson et al., 1993) were observed for ovulatory PCOS compared
waist circumferences for NIH compared with age- and BMI-matched with weight-matched controls. The controls were older, potentially con-

Downloaded from by guest on December 31, 2015

ovulatory PCOS were observed in association with similar fasting founding the results (Robinson et al., 1993; Diamanti-Kandarakis and
insulin (Dewailly et al., 2006). Furthermore, when comparing BMI- Panidis, 2007). Furthermore, as abdominal obesity was not measured, it
matched anovulatory and ovulatory PCOS, ovulatory PCOS had is not possible to rule out different visceral obesity between PCOS and
lower abdominal fat [by dual X-ray absorptiometry (DEXA)] with controls. Where abdominal adiposity (WHR or waist circumference)
lower IR (QUICKI, fasting insulin), hsCRP and higher adiponectin was similar between BMI-matched hyperandrogenic ovulatory PCOS
(Carmina et al., 2008). Thus, even when total adiposity is constant, and controls, similar IR (fasting insulin or HOMA-IR) (Norman et al.,
differences in metabolic features between PCOS phenotypes may 1995a; Welt et al., 2006b; Carmina et al., 2008), OGTT insulin
be due to differences in abdominal or visceral fat. (Norman et al., 1995a), QUICKI (Carmina et al., 2008), OGTT glucose
Conversely, Norman et al. (1995a) reported higher post-OGTT and (Norman et al., 1995a), adiponectin (Carmina et al., 2008), SBP, DBP,
fasting insulin and OGTT glucose for NIH compared with ovulatory lipid profile and prevalence of the metabolic syndrome, IFG and DM2
PCOS despite weight and WHR matching, although no differences in (Welt et al., 2006a) are also seen, although in one study controls demon-
total cholesterol, triglycerides and HDL-C between the subsets were strated a lower BMI and higher HDL-C and apolipoprotein A1 (Norman
observed and the relative age of the different subsets was not stated. et al., 1995a). The literature generally supports similar metabolic profiles
Hence, some research suggests reduced IR in ovulatory PCOS is largely for hyperandrogenic ovulatory PCOS and controls where adiposity or
a function of reduced adiposity. Others report higher IR in the NIH phe- abdominal adiposity are matched.
notypes to be either an inherent feature or related to increased abdominal
fat. The lack of measurement or use of a more imprecise measure of adi- Summary of findings for Question 2
posity distribution [waist circumference or WHR (Norman et al., 1995a;
Do hyperandrogenic ovulatory women with PCOS (phenotype C) present
Dewailly et al., 2006) versus DEXA (Carmina et al., 2008)] in the majority
with elevated metabolic risk compared with controls?
of studies limits conclusions in this area.
There is some evidence that ovulatory PCOS are more adversely
metabolically affected than controls, but this is not universally observed and
Summary of findings for Question 1 appears to be strongly related to the presence of adiposity and specifically
abdominal adiposity.
Do hyperandrogenic ovulatory women with PCOS (phenotype C) present
with similar metabolic risk to NIH PCO (phenotype A/B)?
Non-NIH (hyperandrogenic ovulatory) PCOS (phenotype C) generally
have lower body weight and BMI and better metabolic profiles compared
with NIH PCOS (phenotypes A/B). Non-NIH (hyperandrogenic
Non-NIH PCOS: non-hyperandrogenic
ovulatory) PCOS and weight-matched NIH PCOS appear to present with anovulatory PCOS (phenotype D)
similar metabolic risk profiles, particularly where abdominal fat is similar
The most controversial newer non-NIH PCOS subgroup has been
between subjects.
non-hyperandrogenic anovulatory PCOS (phenotype D), which was
included in the ESHRE/ASRM diagnostic criteria and excluded from
the AES diagnostic criteria in the AES position statement (Azziz
Non-NIH PCOS (hyperandrogenic ovulatory et al., 2006).
PCOS) phenotype C compared with
non-PCOS controls Non-NIH PCOS (non-hyperandrogenic
Further assessment of the metabolic risk associated with ovulatory
anovulatory PCOS) phenotype D compared
PCOS can be demonstrated by a comparison of risk factors for DM2 with NIH PCOS
and CVD with controls (Fig. 1). Where BMI is higher in ovulatory There are limited studies examining non-hyperandrogenic women with
PCOS, there is an increased prevalence of metabolic syndrome but no PCOS (Fig. 1). The majority of these show reduced weight, BMI or
Metabolic features of PCOS phenotypes 483

prevalence of obesity (Broekmans et al., 2006; Welt et al., 2006b; Barber

Non-NIH PCOS (non-hyperandrogenic
et al., 2007; Shroff et al., 2007) and reduced WHR and waist circumfer-
ence (Welt et al., 2006b; Barber et al., 2007) for non-hyperandrogenic
anovulatory PCOS) compared with
compared with NIH PCOS (Broekmans et al., 2006; Welt et al., non-PCOS controls
2006b; Barber et al., 2007; Shroff et al., 2007). Of these studies, two Some evidence exists to suggest that non-hyperandrogenic anovulatory
reported lower blood glucose (Broekmans et al., 2006) and less IR (G/ PCOS have an adverse metabolic profile compared with controls
I , 0.25 mmol/mU) (Barber et al., 2007) for non-hyperandrogenic com- (Fig. 1). Where non-hyperandrogenic anovulatory PCOS displayed elev-
pared to NIH PCOS. This again suggests that where metabolic differences ated BMI compared with controls, increased triglycerides, fasting and
occur between phenotypic subsets, they may be accounted for by adi- OGTT insulin, OGTT glucose, lower HDL-C and apolipoprotein A1
posity. When obese subsets were compared, the prevalence of IR are reported (Norman et al., 1995a). Other research has also shown
remained lower in phenotype D PCOS (Broekmans et al., 2006) and elevated WHR for non-hyperandrogenic women with PCOS compared
differences in IR between the PCOS subsets remained on adjustment with BMI-matched controls (Dewailly et al., 2006), suggesting increased
for age and BMI (Barber et al., 2007) suggesting less metabolic impairment abdominal fat independent of total adiposity, although this was not
for non-hyperandrogenic PCOS independent of adiposity. Conversely, an associated with elevated fasting insulin and no other measures of IR
increase in the prevalence of the metabolic syndrome but no differences or metabolic parameters were made. Moreover, when non-
in IR (Welt et al., 2006b; Shroff et al., 2007), IFG, DM2, fasting glucose, hyperandrogenic PCOS displayed similar age, weight, BMI, waist circum-
total cholesterol, LDL-C, family history of coronary artery disease or ference, metabolic syndrome, IFG and DM2, fasting glucose, SPB, DBP

