- Email: [email protected]
A randomized clinical trial of treatment of clomiphene citrate-resistant anovulation with the use of oral contraceptive pill suppression and repeat clomiphene citrate treatment
A randomized clinical trial of treatment of clomiphene citrate–resistant anovulation with the use of oral contraceptive pill suppression and repeat clomiphene citrate treatment Emmett F Branigan, MD, and M. Antoinette Estes, BS Bellingham, Wash OBJECTIVE: The purpose of this study was to evaluate the effectiveness and endocrine response of oral contraceptive ovarian suppression followed by clomiphene citrate in patients who previously were clomiphene citrate resistant. STUDY DESIGN: Forty-eight patients from a private tertiary infertility clinic were assigned randomly prospectively to either group 1 (oral contraceptive/clomiphene citrate), which received continuous oral contraceptives followed by clomiphene citrate, or to group 2 (control) received no treatment in the cycle before clomiphene citrate treatment. On day 3, 17β-estradiol, follicle-stimulating hormone, luteinizing hormone, and androgens were assayed before and after treatment. Follicle growth, ovulation, and pregnancy were evaluated. The Student t test and analysis of variance were used for statistical significance. RESULTS: The oral contraceptive/clomiphene citrate group had a significantly higher percentage of patients who ovulated and of ovulatory cycles and pregnancies. Significantly lower levels of 17β-estradiol, luteinizing hormone, and androgen levels were seen in the oral contraceptive/clomiphene citrate group, with no significant changes in group 2. CONCLUSION: Suppression of the ovary with oral contraceptives results in excellent rates of ovulation and pregnancy in patients who previously were resistant to clomiphene citrate. The decreases in ovarian androgens, luteinizing hormone, and 17β-estradiol may be responsible for the improved response. (Am J Obstet Gynecol 2003;188:1424-30.)
Key words: Clomiphene citrate, oral contraceptive, chronic anovulation, suppression, androgen
Chronic anovulation is a common cause of infertility in women. Most of these women are treated initially with clomiphene citrate (CC) in doses from 50 to 250 mg. One large study found that 50% of patients ovulate at the 50-mg dosage and that 74% of patients ovulate on the 100-mg dose.1 However, despite successful ovulation, <50% of these women will become pregnant. In our experience >25% of women whom we see with chronic anovulation fail to ovulate on any dose of CC.2 For these women, gonadotropin therapy or a few limited adjunctive therapies used with CC or alone are the next treatment option. Adjunctive therapy options for CC-resistant women include weight loss, insulin-sensitizing agents, bromocrip-
From the Bellingham In Vitro Fertilization and Infertility Center. Presented at the Sixty-Ninth Annual Meeting of the Pacific Coast Obstetrical and Gynecological Society, October 25-27, 2002, Rancho Mirage, Calif. Reprint requests: Emmett F. Branigan, MD, 2980 Squalicum, Parkway #103, Bellingham, WA 98225. © 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1067/mob.2003.459
1424
tine, glucocorticoids, extended doses or continued doses of CC, and surgery. A weight loss of as little as 5% of total body weight has been shown to produce ovulatory cycles.3 The time and effort that is involved or if the patient is lean limits this option. If hyperprolactinemia is present, bromocriptine (Parlodel) either alone or with CC can be successful.4,5 Metformin (an insulin sensitizing agent) used alone or with CC has been used for ovulation induction in an unselected group of patients.6 Other studies have confirmed these results, but which patients might benefit and specific treatment protocols have yet to be formulated.7,8 The use of dexamethasone or other glucocorticoids with CC is useful in hyperandrogenic women with dehydroepiandrosterone sulfate levels of ≥200 µg/mL.9,10 The use of doses of CC beyond the normal 5 days or even continuous doses until ovulation occurs have been used by several investigators.11,12 However, these regimens are limited because of patient complaints of side effects. Surgery (either ovarian wedge resection or laparoscopic ovarian drilling) can result in time-limited ovulatory cycles, but the treatment is expensive and is associated with pelvic adhesion risk.13 Exogenous gonadotropin therapy is usually the next treatment step for CC-resistant anovulation patients. This
Branigan and Estes 1425
Volume 188, Number 6 Am J Obstet Gynecol
Table I. Response to CC after OCs or no treatment intervals
Ovulation (No.) Ovulatory cycles (No.) Endometrium (mm)* Cumulative pregnancy rate (No.)
OC/CC group
Control group
P value
17/24 (71%) 40/62 (64.5%) 9.4 ± 2.4 13/24 (54%)
2/24 (8%) 3/27 (11%) 7.1 ± 2 1/24 (4%)
.001 .001 NS .001
NS, Not significant (P < .05). *Data are given as mean ± SD.
treatment is effective at producing ovulation but is expensive, with significant risks of high-order multiple gestations and ovarian hyperstimulation syndrome. In an observational study, we showed that the use of oral contraceptive (OC) ovarian suppression for 2 months followed by the repeat administration of CC at the 100-mg dose produced effective ovulation and pregnancy rates in patients who had not ovulated with maximum doses of CC alone.2 It suggested that this approach might offer a low-cost and low-risk alternative to gonadotropin therapy for these patients. This study seeks to test the results of the initial study in a more rigorous manner by using the principles of evidence-based clinical medicine. It was also designed to attempt to identify the physiologic mechanism for its effectiveness. The objective of this study then was to evaluate the effectiveness and endocrine response of OC ovarian suppression followed by CC in patients who had not ovulated on maximal doses of CC, compared with a control group who had repeat CC use alone. Material and methods The study design was a prospective randomized clinical trial. Forty-eight patients who met all of the inclusion criteria were randomized into two treatment groups. For inclusion into the study, patients had to meet all of the following criteria: (1) have previously documented anovulation by transvaginal follicular monitoring while receiving CC in doses of ≥150 mg in our clinic, (2) be under the age of 36 years, (3) have documented patent tubes by either hysterosalpingogram or laparoscopy, (4) have normal fasting serum glucose and insulin levels, normal serum prolactin, thyroid-stimulating hormone, and follicle-stimulating hormone levels, dehydroepiandrosterone sulfate levels of ≤200 µg/mL, (5) be normoestrogenic, and (6) have no contraindication to OC use. The male partners had to have a normal semen analysis by World Health Organization criteria.14 The study protocol was approved by an independent institutional review board, and all subjects gave their written informed consent. Twenty-four patients were assigned randomly to group 1 (OC/CC) and received the monophasic low-dose OC Desogen (0.03 mg of ethinyl estradiol and 0.15 mg of desogestrel) continuously, for 42 to 50 days. After withdrawal bleeding, they took 100-mg CC on days 5 to 9 of that cycle.
Table II. Monthly cumulative pregnancy rate Month 1 2 3 4 5 6
Pregnancies (No.)
Cumulative pregnancy rate (%)
5 2 5 1 0 1
21 29 50 54 54 58
Group 2 (control) also had 24 patients and received no treatment for one or two spontaneous cycles (38-56 days), which was followed by 100 mg of CC on days 5 to 9 of their next cycle. This allowed a similar length no-treatment interval to the length of the OC use in group 1. All cycles were spontaneous, and none of the patients required medroxyprogesterone (Provera) for withdrawl bleeding to occur. Day 3 serum 17β-estradiol, follicle-stimulating hormone, luteinizing hormone, testosterone, and androstenedione were assayed in the cycle before the administration of CC; in the cycle, the patients in both groups received CC. Transvaginal ultrasound follicular monitoring was started on day 12 of the CC cycle and repeated every 1 to 2 days until the mean diameter of the lead follicle was ≥20 mm. If the follicle mean diameter failed to grow a minimum of 1 mm per day after a mean diameter of 14 mm was achieved or a 14-mm mean diameter was not achieved, the monitoring was stopped, and the cycle was canceled. Human chorionic gonadotrophin (10,000 IU intramuscularly) was given when the mean diameter of the lead follicle was ≥20 mm. A follow-up ultrasound scan was performed 2 days after human chorionic gonadotropin (hCG) administration to confirm corpus luteum formation. Ultrasound ovulation criteria were the disappearance of the preovulatory follicle, the presence of fluid in the cul-de-sac, and/or the formation of an echogenic cyst that was consistent with a corpus luteum. If only one of these criteria were clearly present on ultrasound scan, a midluteal progesterone sample was drawn to confirm ovulation. Thirty percent of the cycles had ovulation confirmed by midluteal progesterone levels. Follicle growth, endometrial development, and ovulation were confirmed by ultrasound scan. Pregnancy was confirmed by serum hCG levels and 7-week gestation ultrasound scan. Follicle growth, endometrial development, and pregnancy rate were recorded for each
1426 Branigan and Estes
June 2003 Am J Obstet Gynecol
Table III. Day 3 endocrine profiles before and after OC or no treatment intervals OC treatment group
17β-Estradiol (pg/mL)* Follicle-stimulating hormone (IU/L) Luteinizing hormone (IU/L) Testosterone (ng/dL)* Androstenedione (ng/dL)*
Control group
Before
After
P value
Before
After
P value
68 ± 15 6.2 17.8 46.9 ± 10 2.2 ± 0.4
25.7 ± 2.6 5.0 7.1 21.3 ± 3.2 1.0 ± 0.9
.01 NS .001 .001 .03
70 ± 12 6.8 15.2 51.7 ± 13 2.3 ± 0.6
74 ± 16 6.6 17.5 48.6 ± 11 2.7 ± 0.9
NS NS NS NS NS
NS, Not significant. *Data are given as mean ± SD.
