Study to Assess Potential Impairments in Estradiol Augmentation of Gonadotropin Secretion in Polycystic Ovary Syndrome

This is a two-group controlled study to test the following hypothesis: compared to body mass
index (BMI)-matched normal controls, women with polycystic ovary syndrome (PCOS) will
demonstrate blunted LH responses to preovulatory estradiol concentrations. We will study both
normally-cycling controls and women with PCOS. We aim to recruit BMI-matched pairs (PCOS vs.
control within 2 kg/m2). To provoke a gonadotropin surge, subjects will receive graded
transdermal estradiol dosing, and we will use a dose adjustment protocol to maintain serum
estradiol levels of 250-400 pg/ml. To enhance reliability of estradiol delivery, transdermal
estradiol patches will be placed/replaced daily by Clinical Research Unit (CRU) nurses, and
abdominal sites will be rotated. All subjects will begin estradiol on menstrual cycle day 4.
Starting 24 hours before E2 administration, all subjects will collect all urine output in
12-hour time blocks for later urinary LH and FSH analysis; this will continue until the end
of the study. Additionally, subjects will have daily morning blood draws in the CRU for later
hormone measurements. Transvaginal ovarian ultrasound will be performed on study day 6 --
near the time of expected surge initiation -- to document largest follicle sizes. We will
measure serum estradiol daily and employ an estradiol dose-adjustment protocol to maintain
target estradiol levels. We will also measure LH daily. The study will be stopped after
either (a) serum LH increases to 5-fold higher than baseline and subsequently falls to within
200% of baseline, or (b) the subject has received estradiol for a full 7 days, whichever
comes first. The primary endpoint will be estradiol-induced change in 24-hour urinary LH
excretion, defined as 24-hour mean values immediately prior to estradiol administration vs.
peak 24-hour mean values during estradiol administration. A comparison of 24-hour LH changes
between healthy normal and PCOS groups will be conducted by way of a random-effects analysis
of covariance (ANCOVA) model. The ANCOVA model will be specified so that each BMI-matched
pair will represent an independent observational unit with respect to comparing 24-hour LH
change between groups. With regard to hypothesis testing, we will test whether the component
of variability in 24-hour LH change attributed to "Study Group" (healthy normal control vs.
PCOS) is a significant component of the overall variability in 24-hour LH change. Variability
in 24-hour LH change attributed to baseline disparities in LH will be accounted for by
treating subject-specific baseline LH as a covariate in the ANCOVA model. If BMI matching is
inadequate, we will also include BMI as a covariate in the model. If 11 women with PCOS and
11 controls complete study, we expect at least 80% power to detect a 33% difference in
E2-induced augmentation of urinary LH excretion.

Inclusion Criteria:

- PCOS group: post-pubertal (> 4 years post-menarche) adult woman aged 18-30 years with
PCOS, defined as clinical and/or laboratory evidence of hyperandrogenism (hirsutism
and/or elevated serum [calculated] free testosterone concentration) plus ovulatory
dysfunction (irregular menses, fewer than 9 per year), but without evidence for other
potential causes of hyperandrogenism and/or ovulatory dysfunction

- Control group: post-pubertal (> 4 years post-menarche) adult woman aged 18-30 years
with regular menstrual periods (every 26-35 days) and no evidence of hyperandrogenism
(i.e., no hirsutism, normal serum [calculated] free testosterone concentration)

- General good health (excepting overweight, obesity, PCOS, and adequately-treated
hypothyroidism)

- Capable of and willing to provide informed consent

- Willing to strictly avoid pregnancy with use of reliable non-hormonal methods during
the study period

Exclusion Criteria:

- Inability/incapacity to provide informed consent

- Males will be excluded (hyperandrogenism is unique to females)

- Age < 18 years (we do not propose to study children because we have no preliminary
data that would support this particular study in children)

- Age > 30 years (since ovarian reserve may decrease beyond age 30)

- Obesity resulting from a well-defined endocrinopathy or genetic syndrome

- Positive pregnancy test or current lactation

- Evidence for non-physiologic or non-PCOS causes of hyperandrogenism and/or anovulation

- Evidence of virilization (e.g., rapidly progressive hirsutism, deepening of the voice,
clitoromegaly)

- Total testosterone > 150 ng/dl, which suggests the possibility of virilizing ovarian
or adrenal tumor

- DHEA-S greater than upper reference range limit for controls; and DHEA-S elevation >
1.5 times the upper reference range limit for PCOS. Mild elevations may be seen in
PCOS, and will be accepted in this group.

- Early morning 17-hydroxyprogesterone > 200 ng/dl measured in the follicular phase,
which suggests the possibility of congenital adrenal hyperplasia (if elevated during
the luteal phase, the 17-hydroxyprogesterone will be repeated during the follicular
phase). NOTE: If a 17-hydroxyprogesterone > 200 ng/dl is confirmed on repeat testing,
an ACTH stimulated 17-hydroxyprogesterone < 1000 ng/dl will be required for study
participation.

- Abnormal thyroid stimulating hormone (TSH): Note that subjects with stable and
adequately treated primary hypothyroidism, reflected by normal TSH values, will not be
excluded.

- Hyperprolactinemia: Any degree of hyperprolactinemia (confirmed on repeat) will be
grounds for exclusion for subjects without PCOS. Hyperprolactinemia > 20% higher than
the upper limit of normal will be grounds for exclusion for subjects without PCOS.
Mild prolactin elevations may be seen in PCOS, and elevations within 20% higher than
the upper limit of normal will be accepted in this group.

- History and/or physical exam findings suggestive of Cushing's syndrome, adrenal
insufficiency, or acromegaly

- History and/or physical exam findings suggestive of hypogonadotropic hypogonadism
(e.g., symptoms of estrogen deficiency) including functional hypothalamic amenorrhea
(which may be suggested by a constellation of symptoms including restrictive eating
patterns, excessive exercise, psychological stress, etc.)

- Persistent hematocrit < 36% and hemoglobin < 12 g/dl

- Severe thrombocytopenia (platelets < 50,000 cells/microliter) or leukopenia (total
white blood count < 4,000 cells/microliter)

- Previous diagnosis of diabetes, fasting glucose > or = 126 mg/dl, or a hemoglobin A1c
> or = 6.5%

- Persistent liver panel abnormalities, with two exceptions. Mild bilirubin elevations
will be accepted in the setting of known Gilbert's syndrome. Also, mild transaminase
elevations may be seen in obesity/PCOS; therefore, elevations < 1.5 times the upper
limit of normal will be accepted in these groups.

- Significant history of cardiac or pulmonary dysfunction (e.g., known or suspected
congestive heart failure, asthma requiring intermittent systemic corticosteroids,
etc.)

- Decreased renal function evidenced by GFR < 60 ml/min/1.73m2

- A personal history of breast, ovarian, or endometrial cancer

- History of any other cancer diagnosis and/or treatment (with the exception of basal
cell or squamous cell skin carcinoma) unless they have remained clinically disease
free (based on appropriate surveillance) for five years

- History of allergy to transdermal estradiol patches

- BMI < 18 or > 40 kg/m2; BMI < 18 kg/m2 is considered to be underweight, while > 40
kg/m2 is considered to be class 3 obesity -- both may have marked confounding effects
for the outcomes of interest

- Menstrual cycles lasting fewer than 26 days: Cycle frequency < 26 days suggest the
possibility of relatively short follicular phases (e.g., < 12 days). If a subject with
a follicular phase shorter than 12 days participates in Aim 1c, they could experience
an endogenous gonadotropin surge under surveillance. Since we wish to capture only
experimentally-induced surges, we will exclude such subjects.