Clomiphene citrate was first synthesized in 1956, introduced for clinical trials in 1960, and approved for clinical use in the United States in 1967.^27,28 In early clinical trials, 80% of anovulatory women treated with clomiphene achieved ovulation and half of those who ovulated also conceived.^27,28 The collected clinical experience gained in the years since remains consistent with those early observations.

Clomiphene is a non-steroidal triphenylethylene derivative that acts as a selective estrogen receptor modulator (SERM), having both estrogen agonist and antagonist properties.²⁹ However, in almost all circumstances, clomiphene acts purely as an antagonist or antiestrogen; its weak estrogenic actions are clinically apparent only when endogenous estrogen levels are very low. Clomiphene is cleared through the liver and excreted in the stool; approximately 85% is eliminated within a week, but traces can remain in the circulation for longer.³⁰ Clomiphene is a racemic mixture of two different stereoisomers, enclomiphene (62%; originally known as cis-clomiphene) and zuclomiphene (38%; originally known as trans-clomiphene).^29,31 Enclomiphene is the more potent isomer and the one responsible for its ovulation-inducing actions.^29,32 The half-life of enclomiphene is relatively short, so serum concentrations rise and fall quickly during and after treatment.^30,33 Zuclomiphene is cleared much more slowly; serum levels remain detectable for weeks after a single dose³⁰ and may even accumulate gradually over a series of cycles, but there is no evidence that residual zuclomiphene has any important clinical effects or consequences.³³

Structural similarity to estrogen allows clomiphene to compete with endogenous estrogen for nuclear estrogen receptors at sites throughout the reproductive system. However, unlike estrogen, clomiphene binds to nuclear estrogen receptors for an extended interval of time and thereby depletes receptor concentrations by interfering with receptor recycling.²⁹ At the hypothalamic level, estrogen receptor depletion prevents accurate interpretation of
circulating estrogen levels; circulating estrogen levels are perceived as lower than they truly are. Reduced negative estrogen feedback triggers normal compensatory mechanisms that alter the pattern of gonadotropin-releasing hormone (GnRH) secretion and stimulate increased pituitary gonadotropin release which, in turn, drives ovarian follicular development. At the pituitary level, clomiphene also might increase the sensitivity of gonadotrophs to GnRH stimulation.³⁴

When administred to already ovulatory women, clomiphene increases GnRH pulse frequency.³⁵ In anovulatory women with polycystic ovary syndrome (PCOS) who already exhibit an increased GnRH pulse frequency, clomiphene increases only pulse amplitude.³⁶ Serum levels of both FSH and LH rise during clomiphene treatment and fall again soon after the typical 5-day course of therapy is completed.³⁷ In successful treatment cycles, one or more follicles emerge and grow to maturity. In parallel, serum estrogen levels rise progressively, ultimately triggering an LH surge and ovulation. In sum, clomiphene works primarily by stimulating the normal endocrine mechanisms that define the hypothalamic-pituitary-ovarian feedback axis. The importance of other effects it may have on insulinlike growth factors (decrease in IGF-I concentrations) and sex hormone-binding globulin (increase in serum levels) is uncertain.^38,39

In addition to its desirable central actions, clomiphene can exert less desirable anti- estrogenic effects at peripheral sites in the reproductive system, which some have suggested might explain the difference between ovulation and pregnancy rates achieved with clomiphene treatment. Adverse effects of clomiphene on the endocervix, the endometrium, the ovary, the ovum, and the embryo have been described, but there is no compelling evidence to indicate that such effects have important clinical consequences in most women.

On balance, the weight of evidence from controlled trials suggests that the quality and quantity of cervical mucus production can be decreased in clomiphene treatment cycles. Whereas some have observed no significant changes in mucus characteristics during treatment,⁴⁰ others have found clomiphene has dose-dependent adverse effects.^41,42 The conflicting results have several possible explanations. The effect may be more apparent when the interval between the end of treatment and ovulation is short.⁴³ The effect might
often be negated by higher serum estradiol levels resulting from clomiphene-induced multifollicular development.⁴⁴ It also is possible that some individuals may be more sensitive to the effect.⁴⁵ Regardless, any adverse effect that clomiphene may have on cervical mucus now is largely moot (Chapter 27). In recent years, even the evaluation of cervical mucus has all but disappeared from clinical practice because controlled trials have demonstrated that postcoital testing (the traditional test of cervical factors) has little or no predictive value^46,47 and because modern treatment regimens for persistent infertility now routinely incorporate intrauterine insemination (IUI), which bypasses the cervix altogether.^48,49

