American Journal of Obstetrics and Gynecology (2004) 191, 713–7
www.elsevier.com/locate/ajog
CLINICAL OPINION
A practical approach to the diagnosis of polycystic
ovary syndrome
R. Jeffrey Chang, MD
Department of Reproductive Medicine, University of California, San Diego, School of Medicine, La Jolla, Calif
Received for publication April 27, 2004; accepted April 29, 2004
KEY WORDS
Polycystic ovary
syndrome
Hirsutism
Androgens
The diagnosis of polycystic ovary syndrome (PCOS) is primarily achieved through clinical history
and physical findings. The principle features are hirsutism or biochemical evidence of excess
androgen production and irregular menstrual bleeding caused by the chronic anovulation.
Associated findings include insulin resistance with compensatory hyperinsulinemia and obesity.
Ultrasound imaging of the ovary has facilitated the diagnosis. It is important to exclude
conditions that may mimic PCOS, such as hyperthecosis, congenital adrenal hyperplasia, 21hydroxylase deficiency, Cushing’s syndrome, and androgen-producing neoplasms. These
disorders are usually revealed by appropriate laboratory assessment. Screening tests include
measurement of serum total testosterone, DHEA sulfate, and 17-hydroxyprogesterone. In
addition, in the obese individual, determinations of glucose and insulin levels, as well as a lipid
profile, are highly recommended.
Ó 2004 Elsevier Inc. All rights reserved.
Polycystic ovary syndrome (PCOS) is the most common reproductive endocrinopathy of women during their
childbearing years, with a reported prevalence of 4% to
8%.1,2 Classic symptomatology is a result of excessive
ovarian androgen production and chronic anovulation.
In addition, the ovaries display a typical morphologic
pattern that appears to be unique to the disorder. Other
associated clinical features include obesity and insulin
resistance. While the diagnosis is generally indicated by
the clinical presentation, laboratory testing is necessary to
exclude other possible conditions that may mimic PCOS.
Symptomatology of PCOS
In general, the diagnosis of PCOS is largely based on
clinical history. The most distinctive and visible clinical
Reprints not available from the author.
0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.ajog.2004.04.045
feature is hirsutism, the degree of which is variable from
mild to severe. The rate of hair growth is important
clinically because gradual and progressive growth indicates a functional etiology, whereas the rapid appearance of thick, pigmented hair often suggests a neoplastic
source of androgen production. The areas of hirsutism
usually involve the face and chin, as well as an extension
of pubic hair growth in the midline towards the umbilicus. In addition, it is not unusual to document some
degree of temporal balding and acne formation as
a result of androgen overproduction. However, extreme
expression of androgen excess, such as virilization and
clitoromegaly, are not typical findings in this functional
syndrome.
In women with PCOS, menstrual dysfunction is primarily characterized by irregular, infrequent, or absent
menstrual bleeding. In particular, episodes of irregular
bleeding are not typically preceded by premenstrual
714
symptomatology, and thus, not predictable. This important clinical observation is highly suggestive of anovulation. Typically, the inception of irregular bleeding in
PCOS can be traced to menarche and the postpubertal
phase of ovarian activity, during which time there is lack
of monthly menstrual cyclicity. Thus, in the postpubertal adolescent with PCOS, irregular bleeding may be
indistinguishable from the inconsistent bleeding pattern
that accompanies the emergence of regular ovulatory
function following menarche in normal girls. In some
cases, the onset of chronic anovulation is encountered
beyond adolescence, but this is unusual. A small percentage of women will present with amenorrhea. The
disruption of regular menstrual bleeding is not uniform
in all women with PCOS. It is now recognized that some
affected individuals will have normal ovulatory function, which implies that the presence of regular menses
does not preclude the possibility of PCOS.
Obesity has been reported to occur in about half of
PCOS patients.3 However, of recent note, the rate of
obesity associated with PCOS has not been corroborated, and there is a growing impression that the incidence may be greater, at least in the United States,
than that previously described. The obesity of PCOS is
characterized by an increased waist-to-hip ratio, or
‘‘android’’ appearance as opposed to truncal obesity.