Downloaded from by guest on December 31, 2015

DM2 (Shroff et al., 2007) was reported for non-hyperandrogenic PCOS and lipids to controls, they still demonstrated elevated insulin and
compared with NIH PCOS. However, only fasting measures of IR were HOMA-IR (Welt et al., 2006b), suggesting some metabolic impairment
assessed (fasting insulin, HOMA-IR, QUICKI or G/I), which are less independent of total or abdominal fat.
optimal in PCOS than post-OGTT measures (Ciampelli et al., 2005). Conversely, a comparison of non-hyperandrogenic women with
The literature, although limited, suggests that non-hyperandrogenic ano- PCOS and controls with similar BMI (Shroff et al., 2007; Kauffman
vulatory women with PCOS have a more favourable metabolic profile et al., 2008) and waist circumference (Barber et al., 2007) showed
compared with NIH PCOS, an observation again potentially related to no differences in IR (fasting insulin, QUICKI, G/I or HOMA-IR)
greater adiposity and abdominal adiposity in NIH PCOS. (Barber et al., 2007; Kauffman et al., 2008), lipid profile (Barber
Reduced adiposity and abdominal adiposity contributes to a more et al., 2007; Shroff et al., 2007; Kauffman et al., 2008), prevalence
favourable metabolic profile in non-hyperandrogenic anovulatory of IFG, DM2 (Shroff et al., 2007) or the metabolic syndrome
PCOS. This is confirmed by reports that BMI- or WHR-matched non- (Barber et al., 2007; Shroff et al., 2007) and family history of coronary
hyperandrogenic anovulatory PCOS and NIH PCOS (Norman et al., artery disease and DM2 (Shroff et al., 2007). However, in two of these
1995a; Diamanti-Kandarakis and Panidis, 2007; Kauffman et al., 2008) studies, controls were older, which could worsen their metabolic
generally have similar lipid profiles (Kauffman et al., 2008), fasting and profile independent of their non-PCOS status (Barber et al., 2007;
OGTT glucose (Norman et al., 1995a; Kauffman et al., 2008) and IR Shroff et al., 2007) and abdominal fat distribution was not measured
(fasting insulin, OGTT insulin or QUICKI) (Norman et al., 1995a; in one study (Shroff et al., 2007). Chae et al. (2008) reported that
Diamanti-Kandarakis and Panidis, 2007; Kauffman et al., 2008), despite lower waist circumference for non-hyperandrogenic PCOS
although in one study elevated triglycerides were reported for the non- versus BMI-matched controls, PCOS women displayed elevated SBP,
hyperandrogenic PCOS and it was not always possible to determine if DBP, triglycerides, HOMA-IR, fasting insulin but similar total choles-
the groups were age-matched (Norman et al., 1995a). However, con- terol and HDL-C. However, this study comprises a different ethnic
clusions from this literature are limited as sample size for the non- subgroup (Korean) and there is emerging evidence that ethnicity pro-
hyperandrogenic PCOS phenotype was often very small (Norman foundly affects metabolic parameters in PCOS (Norman et al., 1995b;
et al., 1995a). Kauffman et al. (2008) also excluded women with DM2 Kauffman et al., 2002; Wijeyaratne et al., 2002, 2004; Al-Fozan et al.,
and very obese women, potentially removing a subgroup with greater 2005; Carmina et al., 2006a, b; Ehrmann et al., 2006; Kauffman et al.,
metabolic abnormalities that may have differed between the PCOS 2006; Essah et al., 2008).
phenotypic subsets. Furthermore, when non-hyperandrogenic PCOS The evidence for worsened risk factors for DM2 and CVD in non-
present with lower waist circumference or WHR, despite being weight- hyperandrogenic PCOS compared with controls is therefore limited.
matched with NIH PCOS, they also display reduced fasting (Dewailly Limited data report that IR, independent of adiposity or abdominal
et al., 2006; Chae et al., 2008) or OGTT insulin (Chae et al., 2008). adiposity, may exist in the absence of hyperandrogenism in PCOS
Overall, this supports the strong role of abdominal adiposity in the with this IR potentially contributing to the reproductive features of
pathogenesis of IR in PCOS. PCOS. Additionally, despite this subgroup of PCOS being less hyper-
androgenic than the other diagnostic phenotypes, they can still present
Summary of findings for Question 3 with either similar (Norman et al., 1995a; Barber et al., 2007; Chae
et al., 2008; Kauffman et al., 2008) or higher androgens (Dewailly
Do non-hyperandrogenic anovulatory women with PCOS (phenotype D)
et al., 2006; Welt et al., 2006b; Hsu et al., 2007) than controls that
present with similar metabolic risk to NIH PCOS (phenotype A/B)?
could independently impact on IR, reproductive and metabolic par-
Non-hyperandrogenic PCOS (phenotype D) generally present with lower
BMI and less adverse metabolic profiles compared with NIH PCOS ameters. One possibility is that the severity of hyperandrogenism
(phenotype A/B). Non-hyperandrogenic PCOS and NIH PCOS matched for varies in women with PCOS, and while hyperandrogenism drives
total and abdominal obesity generally present with similar metabolic profiles. PCOS in the majority of phenotypes (A, B and C), in women with a
mild abnormality in androgens (phenotype D or non-hyperandrogenic
484 Moran and Teede

