New Methods for Induction of Ovulation

Over the past several years, new developments have occurred in ovulation induction therapy. These developments, which include new information relevant to established methods of inducing ovulation, novel ways of administering these agents, and the recent introduction of new agents such as recombinant follicle-stimulating hormone (rFSH) for routine clinical use. In this chapter, these developments and novel treatment approaches that provide additional options for ovulation induction therapy of anovulation resistant to standard treatment, including indications in World Health Organization (WHO) Groups I and II anovulatory states,¹ are reviewed.

WHO Group I

Women in WHO Group I are estrogen deficient with nonelevated follicle-stimulating hormone (FSH) and prolactin levels and no space-occupying lesion in the hypothalamic pituitary region. They typically have amenorrhea and do not bleed in response to progestin challenge.

WHO Group II

Women in WHO Group II are not estrogen deficient. Their FSH and prolactin levels are not elevated. They typically experience oligomenorrhea, but they may have anovulatory cycles or amenorrhea with bleeding in response to a progestin challenge. Included in this group are women with polycystic ovary syndrome (PCOS).

Pulsatile Gonadotropin-Releasing Hormone

Our ability to select women for treatment with pulsatile gonadotropin-releasing hormone (GnRH) and our understanding of their response to such treatment, including the complications that may develop, have been improved by recent studies.

WHO GROUP I.

The most useful indication for pulsatile GnRH therapy is functional hypothalamic amenorrhea (WHO Group I). In these patients, endogenous GnRH secretion is reduced.^2,3,4

Pulsatile GnRH therapy should restore normal fertility with ovulation rates of 85% to 95% per treatment cycle and pregnancy rates of 20% to 30% per ovulatory cycle in WHO Group I patients. Cumulative conception rates in this group of more than 90% are possible after 6 months of treatment.^2,3 In the 683 treatment cycles in 273 patients reviewed by Filicori and associates,⁴ the average rate of ovulation was 89%, with a pregnancy rate of 27% per cycle. Despite these high success rates with pulsatile GnRH in WHO Group I women, failures of treatment still occur. In these women, human menopausal gonadotropins (hMGs) are the treatment of choice. The recently introduced recombinant gonadotropins provide a novel treatment approach in this group of women. Treatment protocols that are being assessed are discussed later in this chapter.

WHO GROUP II (INCLUDING PCOS).

Pulsatile GnRH has been used with less success in WHO Group II anovulatory women.^2,3,4,5,6,7 Clomiphene citrate and hMG are the agents of choice for ovulation induction therapy in Group II women; however, pulsatile GnRH may be employed when considerations such as safety or cost preclude use of the other agents. In a review of 121 patients with PCOS undergoing 276 cycles of treatment (most were intravenous [IV] treatments), ovulation occurred in 186 cycles (67%), with pregnancy occurring in 45 of 186 ovulatory cycles (24%); 41 of the 121 patients (34%) conceived.³ The spontaneous abortion rate of 36% (16 of 45) included chemical pregnancies. Pregnancy rates were improved using pulse doses between 10 and 20 μg, with at least two treatment periods of 6 to 7 weeks before considering that treatment has failed. In our experience, this duration of uninterrupted IV therapy is poorly accepted.

The optimum dose, pulse frequency, and efficacy of IV versus subcutaneous (SC) administration of GnRH has not as yet been defined. IV administration is usually of shorter duration, requires less medication, and may yield slightly better ovulation and pregnancy rates than SC therapy.^2,3,4 IV administration achieves a narrow and higher serum GnRH peak that more closely approximates physiologic secretion compared with that resulting from SC administration.⁸ Although IV therapy is associated with better absorption and bioavailability, disadvantages include local discomfort, the inconvenience of a forearm IV tube, and the rare potential for thrombophlebitis. The usual IV dose is 50 to 100 ng/kg of body weight (approximately 2.5–7 μg/pulse) every 60 to 120 minutes.^2,4,9

SC administration into the lower abdominal wall offers greater safety, simplicity, and comfort, but drug absorption is slower and lacks a definite serum peak, which may lead to prolongation of the follicular phase or luteal phase defects in some cycles.¹⁰ Treatment time may be extended and dosage greater than that compared with IV route. SC treatment may be initiated at 5 to 10 μg/pulse with increments of 10 μg every 5 days to a maximum of 40 μg/pulse.² The SC route is preferable to the IV route in women with difficult venous access and occupational and other contraindications to a forearm IV site.

In a study of the ovarian response to a fixed dose of 200 ng/kg per pulse of GnRH administered SC at four different pulse frequencies (60, 90, 120, and 180 minutes) in subjects with WHO Group I anovulation (hypothalamic amenorrhea), 20 women were randomly allocated to one of the pulse doses.¹¹ Highest ovulation rates resulted from pulse frequencies of 90 to 120 minutes (60% and 88% of cycles, respectively), less often with pulse frequencies of 60 and 180 minutes (12% and 38%, respectively; p < 0.05). Four of six patients desiring pregnancy conceived at pulse frequencies of 90 and 120 minutes. Pregnancy was not conceived at a frequency of 60 minutes (0 of 1) and 180 minutes (0 of 1). Integrated luteal progesterone levels were significantly higher for 90 and 120 minutes than for 60 and 180 minutes. Although ovulation was induced with a wide range of GnRH pulse frequencies with SC administration, the 90- or 120-minute frequencies more consistently induced follicular maturation, leading to ovulation and normal luteal function, than frequencies of 60 or 180 minutes.