The presence of obesity compounds clinical risk in
PCOS for several reasons. First, obesity is correlated
to decreased sex hormone-binding globulin, which increases circulating free testosterone and estradiol. Second, the likelihood of dyslipidemia is increased in obese
individuals, which raises concern for cardiovascular
risk.4 Third, obesity is associated with insulin resistance,
which may progress to diabetes mellitus in PCOS
patients.5,6 Improvement in these abnormalities with
weight loss underscores the role of obesity in PCOS.
It has been well-documented that women with PCOS,
independent of obesity, are insulin resistant and have
compensatory hyperinsulinemia as a result of their disorder.7 Generally, the degree of insulin resistance is mild,
although the prevalence of glucose intolerance and subsequent diabetes has been reported to be as high as 31%
and 7.5%, respectively.8 Notwithstanding the increased
risk for diabetes, there is indirect evidence to indicate that
insulin resistance may worsen the clinical manifestations
of PCOS. Administration of insulin-lowering drugs has
been shown to improve insulin sensitivity, reduce androgen levels, and restore ovulation in some, but not all
patients with this disorder.9,10 Insulin resistance may also
contribute to metabolic dysfunction in PCOS, including
an increased likelihood of lipid abnormalities.
Acanthosis nigricans is common in obese PCOS, and
often may be found on the nape of the neck, the axilla,
the area beneath the breasts and other intertrigenous
areas, as well as on exposed surfaces such as the elbows
and knuckles. While the skin appears to be pigmented,
Chang
the involved areas of skin do not exhibit increased
number of melanocytes or melanin deposition. Actually,
the lesion arises from epidermal hyperkeratosis and
dermal fibroblast proliferation. The factors responsible
for these findings have not been identified, although the
close association of acanthosis nigricans with marked
insulin resistance suggests a causative relationship. Increased insulin sensitivity following dietary or pharmacologic intervention is associated with a lessening of the
hyperplastic process and cosmetic improvement.
Women with PCOS may experience increased skin
oiliness resulting from excessive stimulation of the
pilosebaceous unit by increased androgen production.
However, increased sebaceous gland activity in PCOS is
neither associated with acne, nor does acne correlate
with increased ovarian androgen production. Therefore,
as an isolated symptom, acne should not be considered
a sign of PCOS.
Differential diagnosis
Ovarian hyperthecosis
Hyperthecosis refers to an unusual proliferative condition in which the ovary contains nests of luteinized theca
cells scattered throughout the stroma.11 The ovary is
enlarged and of an extremely firm texture resulting from
extensive and dense fibroblast growth. The absence of
follicle formation provides a clear morphologic distinction from the PCOS ovary. Because of markedly high
serum androgen concentrations, these individuals suffer
from severe hirsutism, and a significant percentage of
patients exhibit virilizing signs, such as clitoromegaly,
temporal balding, a male body habitus, and a deepening
of the voice. There usually is marked insulin resistance
with substantial elevations of circulating insulin levels.
In addition, these patients are often obese and exhibit
acanthosis nigricans.
Congenital adrenal hyperplasia (CAH)
This disorder of adrenal steroidogenesis is comprised on
several enzymatic defects, the most common of which is
21-hydroxylase deficiency. Compared with the infrequent occurrence of classic CAH-21-hydroxylase deficiency, which is recognized at birth and is notable for
its severity of clinical presentation, the nonclassic or
adult-onset form may simulate the features of PCOS.