PCOS), a greater contribution of inherent or environmentally Limitations of the existing literature

induced abdominal-obesity-related IR may be required to induce
The studies discussed above include a wide range of ethnic groups
reproductive and ovarian dysfunction (Escobar-Morreale and San
(Caucasian of American, Australian, European or the UK origin,
Millan, 2007).
Mexican American, African American, Asian, Non-Mexican American
Hispanic, Asian, Arab/Mediterranean, Afro-Caribbean of UK origin,
Summary of findings for Question 4 Chinese, the Netherlands and Icelandic (Supplementary Appendix
I)]. Overall the literature suggests that different ethnicities of PCOS
Do non-hyperandrogenic anovulatory women with PCOS (phenotype D)
present with elevated metabolic risk compared with controls? present with differences in metabolic features, adiposity, abdominal
adiposity, hyperandrogenism, hirsutism, adrenal hormones, lipid
There is currently little evidence to indicate non-hyperandrogenic women
with PCOS matched for abdominal obesity have a more adverse metabolic profile, homocysteine or IR (Norman et al., 1995b; Kauffman et al.,
profile than controls. 2002, 2006; Wijeyaratne et al., 2002, 2004; Al-Fozan et al., 2005;
Carmina et al., 2006a, b; Ehrmann et al., 2006; Essah et al., 2008),
although this is not universally observed (Welt et al., 2006a).
Further variability in metabolic risk factors in PCOS may relate to
Hyperandrogenic ovulatory PCOS the number of clinical features and the majority of studies do not cat-
phenotype C compared with non-NIH PCOS egorize PCOS populations based on number or severity of features.
(non-hyperandrogenic anovulatory PCOS) For example, those with biochemical and clinical hyperandrogenism

Downloaded from by guest on December 31, 2015

phenotype D and oligo-ovulation were the most severely metabolically affected
and subjects with clinical hyperandrogenism only had lower prevalence
There are even fewer studies comparing the non-NIH phenotypes of
of IR (Yildiz and Gedik, 2001; Chang et al., 2005; Diamanti-Kandarakis
hyperandrogenic ovulatory PCOS (phenotype C) and non-
and Panidis, 2007), although this is not reported by all investigators
hyperandrogenic anovulatory PCOS (phenotype D) and the literature
(Hassa et al., 2006). Control subjects in these studies were also vari-
is very conflicting (Fig. 1). No differences in family histories of coronary
ably defined. A number were precisely defined as regular ovulatory
artery disease, DM2, prevalence of IFG and DM2 (Shroff et al., 2007),
menses with no clinical or biochemical hyperandrogenism; some
IR [G/I, QUICKI, HOMA-IR (Shroff et al., 2007), fasting insulin (Shroff
only specified non-hirsute and did not assess biochemical hyperandro-
et al., 2007; Kauffman et al., 2008) or OGTT insulin (Kauffman et al.,
genism and not all assessed the presence of PCO in control subjects.
2008)] or lipid profile (Kauffman et al., 2008) were reported for
Although not uniformly observed (Michelmore et al., 2001; Legro
heavier (Shroff et al., 2007) or weight-matched (Kauffman et al.,
et al., 2005), women with PCO only may have adverse reproductive
2008) groups. The lack of measurement of abdominal adiposity in
and metabolic parameters reported in association with increased
these studies limits the conclusions that can be drawn. However, in
obesity (Clayton et al., 1992), androgens (Kousta et al., 1999;
another study, the strong effect of abdominal obesity on IR and meta-
Adams et al., 2004), risk factors for CVD including reduced arterial
bolic profiles in the PCOS phenotypes was confirmed by lower waist
compliance (Lakhani et al., 2002) and surrogate markers of IR
circumferences and fasting insulin levels for non-hyperandrogenic
(Adams et al., 2004), potentially influencing the metabolic status of
women with PCOS compared with age- and BMI-matched ovulatory
the control subjects. Furthermore, despite the recognition of IR as a
PCOS (Dewailly et al., 2006). When adiposity and abdominal adiposity
contributor to the metabolic presentation of PCOS, there is a
(WHR or waist circumference) were matched, one study reported
lack of consensus on the agreed defined biochemical cut off, the
similar SBP, DBP, HBA1c, metabolic syndrome, IFG, DM2, fasting
appropriate method to use its measurement and longitudinal
insulin, HOMA-IR and lipid profile (Welt et al., 2006b). An earlier
studies to validate surrogate measures (Ciampelli et al., 2005). The
study with a very small number of non-hyperandrogenic PCOS (n ¼
measurement of IR remains a useful research tool in women with
6) reported elevated triglycerides, fasting insulin and OGTT glucose
PCOS but is not currently recommended for routine clinical practice
and insulin (Norman et al., 1995a) for non-hyperandrogenic PCOS
(Samaras et al., 2006). The impact of ethnicity, more precisely
compared with ovulatory PCOS. This indicates higher rather than
defined PCOS and control populations as well as the use of
lower metabolic risk factors in non-hyperandrogenic ovulatory
more accurate methodologies to examine androgen status, IR
PCOS. From this limited literature, non-hyperandrogenic anovulatory
and abdominal fat distribution are needed to clarify the metabolic
PCOS do not display improved metabolic risk factors compared
implications of PCOS.
with hyperandrogenic ovulatory PCOS. There is thus currently
limited evidence to support the exclusion of non-hyperandrogenic
PCOS as a phenotype of PCOS based on metabolic presentation. Discussion
Summary of findings for Question 5 There is an increasing body of literature devoted to examining the
metabolic implications of the reproductive diagnostic phenotypes of
Do hyperandrogenic ovulatory women with PCOS (non-NIH phenotype PCOS. The evidence currently suggests that those with NIH PCOS,
C) present with similar metabolic risk to non-hyperandrogenic anovulatory
who are hyperandrogenic and generally IR, have the most severe
women with PCOS (non-NIH phenotype D)?
metabolic features. The adverse metabolic profile is related to
There is currently little evidence to indicate non-hyperandrogenic women
with PCOS have a less adverse metabolic profile than ovulatory PCOS.
obesity and abdominal obesity, which appears to be more marked
in NIH PCOS than in other non-NIH phenotypes. Hyperandrogenism
Metabolic features of PCOS phenotypes 485

