Optimizing Ovulation Induction in the Poor Responder

1. Protocol Title: Optimizing ovulation induction in the poor responder: a randomized controlled trial of luteal phase estradiol verses combined oral contraceptive pill (COCP) on ovarian morphology and biomarkers prior to ovulation induction
2. Purpose of the Study: In women undergoing ovulation induction who are at high risk for poor response, we hypothesize that: Null Hypothesis: Mean follicular diameter and coefficient of variation between follicles in patients receiving luteal COCP are statistically equal to follicular diameters and coefficient of variation between follicles in patients receiving estradiol. To test this hypothesis, the specific aims of the present proposal are: Aim 1: Compare the effect of luteal estradiol and COCP pretreatment on follicle synchrony. Mean follicular diameter and the coefficient of variation between follicle measurements will be calculated on cycle day 3 the cycle before and during treatment. Aim 2: Compare the effect of luteal estradiol and COCP pretreatment on follicle recruitment. Markers of follicle recruitment will be measured on cycle day 3 before and during the treatment cycle. Previously identified follicle recruitment markers that positively correlate with follicle recruitment include: antral follicle count (AFC), AMH and androstenedione. Aim 3: Compare the efficacy of luteal estradiol and COCP. The number of mature follicles greater than or equal to 14 mm on day of hCG and pregnancy outcomes will be recorded. The two treatment groups are:

• Group 1: Luteal estradiol administration prior to ovulation induction management per patient's primary physician.
• Group 2: COCP administration for one month prior to ovulation induction management per patient's primary physician. Upon completion, we will have compared follicular and biochemical ovarian response to luteal estradiol and COCPs in poor responders undergoing ovulation induction. The future implications of this study are development of clinical strategies for treating patients at high risk of poor response.

1. Background \& Significance: Ten percent of the United States population, 6.2 million women and their partners, reported infertility in the 1995 National Survey of Family Growth. The number of couples affected by infertility is predicted to increase to 6.5 million by the year 2025.' Among women seeking infertility treatment, a subset will be classified as poor responders due to poor ovarian response to treatment. Though the definition of poor response to treatment is not standardized, studies estimate that 4-26% of patients undergoing treatment are affected.2 The proportion of women classified as poor responders may increase as more women delay fertility until an age of decreased fecundity due to declining ovarian reserve. Establishing effective treatments for poor responders is therefore imperative. The principal goal of ovulation induction in infertile patients with regular menstrual cycles is to achieve two to three mature synchronous follicles to maximize chance of pregnancy and minimize risk of ovarian hyperstimulation syndrome. Poor responders, by definition, have a lower likelihood of sufficient follicular recruitment. Lesser recruitment can be compounded by follicular dysynchrony that necessitates human chorionic gonadotropin (hCG) administration based on larger follicles when lagging smaller follicles are less likely to be mature. Follicular dysynchrony can be compounded by the rise in follicle stimulating hormone (FSH) in the luteal phase. Fanchin et al has demonstrated that luteal estradiol administration suppresses luteal FSH rise, reduces antral follicle size, and synchronizes the follicular cohort.3,4,5,6 Luteal estradiol and GnRH antagonist administration prior to controlled ovarian hyperstimulation for in vitro fertilization (IVF) have been shown to decrease cancellation rate, increase mean number of oocytes retrieved, and improve fertilization rates in poor responders (retrospective study using historic controls presented in abstract form).7 Estradiol has also been used in patients with hypergonadotrophic amenorrhea to suppress elevated endogenous LH and FSH prior to stimulation with exogenous gonadotropins with modest success in achieving ovulation and, in rare cases, pregnancy.8 Combined oral contraceptive pills (COCPs) administered prior to ovulation induction in IVF increase the number of oocytes retrieved and suppress cyst formation.9,10,11 In a small study of poor responders undergoing controlled ovarian hyperstimulation for IVF, pretreatment with COCPs was associated with increased pregnancy rates compared to pretreatment with luteal phase gonadotropin releasing hormone agonist (GnRH-a), follicular phase GnRH-a, and no pretreatment.12 COCPs prior to ovulation induction with microdoses of GnRHa and gonadotropins enhance ovarian responsiveness in poor responders resulting in higher peak estradiol (E2) levels, more mature follicles, and larger number of retrieved oocytes when compared with traditional luteal GnRH-a downregulation followed by gonadotropin stimulation.13 Follicular GnRH-a administration releases endogenous gonadotropins producing a gonadotropin "flare" effect. Smaller, more frequent "microdosing" of GnRH-a minimizes endogenous FSH suppression while preventing premature luteinizing hormone (LH) surges. Oral contraceptive pretreatment prevents formation of a corpus luteum that could produce progesterone prematurely during the flare. This protocol with the addition of growth hormone (GH) has been shown to produce a 50% ongoing pregnancy rate (PR) when used in patients who had prior ovulation induction cycles using luteal phase GnRH-a suppression followed by exogenous gonadotropins and GH cancelled due to poor response.14 While both COCPs and luteal estradiol suppress the luteal rise in FSH and synchronize follicles, they have not been compared against each other. COCPs, with progestin in addition to estrogen, may achieve better suppression of luteal FSH than estradiol and result in better synchronization of the follicles. COCPs, however, have been used to decrease ovarian responsiveness in patients with polycystic ovarian syndrome who are at increased risk of ovarian hyperstimulation syndrome.15 Therefore, COCPs may suppress ovarian responsiveness more than luteal estradiol, an unwanted effect in the poor responder population. One potential mechanism for greater ovarian suppression with COCPs may be the decrease in androstenedione, a marker shown to be positively correlated with total and mature oocyte number obtained during IVF, 16 subsequent to the COCP-induced increase in sex hormone binding globulin. Previously identified follicle recruitment markers include antral follicle count (AFC) and androstenedione which positively correlate with follicle recruitment.16 AFC is a reproducible predictor of follicular recruitment which has been shown to positively correlate with number of oocytes retrieved and IVF pregnancy outcome.18,19, 20 AFC measurements are stable throughout the ovarian cycle and before and after GnRH downregulation.21 AMH may serve as a biomarker for both follicle quantity and quality since both serum and intrafollicular AMH positively correlate with AFC and number of eggs retrieved during IVF.22 The mid-luteal AMH rise may reflect luteal follicle development;23 both estradiol and COCP may affect luteal follicle development therefore CD3 AMH levels before and after treatment will be of great interest. AMH also exhibits superior cycle to cycle reproducibility when compared to FSH, E2, inhibin Band AFC.24
2. Design \& Procedures:

The study is a randomized controlled trial.

Pretreatment Evaluation

Baseline studies preceding ovulation induction:

Cycle day 3 FSH, inhibin B, estradiol, androstenedione, AMH, and AFC will be performed the cycle of OCP or luteal pretreatment and the cycle of ovulation induction.

Transvaginal ultrasound:

• Ovarian morphology and dimensions will be obtained using a vaginal probe ultrasound. All study ultrasounds will be recorded and read by a single investigator (at each center) blinded to patient treatment group.
• Ovarian volume will be calculated using the formula for a prolate ellipsoid (0.5237xD1xD2xD3; D1, D2, and D3 are the maximal longitudinal anteroposterior and transverse diameters, respectively).
• Ovarian volume will also be calculated using the VOCAL (volume calculation) GE software package on the Voluson ultrasound machine.
• Antral follicles will be counted, measured in two dimensions and averaged.

Blood sampling:

• Blood samples will be used to measure FSH, inhibin B, estradiol, androstenedione, AMH, and progesterone during the cycle prior to and of ovarian stimulation.
• Estradiol, FSH, progesterone, inhibin B, androstenedione, and AMH will be transported to, stored and measured in batches at Dr. Sarah Berga's Emory University research laboratory, Duke Fertility Center or the Duke Unievrsity Hospital core laboratory.

Registration/Randomization

Block randomization will occur by computer algorithm to two groups:

• Group 1: Luteal estradiol
• Group 2: COCP

Therapy

1. Ovarian stimulation for ovulation induction: Group 1: Luteal estradiol • Starting cycle day 20, two 0.1 mg/d transdermal estradiol patch will be applied every other day until cycle day 2.

• Patients will present for baseline ultrasound, AFC, FSH, inhibin B, estradiol, progesterone, androstenedione, and AMH measurement on cycle day 3.
• Starting on cycle day 3, ovarian stimulation will begin. The treatment protocol for ovulation induction will be determined by the patient's primary physician and the clinician monitoring the cycle. Group 2: Combined oral contraceptive
• Patients will begin COCP administration on the cycle day one of the cycle prior to ovarian stimulation.
• Patients will present for baseline ultrasound, AFC, FSH, inhibin B, estradiol, progesterone, androstenedione, and AMH measurement on cycle day 3.
• Starting on cycle day 3, ovarian stimulation will begin. The treatment protocol for ovulation induction will be determined by the patient's primary physician and the clinician monitoring the cycle. All participants will be monitored by serum E2 levels and TVUS as determined by the patient's primary physician to assess the rise in circulating estrogen levels and follicular maturation. Follicles will be measured in two dimensions and the average recorded.