The symptoms of adult-onset CAH-21-hydroxylase deficiency reflect the accumulation of serum 17-hyroxyprogesterone (17-OHP), which leads to abnormal
elevations of the hormone compared with circulating
values found in the follicular phase of the menstrual
cycle. Because 17-OHP is an androgen precursor, expression of this defect is associated with increased
production of androstenedione and testosterone with
Chang
resultant hyperandrogenism. While the clinical presentation may be indistinguishable from that of PCOS,
there are several aspects of CAH-21-hydroxylase deficiency which may suggest the diagnosis. These include
severe hirsutism, clitoromegaly, regular menses, familial
tendency, and short stature. The condition is transmitted by an autosomal-recessive inheritance pattern,
while an explanation for regular menses and short
stature is unknown. Morphologically, the ovaries have
been reported to appear similar to those of PCOS.12 The
second most common enzyme deficiency is 11-b-hydroxylase, which may also give rise to mild hirsutism in
association with increases in 17-OHP and 11-deoxycortisol, the immediate precursor for this enzyme. The
accompanying hypertension often distinguishes this disorder from the 21-hydroxylase form of CAH.
Cushing’s syndrome
The clinical features of Cushing’s syndrome primarily
result from excessive cortisol production by an adrenal
neoplasm or from excessive adrenocorticotropin
(ACTH) production. In most cases, ACTH overproduction is caused by a pituitary tumor, although rarely,
ectopic sources of ACTH may be encountered, such as
in adenocarcinoma of the lung. The preponderant findings are obesity, hirsutism, acne, and menstrual irregularity. These suggest the diagnosis of PCOS. However,
additional evidence of moonlike facies, buffalo hump,
hypertension, muscle wasting, abdominal striae, and
osteoporosis indicates a primary problem of cortisol
excess. While circulating androgen levels are elevated,
there is also abnormal cortisol secretion characterized
by increased urinary free cortisol excretion, loss of
circadian rhythm, and failure of suppression in response
to dexamethasone. In contrast to CAH, careful examination of the ovaries does not reveal changes typical of
PCOS in the vast majority of cases. Fortunately, Cushing’s syndrome occurs rarely.
Androgen-producing neoplasms
While seldom encountered, androgen-producing tumors
may arise from the ovary and the adrenal gland. In
contrast to the gradually evolving clinical presentation
associated with functional hyperandrogenism, the neoplastic process can be quite dramatic. Within a matter of
months, these lesions may induce severe hirsutism,
a male body habitus, and virilization with clitoromegaly.
In addition, there may be acne and a lowering of the
voice. Despite the severity of androgenic manifestations,
the early stages of development of these tumors can
mimic PCOS or other functional hyperandrogenic syndromes. Disruption of menstrual cycles varies from
irregular bleeding to amenorrhea. The rapid onset of
symptoms provides an important clue to the diagnosis.
715
In some instances, a pelvic or abdominal mass can be
palpated, which suggests an ovarian tumor.
Imaging studies
In women with PCOS ultrasound imaging of the ovaries
has revealed ovarian enlargement, 10 or more antral
follicles ranging from 2 to 10 mm in diameter arranged
in a peripheral distribution, and increased central
stroma of greater than 25% of the ovarian area.13 This
rather specific description of the polycystic ovary is
distinct from the ultrasound appearance of the multifollicular ovary, which is associated with spontaneous
ovarian follicular activity in a woman recovering from
hypogonadotropic hypogonadism or ovarian stimulation as a result of ovulation induction. The multifollicular ovary has been described as being normal sized or
slightly enlarged, containing 6 or more follicles without
peripheral displacement, and having no increase in
central stroma.14 Nevertheless, addition of the ultrasound description of polycystic ovaries has not necessarily enhanced the ability to make the diagnosis of
PCOS or impact therapy. In the situation of a suspected
ovarian neoplasm, pelvic ultrasound may provide evidence of an ovarian lesion. Similarly, MRI or CT has
been useful in the detection of an adrenal tumor.
Laboratory evaluation
Patients suspected of having PCOS on a clinical basis
should undergo a minimum endocrine evaluation consisting of serum total testosterone, DHEA-sulfate, and,
as indicated, 17-OHP. The primary reason for obtaining
measurements of testosterone and DHEA-S is to exclude the rare possibility of an androgen-producing
tumor of the ovary or adrenal gland, respectively.