patient and treatment cycle. The cumulative monthly pregnancy rate was also recorded. If the patient ovulated in the initial cycle but no pregnancy occurred, the patient was given 100 mg of CC on days 5 to 9 of the next cycle and was monitored again in the same manner to confirm ovulation. This process was repeated for ≤6 cycles, if the patient continued to ovulate, and the results were recorded. The randomization process used random permuted blocks with a block size of 4 to generate the two groups. Before the start of the study, 54 consecutive numbered opaque envelopes contained the group assignments, which were then opened after the patient enrolled in the study. On the basis of our previous observational study sample size, our calculations assumed an ovulation rate of 70% in the OC/CC group and a possible 15% ovulation rate in the CC-only group.2 With this assumption and with 80% power to detect this difference, 18 patients would be needed in each group. Twenty-four patients were enrolled in each group to allow for possibly lower rates. The Student t test for paired groups and analysis of variance were used for statistical significance between the groups and for hormonal comparisons. Results Initially, a total of 51 patients were assigned randomly, but 2 patients in the OC/CC group and 1 patient in the control group did not begin the study, which resulted in two groups of 24 each. All the patients completed the initial treatment cycle, but only those patients who achieved ovulation and who did not become pregnant had further CC cycles. There were 89 total treatment cycles, with 62 cycles in the OC/CC group and 27 cycles in the control group. The baseline demographic data showed no significant differences between the groups. The mean age was 28.2 ± 3.4 years, the mean body mass index was 30.3 ± 5.3 kg/m2, the mean weight was 82.2 ± 2.9 kg, and the mean years of infertility were 2.4 ± 0.8. Table I shows the response to CC in each of the groups. There were significantly higher percentages of patients who initially achieved ovulation (17/24 [71%] vs 2/24 [8%]), which resulted in a significantly higher number of total ovulatory treatment cycles (40/62 [64.5%] vs 3/27 [11%]).
Those patients who achieved ovulation but who did not become pregnant continued to receive CC in their next cycles, with cumulative pregnancy rates in the OC-treatment group compared with the control group. The OCtreatment group had better developed endometrium (9.4 mm vs 7.1 mm) but did not reach statistical significance. Of the 14 pregnancies that occurred during the study, 13 pregnancies were in the OC/CC group, and only 1 pregnancy was in the control group. Table II shows the monthly cumulative pregnancy rate that occurred during the study. Most of the pregnancies occurred in the first three ovulatory cycles, with only 1 pregnancy in the fourth cycle and 1 pregnancy in the sixth cycle. There were two spontaneous abortions (including the only pregnancy in the control group), three twin gestations, and no higher-order gestations. Table III shows the day 3 endocrine profiles before OC treatment in the OC-treatment group and before the notreatment interval in the control group. These same endocrine profiles were repeated on day 3 of the CC treatment cycle in both groups. As might be expected, no significant changes were seen in the control group during the no-treatment interval. Several significant changes were seen in response to OC treatment in that group. Significant decreases in 17β-estradiol, luteinizing hormone, androstenedione, and testosterone were seen. These changes may be responsible for the improved ovulatory response that was noted in that group. To get a clearer idea of how the endocrine changes might be effecting ovulation, we looked at the day 3 endocrine profiles of the subgroup of patients in each of the two groups who demonstrated responses that were opposite to the way the rest of the group responded. This is the subgroup who did not achieve ovulation with OC treatment and the patients who did achieve ovulation in the control group. These results are summarized in Table IV. The two patients who did achieve ovulation in the control group had ovarian androgens below the mean for their group and had significant spontaneous decreases during the no-treatment interval. Of the seven patients who did not achieve ovulation in the OC-treatment group, significant decreases in 17β-estradiol and luteinizing hormone were seen, but their testosterone and androstenedione
Branigan and Estes 1427
Volume 188, Number 6 Am J Obstet Gynecol
Table IV. Day 3 endocrine profiles of the OC/CC group that did not ovulate and the control group that ovulated OC/CC group (n = 7)
17β-Estradiol (pg/mL)* Follicle-stimulating hormone (IU/L) Luteinizing hormone (IU/L) Testosterone (ng/dL)* Androstenedione (ng/dL)*
Control group (n = 2)
Before
After
P value
Before
After
P value
61 ± 10 6 18 54.7 ± 18 2.1 ± 0.4
30 ± 6.4 5.1 8 57.4 ± 19 1.9 ± 0.2
.03 NS .01 NS NS
53 ± 9 6.5 15.2 36 ± 6 1.7 ± 0.1
48 ± 8 6.6 17.5 27 ± 5 1.3 ± 0.2
NS NS NS .05 .05
NS, Not significant. *Data are given as mean ± SD.