Impaired endometrial growth also has been reported in clomiphene-treated women. However, preovulatory endometrial thickness in clomiphene-induced cycles remains well within the range normally observed in spontaneous ovulatory cycles in the large majority of women.^50,51,52,53 and ⁵⁴ Other subtle differences in endometrial morphology have been attributed to the effects of clomiphene, but their clinical relevance, if any, is uncertain.^55,56 It is likely that clomiphene inhibits endometrial growth, at least in some women, for the same reasons that it may inhibit cervical mucus production, but the same caveats apply; the effect is inconsistent, may be offset by the higher estrogen levels in clomiphene-induced cycles, and probably has little clinical importance, except perhaps in those individuals exhibiting grossly poor endometrial growth (peak preovulatory thickness less than 5-6 mm).

Clomiphene does not appear to have any clinically relevant direct effects on the ovary or embryo. Although clomiphene can inhibit steroid hormone production by cultured avian,⁵⁷ ovine,⁵⁸ and human granulosa/luteal cells in vitro,⁵⁹ serum estrogen and progesterone concentrations in clomiphene-induced cycles are typically higher, not lower, than in spontaneous ovulatory cycles. Adverse effects on mouse ovum fertilization and embryo development have been observed in vitro,⁶⁰ but studies in women indicate that serum concentrations of enclomiphene and zuclomiphene never approach the levels required to induce such effects, even after several consecutive treatment cycles.³³

Clomiphene citrate is the traditional drug of choice for ovulation induction in anovulatory infertile women with normal thyroid function, normal serum prolactin levels, and normal endogenous estrogen production, as determined by clinical observations (oligomenorrhea, estrogenic cervical mucus), a serum estradiol determination (greater than approximately 40 pg/mL), or a normal menstrual response to a progestin challenge (WHO Group II).⁶¹ Although the drug also frequently is used empirically to stimulate multi-follicular development in ovulatory women with unexplained infertility (usually in combination with IUI),^48,62,63 and ⁶⁴ the focus here is on ovulation induction in anovulatory women; empiric clomiphene and other treatments for unexplained infertility are discussed in detail in Chapter 27.

Given its mechanism of action, it is not surprising that clomiphene typically is ineffective in women with hypogonadotropic hypogonadism (WHO Group I). Together, low or lownormal FSH levels and low serum estrogen concentrations indicate that the hypothalamic-pituitary-ovarian axis is not functioning normally in women with hypothalamic amenorrhea; if it were, FSH levels would be high because estrogen concentrations are low. If low endogenous estrogen concentrations cannot stimulate increased FSH secretion, there is little reason to think that a clomiphene-induced decrease in the level of negative estrogen feedback will succeed, and it rarely does. Alternative treatments that directly stimulate the pituitary (pulsatile exogenous GnRH) or the ovary (exogenous gonadotropins) usually are required.

Because the corpus luteum derives from the ovulatory follicle, its functional capacity depends, in part, on the quality of preovulatory follicular development. Logically, inadequate follicular development can be expected to cause or predispose to poor luteal function,
if ovulation still occurs. Indeed, the most obvious example of poor luteal function, a short luteal phase, is associated with abnormally low follicular phase FSH levels.^65,66 Consequently, clomiphene is both a logical and effective choice for treatment.^67,68,69 and ⁷⁰ Progesterone levels typically are higher in clomiphene-induced ovulatory cycles than in normal spontaneous cycles, likely because preovulatory follicular development is optimized and because treatment often results in more than one corpus luteum.^71,72

Clomiphene citrate treatment is generally limited to women with demonstrated ovulatory dysfunction but may also be justified in normally ovulating women whose infertility remains unexplained, particularly when they are young and infertility is of short duration, and those unwilling or unable to pursue more aggressive treatments. The efficacy of clomiphene treatment in women with unexplained infertility can be attributed to optimizing follicular development or to the “superovulation” of more than a single ovum.^62,63 Empiric clomiphene treatment is most effective when combined with intrauterine insemination (IUI), in an effort to increase the numbers of both ova and sperm (Chapter 27).^48,73