Threshold values beyond which a neoplasm should be
considered are 200 ng/dL for testosterone and 7,000 ng/
mL for DHEA-S. Determination of serum 17-OHP is
useful for the detection of CAH caused by 21-hydroxylase deficiency. A circulating level less than 3 ng/mL
obtained during the follicular phase in patients with
regular menstrual cycles excludes the diagnosis. Values
in excess of 3 ng/mL warrant further evaluation by an
ACTH stimulation test. If Cushing’s syndrome is a consideration, then the 24-hour urinary free cortisol provides the most sensitive measure. A value in excess of 3
times the normal upper limit for the laboratory performing the test assumes the diagnosis. Intermediate values
warrant a repeat of the test.
As part of the assessment of oligoamenorrhea caused
by anovulation, measurements of prolactin and thyroidstimulating hormone (TSH) have been advocated in
some, but not all instances. In PCOS serum elevations
in prolactin have been reported to be as high as 35%
in patients and probably relates to lactotrope stimulation by chronic estrogen exposure.15 Coexistence of
716
a prolactinoma and PCOS is uncommon. Disorders of
thyroid secretion have been associated with irregular
menstrual bleeding. Usually there are other accompanying clinical features that suggest the diagnosis.
Some comment is deserved regarding the measurement
of serum gonadotropin levels and free testosterone concentrations. Despite the widespread practice of measuring
serum luetinizing hormone (LH) and follicle-stimulating
hormone (FSH), the circulating levels of these glycoproteins really do not contribute significantly to the diagnosis
of PCOS. Increased pituitary LH secretion cannot always
be determined by measurement of the serum concentration because approximately one third of patients will have
circulating levels of LH in the normal range. Accordingly,
the LH to FSH ratio also fails to provide additional useful
information. The role of free testosterone in women
suspected of having androgen excess is slowly evolving.
In women with obvious hirsutism, there is little diagnostic
value. By comparison, in anovulatory women without
excessive hair growth, an elevated free testosterone may
forecast incipient hirsutism. However, the cost/benefit of
obtaining this measurement in this group of women
remains to be determined. The principle utility of free
testosterone may be to monitor the efficacy of treatment
to reduce androgen production.
The observation that women with PCOS are insulin
resistant and have compensatory hyperinsulinemia raises the question of whether assessment of glucose
metabolism and insulin secretion should be evaluated
in these patients. Unfortunately, at the present time the
ability to determine insulin resistance is limited by tests
that lack sensitivity or are impractical for implementation. Based on fasting levels of glucose and insulin,
a variety of indices have been designed to establish
insulin resistance.16-18 While a reasonable correlation
exists between each model and provocative glucose
tolerance tests, normal values do not preclude the
presence of insulin resistance. However, the fasting level
of glucose may be used to distinguish glucose intolerance (110 to 125 mg/dL) or diabetes (O125 mg/dL), and
an elevated fasting insulin level will confer insulin
resistance. Because of significant variability in methodologic techniques of measuring insulin, one should be
mindful of the normal reference range for the laboratory
in which the test was preformed. In the presence of
obesity, it may be prudent to perform an oral glucose
tolerance test, particularly if there is a family history of
diabetes. Whether determination of insulin resistance is
essential to the diagnosis of PCOS seems unlikely.
While not pertinent to the diagnosis of PCOS,
dyslipidemia is a concern given the existence of hyperandrogenemia, insulin resistance with hyperinsulinemia,
and obesity, each of which may independently exert
adverse effects on lipid metabolism. As a result, in the
obese patient with PCOS, a lipid profile should be
obtained.