levels were not changed significantly. From these results, the androgen changes seem to be the most important of the endocrine profile changes for the determination of whether ovulation will occur. Comment For the normogonadotropic anovulatory infertility patient, CC represents the first line of treatment. Most women ovulate with CC, but live birth rates are low, and significant numbers of women do not achieve ovulation on receiving any dose of CC.15 A few adjunctive therapies can be used in these patients in limited situations. Most of the patients are treated with gonadotropins when CC fails to induce ovulation. This treatment is expensive, with significant risk of high-order multiple gestation; the risk of ovarian hyperstimulation syndrome is especially high in this group of patients. There is a need for a low-cost and low-risk alternative to gonadotropin therapy. This study showed that the suppression of the hypothalamic-pituitary-ovarian axis with OCs followed by CC treatment results in an excellent rate of ovulation and pregnancy compared with repeat CC-only treatment in patients who previously had not achieved ovulation on maximal doses of CC. Patients who ovulated but did not become pregnant continued to ovulate in the next cycle with repeat CC treatment. The total cost in our clinic for this treatment (CC, hCG, and ultrasound scanning) averages $200 to $300 per cycle compared with $2000 to $2500 dollars for a typical gonadotropin treatment cycle. This treatment regimen is an effective low-cost alternative without the associated risks of gonadotropins. The use of OCs in this study resulted in significant decreases in early follicular ovarian androgen production and luteinizing hormone and 17β-estradiol levels. These endocrine changes in the early follicular-phase ovarian environment may be responsible for the improved ovarian response to CC. The decrease in ovarian androgens may be the most important of these changes. In the subgroup of patients who received OCs and who did not achieve ovulation while receiving CC, their luteinizing hormone and 17β-estradiol levels were decreased, but not their androgen levels. In the control group during the no-treatment cycle, the subgroup who ovulated had lower
androgen levels to begin with and showed spontaneous decreases in their androgen levels in that time. In either group, if there was no decrease in androgen levels, then the patient did not achieve ovulation while receiving CC. Once ovulation occurred, these patients continued to ovulate with repeat CC treatment. Presumably, the decreased androgen levels continue to be suppressed if ovulation occurs, but this was not measured in this study The adverse effects of androgen level on folliculogenesis have been shown in several papers. It was reported that smaller follicles have increased androgen levels and decreased estrogen levels in their antral fluid compared with preovulatory follicles.16 One of the mechanisms that may explain how ovarian wedge resection works is the marked, but transient, reduction of ovarian androstenedione production and the persistent decrease in testosterone secretion after the surgery.17 This acute decrease of androgen production might result in a decrease of its inhibitory effect on follicular maturation and allow a more favorable microenvironment for folliculogenesis to occur. A decrease in peripheral androgen levels, specifically androstenedione, and testosterone independent of gonadotropin levels is one of the metabolic effects of insulin-sensitizing agents.18 This finding may help explain their usefulness as ovulation induction agents. The idea for this study came from the clinical observation that women with chronic anovulation are often taking OCs when they are not trying to conceive. Many women have a history of a pregnancy that occurred either while they were taking the OCs, especially if they missed a pill or two, or a conception in the first few months after stopping OCs. During the development of the OC, Rock et al in the early 1950s were using the first combination estradiol and progesterone pills, not for contraception but to treat infertility patients. The goal of their treatment was for the rebound effect after the pills were stopped.19 OC treatment does not reduce ovarian activity to zero. Studies have shown that changes in endogenous 17βestradiol and gonadotropins occur and have shown ultrasound evidence of follicular growth and endometrial changes during OC treatment.20 Follicles have been shown to grow to sufficient size and to result in ovulation
1428 Branigan and Estes
during OC treatment cycles.21 There is a rapid return of ovarian function in the cycle after OCs are stopped, especially with low-dose OCs, in women with normal ovulatory cycles before OC treatment.22 Our data suggest that OC treatment in anovulatory patients causes decreases in ovarian androgen levels that allow CC to work more effectively. Those patients who did not respond to OC treatment continued to have elevated androgen levels, which suggests that their androgen levels may not be ovarian in origin. This study and our original observational study showed excellent rates of ovulation and pregnancy. However, this study was randomized, with a control group and sufficient size and power. The cumulative 6-month pregnancy rate was >50% in this study, which is better than the pregnancy rates in patients who achieve ovulation initially while receiving CC.15 The improved pregnancy rates in this study were a result of a decrease in miscarriage rate in these patients; a much larger study would be needed to confirm this finding. Klein and Mishell23 examined daily follicle-stimulating hormone, luteinizing hormone, estradiol, progesterone, and prolactin levels for 2 months in women who had discontinued the use of OCs. These women had normal cyclic patterns that were compatible with normal ovulatory cycles, except for prolonged luteinizing surges 21 to 28 days after stopping the pills. Because of this delayed luteinizing surge, we used hCG to stimulate ovulation when there was adequate follicular development by transvaginal ultrasound scans. The use of single dose of hCG in CC cycles may increase the efficacy of the CC induction of ovulation of any CC cycle.24 The use of ultrasound scanning for the timing of hCG administration in CC cycles has also been established.25 We believe that this is an important part of the protocol. This study clearly demonstrates that the changes in hypothalamic-pituitary-ovarian axis in response to OC resulted in significant decreases in early follicular ovarian androgen production, luteinizing hormone, and 17β-estradiol levels. These endocrine changes in the early follicular phase ovarian environment may be responsible for the improved ovarian response to CC. The improved ovulation rates and resulting pregnancies make this a clear-cut alternative to gonadotropin therapy. It offers a low-cost and low-risk approach to ovulation induction for the CC-resistant anovulation patient. REFERENCES
1. Gysler M, March CM, Mishell DR Jr, Bailey EJ. A decade’s experience with the individualized clomiphene treatment regimen including its effect on the postcoital test. Fertil Steril 1982:37:161-7. 2. Branigan EF, Estes MA. Treatment of chronic anovulation resistant to clomiphene citrate (CC) by using oral contraceptive ovarian suppression followed by repeat CC treatment. Fertil Steril 1999;3:544-6. 3. Pasquali R, Antenucci D, Casimirri F, Venturoli S, Paradisi R, Fabbri R, et al. Clinical and hormonal characteristics of obese amenorrheic hyperandrogenic women before and after weight loss. J Clin Endocrinol Metab 1989;68:173-9.
June 2003 Am J Obstet Gynecol
4. Kletzky OA, Marrs RP, Davajan V. Management of patients with hyperprolactinemia and normal or abnormal tomograms. Am J Obstet Gynecol 1983;147:528-30. 5. Crosignani PG, Ferrari C, Scardnelli C, Picciotti MC, Caldara R, Malinverni A. Spontaneous and induced pregnancies in hyperprolactinemic women. Obstet Gynecol 1981;58:708-11. 6. Nestler JE, Jakubowicz DJ, Evans WS, Pasquali R. Effects of metformin on spontaneous and clomiphene-induced ovulation in the polycystic ovary syndrome. N Engl J Med 1998;338:1876-80. 7. Mitwally MFM, Kuscu NK, Yalcinkaya TM. High ovulatory rates with use of troglitazone in CC-resistant women with polycystic ovary syndrome. Hum Reprod 1999;14:2700-3. 8. Moghetti P, Castello R, Negri C, Tosi F, Perrone F, Caputo M, et al. Metformin effects on clinical features endocrine and metabolic profiles and insulin sensitivity in polycystic ovary syndrome: a randomized, double-blind, placebo-controlled 6 month trial, followed by open long-term clinical evaluation. J Clin Endocrinol Metab 2000;85:139-46. 9. Lobo RA, Wellington P, March CM, Granger L, Kletzky OA. Clomiphene and dexamethasone in women unresponsive to clomiphene alone. Obstet Gynecol 1982;60:497-501. 10. Hoffman D, Lobo RA. Serum dehydroepiandrosterone sulfate and the use of clomiphene citrate in anovulatory women. Fertil Steril 1985;43:196-9. 11. O’Herliky C, Pepperell RJ, Brown JB, Smith MA, Sandri L, McBain JC. Incremental clomiphene therapy: a new method for treating persistent anovulation. Obstet Gynecol 1981;58:535-8. 12. Garcia-Flores RF, Vazquez-Mendez J. Progressive dosages of clomiphene in hypothalamic anovulation. Fertil Steril 1984;42:543-7. 13. Farki J, Soule S, Jacobs HS. Effect of laparoscopic ovarian electrocautery on ovarian response and outcome of treatment with gonadotropins in clomiphene citrate-resistant patients with polycystic ovary syndrome. Fertil Steril 1995;64:930-5. 14. World Health Organization. Laboratory manual for the examination of human semen and sperm-cervical mucus interaction. 3rd ed. New York: Cambridge University Press; 1993. p. 43-4. 15. Sagle M, Bishop K, Ridley N, Alexander FM, Michel M, Bonney RC, et al. Recurrent early miscarriage and polycystic ovaries. BMJ 1988;297:1027-8. 16. Sanyal MK, Berger MJ, Thompson IE, Taymore ML, Horne HW. Development of graafian follicles in adult human ovary: correlation of estradiol and progesterone concentrations in antral fluid with growth of follicles. J Clin Endocrinol Metab 1974;38:828-31. 17. Judd HL, Rigg LA, Anderson DC, Yen SCC. The effects of ovarian wedge resection on circulating gonadotropin and ovarian steroid levels in patients with polycystic ovarian syndrome. J Clin Endocrinol Metab 1976;43:347-50. 18. Ehrmann DA, Schneider DJ, Sobel BE, et al. Troglitazone improves defects in insulin action, insulin secretion, ovarian steroidogenesis and fibrinolysis in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82:2108-12. 19. Wah R, Barbieri RL. Hormonal contraceptive agents. In: Barbieri RL, Schiff I, editors. Reproductive endocrine therapeutics. New York: Alan Liss; 1988. p. 352. 20. Spona J, Elstein M, Sullivan H, Ludicke F, Muller U, Dusterberg B. Shorter pill-free interval in combined oral contraceptives decreases follicular development. Contraception 1996;54:71-7. 21. Spona J, Ferchtinger W, Kinderman C, Wunsch C, Brill K. Inhibition of ovulation by an oral contraceptive containing 100 µg levonorgestrel on combination with 20 µg ethinyl estradiol. Contraception 1996;54:299-304. 22. Killick SR. Ovarian follicles during oral contraceptive cycles: their potential for ovulation. Fertil Steril 1989;52:580-2. 23. Klein TA, Mishell DR. Gonadotropin prolactin and steroid hormone levels after discontinuation of oral contraceptives. Am J Obstet Gynecol 1977;127:585-90. 24. Swyer GI, Radwanska E, McGarrigle HH. Plasma estradiol and progesterone estimation for the monitoring of induction of ovulation with clomiphene and chorionic gonadotropin. Br J Obstet Gynaecol 1975;82:794-9. 25. O’Herliky C, Pepperell RJ, Robinson HP. Ultrasound timing of human chorionic gonadotropin administration in clomiphenestimulated cycles. Obstet Gynecol 1982;59:40-6.