Clomiphene is administered orally, typically beginning on the third to fifth day after the onset of a spontaneous or progestin-induced menses. Ovulation and conception rates and pregnancy outcomes are similar when treatment starts anywhere between cycle days 2 and 5.⁷⁴ In women with amenorrhea, treatment can begin immediately if pregnancy has been excluded. The dose of clomiphene required to induce ovulation correlates with body weight but cannot be predicted confidently for an individual woman.^75,76 Although obese women often require higher doses of clomiphene treatment, the results achieved ultimately are similar to those observed in lean women.^77,78 No clinical or laboratory parameter has proven utility for predicting the dose of clomiphene needed to induce ovulation.⁷⁹

Treatment usually starts with a single 50 mg tablet daily for a 5-day interval and, if necessary, increases by 50 mg increments in subsequent cycles until ovulation is achieved. Most women who respond to clomiphene will respond to either 50 mg (52%) or 100 mg (22%).^80,81 Lower doses (12.5-25 mg daily) deserve consideration for women who prove highly sensitive to the drug or develop large ovarian cysts that prevent continued treatment.⁸² Although not approved by the U.S. Food and Drug Administration, higher doses of clomiphene (150-250 mg daily) sometimes can succeed when lower doses fail (150 mg, 12%; 200 mg 7%; 250 mg 5%).^80,81 We believe that treatment with doses up to 150 mg is reasonable before considering alternatives.⁸³ Longer durations of clomiphene treatment (7-10 days) can succeed in some women when standard treatment does not and occasionally may be useful when there are no practical alternatives.^84,85 and ⁸⁶

The same methods used for diagnosis of anovulation can be used to evaluate the response to treatment. BBT recordings are simple and inexpensive, but can become tedious over time. A serum progesterone level greater than 3 ng/mL provides reliable evidence that ovulation has occurred,^3,4 but must be timed appropriately for confident and correct interpretation. Measuring the serum progesterone concentration between cycle days 22 and 25 will minimize the risk of sampling immediately after ovulation (occurring as late as cycle day 19-20 in cycles lasting up to 35 days) or before menses, when levels less than 3 ng/mL might be observed and misinterpreted. More sophisticated tests of ovulation involving greater costs and logistical demands are unnecessary to determine only if ovulation occurred, but may be justified once successful ovulation induction has been achieved.

Commercial test kits can detect the midcycle urinary LH surge and help to determine not only whether ovulation occurred, but when, and to accurately define the length of the luteal phase.⁸⁷ In clomiphene-induced ovulatory cycles in anovulatory women, the LH surge typically occurs 5-12 days after treatment ends, most often on cycle day 16 or 17 when clomiphene is administered on days 5-9.⁸⁸ Ovulation generally occurs 14-26 hours after surge detection and almost always within 48 hours.⁸⁷ However, in clinical practice, both false negative and false positive results are relatively common.^89,90 An endometrial biopsy yielding secretory endometrium also implies recent ovulation,⁹¹ but the associated costs and discomfort cannot be justified for that purpose alone. Serial transvaginal ultrasonography can demonstrate the size and number of developing follicles, track endometrial growth, and provide presumptive evidence of ovulation,^92,93 but is difficult to justify when less complicated and costly methods can provide the necessary information. Some advocate monitoring at least the first cycle of treatment to identify those who respond excessively.⁹⁴ A study comparing fecundity in clomiphene-induced cycles monitored with BBT, urinary LH excretion, or serial transvaginal ultrasonography found no clear advantage for any one of the three methods.⁹⁵