and IR have both been implicated in adverse metabolic profiles and are In conclusion, abdominal obesity appears to be the primary deter-
both likely to underpin the metabolic phenotypes of women with minant of metabolic abnormalities in PCOS. IR and hyperandrogenism
PCOS, either directly or through an increased predilection for may also, either directly or indirectly, influence metabolic abnormal-
abdominal obesity. The newer reproductive phenotypes (ovulatory ities and potentially contribute to abdominal obesity. However,
PCOS and non-hyperandrogenic PCOS) have milder metabolic phe- more research is needed to provide a greater understanding of the
notypes than NIH PCOS and again these are strongly related to interaction between hyperandrogenism, IR and abdominal adiposity
abdominal adiposity. While the ovulatory subgroup is more hyperan- in PCOS. The optimal screening and treatment for each PCOS pheno-
drogenic, evidence suggests the non-hyperandrogenic group has type and their relatives also needs to be better understood. Overall,
similar IR than ovulatory PCOS. In the setting of either hyperandro- there remain considerable gaps in the literature which require
genism or IR, metabolic abnormalities are observed. This review pro- further research. Ultimately, classification of the metabolic compli-
vides insights into the pathophysiology of metabolic abnormalities in cations for each phenotype will provide an evidence base for screening
PCOS. It is possible that intrinsic PCOS-related abnormalities in of metabolic risk upon diagnosis of PCOS and may guide optimal
insulin signalling and/or a predisposition to visceral weight gain may treatment to prevent metabolic complications of PCOS. Insights in
be greater in the NIH phenotypes contributing to their more severe these areas across the phenotypes will also assist in understanding
reproductive and metabolic presentation. Accumulation of obesity the underlying pathophysiology of metabolic features of this condition.
and abdominal or visceral obesity then further exacerbates the meta-
bolic abnormalities seen in PCOS.
Supplementary data

Downloaded from by guest on December 31, 2015

There is ongoing controversy over the diagnostic criteria for PCOS.
The inclusion of ovulatory PCOS, and even more so of non- Supplementary data are available at http://humupd.oxfordjournals.
hyperandrogenic PCOS, has received criticism (Azziz et al., 2006). org/.
From a metabolic perspective, although the literature is limited, it
appears that both of the newer PCOS phenotypes (ovulatory and
non hyperandrogenic) demonstrate a less adverse reproductive and
metabolic profile, compared with NIH PCOS. However, when weight- Adams JM, Taylor AE, Crowley WF Jr, Hall JE. Polycystic ovarian
matched the metabolic profile of the newer phenotypes is similar to morphology with regular ovulatory cycles: insights into the
the profile seen in NIH phenotypes. The newer phenotypes also pathophysiology of polycystic ovarian syndrome. J Clin Endocrinol
appear to have a more adverse metabolic profile than controls, but Metab 2004;89:4343 – 4350.
this is not universally observed and is strongly related to adiposity Alberti KG, Zimmet P, Shaw J. International Diabetes Federation: a
consensus on type 2 diabetes prevention. Diabet Med 2007;
and abdominal adiposity. Finally, non-hyperandrogenic women
24:451– 463.
appear to have a similar metabolic profile to ovulatory PCOS with
Al-Fozan H, Al-Futaisi A, Morris D, Tulandi T. Insulin responses to the oral
limited evidence even suggesting they present with more severe IR glucose tolerance test in women of different ethnicity with polycystic
and dyslipidaemia. Therefore, we propose that the metabolic litera- ovary syndrome. J Obstet Gynaecol Can 2005;27:33 – 37.
ture in PCOS supports the inclusion of both the newer phenotypes Apridonidze T, Essah PA, Iuorno MJ, Nestler JE. Prevalence and
of PCOS based on the ESHRE/ASRM Rotterdam diagnostic criteria characteristics of the metabolic syndrome in women with polycystic
and suggests these phenotypes are a more mild form of PCOS. ovary syndrome. J Clin Endocrinol Metab 2005;90:1929 – 1935.
The limitations of the literature on metabolic features of PCOS are Asuncion M, Calvo RM, San Millan JL, Sancho J, Avila S,
not insignificant. Ethnic diversity, recruitment sources of participants, Escobar-Morreale HF. A prospective study of the prevalence of the
consistency in the use of end-points and inconsistently defined con- polycystic ovary syndrome in unselected Caucasian women from
trols are but a few of the concerns. Current research has also primarily Spain. J Clin Endocrinol Metab 2000;85:2434 – 2438.
Azziz R. Controversy in clinical endocrinology: diagnosis of polycystic
focused on metabolic syndrome features or on surrogate markers of
ovarian syndrome: the Rotterdam criteria are premature. J Clin
DM2 and CVD. No studies have adequately examined the natural
Endocrinol Metab 2006;91:781 – 785.
history of metabolic disease in PCOS across the different diagnostic Azziz R, Woods KS, Reyna R, Key TJ, Knochenhauer ES, Yildiz BO. The
phenotypes of women, especially in the newer phenotypes. The prevalence and features of the polycystic ovary syndrome in an
majority of research has additionally studied risk factors for metabolic unselected population. J Clin Endocrinol Metab 2004;89:2745– 2749.
disease and only a few papers examined clinical disease outcomes (e.g. Azziz R, Marin C, Hoq L, Badamgarav E, Song P. Health care-related
DM2, coronary artery disease, subclinical or clinical atherosclerosis) economic burden of the polycystic ovary syndrome during the
(Welt et al., 2006b; Barber et al., 2007; Pehlivanov and Orbetzova, reproductive life span. J Clin Endocrinol Metab 2005;90:4650– 4658.
2007; Shroff et al., 2007; Chae et al., 2008). Robust, well designed Azziz R, Carmina E, Dewailly D, Diamanti-Kandarakis E,
studies with well-defined controls and longitudinal follow-up to Escobar-Morreale HF, Futterweit W, Janssen OE, Legro RS,
capture clinical outcomes are needed to resolve many of the outstand- Norman RJ, Taylor AE et al. Position statement: criteria for defining
polycystic ovary syndrome as a predominantly hyperandrogenic
ing questions on the metabolic phenotype of PCOS. Furthermore, the
syndrome: an Androgen Excess Society guideline. J Clin Endocrinol
majority of the literature reports elevated risk factors for DM2 and
Metab 2006;91:4237 – 4245.
CVD in family members of women with PCOS (Yilmaz et al., 2005; Azziz R, Carmina E, Dewailly D, Diamanti-Kandarakis E,
Sam et al., 2006; Baillargeon and Carpentier, 2007; Crisosto et al., Escobar-Morreale HF, Futterweit W, Janssen OE, Legro RS,
2007; Unluhizarci et al., 2007), but this is poorly examined across Norman RJ, Taylor AE et al. The Androgen Excess and PCOS Society
the diagnostic phenotypes of PCOS (Shi et al., 2008; Shroff et al., criteria for the polycystic ovary syndrome: the complete task force
2007; Chae et al., 2008). report. Fertil Steril 2009;91:456– 488.
486 Moran and Teede