Volume 188, Number 6 Am J Obstet Gynecol
Editors’ note: This manuscript was revised after these discussions were presented. Discussion DR MICHAEL KETTEL, San Diego, Calif. Drs Branigan and Estes have accomplished something that is seen rarely in today’s world of clinical investigation. Their project began with an idea; the idea was researched effectively through the literature and investigated initially through a well-designed pilot study. After analyzing the data from the pilot study, they followed with the randomized, prospective clinical trial for which we have heard the results of today. This study is powerful and clinically relevant. Treatment of anovulation with CC was described initially >50 years ago, yet there are still patients for whom treatment is ineffective or who achieve ovulation and do not conceive. A variety of choices are available for the physician who treats these patients who do not achieve ovulation with the use of CC. There are several adjuvant medications that can be added to the CC treatment, which include corticosteroids, dopamine agonists such as bromocriptine, or insulin-sensitizing agents such as metformin. These adjuvants work inconsistently, and are few instances in the literature that describe the optimal patients for consider for these choices. When CC fails to induce ovulation, the alternative that is often chosen is to switch to injectable gonadotropins. Injectable gonadotropins are expensive, more invasive, and costly. They also carry a higher risk for multiple gestation and ovarian hyperstimulation syndrome. The average cost of a cycle of clomiphene is between $200 and $300, and the average cost of a cycle of gonadotropins usually exceeds $2000. The risk for a multiple gestation is also increased with gonadotropin therapy. Different series have reported different risks; in general, clomiphene treatment carries a 10% risk for twins, and gonadotropin treatment carries a 20% risk for multiple gestations. However, with gonadotropin therapy, there is a 5% risk for high-order multiple gestations. High-order multiple gestations are defined as triplets or greater. We are all acutely aware of the high-order pregnancies that are reported in the media and the political attention that has been generated in the efforts to reduce this complication of infertility care. This study describes a tool that can decrease the number of failed CC therapy and can decrease our need to rely on injectable gonadotropins. 1. How did you arbitrarily decide on a dose of 100 mg for the CC treatment when these patients had failed previously to achieve ovulation with 150 mg? 2. Anovulatory women can wait months for a spontaneous menstrual bleed. Did you induce menstrual withdrawal in any of these patients with medroxyprogesterone (Provera) to expedite treatment? If so, did
Branigan and Estes 1429
you measure the androgen levels after Provera therapy and were they suppressed? 3. Desogestrel (Desogen) was selected as the OC for suppression in this study. Was there a specific reason that this contraceptive was chosen? Did the low androgenic effect of desogestrel play an important role in the selection of this product or do you believe that any OC will do? 4. Many general obstetrician-gynecologists use clomiphene and have a wealth of clinical experience with this tool. You performed serial transvaginal ultrasound scans and triggered ovulation with hCG. Do you believe that this protocol should become the standard for prescribing clomiphene, or were these tools used only in the context of the research evaluation? 5. Was there a reason that you chose 42 to 50 days for continuous OC suppression instead of a standard 21day cycle pack? 6. Why do you think your pregnancy rates in the OCsuppressed group were so high? Has your clinical experience with this treatment since the completion of the study continued to result in these high success rates? DR GRETCHEN LENTZ, Seattle, Wash. Have you expanded this practice to women over age 36 years, or are you planning to study that group as well? DR FRANK GAMBERDELLA, Santa Barbara, Calif. Your paper is particularly timely for generalists who do infertility. 1. Do you think your patients would have benefitted from luteal progesterone support? 2. Did you look at cervical mucus at mid cycle as a parameter of ovulation and effective cervical mucus and sperm transport? 3. What are your plans for those who did not achieve ovulation with gonadotropin therapy or ovarian drilling? DR PHILLIP PATTON, Portland, Ore. Dr Robert Casper has treated a similar group of patients who were CC resistant with aromatase inhibitors. Rather than changes in androgen concentrations as a mechanism for CC responsiveness, could changes in estrogen concentration be an alternative mechanism to explain your results? DR LEE LEARMAN, San Francisco, Calif. In your protocol, you used hCG to trigger ovulation at a follicle size of 20 mm. Do you have any information on how often that follicle size was achieved in the treatment and control groups? If more follicles grew to that size in the treatment group, then hCG would become a cointervention in that group and would make us wonder whether we can generalize to a setting in which we do not use hCG to trigger ovulation. DR SIMON HENDERSON, San Francisco, Calif. Congratulations on your pregnancy rates, these were extraordinarily high. Did you do this protocol for each cycle in a
1430 Branigan and Estes
repetitive way? If the patient ovulated in the first cycle and did not get pregnant, did you also in the second cycle put her on birth control pills for a month or two or did you just go straight to your next cycle of CC (Clomid)? In my practice, patients are not patient unfortunately, they are usually impatient, and the suggestion of being on birth control pills for 2 months often is not a terribly successful suggestion. I put most of my patients on metformin therapy and actually that has worked out very well, but I cannot claim such a good pregnancy rate as yours. DR BRANIGAN (Closing). Regarding Dr Kettel’s questions: 1. In a pilot study that was done before our initial observational study, we tried CC in doses of 50, 100, and 150 mg. We found that only 50% of the patients who will achieve ovulation eventually do so at the 50-mg dosage. At the 100-mg dose, virtually all of those patients who will achieve ovulation will do so with only a few more at the 150-mg dosage. We gave the control group the 100-mg dose so that there would not be a dose effect that might alter the results. 2. We used only spontaneous menstrual bleeding and no Provera-induced withdrawal bleeding. Therefore, androgen levels would not be affected by Provera in our study. 3. In our original study, we chose Desogen because it is monophasic and has a low androgenic effect. However, we did not know until the completion of this study that this might have been a serendipitous choice. We have used other monophasic OCs in a few patients and have had success, but the numbers are small. We prefer to use Desogen because of the desogestrel component. 4. We use transvaginal ultrasound follicular monitoring in all our patients who are receiving CC. The information that can be obtained from an ultrasound scan allows one to adjust the cycle on the basis of follicular growth and endometrial development to maximize each cycle and to adjust future cycles in ways that cannot be duplicated with other ovulation monitoring tools. The ability of an individual obstetrician-gynecologist to perform the ultrasound scans and to interpret the data would be important in the determination of its clinical usefulness. An obstetrician-gynecologist who performs and is comfortable with transvaginal ultrasound scans should be able to adopt and use our protocol easily. 5. We chose 2 months of the continuous OCs that were used in this study on the basis of a pilot study that was done before our observational study. We found that,
June 2003 Am J Obstet Gynecol
if only 1 month of suppression is used, only approximately one third of the patients who will respond eventually will do so. When this is increased to 2 months of continuous suppression, almost all of the patients who will respond do, with very few added with longer suppression. 6. We believe that our pregnancy rates were so good because we chose young patients with isolated ovulation problems and no male factor problems. If you can get them to ovulate, they would be expected to have good pregnancy rates. We have used this protocol for >6 years in our clinic, and the pregnancy rates have remained high for these patients. Dr Lentz, we do use this protocol in older women. However, in patients >38 years old, more aggressive therapy (gonadotropins or in vitro fertilization) may be a better option because of the age-related decline in fertility. Dr Gamberdella, we do occasionally use luteal support in some of these types of patients. However, none was used in the study. The hCG we use to trigger ovulation is, in our opinion, more important than luteal support, which we rarely use. We do not use any mucus parameters in these patients. The endometrial development seen on ultrasound monitoring is a good indicator of estrogen effect. Dr Patton, our data suggest that androgen levels are more important than estrogen levels in the prediction of ovulation success. This was seen in the OC group who did not ovulate. This subgroup showed significant drops in their estrogen levels but not in their androgen levels. If no significant drop in androgen levels occurs, the patient does not ovulate, despite estrogen changes. We have done a workup of some of these patients and have found several cases of nonclassic (late onset) congenital adrenal hyperplasia and one case of Cushing syndrome. The patient with Cushing syndrome had vaginal spotting the entire time that she was receiving the OC treatment. Dr Learman, we have found that giving hCG at a mean diameter of 20 mm reliably produces ovulation. Dr Henderson, those patients who achieve ovulation with OC treatment, but who do not become pregnant, are not treated again with OCs in the next cycle. They are given 100 mg of CC on days 5 to 9 of their next cycle and are monitored in the same manner. We believe that an ovulatory cycle keeps androgen levels low, but this was not measured specifically in this study. However, almost all those patients who did achieve ovulation in the initial cycle, but who did not become pregnant, continue to ovulate with repeat CC for ≥6 cycles.