Traditionally, when the serum progesterone concentration reveals persistent anovulation after clomiphene treatment, a progestin is prescribed (e.g., medroxyprogesterone acetate, 5-10 mg daily for 5-7 days) to induce menses before the next cycle begins at a higher dosage. Although effective, the sequence takes time and several months may pass before a patient is proven unresponsive to clomiphene. A “stair-step” treatment protocol is an alternative that can shorten the time required to achieve ovulation and to identify those who require different treatment. The regimen involves treatment with clomiphene (50 mg) on cycle days 5-9 after a spontaneous or induced menses, ultrasonography on day 11-14, immediate treatment at the next higher dose level (100 mg) if no dominant follicle (≥15 mm) has emerged, repeated ultrasonography 1 week later, and if still no dominant follicle is observed, immediate treatment at the highest dose level (150 mg) and ultrasonography again 1 week later.⁹⁶ In a case series involving 31 anovulatory infertile women with PCOS treated with a stair-step regimen, the time to ovulation was 23-35 days, cycle fecundability was 13%, and estimated costs were increased only modestly.⁹⁶

In the past, monthly pelvic examinations to exclude residual ovarian enlargement were recommended before each new clomiphene treatment cycle began. It is still prudent to
postpone further treatment when symptoms lead to discovery of a large cyst or grossly enlarged ovaries, but clinical studies and experience indicate that routine baseline physical examination or ultrasonography is unnecessary.⁸⁸ Nevertheless, regular contact should be maintained to review the progress of treatment and to ensure that any additional evaluation needed is accomplished efficiently.

Clomiphene will induce ovulation successfully in 70-80% of properly selected women.^97,98 The likelihood of response decreases with increasing age and body mass index and with the extent of any associated hyperandrogenemia in anovulatory women. Interestingly, women with amenorrhea are more likely to conceive than those with oligomenorrhea,⁹⁷ possibly because infertile women who menstruate also likely ovulate, albeit infrequently, and are more likely to have other co-existing infertility factors.

Among anovulatory infertile women who respond to clomiphene treatment, the overall cycle fecundability is approximately 15%. In women with no other infertility factors, cycle fecundability may reach as high as 22%, comparable to that observed in normal fertile couples after discontinuation of barrier contraception and those with male factor infertility receiving therapeutic donor inseminations.⁹⁷ Cumulative pregnancy rates of 70-75% can be expected over six to nine cycles of treatment.^99,100 Thereafter, cycle fecundability falls substantially. When pregnancy is not achieved within 3-6 clomiphene-induced ovulatory cycles, the infertility investigation should be expanded to exclude other infertility factors not yet evaluated, or to change the overall treatment strategy if evaluation is already complete. Prolonged treatment with clomiphene is inappropriate, particularly for women over age 35.

Clomiphene treatment generally is very well tolerated. Minor side effects are relatively common, but rarely are persistent or severe enough to require that treatment be discontinued.

Transient hot flashes, usually limited to the short interval of treatment, occur in 10-20% of women.⁸³ Considering that clomiphene causes a central misperception that endogenous estrogen levels are low, vasomotor symptoms are not difficult to understand. Mood swings also are relatively common. Other mild and less common side effects include headache, breast tenderness, pelvic pressure or pain, and nausea. Visual disturbances (blurred or double vision, scotomata, light sensitivity) are uncommon (1-2%) and reversible, but reports of persistent “afterimages” (palinopsia) and light sensitivity (photopobia) make them nonetheless unnerving;¹⁰¹ when such symptoms appear, prudence dictates that treatment be abandoned in favor of alternative methods for ovulation induction.

Clomiphene treatment has risks, but serious complications are rare. Inevitably, questions arise concerning the risks for multiple pregnancy, congenital anomalies, and other potential adverse outcomes associated with clomiphene treatment.

The principal risk associated with clomiphene treatment is an increased risk for conceiving a conceiving a multiple pregnancy. The clomiphene-induced increase in FSH secretion is only transient, so normal selection mechanisms still operate to yield only a single mature follicle in most treatment cycles in anovulatory women. Nevertheless, multi-follicular development is relatively common and the overall risk of multiple pregnancy is increased to approximately 7-10%.^102,103 and ¹⁰⁴ The large majority of multiple pregnancies conceived in clomiphene-induced cycles are twins; the risk for triplets is 0.3-0.5%, for quadruplets 0.3%, and for quintuplets 0.1%.²⁸ The higher risk of multi-fetal gestation is another reason to treat with the lowest effective dose of clomiphene; higher doses do not improve results and only increase the risk of superovulation and multiple pregnancy, with all of the attendant antenatal and neonatal complications.