In a recent report of 600 GnRH-treated cycles,¹² IV GnRH was given at a dose of 1.25 to 20 μg every 30 to 120 minutes, with most cycles receiving 2.5 to 5 μg/pulse every 60 to 90 minutes continued until menses occurred or until human chorionic gonadotropin (hCG) was detectable, whereas in the remaining cycles, 2000 IU of hCG was given at 3-day intervals for luteal support. Ovulation was achieved in 447 (75%) treatment cycles; 105 women conceived, with a pregnancy rate per treatment cycle and per ovulatory cycle of 18% and 23%, respectively. Ovulatory and pregnancy rates were lowest in women with PCOS. The ovulation rates in the hypothalamic groups were 79% to 83%, whereas in the PCOS and hyperandrogenic groups, the ovulation rates were 65% to 72% (p < 0.05). The pregnancy rates per treatment cycle were 19% to 22% for hypothalamic groups and 13% for those with PCOS. Higher weight and insulin levels were associated with lower ovulatory and pregnancy rates, higher luteinizing hormone (LH) and testosterone levels were associated with lower ovulation rates only. The multiple pregnancy rate was 3.8% (4 of 105 pregnancies). The abortion rate was 30% and highest in the PCOS group at 45%. Severe ovarian hyperstimulation did not develop in any patients. Gonadotropin-releasing hormone agonist (GnRHa) pretreatment improved ovulatory rates only in women with PCOS (from 49% to 71%; p < 0.001), but it had no significant effect on pregnancy and abortion rates in any group. Because the disordered gonadotropin secretion recurs within a few weeks after discontinuing the GnRH agonist, retreating with the agonist is required before each treatment cycle, such protocols are cumbersome for routine clinical use.

hMG and pulsatile GnRH in the treatment of WHO Group I patients were compared in a retrospective study of 30 patients receiving 111 cycles of hMG and 41 patients treated over 118 cycles of pulsatile GnRH.¹³ The dose of GnRH ranged from 75 to 250 ng/kg (IV). Treatment with pulsatile GnRH was continued until completion of the cycle. Overall, ovulatory and conception rates were similar between the two groups. However, the cumulative conception rates after six cycles were higher in the GnRH group (96%) than in the hMG group (72%). Multiple pregnancy was higher with hMG (14.8% versus 8.3%, not significant). High-order multiple gestation (triplets or more) occurred only with hMG. Severe ovarian hyperstimulation was not observed in either group of patients. Compared with hMG, pulsatile GnRH resulted in high rates of ovulation with decreased risk of high-order multiple pregnancy in women with hypothalamic amenorrhea. Pulsatile GnRH should be the treatment of choice for this group of women, before using hMG.

LUTEAL SUPPORT.

Many patients ovulate spontaneously, thus the pump can be discontinued once ovulation has been documented. When pulsatile GnRH is continued to support the luteal phase, spontaneous menses occur unless the patient is pregnant. Alternately, the pump can be discontinued after ovulation, and divided doses of hCG (1,000–2,000 IU) given at 3- to 4-day intervals for three doses to support the luteal phase.⁴ Another option is a single injection of hCG (5000–10,000 IU) both to trigger ovulation and to support the luteal phase¹⁴ once a dominant follicle (18-mm diameter) and preovulatory E2 levels (about 300 pg/mL or 1100 pmol/L) have been documented. Exogenous progesterone also may be used for luteal support, but this has not been widely used.

COMPLICATIONS.

Mild ovarian hyperstimulation occurs in approximately 2.5% of treatment cycles. In earlier reports, multiple gestations (twins and rarely triplets or quadruplets) were observed in 13% (30 of 231 conceptions, all in Group I, none in Group II anovulation).³ In more recent reports, multiple pregnancy rates were 3.8%¹² and 8.3%,¹³ with no cases of severe ovarian hyperstimulation syndrome (OHSS). The apparently physiologic pattern of GnRH administration and the maintenance of normal negative feedback mechanisms tend to promote the development of only a single dominant follicle. However, these intrinsic control mechanisms can be overridden in some individuals, either by the use of large IV pulse doses or by hCG given to trigger ovulation. hCG also may provide a sufficient stimulus to induce ovulation from follicles smaller than the dominant one.¹⁵ Earlier studies showed spontaneous abortion rates comparable to the general population.³ Recent studies report a spontaneous abortion rate of 30% with a loss rate of 45% in PCOS.¹² No change in the incidence of congenital abnormalities has been noted in GnRH-induced pregnancies when compared with those in the general population.³

A small risk of thrombophlebitis, cellulitis, or sepsis may occur. The use of sterile techniques, occlusive dressings and a “closed” infusion system should reduce the risk of these complications to less than 1%.¹⁶

Urticaria, anaphylaxis, and refractory responses to therapy associated with circulating antibodies to GnRH have been detected sporadically, usually in association with high doses or prolonged therapy.^2,3

Exogenous Gonadotropins

FSH is a glycoprotein hormone of approximately 35 kD, synthesized and secreted by the pituitary gland. FSH is essential for follicular growth, oocyte maturation, and sex steroid hormone secretion from the ovary. The use of FSH is well established for the induction of ovulation in anovulatory women and for the induction of multiple oocyte development in preparation for various assisted reproductive technologies. Until the recent advent of recombinant FSH, gonadotropin preparations were available only from urinary sources. hMGs contain equal amounts of FSH and LH, whereas purified urinary preparations of FSH (urofollitropin or “pure” FSH [pFSH]) have only trace amounts of LH that remain as contaminants from the purification process. The traditional urinary preparations were associated with several disadvantages, including low purity, a lack of absolute source control (thus availability), a need for the cumbersome process of collecting urine on an extensive scale, LH contamination in “purified” FSH preparations, and batch-to-batch variations in the commercial product. The availability of recombinant sources of FSH derived from Chinese hamster ovarian cells, in which the alpha and beta subunit genes of FSH were inserted, can overcome these problems and provide an unlimited amount of hormone.^17,18