Key words: Clomiphene citrate, oral contraceptive, chronic anovulation, suppression, androgen
Chronic anovulation is a common cause of infertility in women. Most of these women are treated initially with clomiphene citrate (CC) in doses from 50 to 250 mg. One large study found that 50% of patients ovulate at the 50-mg dosage and that 74% of patients ovulate on the 100-mg dose.1 However, despite successful ovulation, <50% of these women will become pregnant. In our experience >25% of women whom we see with chronic anovulation fail to ovulate on any dose of CC.2 For these women, gonadotropin therapy or a few limited adjunctive therapies used with CC or alone are the next treatment option. Adjunctive therapy options for CC-resistant women include weight loss, insulin-sensitizing agents, bromocrip-
From the Bellingham In Vitro Fertilization and Infertility Center. Presented at the Sixty-Ninth Annual Meeting of the Pacific Coast Obstetrical and Gynecological Society, October 25-27, 2002, Rancho Mirage, Calif. Reprint requests: Emmett F. Branigan, MD, 2980 Squalicum, Parkway #103, Bellingham, WA 98225. © 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1067/mob.2003.459
1424
tine, glucocorticoids, extended doses or continued doses of CC, and surgery. A weight loss of as little as 5% of total body weight has been shown to produce ovulatory cycles.3 The time and effort that is involved or if the patient is lean limits this option. If hyperprolactinemia is present, bromocriptine (Parlodel) either alone or with CC can be successful.4,5 Metformin (an insulin sensitizing agent) used alone or with CC has been used for ovulation induction in an unselected group of patients.6 Other studies have confirmed these results, but which patients might benefit and specific treatment protocols have yet to be formulated.7,8 The use of dexamethasone or other glucocorticoids with CC is useful in hyperandrogenic women with dehydroepiandrosterone sulfate levels of ≥200 µg/mL.9,10 The use of doses of CC beyond the normal 5 days or even continuous doses until ovulation occurs have been used by several investigators.11,12 However, these regimens are limited because of patient complaints of side effects. Surgery (either ovarian wedge resection or laparoscopic ovarian drilling) can result in time-limited ovulatory cycles, but the treatment is expensive and is associated with pelvic adhesion risk.13 Exogenous gonadotropin therapy is usually the next treatment step for CC-resistant anovulation patients. This
Branigan and Estes 1425
Volume 188, Number 6 Am J Obstet Gynecol
Table I. Response to CC after OCs or no treatment intervals
Ovulation (No.) Ovulatory cycles (No.) Endometrium (mm)* Cumulative pregnancy rate (No.)
OC/CC group
Control group
P value
17/24 (71%) 40/62 (64.5%) 9.4 ± 2.4 13/24 (54%)
2/24 (8%) 3/27 (11%) 7.1 ± 2 1/24 (4%)
.001 .001 NS .001
NS, Not significant (P < .05). *Data are given as mean ± SD.
treatment is effective at producing ovulation but is expensive, with significant risks of high-order multiple gestations and ovarian hyperstimulation syndrome. In an observational study, we showed that the use of oral contraceptive (OC) ovarian suppression for 2 months followed by the repeat administration of CC at the 100-mg dose produced effective ovulation and pregnancy rates in patients who had not ovulated with maximum doses of CC alone.2 It suggested that this approach might offer a low-cost and low-risk alternative to gonadotropin therapy for these patients. This study seeks to test the results of the initial study in a more rigorous manner by using the principles of evidence-based clinical medicine. It was also designed to attempt to identify the physiologic mechanism for its effectiveness. The objective of this study then was to evaluate the effectiveness and endocrine response of OC ovarian suppression followed by CC in patients who had not ovulated on maximal doses of CC, compared with a control group who had repeat CC use alone. Material and methods The study design was a prospective randomized clinical trial. Forty-eight patients who met all of the inclusion criteria were randomized into two treatment groups. For inclusion into the study, patients had to meet all of the following criteria: (1) have previously documented anovulation by transvaginal follicular monitoring while receiving CC in doses of ≥150 mg in our clinic, (2) be under the age of 36 years, (3) have documented patent tubes by either hysterosalpingogram or laparoscopy, (4) have normal fasting serum glucose and insulin levels, normal serum prolactin, thyroid-stimulating hormone, and follicle-stimulating hormone levels, dehydroepiandrosterone sulfate levels of ≤200 µg/mL, (5) be normoestrogenic, and (6) have no contraindication to OC use. The male partners had to have a normal semen analysis by World Health Organization criteria.14 The study protocol was approved by an independent institutional review board, and all subjects gave their written informed consent. Twenty-four patients were assigned randomly to group 1 (OC/CC) and received the monophasic low-dose OC Desogen (0.03 mg of ethinyl estradiol and 0.15 mg of desogestrel) continuously, for 42 to 50 days. After withdrawal bleeding, they took 100-mg CC on days 5 to 9 of that cycle.
Table II. Monthly cumulative pregnancy rate Month 1 2 3 4 5 6
Pregnancies (No.)
Cumulative pregnancy rate (%)
5 2 5 1 0 1
21 29 50 54 54 58
Group 2 (control) also had 24 patients and received no treatment for one or two spontaneous cycles (38-56 days), which was followed by 100 mg of CC on days 5 to 9 of their next cycle. This allowed a similar length no-treatment interval to the length of the OC use in group 1. All cycles were spontaneous, and none of the patients required medroxyprogesterone (Provera) for withdrawl bleeding to occur. Day 3 serum 17β-estradiol, follicle-stimulating hormone, luteinizing hormone, testosterone, and androstenedione were assayed in the cycle before the administration of CC; in the cycle, the patients in both groups received CC. Transvaginal ultrasound follicular monitoring was started on day 12 of the CC cycle and repeated every 1 to 2 days until the mean diameter of the lead follicle was ≥20 mm. If the follicle mean diameter failed to grow a minimum of 1 mm per day after a mean diameter of 14 mm was achieved or a 14-mm mean diameter was not achieved, the monitoring was stopped, and the cycle was canceled. Human chorionic gonadotrophin (10,000 IU intramuscularly) was given when the mean diameter of the lead follicle was ≥20 mm. A follow-up ultrasound scan was performed 2 days after human chorionic gonadotropin (hCG) administration to confirm corpus luteum formation. Ultrasound ovulation criteria were the disappearance of the preovulatory follicle, the presence of fluid in the cul-de-sac, and/or the formation of an echogenic cyst that was consistent with a corpus luteum. If only one of these criteria were clearly present on ultrasound scan, a midluteal progesterone sample was drawn to confirm ovulation. Thirty percent of the cycles had ovulation confirmed by midluteal progesterone levels. Follicle growth, endometrial development, and ovulation were confirmed by ultrasound scan. Pregnancy was confirmed by serum hCG levels and 7-week gestation ultrasound scan. Follicle growth, endometrial development, and pregnancy rate were recorded for each
1426 Branigan and Estes
June 2003 Am J Obstet Gynecol
Table III. Day 3 endocrine profiles before and after OC or no treatment intervals OC treatment group
17β-Estradiol (pg/mL)* Follicle-stimulating hormone (IU/L) Luteinizing hormone (IU/L) Testosterone (ng/dL)* Androstenedione (ng/dL)*
Control group
Before
After
P value
Before
After
P value
68 ± 15 6.2 17.8 46.9 ± 10 2.2 ± 0.4
25.7 ± 2.6 5.0 7.1 21.3 ± 3.2 1.0 ± 0.9
.01 NS .001 .001 .03
70 ± 12 6.8 15.2 51.7 ± 13 2.3 ± 0.6
74 ± 16 6.6 17.5 48.6 ± 11 2.7 ± 0.9
NS NS NS NS NS
NS, Not significant. *Data are given as mean ± SD.