There is no evidence that clomiphene treatment increases the overall risk of birth defects or of any one anomaly in particular.¹⁰⁵ Several large series have examined the question and have drawn the same conclusion.^102,103,104 and ¹⁰⁵ In a series of 1,034 pregnancies resulting from clomiphene treatment, 14.2% ended in miscarriage, 0.5% in ectopic pregnancy, 0.1% in molar pregnancy, and 1.6% in stillbirth, and among 935 live born infants, malformations were detected in 21 (2.3%).¹⁰⁵ Earlier suggestions that the risk of neural tube defects might be higher in pregnancies conceived after clomiphene treatment were not confirmed in later investigations.^106,107 A small study of pregnancy outcomes in women exposed inadvertently to clomiphene during the first trimester of pregnancy found no evidence of teratogenicity.¹⁰⁸ There also is no evidence that clomiphene treatment increases the risk of developmental delay or learning disability in children conceived during clomiphene treatment.¹⁰⁹

Early studies suggested that the incidence of spontaneous miscarriage in pregnancies resulting from clomiphene treatment might be increased, but a number of others have observed miscarriage rates no different from those in pregnancies conceived without treatment.^80,110

The incidence of ovarian hyperstimulation syndrome (OHSS) in clomiphene-induced cycles is difficult to determine confidently because definitions of the syndrome vary widely among studies. In general, mild symptoms of ovarian hyperstimulation (transient abdominal discomfort, mild nausea, vomiting, diarrhea, and abdominal distention) are not altogether uncommon but require only expectant management. When induction of ovulation proceeds in the recommended incremental fashion to establish the minimum effective dose, the risk of clinically significant OHSS (massive ovarian enlargement, progressive weight gain, severe abdominal pain, intractable nausea and vomiting, gross ascites, oliguria) is remote.

The incidence of ovarian cancer is decreased among parous women and those using hormonal contraception for prolonged intervals, suggesting that “incessant ovulation” (repeated epithelial disruption and repair) predisposes to development of ovarian cancer and that treatment with ovulation-inducing drugs might increase the risk.¹¹¹ The results of case-control studies conducted in the 1990s lent credence to the notion and raised considerable concern,^112,113 although their conclusions were challenged because of important methodologic flaws. One study compared infertile treated women to fertile women rather than to infertile untreated women, even though infertility and nulliparity were known risk factors for ovarian cancer.¹¹² Another included cancers of all types and tumors of low malignant potential despite their differing pathophysiology.¹¹³ Since then, numerous studies have con-firmed that the incidence of ovarian cancer is increased in infertile women, but have failed to find any substantive evidence that ovulation-inducing drugs increase the risk.^{114,115,116,117,118,119,120} and ¹²¹ The results of studies of the risk for breast cancer have been conflicting, with some showing no association with ovulation-inducing drugs and others suggesting possible increases in risk in certain subgroups.¹²⁰ No causal relationship between ovulation-inducing drugs and ovarian or breast cancer has been established, but prolonged treatment with clomiphene nonetheless should be avoided, primarily because it has little hope of success.

Clomiphene failure describes women who do not ovulate in response to clomiphene treatment, not those who fail to conceive despite successful clomiphene-induced ovulation. In the latter group, additional evaluation is indicated to identify other potential co-existing infertility factors not already excluded. If or when that has been accomplished, persistent infertility is best regarded and treated as unexplained infertility (Chapter 27).

Although most properly selected women will ovulate in response to clomiphene treatment, many do not. Direct ovarian stimulation with exogenous gonadotropins, discussed later in detail, is an obvious alternative, but it is by no means the only option that merits consideration. Many clomiphene-resistant anovulatory infertile women will respond to supplemental or combination treatment regimens. Options include adjuvant treatment with glucocorticoids, exogenous human chorionic gonadotropin (hCG), or metformin, and preliminary suppressive therapy (hormonal contraceptives). It is helpful to be familiar with these less common ovulation induction strategies because many couples are understandably reluctant or unable to pursue the obvious alternative of gonadotropin treatment once fully advised of the associated costs, logistical demands, and risks.