The FSH and LH contained in commercially available urinary preparations exhibit heterogeneity in their biologic and immunologic activities. Variations occur in FSH to LH contents, isoelectric points, in vivo potencies, circulating half-lives, and batch-to-batch clinical responses to these preparations.¹⁹ Commercial preparations of hMG include Pergonal, (Serono, Randolph, MA) Humegon (Organon, West Orange, NJ), and Repronex (Ferring, Tarrytown, NY). hMG preparations contain an equal amount of FSH and LH, 75 IU of each per vial. Metrodin (Serono) has negligible amounts (<1 IU of LH per vial) and has been referred to as “pure” FSH. These preparations have low FSH-specific activities of between 50 and 150 IU of FSH per milligram of protein with the content of contaminating non-FSH human proteins exceeding 95%. Metrodin-HP is a highly purified urinary FSH containing less than 0.001 IU of LH per vial, with a specific activity of 10,000 IU of FSH per milligram of protein and less than 1% contaminating non-FSH protein. rFSH contains a high FSH-specific activity of 10,000 IU of FSH per milligram of protein, and in contrast to the traditional urinary sources of FSH, rFSH is devoid of LH and contaminating non-FSH human proteins.¹⁸ Commercial preparations of recombinant FSH include Gonal-F (Serono) and Puregon (also known as Follistatin, Organon).

In human female volunteers, the total clearance and terminal half-lives of rFSH and urinary FSH (uFSH) were similar.¹⁸ After single intramuscular (IM) and SC injections of rFSH, the absolute bioavailability was about 77%. As with uFSH, multiple doses of rFSH resulted in a large variation in the individual response to a fixed dose, confirming the necessity for adjusting the dose of FSH to the individual patient's response. The variability occurring in the response of different individuals' to treatment was mainly caused by individual ovarian sensitivity to FSH rather than by differences in FSH pharmacokinetics.¹⁸ Relatively high concentrations of FSH are reached within about 12 hours, with dose proportional increases during the course of treatment and a steady state at each dose level reached by 3 to 5 days. After IM and SC administration, the pharmocokinetic characteristics of rFSH and uFSH are similar. Clinically, the pharmokinetic studies suggest that rFSH could be used in treatment protocols with the regimens and doses currently in use for uFSH without major practical changes.¹⁸ However, recent clinical trials have shown that under certain circumstances, ovarian stimulation (multifollicular development for in vitro fertilization [IVF] and ovulation induction in WHO Group II patients) can be achieved with lower doses of rFSH compared with uFSH (for further details, see subsequent discussion on rFSH versus uFSH). The practical implication of this higher efficiency of rFSH (lower dose requirements) seen in some clinical studies to date needs to be evaluated in further clinical study and with more extensive clinical use of these agents.

Exogenous gonadotropins constitute the most potent available means of stimulating ovulation. This, coupled with the well-recognized variation in the response of women, requires special monitoring of the ovarian response to treatment. Despite careful and improved monitoring, multiple pregnancies and ovarian hyperstimulation remain significant complications. In anovulatory women, first-line therapy includes clomiphene citrate given either alone or combined with low-dose glucocorticoids to suppress androgen production. The latter combinations can be successful in inducing ovulation in women with hyperandrogenic anovulation or PCOS.^7,20,21,22 Bromocriptine remains the treatment of choice in hyperprolactinemic anovulation. Failure to respond or intolerance to these agents are indications for exogenous gonadotropin therapy.

PULSATILE ADMINISTRATION OF HUMAN MENOPAUSAL GONADOTROPIN.

Whereas pituitary gonadotropins typically are secreted in a pulsatile pattern, gonadotropins usually are administered by bolus IM or SC injections. Using portable automated pumps, these hormones can be given in a manner that simulates the episodic pattern of release from the pituitary.

During pulsatile SC administration of hMG in 15 infertile women, ovulation resulted in 26 of 28 treatment cycles, with 2 women conceiving and ovarian hyperstimulation syndrome developing in one treatment cycle. The use of a daily dose of 150 IU of hMG (9.38 IU/pulse every 90 minutes) was superior to 75 IU per day (4.69 IU/pulse every 90 minutes).²³ In 212 cycles using SC pulsatile treatment in 61 patients (WHO Groups I and II, in most cycles hMG was employed but cycles with pFSH were included), ovulation occurred in 185 cycles (87.3%) in 60 patients (98.4%). Pregnancies occurred in 25 patients (41%) and 26 cycles (12.3%), with one twin gestation (3.8%) and low rates of severe ovarian hyperstimulation syndrome and spontaneous abortion.²⁴ In a randomized study of 88 pituitary-suppressed in vitro fertilization patients with a previous failed IVF cycle compared with bolus hMG, pulsatile SC hMG treatment resulted in improved clinical but not ongoing pregnancy rates; there were significantly more embryos with superior morphology, suggesting that in women refractory to bolus hMG, pulsatile treatment may improve pregnancy rates by improving embryo quality.²⁵ Another study²⁶ showed that pregnancy and ovulation occurred more frequently after IM hMG therapy compared with intermittent SC treatment. Constant SC administration using an infusion pump resulted in ovulation but no pregnancies.²⁷

The use of small pulses of hMG administered intravenously using an automated pump was shown to be successful in inducing ovulation in anovulatory women refractory to other agents, including clomiphene citrate, GnRH, and bolus hMG.^28,29 For IV administration, starting doses of about 6 to 9 IU per pulse may be used at a pulse frequency of 90 to 120 minutes. Doses are adjusted depending on the estradiol and ultrasound response. With the progress of follicular maturation, the ultrasound findings and estradiol levels are used to decide when to stop the stimulation and when to administer the ovulating dose of hCG.^28,29