Baillargeon J, Carpentier AC. Brothers of women with polycystic ovary characterization of the three major clinical subgroups. Fertil Steril 2005;
syndrome are characterised by impaired glucose tolerance, reduced 83:1717 – 1723.
insulin sensitivity and related metabolic defects. Diabetologia 2007; Ciampelli M, Leoni F, Cucinelli F, Mancuso S, Panunzi S, De Gaetano A,
50:2424 – 2432. Lanzone A. Assessment of insulin sensitivity from measurements in
Balen AH, Conway GS, Kaltsas G, Techatrasak K, Manning PJ, West C, the fasting state and during an oral glucose tolerance test in polycystic
Jacobs HS. Polycystic ovary syndrome: the spectrum of the disorder ovary syndrome and menopausal patients. J Clin Endocrinol Metab
in 1741 patients. Hum Reprod 1995;10:2107 – 2111. 2005;90:1398 – 1406.
Balen AH, Laven JS, Tan SL, Dewailly D. Ultrasound assessment of the Clayton RN, Ogden V, Hodgkinson J, Worswick L, Rodin DA, Dyer S,
polycystic ovary: international consensus definitions. Hum Reprod Meade TW. How common are polycystic ovaries in normal women
Update 2003;9:505– 514. and what is their significance for the fertility of the population? Clin
Barber TM, Wass JA, McCarthy MI, Franks S. Metabolic characteristics of Endocrinol (Oxf) 1992;37:127 – 134.
women with polycystic ovaries and oligo-amenorrhoea but normal Corbould A. Chronic testosterone treatment induces selective insulin
androgen levels: implications for the management of polycystic ovary resistance in subcutaneous adipocytes of women. J Endocrinol 2007;
syndrome. Clin Endocrinol (Oxf) 2007;66:513 – 517. 192:585– 594.
Belosi C, Selvaggi L, Apa R, Guido M, Romualdi D, Fulghesu AM, Crisosto N, Codner E, Maliqueo M, Echiburu B, Sanchez F, Cassorla F,
Lanzone A. Is the PCOS diagnosis solved by ESHRE/ASRM. Hum Sir-Petermann T. Anti-Mullerian hormone levels in peripubertal
Reprod 2006;21:3108 – 3115. daughters of women with polycystic ovary syndrome. J Clin Endocrinol
Bhattacharya SM. Metabolic syndrome in females with polycystic ovary Metab 2007;92:2739 – 2743.
syndrome and International Diabetes Federation criteria. J Obstet Cussons AJ, Watts GF, Burke V, Shaw JE, Zimmet PZ, Stuckey BG.

Downloaded from by guest on December 31, 2015

Gynaecol Res 2008;34:62– 66. Cardiometabolic risk in polycystic ovary syndrome: a comparison of
Boomsma CM, Eijkemans MJ, Hughes EG, Visser GH, Fauser BC, different approaches to defining the metabolic syndrome. Hum Reprod
Macklon NS. A meta-analysis of pregnancy outcomes in women 2008;23:2352 – 2358.
with polycystic ovary syndrome. Hum Reprod Update 2006;12:673– 683. Dewailly D, Catteau-Jonard S, Reyss AC, Leroy M, Pigny P.
Boudreaux MY, Talbott EO, Kip KE, Brooks MM, Witchel SF. Risk of Oligoanovulation with polycystic ovaries but not overt
T2DM and impaired fasting glucose among PCOS subjects: results of hyperandrogenism. J Clin Endocrinol Metab 2006;91:3922 – 3927.
an 8-year follow-up. Curr Diab Rep 2006;6:77 – 83. Diamanti-Kandarakis E, Panidis D. Unravelling the phenotypic map of
Broekmans FJ, Knauff EA, Valkenburg O, Laven JS, Eijkemans MJ, polycystic ovary syndrome (PCOS): a prospective study of 634
Fauser BC. PCOS according to the Rotterdam consensus criteria: women with PCOS. Clin Endocrinol (Oxf) 2007;67:735 –742.
change in prevalence among WHO-II anovulation and association Diamanti-Kandarakis E, Papavassiliou AG. Molecular mechanisms of insulin
with metabolic factors. BJOG 2006;113:1210 – 1217. resistance in polycystic ovary syndrome. Trends Mol Med 2006;
Carmina E, Lobo RA. Do hyperandrogenic women with normal menses 12:324 – 332.
have polycystic ovary syndrome? Fertil Steril 1999;71:319 – 322. Diamanti-Kandarakis E, Kouli CR, Bergiele AT, Filandra FA, Tsianateli TC,
Carmina E, Lobo RA. Polycystic ovaries in hirsute women with normal Spina GG, Zapanti ED, Bartzis MI. A survey of the polycystic ovary
menses. Am J Med 2001;111:602 – 606. syndrome in the Greek island of Lesbos: hormonal and metabolic
Carmina E, Longo A, Lobo RA. Does ovarian blood flow distinguish profile. J Clin Endocrinol Metab 1999;84:4006 – 4011.
between ovulatory and anovulatory patients with polycystic ovary Ehrmann DA, Liljenquist DR, Kasza K, Azziz R, Legro RS, Ghazzi MN,
syndrome? Am J Obstet Gynecol 2003;189:1283 – 1286. Group PCTS. Prevalence and predictors of the metabolic syndrome in
Carmina E, Chu MC, Longo RA, Rini GB, Lobo RA. Phenotypic variation in women with polycystic ovary syndrome. J Clin Endocrinol Metab 2006;
hyperandrogenic women influences the findings of abnormal metabolic 91:48 – 53.
and cardiovascular risk parameters. J Clin Endocrinol Metab 2005; Escobar-Morreale HF, San Millan JL. Abdominal adiposity and the
90:2545 – 2549. polycystic ovary syndrome. Trends Endocrinol Metab 2007;18:
Carmina E, Napoli N, Longo RA, Rini GB, Lobo RA. Metabolic syndrome 266 – 272.
in polycystic ovary syndrome (PCOS): lower prevalence in southern Italy ESHRE/ASRM. Revised 2003 consensus on diagnostic criteria and
than in the USA and the influence of criteria for the diagnosis of PCOS. long-term health risks related to polycystic ovary syndrome.
Eur J Endocrinol 2006a;154:141– 145. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop
Carmina E, Rosato F, Janni A, Rizzo M, Longo RA. Extensive clinical Group. Fertil Steril 2004a;81:19 – 25.
experience: relative prevalence of different androgen excess disorders ESHRE/ASRM. Revised 2003 consensus on diagnostic criteria and
in 950 women referred because of clinical hyperandrogenism. J Clin long-term health risks related to polycystic ovary syndrome (PCOS).
Endocrinol Metab 2006b;91:2– 6. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop
Carmina E, Bucchieri S, Mansueto P, Rini G, Ferin M, Lobo RA. Circulating Group. Hum Reprod 2004b;19:41 –47.
levels of adipose products and differences in fat distribution in the Essah PA, Nestler JE, Carmina E. Differences in dyslipidemia between
ovulatory and anovulatory phenotypes of polycystic ovary syndrome. American and Italian women with polycystic ovary syndrome.
Fertil Steril 2008 Epub ahead of print. J Endocrinol Invest 2008;31:35 – 41.
Cenk Sayin N, Gucer F, Balkanli-Kaplan P, Ali Yuce M, Yardim T. Insulin Farquhar CM, Birdsall M, Manning P, Mitchell JM, France JT. The prevalence
resistance and lipid profile in women with polycystic appearing of polycystic ovaries on ultrasound scanning in a population of randomly
ovaries: implications with regard to polycystic ovary syndrome. selected women. Aust N Z J Obstet Gynaecol 1994;34:67– 72.
Gynecol Endocrinol 2003;17:387 – 396. Ferriman D, Gallwey JD. Clinical assessment of body hair growth in
Chae SJ, Kim JJ, Choi YM, Hwang KR, Jee BC, Ku SY, Suh CS, Kim SH, women. J Clin Endocrinol Metab 1961;21:1440 – 1447.
Kim JG, Moon SY. Clinical and biochemical characteristics of polycystic Folsom AR, Szklo M, Stevens J, Liao F, Smith R, Eckfeldt JH. A prospective
ovary syndrome in Korean women. Hum Reprod 2008;23:1924–1931. study of coronary heart disease in relation to fasting insulin, glucose, and
Chang WY, Knochenhauer ES, Bartolucci AA, Azziz R. Phenotypic diabetes. The Atherosclerosis Risk in Communities (ARIC) Study.
spectrum of polycystic ovary syndrome: clinical and biochemical Diabetes Care 1997;20:935 – 942.
Metabolic features of PCOS phenotypes 487