Failure to respond to one or more of these less commonly employed treatment regimens is not a prerequisite for exogenous gonadotropin therapy. They are simply useful alternatives for those unwilling or unable to pursue gonadotropin therapy and, for some, may be the only options when clomiphene treatment has failed. A choice among them is not entirely arbitrary, but should consider specific elements of the patient’s history, the results of laboratory evaluation, and observations in previous unsuccessful clomiphene treatment cycles.

In almost all studies conducted to date, letrozole (2.5-7.5 mg daily) and anastrazole (1 mg daily) have been administered for a 5-day interval in a manner very similar to that typical for clomiphene treatment (e.g., cycle day 3-7). One trial comparing outcomes achieved with 5 or 10 days of letrozole treatment (2.5 mg daily, beginning on cycle day 1) in clomiphene-resistant women (100 mg daily) observed significantly more preovulatory follicles (3.0 vs. 1.8) and higher pregnancy rates (17.4% vs. 12.4%) in women receiving the longer course of treatment.¹⁸⁴ A single-dose treatment regimen of letrozole (20 mg on cycle day 3) also has been described, with preliminary data suggesting it can achieve results similar to those observed with a lower multi-dose treatment protocol.¹⁸⁵

The optimal dosage of letrozole and anastrazole has not been established firmly. In most trials involving anovulatory women, 2.5 mg of letrozole or 1 mg of anastrozole has been administered. In a trial comparing outcomes achieved with a 2.5 mg or 5 mg dosage of letrozole in ovulatory women with unexplained infertility, the 5 mg dose yielded significantly greater numbers of follicles and a higher pregnancy rate (26.3% vs. 5.9%).¹⁸⁶ In another trial comparing outcomes in women with unexplained infertility randomly assigned to receive treatment with letrozole (7.5 mg daily) or clomiphene (100 mg daily), followed by intrauterine insemination, the two treatments yielded similar numbers of preovulatory follicles (2.1 vs. 1.7) and pregnancy rates (11.5% vs. 8.9%); results also suggested that such higher doses of letrozole may result in greater and longer suppression of estrogen production that could limit endometrial proliferation.¹⁸⁷

Altogether, the available data suggest that the optimal dose of letrozole probably ranges between 2.5 mg and 5 mg daily; outcomes achieved with doses of anastrazole greater than 1 mg daily remain to be evaluated.

Evidence for the efficacy of aromatase inhibitors for ovulation induction is accumulating rapidly. Early studies exploring the use of aromatase inhibitors focused on anovulatory women considered resistant to clomiphene, because they failed to ovulate or exhibited poor endometrial proliferation during treatment. More recent studies have sought to compare the effectiveness of aromatase inhibitors to that of clomiphene in unselected populations of anovulatory infertile women.

In an early proof of concept trial involving anovulatory clomiphene-resistant women, 9/12 patients (75%) ovulated after treatment with letrozole (2.5 mg daily) and hCG (lead follicle follicle > 20 mm), three conceived (resulting in two singleton births), and normal endometrial proliferation was observed in all.¹⁸⁸ In a subsequent trial involving clomiphene-resistant (150 mg daily) anovulatory women with polycystic ovary syndrome (PCOS), 22/44 patients (50%) ovulated after treatment with letrozole (2.5 mg daily) and hCG (lead follicle > 18 mm) and six conceived; response did not relate to age, BMI, or menstrual pattern (amenorrhea vs. oligomenorrhea) and mean endometrial thickness was 10.2 mm.¹⁸⁹ In a third involving 64 women with clomiphene-resistant anovulation (100 mg daily), patients were randomized to receive treatment with letrozole (7.5 mg daily) or clomiphene (150 mg daily), followed by hCG (lead follicle ≥ 18 mm); women receiving letrozole had higher ovulation (62.5% vs. 37.5%) and pregnancy rates (40.1% vs. 18.8%) but the differences were not significant.¹⁹⁰