In an experience of 107 treatment cycles in 30 refractory anovulatory women,²⁹ 4 WHO Group I anovulatory women ovulated in all 8 treatment cycles; the pregnancy rate per cycle was 63%, and the cumulative pregnancy rate was 100%. Of five pregnancies conceived, one fetus aborted spontaneously; of the remaining four, two were single pregnancies, one was twins, and one was triplets. In one patient, ovarian hyperstimulation developed but subsided after hospital admission.²⁹ In the 26 WHO Group II patients (99 cycles of treatment), 42 treatment cycles (42.4%) were in 12 women previously refractory to the IM route of administration. In the 99 IV treatment cycles administered to refractory Group II women, the ovulation rate was 86%, the pregnancy rate per ovulatory cycle was 14%, and the cumulative pregnancy rate was 56%. Of 12 pregnancies conceived in this group, four fetuses aborted spontaneously; of the remaining 8 pregnancies, five were single and 3 were twins. Ovarian hyperstimulation occurred in 3 of the 99 treatment cycles; in all instances, hyperstimulation resolved after admission to hospital. Mild phlebitis occurred at the site of the IV catheter in 24% of treatment cycles. These findings demonstrate the efficacy of pulsatile hMG in the treatment of anovulatory infertile patients, including those refractory to bolus hMG and other conventional methods of ovulation induction. In addition, there was a dose-sparing effect (about 40%) with pulsatile IV as compared with bolus IM treatment. A dose-sparing effect of SC administration has not been demonstrated, suggesting that this action is unique to the IV route.^29,30 Considering our overall experience with the use of hMG,³¹ in Group I, the rate for ovulation was 96%, the per cycle fecundity rate (clinical pregnancy per ovulatory cycle) was 26%, and cumulative conception rate was 89% (over 6 cycles of treatment). For Group II (which was considered in 3 subgroups), the average ovulation rate was 96% (range 90%–100%), clinical pregnancy rate per ovulatory cycle was 10% (range 5%–16%), and cumulative pregnancy rate was 43% (range 30%–63%), respectively. The spontaneous abortion rate was 6% in Group I and 31% in Group II (range 25%–50%). Incidentally, we found that the best outcomes in WHO Group II were in women experiencing oligomenorrhea with normal androgen levels. The 63% cumulative conception rate in this subgroup of WHO Group II was not significantly different from the cumulative conception rate of 89% in WHO Group I patients. Women with elevated androgen levels and luteal defects experienced significantly poorer outcomes.³¹ Although our retrospective data do not suggest that pulsatile IV administration is superior in terms of pregnancy outcomes, our clinical results with pulsatile hMG are encouraging because a large proportion of cycles were in women proven to have various refractory responses to the bolus IM route and to other conventional approaches. The potent effect in small numbers of resistant Group I women require both caution and further study.

Controlled trials evaluating the risks and benefits of the various routes of administering hMG and new FSH preparations in specific diagnostic categories of anovulatory women have not yet been performed. Pulsatile IV administration may be considered an alternative method of treatment in selected women experiencing refractory responses and failure with other methods, including IM treatment.^29,31,32 However, considering that complications including multiple pregnancies and ovarian hyperstimulation syndrome occur, caution and experience in the use of this method of administering exogenous gonadotropins and careful monitoring are important prerequisites for a successful outcome to treatment.

The fact that Group I women have far superior responses to ovulation induction than Group II women is well documented, but the reasons for this difference are not fully understood. Excess androgens and ovulation stimulation have been associated with oocyte degeneration, delayed implantation, and infertility.^33,34,35,36 Another factor in PCOS is high endogenous LH levels, which may cause oocyte aging due to precocious oocyte maturation.³⁷ High LH excretion also has been associated with poor oocyte quality, failed fertilization, and low pregnancy rates.³⁸

PURE FOLLICLE-STIMULATING HORMONE.

Because of increased LH and the elevated ratio of LH to FSH in PCOS, the low LH content in pFSH provides a theoretical advantage for the induction of ovulation in anovulatory patients with PCOS.

In an observational study, results of treatment with pFSH and hMG were compared in groups of 21 and 22 women with PCOS, respectively.³⁹ All women failed to conceive after prior treatment with clomiphene citrate. Rates of ovulation were 95.2% for pFSH and 100% for hMG. Eight (38.1%) and 11 (50%) patients conceived, 6 (28.5%) and 8 (36.3%) delivered, and 2 (9.5%) and 3 (13.6)% aborted with pFSH and hMG, respectively. Multiple pregnancies were not conceived. Ovarian hyperstimulation occurred more frequently after pFSH (40%) than hMG treated cycles (22%). This study did not demonstrate any significant advantages of pFSH over that of hMG in ovulation induction therapy for anovulatory patients with PCOS.

In a comparison of a conventional protocol with a slow protocol using pFSH in women with PCOS, the slow protocols resulted in longer durations of stimulation, development of more single versus multiple follicles, and fewer canceled cycles due to concerns for hyperstimulation with a higher per cycle fecundity (23% vs. 15%, not statistically significant, slow versus conventional protocol, respectively).⁴⁰ In a more recent study,⁴¹ a chronic low-dose protocol using urinary or recombinant FSH was compared with a conventional protocol. The low-dose protocol resulted in higher pregnancy rates (40% vs. 24%, not statistically significant) compared with the conventional protocol. The low-dose protocol eliminated the problem of multiple pregnancy and reduced the rate of ovarian hyperstimulation syndrome (11% and 33% in low-dose vs. the conventional protocol). In these observational studies, slow protocols were found to be associated with lower complication rates than the conventional protocols, but the pregnancy rates were not improved.

In a study of five patients with PCOS, SC pulsatile pFSH had a dose-sparing action of 20% compared with hMG given in a similar manner.²⁴ Four of the five women who failed to conceive on the hMG regimen became pregnant with pFSH given in a pulsatile fashion by means of the SC route.

RECOMBINANT DNA-DERIVED SOURCES OF HUMAN GONADOTROPINS.