Chang
Summary recommendation
The diagnosis of PCOS is primarily dictated by the
development of hirsutism and anovulation, which is
consistent with the majority recommendation of the
1990 PCOS conference held at the NIH.19 Ultrasound
imaging revealing polycystic ovaries combined with
these features is essentially confirmatory. Recently, an
attempt was made to revise the diagnostic criteria for
PCOS based on discussions conducted at the 2003
Rotterdam ESHRE/ASRM-sponsored PCOS consensus
workshop, which were published in Fertility and Sterility
and Human Reproduction in 2004.20,21 It was recommended that the diagnosis of PCOS may be fulfilled
when 2 of the following 3 clinical features were present:
clinical or biochemical evidence of hyperandrogenism,
chronic anovulation, and imaging of polycystic ovaries.
While the new guidelines are more flexible and address
the problem of heterogeneous presentations, they also
pose additional confusion. For example, according to
these criteria, a woman with irregular menstrual bleeding and polycystic ovaries may be designated as having
PCOS without any clinical or biochemical evidence of
androgen excess. That PCOS could be diagnosed without hyperandrogenism is inconsistent with the original
description of the syndrome and difficult to reconcile.
Nevertheless, the conclusions of the workshop represent
a credible first step to account for the patient that
exhibits some, but not all of the previous criteria
established for this disorder. It may well be that a better
understanding of the heterogeneity of PCOS is necessary
before these revised diagnostic criteria are embraced in
clinical practice. Moreover, it is likely that this topic will
undergo further debate and discussion until the time
when uniform and definitive criteria or subclasses of
criteria are achieved.
For women suspected of having PCOS, the laboratory screening tests include measurement of a serum
total testosterone, DHEA-S, and 17-OHP. Measurement of serum gonadotropins is not necessary for
diagnosis. In patients who exhibit moderate to severe
hirsutism associated with a rapid onset of symptoms,
assessment should be directed toward determining the
presence of an androgen-producing neoplasm. Serum
total testosterone and DHEA-S are essential. Should
these values exceed the threshold levels for tumor, then
imaging studies such as ultrasound and MRI or CT are
warranted to locate the lesion. Occasionally, high circulating androgen levels may not be associated with
a distinct lesion, but rather bilateral noncystic ovarian
enlargement. If accompanied with a gradual onset of
symptoms, this presentation would suggest the diagnosis
of hyperthecosis.
The diagnoses of adrenal disorders include CAH,
primarily 21-hydroxylase deficiency, Cushing’s syndrome, and adrenal tumors. CAH-21-hydroxylase
Chang
deficiency is suspected in patients with an elevated 17OHP level beyond 3 ng/mL, and an ACTH stimulation
test should be done. Serum 17-OHP responses to ACTH
that exceed 10 ng/mL are indicative of the enzyme
defect. Laboratory screening for Cushing’s syndrome is
best achieved with a 24-hour urinary free cortisol.
Abnormal responses require further testing, including
low-dose, high-dose dexamethasone suppression tests,
as well as imaging studies, to determine the existence of
adrenal hyperplasia, Cushing’s syndrome, adrenal adenoma, or an ectopic ACTH production site.
References
1. Knochenhauer ES, Key TJ, Kahsar-Miller M, Waggoner M,
Boots LR, Azziz R. Prevalence of the polycystic ovary syndrome
in unselected black and white women of the southeastern United
States: a prospective study. J Clin Endocrinol Metab 1998;83:
3078-82.
2. Farah L, Lazenby AJ, Boots LR, Azziz R. Prevalence of polycystic
ovary syndrome in women seeking treatment from community
electrologists. Alabama Professional Electrology Association
Study Group. J Reprod Med 1999;44:870-4.
3. Goldzieher JW, Green JA. The polycystic ovary. I. Clinical and
histologic features. J Clin Endocrinol Metab 1962;22:325-38.
4. Guzick DS. Cardiovascular risk in women with polycystic ovarian
syndrome. Semin Reprod Endocrinol 1996;14:45-9.
5. Dunaif A, Graf M, Mandeli J, Laumas V, Dobrjansky A.
Characterization of groups of hyperandrogenic women with
acanthosis nigricans, impaired glucose tolerance, and/or hyperinsulinemia. J Clin Endocrinol Metab 1987;65:499-507.