Two randomized trials have compared the effectiveness of letrozole (2.5 mg daily) and anastrazole (1 mg daily) for ovulation induction in clomiphene-resistant anovulatory women with PCOS also receiving hCG (lead follicle follicle ≥ 18 mm). One involved 40 clomiphene-resistant patients (200 mg daily, or endometrial thickness ≤ 5 mm) and observed significantly greater endometrial thickness (8.2 vs. 6.5 mm), ovulation rates (84% vs. 60%), and pregnancy rates (19% vs. 10%) in women receiving letrozole.¹⁹¹ In the larger trial involving 220 clomiphene-resistant patients (100 mg daily, or endometrial thickness < 5mm), ovulation rates (62% vs. 63%), pregnancy rates (12 % vs. 15%), and endometrial thickness (9.1 vs. 10.2 mm) in the two groups were similar.¹⁹² Taken together, these observations suggest strongly that aromatase inhibitors can be effective in anovulatory women who fail to ovulate in response to clomiphene treatment. Aromatase inhibitors also might be considered for women who respond to clomiphene but exhibit grossly poor endometrial proliferation. Virtually all studies have included monitoring with ultrasonography and adjuvant treatment with hCG, so it remains unclear whether these additions
to the treatment regimen are necessary to achieve success. If not, the substantially lower complexity, risks, and costs of treatment, compared to the alternative of gonadotropin therapy, make it easy to justify a trial of treatment with an aromatase inhibitor for clomiphene-resistant anovulatory women.

The results achieved with aromatase inhibitors in clomiphene-resistant anovulatory women suggested that aromatase inhibitors might be considered a first-line treatment for ovulation induction. Three trials have compared letrozole and clomiphene in treatment-naïve anovulatory women. The first involved 106 patients who were randomized to receive letrozole (2.5 mg daily) or clomiphene (100 mg daily), followed by hCG (lead follicle ≥ 18 mm); a significantly greater endometrial thickness (8.4 vs. 5.2 mm), ovulation rate (82% vs. 64%), and pregnancy rate (22% vs. 9%) was observed in women receiving letrozole.¹⁹³ In a second trial with the same design, ovulation occurred in 65/99 letrozole-induced cycles (66%) and in 71/95 (75%) clomiphene-induced cycles, but endometrial thickness (8 mm vs. 8 mm) and pregnancy rates (9% vs. 7%) were not different.¹⁹⁴ The largest of the three randomized trials involved a total of 438 women who received treatment with letrozole (5 mg daily) or clomiphene (100 mg daily) and hCG (lead follicle follicle ≥ 18 mm); ovulation rates (365/540, 68% vs. 371/523, 71%) and pregnancy rates (15% vs. 18%) were not different and endometrial thickness was significantly greater in women receiving clomiphene (9.2 vs. 8.1 mm).¹⁹⁵ In the only trial involving anastrazole, 115 patients received treatment with anastrazole (1 mg daily, cycle days 3-7, 243 cycles) and hCG (lead follicle follicle ≥ 18 mm) and outcomes were compared to those in matched historical controls treated with clomiphene (100 mg daily, 226 cycles); endometrial thickness was significantly greater in anastrazole cycles (10.1 vs. 8.2 mm), but ovulation rates (68% vs. 69%) and pregnancy rates (10.2% vs. 7.9%) were similar in the two groups.¹⁹⁶

A few studies have examined the outcomes of pregnancies conceived after treatment with aromatase inhibitors. A cohort study found that pregnancies conceived in letrozole-induced cycles are significantly more likely to be singletons than those conceived in cycles involving treatment with clomiphene or gonadotropins.¹⁹⁷ One case series comparing the incidence of congenital malformations in 911 newborns of women who conceived after treatment with letrozole (14/514, 2.4%) or clomiphene (19/397, 3.0%) found no difference.¹⁷¹ Another comparing the incidence of birth defects in children born to mothers treated with letrozole or clomiphene to that in pregnancies conceived without treatment also observed no differences.¹⁷²

Surgical treatments aimed at restoring ovulatory function in anovulatory infertile women date back to the classic bilateral ovarian wedge resection described originally by Stein and Leventhal in 1935.¹⁹⁸ The procedure understandably fell out of favor after the introduction
of clomiphene citrate and gonadotropins for ovulation induction. Advances in laparoscopic surgery sparked renewed interest in the procedure, with ovarian “drilling” now representing the modern equivalent of the classical wedge resection and another treatment option for clomiphene-resistant, hyperandrogenic, anovulatory women.