Franks S. Polycystic ovary syndrome: a changing perspective. Clin Endocrinol Knochenhauer ES, Key TJ, Kahsar-Miller M, Waggoner W, Boots LR,
(Oxf) 1989;31:87 – 120. Azziz R. Prevalence of the polycystic ovary syndrome in unselected
Franks S. Controversy in clinical endocrinology: diagnosis of polycystic black and white women of the southeastern United States: a
ovarian syndrome: in defense of the Rotterdam criteria. J Clin prospective study. J Clin Endocrinol Metab 1998;83:3078 – 3082.
Endocrinol Metab 2006;91:786 – 789. Kousta E, White DM, Cela E, McCarthy MI, Franks S. The prevalence of
Gambineri A, Patton L, Vaccina A, Cacciari M, Morselli-Labate AM, polycystic ovaries in women with infertility. Hum Reprod 1999;
Cavazza C, Pagotto U, Pasquali R. Treatment with flutamide, 14:2720– 2723.
metformin, and their combination added to a hypocaloric diet in Lakhani K, Seifalian AM, Hardiman P. Impaired carotid viscoelastic
overweight-obese women with polycystic ovary syndrome: a properties in women with polycystic ovaries. Circulation 2002;
randomized, 12-month, placebo-controlled study. J Clin Endocrinol 106:81– 85.
Metab 2006;91:3970 – 3980. Legro RS, Kunselman AR, Dodson WC, Dunaif A. Prevalence and
Ganie MA, Khurana ML, Eunice M, Gupta N, Gulati M, Dwivedi SN, predictors of risk for type 2 diabetes mellitus and impaired glucose
Ammini AC. Comparison of efficacy of spironolactone with metformin tolerance in polycystic ovary syndrome: a prospective, controlled
in the management of polycystic ovary syndrome: an open-labeled study in 254 affected women. J Clin Endocrinol Metab 1999;84:165– 169.
study. J Clin Endocrinol Metab 2004;89:2756– 2762. Legro RS, Kunselman AR, Dunaif A. Prevalence and predictors of
Geisthovel F, Rabe T. The ESHRE/ASRM consensus on polycystic ovary dyslipidemia in women with polycystic ovary syndrome. Am J Med
syndrome (PCOS) – an extended critical analysis. Reprod Biomed Online 2001;111:607 – 613.
2007;14:522 – 535. Legro RS, Chiu P, Kunselman AR, Bentley CM, Dodson WC, Dunaif A.
Goldzieher JW. Polycystic ovarian disease. Fertil Steril 1981;35:371 – 394. Polycystic ovaries are common in women with hyperandrogenic