patient and treatment cycle. The cumulative monthly pregnancy rate was also recorded. If the patient ovulated in the initial cycle but no pregnancy occurred, the patient was given 100 mg of CC on days 5 to 9 of the next cycle and was monitored again in the same manner to confirm ovulation. This process was repeated for ≤6 cycles, if the patient continued to ovulate, and the results were recorded. The randomization process used random permuted blocks with a block size of 4 to generate the two groups. Before the start of the study, 54 consecutive numbered opaque envelopes contained the group assignments, which were then opened after the patient enrolled in the study. On the basis of our previous observational study sample size, our calculations assumed an ovulation rate of 70% in the OC/CC group and a possible 15% ovulation rate in the CC-only group.2 With this assumption and with 80% power to detect this difference, 18 patients would be needed in each group. Twenty-four patients were enrolled in each group to allow for possibly lower rates. The Student t test for paired groups and analysis of variance were used for statistical significance between the groups and for hormonal comparisons. Results Initially, a total of 51 patients were assigned randomly, but 2 patients in the OC/CC group and 1 patient in the control group did not begin the study, which resulted in two groups of 24 each. All the patients completed the initial treatment cycle, but only those patients who achieved ovulation and who did not become pregnant had further CC cycles. There were 89 total treatment cycles, with 62 cycles in the OC/CC group and 27 cycles in the control group. The baseline demographic data showed no significant differences between the groups. The mean age was 28.2 ± 3.4 years, the mean body mass index was 30.3 ± 5.3 kg/m2, the mean weight was 82.2 ± 2.9 kg, and the mean years of infertility were 2.4 ± 0.8. Table I shows the response to CC in each of the groups. There were significantly higher percentages of patients who initially achieved ovulation (17/24 [71%] vs 2/24 [8%]), which resulted in a significantly higher number of total ovulatory treatment cycles (40/62 [64.5%] vs 3/27 [11%]).
Those patients who achieved ovulation but who did not become pregnant continued to receive CC in their next cycles, with cumulative pregnancy rates in the OC-treatment group compared with the control group. The OCtreatment group had better developed endometrium (9.4 mm vs 7.1 mm) but did not reach statistical significance. Of the 14 pregnancies that occurred during the study, 13 pregnancies were in the OC/CC group, and only 1 pregnancy was in the control group. Table II shows the monthly cumulative pregnancy rate that occurred during the study. Most of the pregnancies occurred in the first three ovulatory cycles, with only 1 pregnancy in the fourth cycle and 1 pregnancy in the sixth cycle. There were two spontaneous abortions (including the only pregnancy in the control group), three twin gestations, and no higher-order gestations. Table III shows the day 3 endocrine profiles before OC treatment in the OC-treatment group and before the notreatment interval in the control group. These same endocrine profiles were repeated on day 3 of the CC treatment cycle in both groups. As might be expected, no significant changes were seen in the control group during the no-treatment interval. Several significant changes were seen in response to OC treatment in that group. Significant decreases in 17β-estradiol, luteinizing hormone, androstenedione, and testosterone were seen. These changes may be responsible for the improved ovulatory response that was noted in that group. To get a clearer idea of how the endocrine changes might be effecting ovulation, we looked at the day 3 endocrine profiles of the subgroup of patients in each of the two groups who demonstrated responses that were opposite to the way the rest of the group responded. This is the subgroup who did not achieve ovulation with OC treatment and the patients who did achieve ovulation in the control group. These results are summarized in Table IV. The two patients who did achieve ovulation in the control group had ovarian androgens below the mean for their group and had significant spontaneous decreases during the no-treatment interval. Of the seven patients who did not achieve ovulation in the OC-treatment group, significant decreases in 17β-estradiol and luteinizing hormone were seen, but their testosterone and androstenedione
Branigan and Estes 1427
Volume 188, Number 6 Am J Obstet Gynecol
Table IV. Day 3 endocrine profiles of the OC/CC group that did not ovulate and the control group that ovulated OC/CC group (n = 7)
17β-Estradiol (pg/mL)* Follicle-stimulating hormone (IU/L) Luteinizing hormone (IU/L) Testosterone (ng/dL)* Androstenedione (ng/dL)*
Control group (n = 2)
Before
After
P value
Before
After
P value
61 ± 10 6 18 54.7 ± 18 2.1 ± 0.4
30 ± 6.4 5.1 8 57.4 ± 19 1.9 ± 0.2
.03 NS .01 NS NS
53 ± 9 6.5 15.2 36 ± 6 1.7 ± 0.1
48 ± 8 6.6 17.5 27 ± 5 1.3 ± 0.2
NS NS NS .05 .05
NS, Not significant. *Data are given as mean ± SD.
levels were not changed significantly. From these results, the androgen changes seem to be the most important of the endocrine profile changes for the determination of whether ovulation will occur. Comment For the normogonadotropic anovulatory infertility patient, CC represents the first line of treatment. Most women ovulate with CC, but live birth rates are low, and significant numbers of women do not achieve ovulation on receiving any dose of CC.15 A few adjunctive therapies can be used in these patients in limited situations. Most of the patients are treated with gonadotropins when CC fails to induce ovulation. This treatment is expensive, with significant risk of high-order multiple gestation; the risk of ovarian hyperstimulation syndrome is especially high in this group of patients. There is a need for a low-cost and low-risk alternative to gonadotropin therapy. This study showed that the suppression of the hypothalamic-pituitary-ovarian axis with OCs followed by CC treatment results in an excellent rate of ovulation and pregnancy compared with repeat CC-only treatment in patients who previously had not achieved ovulation on maximal doses of CC. Patients who ovulated but did not become pregnant continued to ovulate in the next cycle with repeat CC treatment. The total cost in our clinic for this treatment (CC, hCG, and ultrasound scanning) averages $200 to $300 per cycle compared with $2000 to $2500 dollars for a typical gonadotropin treatment cycle. This treatment regimen is an effective low-cost alternative without the associated risks of gonadotropins. The use of OCs in this study resulted in significant decreases in early follicular ovarian androgen production and luteinizing hormone and 17β-estradiol levels. These endocrine changes in the early follicular-phase ovarian environment may be responsible for the improved ovarian response to CC. The decrease in ovarian androgens may be the most important of these changes. In the subgroup of patients who received OCs and who did not achieve ovulation while receiving CC, their luteinizing hormone and 17β-estradiol levels were decreased, but not their androgen levels. In the control group during the no-treatment cycle, the subgroup who ovulated had lower