Recombinant FSH and purified FSH isolated from pituitary glands stimulated steroidogenesis in a similar manner.⁴² rFSH averts the problems associated with heterogeneity characteristic of the preparations obtained from urinary extraction. Recombinant human LH (rLH) also has been shown to be effective in inducing ovulation.⁴³ The shorter half-life of this source of bioactive LH (compared with hCG) has the potential to be used in a more physiologic manner and perhaps overcome complications such as the ovarian hyperstimulation syndrome and multiple pregnancy.⁴⁴ Much of the available clinical studies with rFSH were carried out in clinical trials of patients undergoing IVF and embryo transfer (ET). A review of these data and data from studies employing rFSH in the induction of ovulation in WHO Groups I and II anovulatory women is provided.

RECOMBINANT FOLLICLE-STIMULATING HORMONE VERSUS URINARY FOLLICLE-STIMULATING HORMONE.

In a large prospective, randomized, assessor-blind, multicenter study, rFSH and uFSH were compared in infertile women undergoing IVF and ET,⁴⁵ 585 subjects received rFSH (Puregon) and 396 received uFSH. Pituitary suppression was achieved with buserelin. A maximum of three embryos were replaced. More oocytes were retrieved after rFSH (mean adjusted for center 10.84 vs. 8.95, p < 0.0001), with more high-quality embryos (p = 0.003). Ongoing pregnancy rates per attempt and per transfer in the rFSH group were 22.17% and 25.97%, respectively, and in the uFSH group, 18.22% and 22.2% (not significant). Ongoing pregnancy rates including pregnancies conceived after frozen-thawed embryo cycles were 25.7% for rFSH and 20.4% for uFSH (p = 0.05). The dose of FSH was lower with rFSH (2,138 vs. 2,385 IU, p < 0.001) compared with uFSH, with a shorter treatment period (10.7 vs. 11.3 days, p < 0.0001). OHSS leading to hospitalization occurred in 19 of 585 (3.2%) subjects treated with rFSH and 8 of 396 (2.0%) subjects who received uFSH. No significant increases in serum FSH antibodies or Chinese hamster ovary-derived proteins were detected. This study showed that rFSH was more effective than urinary FSH in inducing multifollicular development and in achieving an ongoing pregnancy when frozen thaw embryo replacement cycles were included.

In a prospective, randomized, assessor-blind study comparing efficacy and safety of rFSH (Gonal-F) versus highly purified uFSH (Metrodin HP, uFSH HP), both given subcutaneously for IVF including intracytoplasmic sperm injection with a long GnRH agonist protocol, there were 119 women in the rFSH group and 114 in the uFSH HP group.⁴⁶ Usually two embryos were replaced with two patients in each group receiving three embryos. The mean number of oocytes recovered was 12.2 versus 7.6 (p < 0.0001), the mean number of embryos was 8.1 versus 4.2 (p < 0.0001), the mean number of treatment days was 11 versus 13.5 (p < 0.0001), and the mean number of 75-IU ampules was 21.9 versus 31.9 (p < 0.0001) in the rFSH group and uFSH HP groups, respectively. Clinical pregnancy rates per started cycle were 45% and 37% (not significant) for rFSH and uFSH HP, respectively. Clinical pregnancy rates per embryo transfer were 48% and 47% (not significant) for rFSH and uFSH HP, respectively. Spontaneous abortion rate was 25% versus 14%, and multiple pregnancy rate was 42% versus 25%, for rFSH and uFSH HP, respectively. No triplets were conceived. OHSS occurred in 5.2% and 1.7% with rFSH and uFSH HP, respectively. rFSH was more effective than uFSH HP in inducing multiple follicular development in this study.

RECOMBINANT FOLLICLE-STIMULATING HORMONE VERSUS HUMAN MENOPAUSAL GONADOTROPIN.

A randomized study in which rFSH was compared with hMG in 89 nonpituitary-suppressed women showed that the number of oocytes retrieved was 11.2 versus 8.3, whereas ongoing pregnancy rates per attempt and per transfer were 22.2% versus 17.1% and 30.8% versus 22.2%, respectively. None of these differences were significant.⁴⁷ rFSH administered subcutaneously versus intramuscularly showed no significant differences in efficacy, efficiency, or safety endpoints, except for bruising, which occurred more often after SC injection compared with the IM route.

In hypogonadal women given daily IM injections of rFSH in increasing doses starting over 3 weeks (75 IU/day for 1 week, 150 IU/day in the second week and 225 IU/day in the third week), maximum FSH levels were between 7.1 and 11.8 IU/L, but LH levels remained unchanged.⁴⁸ Serum androstenedione showed no significant increases, and estradiol levels showed only slight increases with, inhibin increasing to late follicular phase levels. Despite the low levels of estrogen, ovarian follicles grew to preovulatory sizes (> 15 mm). These findings confirm that rFSH has no intrinsic LH effects and suggest that granulosa cells function normally in the presence of a reduced availability of androgens for aromatization to estrogen. Despite the attenuated estrogen increase, ovarian follicles develop to the preovulatory stage in response to rFSH in hypogonadal women.

A combination of rFSH and rLH resulted in the successful conception and completion of pregnancy in a woman with Kallmann's syndrome.⁴³ The first cycle of treatment was initiated with 150 IU daily of rFSH (Gonal-F) and 75 IU daily of rLH (Lhadi, Serono). Ovulation was induced with 10,000 IU of hCG, and hCG was used for luteal support. In the second cycle of treatment, the same dose of rFSH was used, but the dose of rLH was increased to 225 IU/day. Pregnancy was conceived successfully in this cycle, with delivery of a healthy infant. The availability of rFSH and LH provides another treatment option to hMG when there is a failure to respond to pulsatile GnRH therapy in women with hypogonadotropic (WHO Group I) anovulation.