Downloaded from by guest on December 31, 2015

Haffner SM, Stern MP, Mitchell BD, Hazuda HP, Patterson JK. Incidence of type chronic anovulation but do not predict metabolic or reproductive
II diabetes in Mexican Americans predicted by fasting insulin and glucose phenotype. J Clin Endocrinol Metab 2005;90:2571 – 2579.
levels, obesity, and body-fat distribution. Diabetes 1990;39:283–288. Lehto S, Ronnemaa T, Pyorala K, Laakso M. Cardiovascular risk factors
Hahn S, Bering van Halteren W, Roesler S, Schmidt M, Kimmig R, Tan S, clustering with endogenous hyperinsulinaemia predict death from
Mann K, Janssen OE. The combination of increased ovarian volume and coronary heart disease in patients with Type II diabetes. Diabetologia
follicle number is associated with more severe hyperandrogenism in 2000;43:148 – 155.
German women with polycystic ovary syndrome. Exp Clin Endocrinol Lillioja S, Mott DM, Spraul M, Ferraro R, Foley JE, Ravussin E,
Diabetes 2006;114:175– 181. Knowler WC, Bennett PH, Bogardus C. Insulin resistance and insulin
Hart R, Hickey M, Franks S. Definitions, prevalence and symptoms of secretory dysfunction as precursors of non-insulin-dependent diabetes
polycystic ovaries and polycystic ovary syndrome. Best Pract Res Clin mellitus. Prospective studies of Pima Indians. N Engl J Med 1993;
Obstet Gynaecol 2004;18:671 – 683. 329:1988– 1992.
Hassa H, Tanir HM, Yildiz Z. Comparison of clinical and laboratory Loucks TL, Talbott EO, McHugh KP, Keelan M, Berga SL, Guzick DS. Do
characteristics of cases with polycystic ovarian syndrome based on polycystic-appearing ovaries affect the risk of cardiovascular disease
Rotterdam’s criteria and women whose only clinical signs are oligo/ among women with polycystic ovary syndrome? Fertil Steril 2000;
anovulation or hirsutism. Arch Gynecol Obstet 2006;274:227 –232. 74:547– 552.
Hatch R, Rosenfield RL, Kim MH, Tredway D. Hirsutism: implications, Lundgren H, Bengtsson C, Blohme G, Lapidus L, Waldenstrom J. Fasting
etiology, and management. Am J Obstet Gynecol 1981;140:815 – 830. serum insulin concentration and early insulin response as risk
Himelein MJ, Thatcher SS. Polycystic ovary syndrome and mental health: a determinants for developing diabetes. Diabet Med 1990;7:407 – 413.
review. Obstet Gynecol Surv 2006;61:723 – 732. Meyer C, McGrath BP, Cameron J, Kotsopoulos D, Teede HJ. Vascular
Hsu MI, Liou TH, Chou SY, Chang CY, Hsu CS. Diagnostic criteria dysfunction and metabolic parameters in polycystic ovary syndrome.
for polycystic ovary syndrome in Taiwanese Chinese women: J Clin Endocrinol Metab 2005a;90:4630– 4635.
comparison between Rotterdam 2003 and NIH 1990. Fertil Steril Meyer C, McGrath BP, Teede HJ. Overweight women with polycystic
2007;88:727 – 729. ovary syndrome have evidence of subclinical cardiovascular disease.
Janssen I, Powell LH, Crawford S, Lasley B, Sutton-Tyrrell K. Menopause J Clin Endocrinol Metab 2005b;90:5711– 5716.
and the metabolic syndrome: the Study of Women’s Health across Michelmore K, Ong K, Mason S, Bennett S, Perry L, Vessey M, Balen A,
the Nation. Arch Intern Med 2008;168:1568– 1575. Dunger D. Clinical features in women with polycystic ovaries:
Kauffman RP, Baker VM, Dimarino P, Gimpel T, Castracane VD. Polycystic relationships to insulin sensitivity, insulin gene VNTR and birth weight.
ovarian syndrome and insulin resistance in white and Mexican American Clin Endocrinol (Oxf) 2001;55:439 – 446.
women: a comparison of two distinct populations. Am J Obstet Gynecol Najmabadi S, Wilcox JG, Acacio BD, Thornton MH, Kolb BA, Paulson RJ.
2002;187:1362 – 1369. The significance of polycystic-appearing ovaries versus normal-appearing
Kauffman RP, Baker VM, DiMarino P, Castracane VD. Hyperinsulinemia ovaries in patients with polycystic ovary syndrome. Fertil Steril 1997;
and circulating dehydroepiandrosterone sulfate in white and Mexican 67:631– 635.
American women with polycystic ovary syndrome. Fertil Steril 2006; Norman RJ, Hague WM, Masters SC, Wang XJ. Subjects with polycystic
85:1010– 1016. ovaries without hyperandrogenaemia exhibit similar disturbances in
Kauffman RP, Baker TE, Baker VM, DiMarino P, Castracane VD. Endocrine and insulin and lipid profiles as those with polycystic ovary syndrome.
metabolic differences among phenotypic expressions of polycystic ovary Hum Reprod 1995a;10:2258– 2261.
syndrome according to the 2003 Rotterdam consensus criteria. Am Norman RJ, Mahabeer S, Masters S. Ethnic differences in insulin and
J Obstet Gynecol 2008;198:670.e1–670.e7; discussion 670.e7–670.e10. glucose response to glucose between white and Indian women with
Kiddy DS, Sharp PS, White DM, Scanlon MF, Mason HD, Bray CS, polycystic ovary syndrome. Fertil Steril 1995b;63:58 – 62.
Polson DW, Reed MJ, Franks S. Differences in clinical and endocrine Norman RJ, Dewailly D, Legro RS, Hickey TE. Polycystic ovary syndrome.
features between obese and non-obese subjects with polycystic ovary Lancet 2007;370:685 – 697.
syndrome: an analysis of 263 consecutive cases. Clin Endocrinol (Oxf) Oberemok SS, Shalita AR. Acne vulgaris, I: pathogenesis and diagnosis.
1990;32:213 – 220. Cutis 2002;70:101 – 105.
488 Moran and Teede