androgen levels to begin with and showed spontaneous decreases in their androgen levels in that time. In either group, if there was no decrease in androgen levels, then the patient did not achieve ovulation while receiving CC. Once ovulation occurred, these patients continued to ovulate with repeat CC treatment. Presumably, the decreased androgen levels continue to be suppressed if ovulation occurs, but this was not measured in this study The adverse effects of androgen level on folliculogenesis have been shown in several papers. It was reported that smaller follicles have increased androgen levels and decreased estrogen levels in their antral fluid compared with preovulatory follicles.16 One of the mechanisms that may explain how ovarian wedge resection works is the marked, but transient, reduction of ovarian androstenedione production and the persistent decrease in testosterone secretion after the surgery.17 This acute decrease of androgen production might result in a decrease of its inhibitory effect on follicular maturation and allow a more favorable microenvironment for folliculogenesis to occur. A decrease in peripheral androgen levels, specifically androstenedione, and testosterone independent of gonadotropin levels is one of the metabolic effects of insulin-sensitizing agents.18 This finding may help explain their usefulness as ovulation induction agents. The idea for this study came from the clinical observation that women with chronic anovulation are often taking OCs when they are not trying to conceive. Many women have a history of a pregnancy that occurred either while they were taking the OCs, especially if they missed a pill or two, or a conception in the first few months after stopping OCs. During the development of the OC, Rock et al in the early 1950s were using the first combination estradiol and progesterone pills, not for contraception but to treat infertility patients. The goal of their treatment was for the rebound effect after the pills were stopped.19 OC treatment does not reduce ovarian activity to zero. Studies have shown that changes in endogenous 17βestradiol and gonadotropins occur and have shown ultrasound evidence of follicular growth and endometrial changes during OC treatment.20 Follicles have been shown to grow to sufficient size and to result in ovulation
1428 Branigan and Estes
during OC treatment cycles.21 There is a rapid return of ovarian function in the cycle after OCs are stopped, especially with low-dose OCs, in women with normal ovulatory cycles before OC treatment.22 Our data suggest that OC treatment in anovulatory patients causes decreases in ovarian androgen levels that allow CC to work more effectively. Those patients who did not respond to OC treatment continued to have elevated androgen levels, which suggests that their androgen levels may not be ovarian in origin. This study and our original observational study showed excellent rates of ovulation and pregnancy. However, this study was randomized, with a control group and sufficient size and power. The cumulative 6-month pregnancy rate was >50% in this study, which is better than the pregnancy rates in patients who achieve ovulation initially while receiving CC.15 The improved pregnancy rates in this study were a result of a decrease in miscarriage rate in these patients; a much larger study would be needed to confirm this finding. Klein and Mishell23 examined daily follicle-stimulating hormone, luteinizing hormone, estradiol, progesterone, and prolactin levels for 2 months in women who had discontinued the use of OCs. These women had normal cyclic patterns that were compatible with normal ovulatory cycles, except for prolonged luteinizing surges 21 to 28 days after stopping the pills. Because of this delayed luteinizing surge, we used hCG to stimulate ovulation when there was adequate follicular development by transvaginal ultrasound scans. The use of single dose of hCG in CC cycles may increase the efficacy of the CC induction of ovulation of any CC cycle.24 The use of ultrasound scanning for the timing of hCG administration in CC cycles has also been established.25 We believe that this is an important part of the protocol. This study clearly demonstrates that the changes in hypothalamic-pituitary-ovarian axis in response to OC resulted in significant decreases in early follicular ovarian androgen production, luteinizing hormone, and 17β-estradiol levels. These endocrine changes in the early follicular phase ovarian environment may be responsible for the improved ovarian response to CC. The improved ovulation rates and resulting pregnancies make this a clear-cut alternative to gonadotropin therapy. It offers a low-cost and low-risk approach to ovulation induction for the CC-resistant anovulation patient. REFERENCES
1. Gysler M, March CM, Mishell DR Jr, Bailey EJ. A decade’s experience with the individualized clomiphene treatment regimen including its effect on the postcoital test. Fertil Steril 1982:37:161-7. 2. Branigan EF, Estes MA. Treatment of chronic anovulation resistant to clomiphene citrate (CC) by using oral contraceptive ovarian suppression followed by repeat CC treatment. Fertil Steril 1999;3:544-6. 3. Pasquali R, Antenucci D, Casimirri F, Venturoli S, Paradisi R, Fabbri R, et al. Clinical and hormonal characteristics of obese amenorrheic hyperandrogenic women before and after weight loss. J Clin Endocrinol Metab 1989;68:173-9.
June 2003 Am J Obstet Gynecol
4. Kletzky OA, Marrs RP, Davajan V. Management of patients with hyperprolactinemia and normal or abnormal tomograms. Am J Obstet Gynecol 1983;147:528-30. 5. Crosignani PG, Ferrari C, Scardnelli C, Picciotti MC, Caldara R, Malinverni A. Spontaneous and induced pregnancies in hyperprolactinemic women. Obstet Gynecol 1981;58:708-11. 6. Nestler JE, Jakubowicz DJ, Evans WS, Pasquali R. Effects of metformin on spontaneous and clomiphene-induced ovulation in the polycystic ovary syndrome. N Engl J Med 1998;338:1876-80. 7. Mitwally MFM, Kuscu NK, Yalcinkaya TM. High ovulatory rates with use of troglitazone in CC-resistant women with polycystic ovary syndrome. Hum Reprod 1999;14:2700-3. 8. Moghetti P, Castello R, Negri C, Tosi F, Perrone F, Caputo M, et al. Metformin effects on clinical features endocrine and metabolic profiles and insulin sensitivity in polycystic ovary syndrome: a randomized, double-blind, placebo-controlled 6 month trial, followed by open long-term clinical evaluation. J Clin Endocrinol Metab 2000;85:139-46. 9. Lobo RA, Wellington P, March CM, Granger L, Kletzky OA. Clomiphene and dexamethasone in women unresponsive to clomiphene alone. Obstet Gynecol 1982;60:497-501. 10. Hoffman D, Lobo RA. Serum dehydroepiandrosterone sulfate and the use of clomiphene citrate in anovulatory women. Fertil Steril 1985;43:196-9. 11. O’Herliky C, Pepperell RJ, Brown JB, Smith MA, Sandri L, McBain JC. Incremental clomiphene therapy: a new method for treating persistent anovulation. Obstet Gynecol 1981;58:535-8. 12. Garcia-Flores RF, Vazquez-Mendez J. Progressive dosages of clomiphene in hypothalamic anovulation. Fertil Steril 1984;42:543-7. 13. Farki J, Soule S, Jacobs HS. Effect of laparoscopic ovarian electrocautery on ovarian response and outcome of treatment with gonadotropins in clomiphene citrate-resistant patients with polycystic ovary syndrome. Fertil Steril 1995;64:930-5. 14. World Health Organization. Laboratory manual for the examination of human semen and sperm-cervical mucus interaction. 3rd ed. New York: Cambridge University Press; 1993. p. 43-4. 15. Sagle M, Bishop K, Ridley N, Alexander FM, Michel M, Bonney RC, et al. Recurrent early miscarriage and polycystic ovaries. BMJ 1988;297:1027-8. 16. Sanyal MK, Berger MJ, Thompson IE, Taymore ML, Horne HW. Development of graafian follicles in adult human ovary: correlation of estradiol and progesterone concentrations in antral fluid with growth of follicles. J Clin Endocrinol Metab 1974;38:828-31. 17. Judd HL, Rigg LA, Anderson DC, Yen SCC. The effects of ovarian wedge resection on circulating gonadotropin and ovarian steroid levels in patients with polycystic ovarian syndrome. J Clin Endocrinol Metab 1976;43:347-50. 18. Ehrmann DA, Schneider DJ, Sobel BE, et al. Troglitazone improves defects in insulin action, insulin secretion, ovarian steroidogenesis and fibrinolysis in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82:2108-12. 19. Wah R, Barbieri RL. Hormonal contraceptive agents. In: Barbieri RL, Schiff I, editors. Reproductive endocrine therapeutics. New York: Alan Liss; 1988. p. 352. 20. Spona J, Elstein M, Sullivan H, Ludicke F, Muller U, Dusterberg B. Shorter pill-free interval in combined oral contraceptives decreases follicular development. Contraception 1996;54:71-7. 21. Spona J, Ferchtinger W, Kinderman C, Wunsch C, Brill K. Inhibition of ovulation by an oral contraceptive containing 100 µg levonorgestrel on combination with 20 µg ethinyl estradiol. Contraception 1996;54:299-304. 22. Killick SR. Ovarian follicles during oral contraceptive cycles: their potential for ovulation. Fertil Steril 1989;52:580-2. 23. Klein TA, Mishell DR. Gonadotropin prolactin and steroid hormone levels after discontinuation of oral contraceptives. Am J Obstet Gynecol 1977;127:585-90. 24. Swyer GI, Radwanska E, McGarrigle HH. Plasma estradiol and progesterone estimation for the monitoring of induction of ovulation with clomiphene and chorionic gonadotropin. Br J Obstet Gynaecol 1975;82:794-9. 25. O’Herliky C, Pepperell RJ, Robinson HP. Ultrasound timing of human chorionic gonadotropin administration in clomiphenestimulated cycles. Obstet Gynecol 1982;59:40-6.