1. The mode of ovulation trigger, and IUI or IVF will be at the discretion of the patient's primary physician and the clinician monitoring the cycle. Patient Assessment Patients will return 14 days after IUI or IVF for serum beta hCG testing. In the event of a negative pregnancy test, patients will be given the option of crossing over to the other treatment group. Data Collection Clinical Information/Clinical Data Items:

• Age
• Infertility diagnoses
• Prior dates and specifics of infertility treatments and results
• Prior cycle characteristics: antral follicle count (AFC) and dimensions, ovarian volume, FSH, inhibin A, E2, progesterone, androstenedione, and AMH before and during treatment
• Peak E2
• Outcome: Mean follicular diameter and the coefficient of variation between follicles on cycle day 3 before and during treatment
• Number of mature follicles greater than or equal to 14 mm on day of hCG
• β-hCG result

1. Selection of Subjects:

The Emory Reproductive Center (ERC) and Duke Fertility Center (DFC) are assisted reproductive technology centers offering ovulation induction, IVF, gamete micromanipulation, and embryo cryopreservation. Between the two centers, approximately 250-300 stimulated cycles are completed each year. While the study is open, all women at risk for poor response to ovulation induction at the two centers will be offered participation by the PI, Co-PI, Co-Investigator or coordinating research nurse.

Inclusion Criteria:

• Antral follicle count < 8
• Antimullerian hormone (AMH) <1.3
• Follicle stimulating hormone (FSH)>10
• History of follicular dysynchrony or poor response

Exclusion Criteria:

• Menopause
• FSH >40
• Age > 50
• Contraindication to estradiol or COCP (including pre-existing cardiovascular disease, familial thrombophilia (factor V Leiden), severe hypercholesterolemia, smoker over age 35) Ovulation induction during month of estradiol treatment

1. Subject Recruitment \& Compensation:

Please see section #5. There will be no patient compensation.

1. Consent Process:

An IRB-approved written informed consent will be obtained from each subject at entry into the study; elements of informed consent will include: (a) having the subject review the study consent form; (b) having the investigator(s) or study staff meet with the subject to review the consent, confirm understanding, and answer any questions; and (c) once the investigator(s) or study staff are convinced that the protocol is understood and that there is agreement to participate, having the consent signed in the presence of a witness. The potential participant will have until the beginning of her first stimulated cycle to decide whether or not she is participating in the study. The consent will be discussed and signed in person at Emory Reproductive Center or Duke Fertility Center. The consent process, as outlined above, will occur in private consultation rooms. There is no time minimum or maximum for the consent process. The consent process is determined to be complete when the participant verbalizes understanding, states that she has no further questions and signs the consent forms. The PI, Co-PI, Co-Investigators and coordinating research nurse will be readily available by telephone or in person at ERC and/or DFC to answer questions concerning the consent. In order to reduce potential coercion and perceived influence by the study personnel, the study details and consent forms will be discussed in a matter-of-fact and consistent manner. Only patients who speak and read English will be admitted into the study.

1. Subject's Capacity to Give Legally Effective Consent:

Only patients with the capacity to give legally effective consent will be included in the study.

1. Study Interventions:

Please see section #4.

1. Risk/Benefit Assessment:

The treatments used in this study have been studied individually and are used in clinical practice. Our goal is to see which one is better. Therefore the risks of this study are the same as the risk of ovulation induction to which the subjects will be exposed whether they are enrolled in the study or not.

1. Costs to the Subject:

There is no additional cost to the subject above the regular fees assessed for ART.

1. Data Analysis \& Statistical Considerations:

Coefficient of variation will be used to evaluate follicular size discrepancy by the following formula: Coefficient of variation = 100 * Standard deviation / mean of the set

Sample Size

We predict that a difference of 1.2 mm between mean follicular size before and after COCPs will be clinically significant. This is based on observations by Fanchin in which women treated with luteal estradiol displayed a decrease in mean follicular size on day 3 of 1.2 mm that was associated with a significant decrease in coefficient of variation between follicles and an increase in mature follicles >16 mm number on day of hCG administration.

Power to detect a difference between groups Number cycles/group to demonstrate COCP equivalence Number cycles/group to demonstrate COCP superiority 80% 7 25 85% 7 29 90% 9 33 95% 11 41

To detect equivalence of COCP to luteal estradiol with a power of 90% at an alpha of .05 assuming a 10% attrition rate, we will enroll women into each study group so as to achieve final number of 10 cycles per treatment group. We assume that an additional 50% decrease in follicular size would prove COCP superiority. To detect superiority of COCP to luteal estradiol with a power of 80% at an alpha of 0.05 assuming 10% attrition rate, we will enroll women into each study group so as to achieve a final number of 28 cycles per treatment group. Only newly initiated treatment cycles will be included; only the first cycle of patients' crossing over to the other treatment group will count toward this recruitment goal. The second cross over cycle will be compared within patient.