6. Legro RS, Kunselman AR, Dodson WC, Dunaif A. Prevalence
and predictors of risk for type 2 diabetes mellitus and impaired
glucose tolerance in polycystic ovary syndrome: a prospective,
controlled study in 254 affected women. J Clin Endocrinol Metab
1999;84:165-9.
7. Chang RJ, Nakamura RM, Judd HL, Kaplan SA. Insulin resistance in nonobese patients with polycystic ovarian disease. J
Clin Endocrinol Metab 1983;57:356.
8. Ehrmann DA, Barnes RB, Rosenfield RL, Cavaghan MK,
Imperial J. Prevalence of impaired glucose tolerance and diabetes
in women with polycystic ovary syndrome. Diabetes Care
1999;22:141-6.
9. Azziz R, Ehrmann D, Legro RS, Whitcomb RW, Hanley R,
Fereshetian AG, et al. Troglitazone improves ovulation and
717
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
hirsutism in the polycystic ovary syndrome: a multicenter, double
blind, placebo-controlled trial. J Clin Endocrinol Metab 2001;86:
1626-32.
Velazquez EM, Mendoza S, Hamer T, Sosa F, Glueck CJ.
Metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic
blood pressure, while facilitating normal menses and pregnancy.
Metabolism 1994;43:647-54.
Culiner A, Shippel S. Virilism and theca cell hyperplasia of
the ovary syndrome. J Obstet Gynaecol Br Commonw 1949;56:
439.
Barnes RB, Rosenfield RL, Ehrmann DA, Cara JF, Cuttler L,
Levitsky LL, et al. Ovarian hyperandrogynism as a result of
congenital adrenal virilizing disorders: evidence for perinatal
masculinization of neuroendocrine function in women. J Clin
Endocrinol Metab 1994;79:1328-33.
Adams J, Polson DW, Franks S. Prevalence of polycystic ovaries
in women with anovulation and idiopathic hirsutism. Br Med J
Clin Res Ed 1986;293:355.
Adams J, Franks S, Polson DW, Mason HD, Abdulwahid N,
Tucker M, et al. Multifollicular ovaries: clinical and endocrine
features and response to pulsatile gonadotropin releasing hormone. Lancet 1985;2:1375-9.
Corenblum B, Taylor PJ. The hyperprolactinemic polycystic
ovary syndrome may not be a distinct entity. Fertil Steril 1982;38:
549-52.
Legro RS, Finegood D, Dunaif A. A fasting glucose to insulin ratio
is a useful measure of insulin sensitivity in women with polycystic
ovary syndrome. J Clin Endocrinol Metab 1998;83:2694-8.
Matthews DR, Hosker JP, Rudenski AS, Naylor BA,
Treacher DF, Turner RC. Homeostasis model assessment: insulin
resistance and beta-cell function from fasting plasma glucose and
insulin concentrations in man. Diabetologia 1985;28:412-9.
Katz A, Nambi SS, Mather K, Baron AD, Follmann DA,
Sullivan G, et al. Quantitative insulin sensitivity check index:
a simple, accurate method for assessing insulin sensitivity in
humans. J Clin Endocrinol Metab 2000;85:2402-10.
Zawadski JS, Dunaif A. Diagnostic criteria for polycystic ovary
syndrome: toward a rational approach. In: Dunaif A, Givens JR,
Haseltine FP, Merriam GR, editors. Polycystic ovary syndrome.
Boston: Blackwell Scientific; 1992. p. 377-84.
Revised 2003 consensus on diagnostic criteria and long-term
health risks related to polycystic ovary syndrome. Rotterdam
ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group.
Fertil Steril 2004;81:19-25.
Revised 2003 consensus on diagnostic criteria and long-term
health risks related to polycystic ovary syndrome. Rotterdam
ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group.
Hum Reprod 2004;19:41-7.