In women with WHO Group I anovulation, endogenous LH production is below the threshold required for optimal follicular development. Thus, LH is required during FSH administration for follicular growth and ovarian hormone production. rLH was evaluated as an adjunct to rFSH in hypogonadotropic hypogonadal women in a randomized study of 38 patients.⁴⁹ Fixed doses of either 25 IU, 75 IU, or 225 IU per day of rLH combined with a fixed dose of 150 IU rFSH were compared with 150 IU/day of rFSH alone. SC doses of LH from 75 to 225 IU/day resulted in optimal ovarian estradiol secretion.

In a study comparing SC rFSH (Gonal-F) with IM uFSH (Metrodin) in a randomized multicenter trial in WHO Group II anovulatory women,⁵⁰ women were allocated randomly to receive daily Gonal-F or Metrodin, starting with 75 IU per day and increasing in increments of 37.5 IU. The first increase was performed after 14 days only if no significant increase in serum estradiol occurred. Further increases were made after a period of 7 days. If an ovarian response failed to occur during 35 days of a previous cycle, the starting dose was increased to 150 IU of FSH. The dose was decreased to 37.5 IU if the patient was considered at risk for ovarian hyperstimulation. Five thousand international units of hCG were given to induce ovulation when one follicle reached a mean diameter of 17 mm; there was a maximum of three follicles greater than 15 mm with estradiol greater than 150 pg/ml per follicle greater than 16mm.

Eighty-seven patients were treated over 179 cycles, with 41 receiving Gonal-F and 46 receiving Metrodin. hCG was given in 71 cycles (83.5%) and in 78 cycles (82.9%) of Gonal-F and Metrodin, respectively. There were no differences in the doses used or the duration of treatment. Monofollicular development occurred in 63.4% and 61.5% for Gonal-F and Metrodin, respectively. Ovulation occurred in 89% and 97% of patients with clinical pregnancies in 15% and 18% per initiated cycle and 25% and 29% per ovulatory cycle for Gonal-F and Metrodin, respectively. OHSS did not occur in either group. Multiple pregnancy was conceived in none of the 35 patients completing the study in the Gonal-F group, whereas in the Metrodin group, 2 of 41 patients conceived multiple pregnancies. The treatment outcomes between Gonal-F and Metrodin revealed no statistically significant differences. This preliminary study showed no differences in follicular development, ovarian steroid production, ovulation, and pregnancy rates between Gonal-F and Metrodin. The potential clinical advantages of Gonal-F over Metrodin are the high purity of rFSH, which allows SC injection by self-administration and the consistent batch-to-batch purity of the recombinant hormone. Gonal-F was well tolerated, but SC administration was associated with several instances of mild swelling at the injection site.

In a study using the rFSH Puregon, women with WHO Group II anovulation who failed to respond to clomiphene citrate were allocated randomly to Puregon (n = 105) or Metrodin (n = 67) over a maximum of three treatment cycles.⁵¹ A low-dose step-up regimen was used, with 75 IU of FSH given intramuscularly daily for a maximum of 14 days. If needed, weekly increments of half an ampule were given thereafter until the threshold dose for follicular development was achieved. The maximal daily dose of FSH was 3 ampules daily. Treatment was discontinued after 6 weeks. In the second and third treatment cycles, upward adjustments were permitted after 1 week of treatment. When a follicle reached a mean diameter of 18 mm or two to three follicles reached a diameter of 15 mm or greater, ovulation was triggered with 10,000 IU of HCG. Treatment with Puregon resulted in a significantly greater number of follicles 12 mm or greater. The dose of FSH was lower with Puregon (median dose 750 IU vs. 1,035 IU, p < 0.001), which was given over a shorter time period (10 days versus 13 days, p < 0.001) than for Metrodin. Cumulative ovulation rates after three cycles were 95% for Puregon and 96% for Metrodin. The cumulative pregnancy rates were 27% for Puregon and 24% for Metrodin. The ovulation and pregnancy rates were not significantly different. OHSS occurred in eight women receiving Puregon (7.6%) and three receiving Metrodin (4.5%). Each treatment group had a twin pregnancy. One triplet pregnancy occurred in the rFSH group only. Higher cancellation rates for ovarian hyper-responsiveness (> 3 follicles 15 mm in size or risk of ovarian hyperstimulation) were seen in the second and third cycles in the rFSH group compared with the uFSH group. This suggests that close surveillance with lower starting doses and slow increases in doses are required to compensate for the apparently higher bioactivity of rFSH. Serum anti-FSH antibodies were not detected. Thus, in this study, no significant differences were observed in efficacy and safety of Puregon and Metrodin in WHO Group II clomiphene-resistant chronic anovulation. However, Puregon resulted in ovulation with a lower dose and over a shorter time interval, indicating a higher efficiency than Metrodin. Because of its greater efficiency in WHO Group II women, it was suggested that Puregon could be started at a lower dose than Metrodin and women could be kept at this dose for a longer period of time.

Ovarian Electrocautery

In a study comparing laparoscopic ovarian cautery with hMG and pFSH in PCOS, 88 clomiphene-resistant infertile women with PCOS were randomized into three groups: ovarian electrocautery, hMG, and pFSH. Ovulation was induced successfully in 71.4%, 70.6%, and 66.7%, respectively; six-cycle cumulative conception rates were 52.1%, 55.4%, and 33.8%, respectively; pregnancy wastage occurred in 21.4%, 53.3%, and 40%, in these groups, respectively. It was concluded that electrocautery was as effective as hMG and pFSH in clomiphene-resistant anovulatory patients with PCOS.⁵²