µ1-µ 2 = 4.9 - 3.7 mm = 1.2 (µ1-µ 2)2 = 1.44 (COCP equivalence) µ1-µ 2 = 1.8 - 1.2 mm = 0.6 (µ1-µ 2)2 = 0.36 (COCP superiority) SD = б=0.75 б2 = 0.5625 2 б2 = 1.125 α = 0.05 Zα = 1.96 (2-sided test) β = 0.20 1- β = 0.80 Z β = 0.84 (1-sided) (Zα +Zβ)2= 7.84 β =0.15 1- β = 0.85 Z β = 1.03 (1-sided) (Zα +Zβ) 2= 8.94 β = 0.10 1- β = 0.90 Z β = 1.28 (1-sided) (Zα +Zβ) 2= 10.498 β = 0.05 1- β = 0.95 Z β = 1.645 (1-sided) (Zα +Zβ) 2 =12.996

Calculation: n= (Zα +Zβ)2* 2(б2) / (µ1-µ 2)2

Methods of Analysis

Chi square analysis will be used to compare categorical variables between treatment groups. Student's t test will be used to compare the mean follicle measurements between treatment groups.

1. Data \& Safety Monitoring:

Adherence and Monitoirng Statement: The Data Safety Monitoring Plan (DSMP) outlined below will adhere to the protocol approved by the IRB at both Emory University School of Medicine and Duke University School of Medicine. The Principal investigator (PI) will review all data collection forms at least annually for completeness and accuracy of the data as well as protocol compliance. The PI will review this protocol on a continuing basis for subject safety and include the results of the review in annual progress reports submitted to the IRB. Adverse events and serious adverse events will also be reviewed by the PI weekly.

Patient Monitoring: Performed by the P.I., the Co-P.I. and/or Co Investigators.

Patient safety data examination, monitoring procedures/oversight: All adverse events (AEs) will be graded as to their attribution (unrelated to protocol, or possibly, probably, or definitely related to protocol). Any AE that is reported to either the PI or her designated research associates by a study subject or by medical staff caring for the subject and which meets the criteria will be documented as such.

Serious adverse events (SAEs) are predefined as any experience that suggests a significant hazard, such as events which: a) are fatal, b) are life threatening, c) result in permanent disability, d) require inpatient hospitalization, or e) involve cancer, a congenital anomaly, or drug overdose.

Any AEs will be reported to the IRB at Emory and Duke within 24-48 hours of the event. The standard Emory and Duke IRB reporting guidelines for AE and SAE reporting will be followed. The investigators and staff will evaluate the SAEs in close coordination with the Emory and Duke IRB.

Expected adverse events are detailed in the Consent Form and include the following: side effects of estradiol and combined oral contraceptive pills (breakthrough bleeding, nausea, headaches, depression, breast tenderness, changes in blood pressure, and increase in risk of blood clotting); risks of ovulation induction (multiple birth, ovarian hyperstimulation syndrome, ovarian cyst development, failure to achieve pregnancy); risks of multiple birth and ovarian hyperstimulation syndrome are less in poor responders regardless of treatment used; risk of failure to achieve pregnancy is greater in poor responders regardless of treatment used.

Procedures for minimizing risks: Monitoring with serial ultrasound and blood work will be used to minimize risks.

Plans for transmission of temporary or permanent suspension actions: Any actions that mandate temporary or permanent suspension of study will be transmitted to the Emory and Duke IRB, and, if appropriate, to the FDA and the National Institutes of Health.

Plans for assuring data accuracy and protocol human safety compliance: The above detailed plans should assure data accuracy and protocol human safety compliance. These include computerized database management, and IRB oversight and communication. This plan, together with proposed monitoring by the IRB, should be sufficient without the addition of more faculty members to constitute a DSMB.

1. Privacy, Data Storage \& Confidentiality:

All information and materials will be obtained for research purposes only and the data will be kept in strict confidence. Confidentiality will be assured by the use of subject codes rather than personal identifiers. The study database will be secured, and information will only be entered using subject identifier codes rather than personal identifiers. Electronic communication will involve only coded, unidentifiable information. All adverse event reports and annual summaries will not include subject-identifiable material.