Volume 188, Number 6 Am J Obstet Gynecol
Editors’ note: This manuscript was revised after these discussions were presented. Discussion DR MICHAEL KETTEL, San Diego, Calif. Drs Branigan and Estes have accomplished something that is seen rarely in today’s world of clinical investigation. Their project began with an idea; the idea was researched effectively through the literature and investigated initially through a well-designed pilot study. After analyzing the data from the pilot study, they followed with the randomized, prospective clinical trial for which we have heard the results of today. This study is powerful and clinically relevant. Treatment of anovulation with CC was described initially >50 years ago, yet there are still patients for whom treatment is ineffective or who achieve ovulation and do not conceive. A variety of choices are available for the physician who treats these patients who do not achieve ovulation with the use of CC. There are several adjuvant medications that can be added to the CC treatment, which include corticosteroids, dopamine agonists such as bromocriptine, or insulin-sensitizing agents such as metformin. These adjuvants work inconsistently, and are few instances in the literature that describe the optimal patients for consider for these choices. When CC fails to induce ovulation, the alternative that is often chosen is to switch to injectable gonadotropins. Injectable gonadotropins are expensive, more invasive, and costly. They also carry a higher risk for multiple gestation and ovarian hyperstimulation syndrome. The average cost of a cycle of clomiphene is between $200 and $300, and the average cost of a cycle of gonadotropins usually exceeds $2000. The risk for a multiple gestation is also increased with gonadotropin therapy. Different series have reported different risks; in general, clomiphene treatment carries a 10% risk for twins, and gonadotropin treatment carries a 20% risk for multiple gestations. However, with gonadotropin therapy, there is a 5% risk for high-order multiple gestations. High-order multiple gestations are defined as triplets or greater. We are all acutely aware of the high-order pregnancies that are reported in the media and the political attention that has been generated in the efforts to reduce this complication of infertility care. This study describes a tool that can decrease the number of failed CC therapy and can decrease our need to rely on injectable gonadotropins. 1. How did you arbitrarily decide on a dose of 100 mg for the CC treatment when these patients had failed previously to achieve ovulation with 150 mg? 2. Anovulatory women can wait months for a spontaneous menstrual bleed. Did you induce menstrual withdrawal in any of these patients with medroxyprogesterone (Provera) to expedite treatment? If so, did
Branigan and Estes 1429
you measure the androgen levels after Provera therapy and were they suppressed? 3. Desogestrel (Desogen) was selected as the OC for suppression in this study. Was there a specific reason that this contraceptive was chosen? Did the low androgenic effect of desogestrel play an important role in the selection of this product or do you believe that any OC will do? 4. Many general obstetrician-gynecologists use clomiphene and have a wealth of clinical experience with this tool. You performed serial transvaginal ultrasound scans and triggered ovulation with hCG. Do you believe that this protocol should become the standard for prescribing clomiphene, or were these tools used only in the context of the research evaluation? 5. Was there a reason that you chose 42 to 50 days for continuous OC suppression instead of a standard 21day cycle pack? 6. Why do you think your pregnancy rates in the OCsuppressed group were so high? Has your clinical experience with this treatment since the completion of the study continued to result in these high success rates? DR GRETCHEN LENTZ, Seattle, Wash. Have you expanded this practice to women over age 36 years, or are you planning to study that group as well? DR FRANK GAMBERDELLA, Santa Barbara, Calif. Your paper is particularly timely for generalists who do infertility. 1. Do you think your patients would have benefitted from luteal progesterone support? 2. Did you look at cervical mucus at mid cycle as a parameter of ovulation and effective cervical mucus and sperm transport? 3. What are your plans for those who did not achieve ovulation with gonadotropin therapy or ovarian drilling? DR PHILLIP PATTON, Portland, Ore. Dr Robert Casper has treated a similar group of patients who were CC resistant with aromatase inhibitors. Rather than changes in androgen concentrations as a mechanism for CC responsiveness, could changes in estrogen concentration be an alternative mechanism to explain your results? DR LEE LEARMAN, San Francisco, Calif. In your protocol, you used hCG to trigger ovulation at a follicle size of 20 mm. Do you have any information on how often that follicle size was achieved in the treatment and control groups? If more follicles grew to that size in the treatment group, then hCG would become a cointervention in that group and would make us wonder whether we can generalize to a setting in which we do not use hCG to trigger ovulation. DR SIMON HENDERSON, San Francisco, Calif. Congratulations on your pregnancy rates, these were extraordinarily high. Did you do this protocol for each cycle in a
1430 Branigan and Estes
repetitive way? If the patient ovulated in the first cycle and did not get pregnant, did you also in the second cycle put her on birth control pills for a month or two or did you just go straight to your next cycle of CC (Clomid)? In my practice, patients are not patient unfortunately, they are usually impatient, and the suggestion of being on birth control pills for 2 months often is not a terribly successful suggestion. I put most of my patients on metformin therapy and actually that has worked out very well, but I cannot claim such a good pregnancy rate as yours. DR BRANIGAN (Closing). Regarding Dr Kettel’s questions: 1. In a pilot study that was done before our initial observational study, we tried CC in doses of 50, 100, and 150 mg. We found that only 50% of the patients who will achieve ovulation eventually do so at the 50-mg dosage. At the 100-mg dose, virtually all of those patients who will achieve ovulation will do so with only a few more at the 150-mg dosage. We gave the control group the 100-mg dose so that there would not be a dose effect that might alter the results. 2. We used only spontaneous menstrual bleeding and no Provera-induced withdrawal bleeding. Therefore, androgen levels would not be affected by Provera in our study. 3. In our original study, we chose Desogen because it is monophasic and has a low androgenic effect. However, we did not know until the completion of this study that this might have been a serendipitous choice. We have used other monophasic OCs in a few patients and have had success, but the numbers are small. We prefer to use Desogen because of the desogestrel component. 4. We use transvaginal ultrasound follicular monitoring in all our patients who are receiving CC. The information that can be obtained from an ultrasound scan allows one to adjust the cycle on the basis of follicular growth and endometrial development to maximize each cycle and to adjust future cycles in ways that cannot be duplicated with other ovulation monitoring tools. The ability of an individual obstetrician-gynecologist to perform the ultrasound scans and to interpret the data would be important in the determination of its clinical usefulness. An obstetrician-gynecologist who performs and is comfortable with transvaginal ultrasound scans should be able to adopt and use our protocol easily. 5. We chose 2 months of the continuous OCs that were used in this study on the basis of a pilot study that was done before our observational study. We found that,
June 2003 Am J Obstet Gynecol
if only 1 month of suppression is used, only approximately one third of the patients who will respond eventually will do so. When this is increased to 2 months of continuous suppression, almost all of the patients who will respond do, with very few added with longer suppression. 6. We believe that our pregnancy rates were so good because we chose young patients with isolated ovulation problems and no male factor problems. If you can get them to ovulate, they would be expected to have good pregnancy rates. We have used this protocol for >6 years in our clinic, and the pregnancy rates have remained high for these patients. Dr Lentz, we do use this protocol in older women. However, in patients >38 years old, more aggressive therapy (gonadotropins or in vitro fertilization) may be a better option because of the age-related decline in fertility. Dr Gamberdella, we do occasionally use luteal support in some of these types of patients. However, none was used in the study. The hCG we use to trigger ovulation is, in our opinion, more important than luteal support, which we rarely use. We do not use any mucus parameters in these patients. The endometrial development seen on ultrasound monitoring is a good indicator of estrogen effect. Dr Patton, our data suggest that androgen levels are more important than estrogen levels in the prediction of ovulation success. This was seen in the OC group who did not ovulate. This subgroup showed significant drops in their estrogen levels but not in their androgen levels. If no significant drop in androgen levels occurs, the patient does not ovulate, despite estrogen changes. We have done a workup of some of these patients and have found several cases of nonclassic (late onset) congenital adrenal hyperplasia and one case of Cushing syndrome. The patient with Cushing syndrome had vaginal spotting the entire time that she was receiving the OC treatment. Dr Learman, we have found that giving hCG at a mean diameter of 20 mm reliably produces ovulation. Dr Henderson, those patients who achieve ovulation with OC treatment, but who do not become pregnant, are not treated again with OCs in the next cycle. They are given 100 mg of CC on days 5 to 9 of their next cycle and are monitored in the same manner. We believe that an ovulatory cycle keeps androgen levels low, but this was not measured specifically in this study. However, almost all those patients who did achieve ovulation in the initial cycle, but who did not become pregnant, continue to ovulate with repeat CC for ≥6 cycles.