Pehlivanov B, Orbetzova M. Characteristics of different phenotypes of Unluhizarci K, Ozocak M, Tanriverdi F, Atmaca H, Kelestimur F.
polycystic ovary syndrome in a Bulgarian population. Gynecol Investigation of hypothalamo-pituitary-gonadal axis and glucose
Endocrinol 2007;23:604 – 609. intolerance among the first-degree female relatives of women with
Pierpoint T, McKeigue PM, Isaacs AJ, Wild SH, Jacobs HS. Mortality of polycystic ovary syndrome. Fertil Steril 2007;87:1377– 1382.
women with polycystic ovary syndrome at long-term follow-up. J Clin Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple
Epidemiol 1998;51:581 – 586. methods for the estimation of free testosterone in serum. J Clin
Robinson S, Kiddy D, Gelding SV, Willis D, Niththyananthan R, Bush A, Endocrinol Metab 1999;84:3666 – 3672.
Johnston DG, Franks S. The relationship of insulin insensitivity to Vrbikova J, Hill M, Dvorakova K, Stanicka S, Vondra K, Starka L. Flutamide
menstrual pattern in women with hyperandrogenism and polycystic suppresses adrenal steroidogenesis but has no effect on insulin
ovaries. Clin Endocrinol (Oxf) 1993;39:351 – 355. resistance and secretion and lipid levels in overweight women with
Ruige JB, Assendelft WJ, Dekker JM, Kostense PJ, Heine RJ, Bouter LM. polycystic ovary syndrome. Gynecol Obstet Invest 2004;58:36 – 41.
Insulin and risk of cardiovascular disease: a meta-analysis. Circulation Welt CK, Arason G, Gudmundsson JA, Adams J, Palsdottir H,
1998;97:996 – 1001. Gudlaugsdottir G, Ingadottir G, Crowley WF. Defining constant
Rutter MK, Meigs JB, Sullivan LM, D’Agostino RB Sr, Wilson PW. Insulin versus variable phenotypic features of women with polycystic ovary
resistance, the metabolic syndrome, and incident cardiovascular syndrome using different ethnic groups and populations. J Clin
events in the Framingham Offspring Study. Diabetes 2005; Endocrinol Metab 2006a;91:4361– 4368.
54:3252 – 3257. Welt CK, Gudmundsson JA, Arason G, Adams J, Palsdottir H,
Sahin I, Serter R, Karakurt F, Demirbas B, Culha C, Taskapan C, Kosar F, Gudlaugsdottir G, Ingadottir G, Crowley WF. Characterizing discrete
Aral Y. Metformin versus flutamide in the treatment of metabolic subsets of polycystic ovary syndrome as defined by the Rotterdam

Downloaded from by guest on December 31, 2015

consequences of non-obese young women with polycystic ovary criteria: the impact of weight on phenotype and metabolic features.
syndrome: a randomized prospective study. Gynecol Endocrinol 2004; J Clin Endocrinol Metab 2006b;91:4842 – 4848.
19:115 – 124. Wijeyaratne CN, Balen AH, Barth JH, Belchetz PE. Clinical manifestations
Sam S, Legro RS, Essah PA, Apridonidze T, Dunaif A. Evidence for and insulin resistance (IR) in polycystic ovary syndrome (PCOS) among
metabolic and reproductive phenotypes in mothers of women with South Asians and Caucasians: is there a difference? Clin Endocrinol (Oxf)
polycystic ovary syndrome. Proc Natl Acad Sci USA 2006; 2002;57:343 – 350.
103:7030– 7035. Wijeyaratne CN, Nirantharakumar K, Balen AH, Barth JH, Sheriff R,
Samaras K, McElduff A, Twigg SM, Proietto J, Prins JB, Welborn TA, Belchetz PE. Plasma homocysteine in polycystic ovary syndrome: does
Zimmet P, Chisholm DJ, Campbell LV. Insulin levels in insulin resistance: it correlate with insulin resistance and ethnicity? Clin Endocrinol (Oxf)
phantom of the metabolic opera? Med J Aust 2006;185:159– 161. 2004;60:560 – 567.
Sharp PS, Kiddy DS, Reed MJ, Anyaoku V, Johnston DG, Franks S. Wild S, Pierpoint T, McKeigue P, Jacobs H. Cardiovascular disease in
Correlation of plasma insulin and insulin-like growth factor-I with women with polycystic ovary syndrome at long-term follow-up: a
indices of androgen transport and metabolism in women with retrospective cohort study. Clin Endocrinol (Oxf) 2000;52:595 – 600.
polycystic ovary syndrome. Clin Endocrinol (Oxf) 1991;35:253– 257. Witkowski JA, Parish LC. The assessment of acne: an evaluation of grading
Shaw LJ, Bairey Merz CN, Azziz R, Stanczyk FZ, Sopko G, Braunstein GD, and lesion counting in the measurement of acne. Clin Dermatol 2004;
Kelsey SF, Kip KE, Cooper-Dehoff RM, Johnson BD et al. 22:394 – 397.
Postmenopausal women with a history of irregular menses and Yildiz BO, Gedik O. Insulin resistance in polycystic ovary syndrome:
elevated androgen measurements at high risk for worsening hyperandrogenemia versus normoandrogenemia. Eur J Obstet Gynecol
cardiovascular event-free survival: results from the National Institutes Reprod Biol 2001;100:62 – 66.
of Health – National Heart, Lung, and Blood Institute sponsored Yilmaz M, Bukan N, Ersoy R, Karakoc A, Yetkin I, Ayvaz G, Cakir N,
Women’s Ischemia Syndrome Evaluation. J Clin Endocrinol Metab 2008; Arslan M. Glucose intolerance, insulin resistance and cardiovascular
93:1276 – 1284. risk factors in first degree relatives of women with polycystic ovary
Shi Y, Gao X, Sun X, Zhang P, Chen Z. Clinical and metabolic syndrome. Hum Reprod 2005;20:2414– 2420.
characteristics of polycystic ovary syndrome without polycystic ovary: Zawdaki J, Dunaif A. Diagnostic criteria for polycystic ovary syndrome:
a pilot study on Chinese women. Fertil Steril 2008;90:1139 – 1143. towards a rational approach. In: Dunaif A, Givens J, Haseltine F,
Shroff R, Syrop CH, Davis W, Van Voorhis BJ, Dokras A. Risk of metabolic Marrian G (eds). Polycystic Ovary Syndrome. Current Issues in
complications in the new PCOS phenotypes based on the Rotterdam Endocrinology and Metabolism, Vol. 4. Boston: Blackwell Scientific,
criteria. Fertil Steril 2007;88:1389 – 1395. 1992,377 –384.
Teede H, Hutchison SK, Zoungas S. The management of insulin resistance Submitted on August 8, 2008; resubmitted on January 29, 2009; accepted on
in polycystic ovary syndrome. Trends Endocrinol Metab 2007;18:273–279. February 3, 2009

View publication stats