In an extensive review of the subject by Donesky and Adashi⁵³ involving 27 studies reporting on 729 patients, 614 ovulated (ovulation rate of 84.2%) and 406 conceived (pregnancy rate of 55.7%, range 20%–87.5%). The spontaneous loss rates appear to be lower after ovarian cautery compared with hMG, clomiphene, or FSH. In a follow-up study of 206 women,⁵⁴ the early miscarriage rate was found to be 18%. Other potential advantages included reduction of multiple pregnancy, elimination of OHSS, less cost, and no need for cycle monitoring compared with the use of gonadotropins. However, the procedure may only provide transient effects, and adnexal adhesions may compromise the pregnancy rates. Postoperative complications include adhesion formation, which appears frequent. However, these adhesions do not necessarily preclude pregnancy. Periovarian adhesions of varying severity at second-look laparoscopy were found in all patients examined after laparoscopic ovarian cautery,⁵⁵ with seven of the eight women conceiving. When comparing gonadotropin treatment before and after ovarian electrocautery in clomiphene-resistant women with PCOS, a higher rate of ovulation and pregnancy was demonstrated when gonadotropins were given after the procedure.⁵⁶ In women who did not ovulate before laparoscopic ovarian cautery, normal menstrual cyclicity (cycles occurring every 27 to 35 days) resumed in 41% of patients, with this pattern lasting for between 3 to 15 months, with an average of 6.21 months.⁵⁶

There is a need for prospective randomized trials to evaluate the efficacy of ovarian cautery in clomiphene-resistant women with PCOS compared with gonadotropins. There are no long-term data evaluating the safety of this approach. It has been suggested that laparoscopic ovarian cautery be viewed as experimental until adequate long-term follow-up has established its safety.⁵⁷ The treatment of choice for clomiphene-resistant women with PCOS is hMG or rFSH.

Gonadotropin-Releasing Hormone Agonists Combined with Human Menopausal Gonadotropin

Gonadoptropin-releasing hormone agonists downregulate gonadotropin secretion and suppress ovarian function, resulting in a hypoestrogenic state analogous to the hypothalamic hypogonadal state of women with WHO Group I anovulation. Downregulation of the hypothalamic pituitary gonadal axis of women with WHO Group II anovulation with these agonists has been used as an adjunct to the induction of ovulation with exogenous gonadotropins.⁵⁸

To date, the adjunctive use of GnRH agonists with exogenous gonadotropins in WHO Group II women have not been shown to improve the outcome, in terms of pregnancy rates or complications such as ovarian hyperstimulation.^58,59 After pituitary suppression with buserelin acetate, stimulation with hMG or FSH (Metrodin) in 34 patients with PCOS resulted in no differences in days of stimulation, doses of hMG or FSH, number of oocytes retrieved, proportion of mature oocytes, or cleavage rates or pregnancy rates between hMG or FSH.⁶⁰

In a retrospective study,⁶¹ it was found that the use of a GnRH agonist in women with PCOS before starting ovulation induction with hMG reduced the miscarriage rates to 16.7% compared with 39.4% without GnRH agonist. These results were similar in women with PCOS undergoing IVF. The cumulative live birth rate was 64% in cotreatment with the agonist and 26% with hMG alone. Randomized trials are needed to better evaluate the role of agonist cotreatment with hMG or FSH in improving the spontaneous loss rates in WHO Group II anovulation and PCOS. Until further data are available, GnRH agonist cotreatment with gonadotropins cannot be recommended for routine use. However, it seems reasonable to consider such a protocol in WHO Group II women with a history of recurrent spontaneous abortions in response to ovulation induction therapy.

Growth Hormone Combined with Exogenous Gonadotropins

In a patient with panhypopituitarism with undetectable growth hormone (GH) levels who required high doses of hMG to stimulate ovulation, cotreatment with GH decreased the dose of hMG per cycle to induce ovulation from 76 to 96 ampules to 35 to 36 ampules per cycle. The patient conceived after receiving combined GH and hMG treatment.⁶²

In a randomized study of 16 hypogonadotropic women,⁶³ cotreatment with GH resulted in a reduction in the dose and duration of hMG used. In a randomized trial comparing the effects of GH as an adjunct to GnRH and hMG for ovulation induction in 30 patients with PCOS, no clinically beneficial effects of GH cotreatment could be demonstrated. There were no improvements in the dose of hMG required to achieve ovulation, the duration of treatment, or ovulation and pregnancy rates.⁶⁴ Cotreatment with GH and hMG in 64 WHO Group I patients (women with hypogonadotropic hypogonadism) in a randomized study showed that the addition of GH reduced the gonadotropin dose needed to achieve ovulation. However, GH did not improve the rate of ovulation and the pregnancy rates were lower in the GH group compared with placebo.⁶⁵ Until further data on the cost-effectiveness of GH cotreatment and the effect of GH on pregnancy rates are available, GH remains of experimental interest at this time and cannot be recommended for routine use.

Hyperinsulinemia

Hyperinsulinemia and insulin resistance, which occur in women with PCOS,⁶⁶ persist even after androgen levels have been suppressed with the use of GnRH agonists, suggesting that insulin hypersecretion is independent of androgens.⁶⁷ Insulin has synergistic effects with LH to stimulate androgen production in the ovary.⁶⁸ Insulin also stimulates ovarian cytochrome P450c17α enzyme activity to increase LH-induced androgen synthesis in the ovary and by reducing sex hormone-binding globulin levels, increases biologically active unbound testosterone. Pituitary LH secretion also may be increased in women with PCOS, and these defects in LH secretion may be explained by hyperinsulinemia.⁶⁶ Because of the important role of insulin in PCOS, interventions to lower insulin levels may improve androgen function. One such intervention is weight loss, which improves insulin sensitivity and decreases circulating insulin levels with improvement in ovulatory function. Overweight women are prone to infertility.⁶⁹ Because modest weight reduction in overweight women can restore ovulation, it is recommended that these women undergo appropriate counseling on dietary measures and exercise to reduce weight. Reduction of approximately 5% of body weight can be followed by ovulation and conception of pregnancy.⁷⁰

In addition to diet and exercise programs to achieve and maintain weight loss, pharmacologic agents that reduce insulin levels or improve insulin sensitivity may be useful in the treatment of PCOS, where weight loss and other measures may have failed.⁶⁶

INSULIN SENSITIZING AGENTS.

Treatment with metformin, a biguanide agent that enhances the action of insulin at the cellular level, may improve the hyperinsulinemia and hyperandrogenemia with improvement in the menstrual abnormalities and subsequent conception of pregnancy in women with PCOS.⁷¹ Twenty-two women were treated with metformin 500 mg three times daily. Testing after 8 weeks showed significant reduction in fasting and integrated insulin responses to a glucose load, serum LH, free testosterone, and LH/FSH ratio and a small but significant reduction in body mass index of the subjects. Twenty-one of 22 women (95%) experienced resumption of menstrual cyclicity, with four of the women (19%) achieving pregnancy within 6 and 7 months of metformin treatment. The authors concluded that a 6-month course of metformin improves menstrual cyclicity and fertility in women with PCOS and that insulin sensitizing agents provided a rational treatment of metabolic and endocrine dysfunction in PCOS.

In a study of 59 obese women with PCOS randomly allocated to metformin (500 mg 3 times daily, n = 36) or placebo (n = 23), ovulation occurred in 28 of the 36 women (78%) while they were receiving metformin alone or a combination of metformin and clomiphene; only 1 of 23 women (4%) receiving placebo ovulated.⁷² In this study, metformin augmented spontaneous and clomiphene-induced ovulation in obese women with PCOS.

The metabolic effects on insulin action, LH levels, androgens and sex hormone-binding globulin levels for metformin⁷¹ also were reported for troglitazone, a new thiazolidinedione derivative.⁷³ Patients taking troglitazone have experienced severe hepatic toxicity.⁷⁴

In a study of obese women with PCOS, metformin had no effects on insulin function, free testosterone, basal and stimulated LH and FSH, 17-hydroxyprogesterone, and other steroids including dehydroepiandrosterone sulfate (DHEAS), androstenedione, and estradiol.⁷⁵ This and another study⁷⁶ were unable to demonstrate any effect of metformin on the hyperinsulinemia and androgen excess in obese nondiabetic women with PCOS. Metformin also had no direct effect on ovarian thecal androgen secretion in in vitro studies.⁷⁷ Methodological factors and effects of study design, such as the effects of weight maintenance during the study⁷⁵ or weight loss that occurred in the subjects while they were receiving metformin,⁷¹ may explain these divergent findings. Although improvement in menstrual cyclicity and ovulatory function by insulin-sensitizing agents has been documented, their use in ovulation induction therapy is a problem. Teratogenicity with these agents, including metformin, has not been excluded, and because of the potentially serious toxicity of troglitazone, these agents should not be used for routine ovulation induction therapy at this time. Until further studies are done, the best option for ovulation in patients with PCOS wishing to conceive is a combination of weight loss, exercise, and ovulation induction therapy with clomiphene citrate, hMG, or rFSH.⁷⁴

Pulsatile Gonadotropin-Releasing Hormone Combined with Exogenous Gonadotropins

In a study of eight WHO Group II women who underwent ovulation induction with combined treatment (pulsatile GnRH followed by pFSH) and pFSH alone, combined treatment resulted in development of a single follicle with ovulation in eight of eight cycles versus three of seven cycles and pregnancy in three of eight cycles versus one of seven cycles, respectively.⁷⁸ In 18 women with hypothalamic amenorrhea treated with sequential FSH and pulsatile GnRH compared with conventional treatment with FSH, the rates of ovulation and pregnancy were not significantly different—94% versus 89.7% and 18.8% versus 22.2% for FSH and pulsatile GnRH, respectively. Sequential treatment resulted in few follicles developing with a high rate of single follicles, 80% versus 0% for FSH alone. Sequential treatment provided a better chance of single dominant follicle development and ovulation without complications.⁷⁹ This approach seems promising for inducing single follicular development and for limiting the rates of complications, such as multiple pregnancy and OHSS, but the numbers of patients are too small to assess the true success and complication rates.

Gonadotropin-Releasing Hormone Antagonists

Unlike GnRH agonists, these compounds suppress gonadotropin secretion immediately and without any stimulatory effects that occur with the agonists. In earlier antagonist studies, problems were encountered with side-effects due to histamine release and high-dose requirements that limited their clinical use. New potent and safer GnRH antagonists recently have become available. The GnRH antagonist Cetrorelix (ASTA Medica AG, Frankfurt am Main, Germany) administered on day 8 of the cycle resulted in LH and estradiol being reduced to subnormal levels within 24 hours of administration.⁸⁰ In a study of 20 women undergoing IVF,⁸¹ hMG was started on day 2 and from day 7 until hCG, 1 or 3 mg of Cetrorelix was given daily. An endogenous LH surge was not observed in any patient. A median of 27 ampules of hMG was used to complete the stimulation. The time for completion of the stimulation and the dose of hMG used was less than expected with Cetrorelix as would have been the case if hMG was administered with a long GnRH agonist protocol. In another study, a single injection of 3 mg of Cetrorelix⁸² prevented the LH surge in all 11 patients studied. A mean of 24.5 ampules of hMG was used to complete the stimulation. In a study of the antagonist Nal-Glu in oocyte donors stimulated with hMG, compared with cycles in the same donors in whom GnRH agonists were used, the antagonist reduced significantly the dose of hMG (24.7 vs. 31.3 ampules). In addition, there was also a significant reduction in the number of patient visits per cycle and the required duration of treatment to complete the cycle (11.3 days for the antagonist cycles vs. 21.7 for the agonist cycles). No differences were found in the number of oocytes retrieved, the fertilization rate of oocytes or the number of embryos available per replacement.⁸³ GnRH antagonists thus have been shown to be effective in suppressing the LH surge, and studies in IVF patients undergoing pituitary suppression show that there are advantages over the use of GnRH agonists in ovulation induction in these patients.

The author thanks Regina Whorms, whose search of the literature is greatly appreciated.