Several methods for ovarian drilling have been described, including electrocautery or laser vaporization (approximately four to six sites per ovary, avoiding surfaces near the tubo-ovarian interface to minimize the risk for adhesions that might adversely affect ovum capture) and multiple biopsy.^199,200 and ²⁰¹ All are intended to cause focal destruction of the ovarian stroma in efforts to decrease both intraovarian and systemic androgen concentrations. There is no evidence for the superiority of any one method, but the most common technique involves electrocautery using a unipolar needle electrode insulated above the distal 1-2 cm.

Postoperative serum concentations of androstenedione and testosterone decrease, at least for a time,^202,203 and ²⁰⁴ and inhibin concentrations also decline.^205,206 Both changes likely contribute to an associated increase in FSH levels. As with the classical surgical procedure, the principal risk associated with laparoscopic ovarian drilling is postoperative adnexal adhesion formation that may decrease overall fertility, although the risk and severity of adhesions are lower;^207,208,209 and ²¹⁰ second-look laparoscopy and adhesiolysis do not appear necessary or useful.^211,212 There is one report of unilateral ovarian atrophy after ovarian drilling with electrocautery.²¹³ Whether ovarian drilling might adversely affect ovarian reserve and predispose to early menopause has not been investigated specifically, but other data indicate that reductive ovarian surgery, including ovarian wedge resection, increases risk for early menopause.²¹⁴

In numerous uncontrolled observational studies, 40-90% of women have ovulated after laparoscopic ovarian drilling and approximately half of those have conceived.^199,200 and ^201,215,216 In truly clomiphene-resistant women, ovarian drilling can improve clomiphene sensitivity or response to exogenous gonadotropins when it does not restore spontaneous ovulatory cycles.²¹⁷ When considering laparoscopic ovarian drilling as a treatment option in clomiphene-resistant anovulatory infertile women, the most relevant data derive from randomized controlled trials comparing surgical treatment with ovulation induction using exogenous gonadotropins.^211,218,219 A 2007 systematic review and meta-analysis including nine trials found no evidence of a difference in live birth (OR=1.04, CI=0.59-1.85) or clinical pregnancy rate (OR=1.08, CI=0.69-1.71).²²⁰ After 12 months, the ovulation rate achieved with drilling (52%) was similar to that with gonadotropin therapy (62%). However, as might be expected, laparoscopic ovarian drilling yields far fewer multiple pregnancies than gonadotropin treatment (1% vs. 16%; OR=0.13, CI=0.03-0.52); the miscarriage rates associated with surgical and medical treatment are comparable.²²⁰

The ideal candidate for ovarian drilling is not obese and has no other infertility factors. The procedure often is unsuccessful in obese women (BMI >30 kg/m²),^221,222 and whereas over 80% of women can be expected to conceive after surgery when clomiphene-resistant anovulation is the sole cause of infertility,^222,223 only 15-30% achieve pregnancy when there is a coexisting tubal, male, or other infertility factor.^203,223

Gonadotropin preparations have evolved gradually over the years, from relatively crude urinary extracts, to more highly purified urinary extracts, to the recombinant preparations in common use today.^224,225

For almost 30 years, the only exogenous gonadotropins available were human menopausal gonadotropins (hMG, menotropins), an extract prepared from the urine of postmenopausal women containing equivalent amounts (75 IU) of FSH and LH per ampule or vial and requiring intramuscular injection. Originally, the urinary source was a single convent in Italy, but later collections were expanded to a number of centers in other countries.²²⁶ Urinary menotropins also contain small but measurable and varying amounts of hCG, most of it added intentionally during the manufacturing process to provide the appropriate amount of LH activity and some derived from other sources.²²⁷ Clinical use of hMG began in 1950 but the clinical trials did not begin until after 1960.^228,229 Relatively crude gonadotropin extracts like traditional hMG also contained significant amounts of uncharacterized urinary protein that may be antigenic.²³⁰ Contemporary hMG preparations are more highly purified than in the past and can be administered subcutaneously.²³¹

Beginning about 25 years ago, more purified urinary FSH preparations (urofollitropin) were developed by removing LH from urinary extracts using immunoaffinity columns containing polyclonal anti-hCG antibodies.²³² Early preparations of purified urinary FSH (75 IU) contained less than 1 IU LH but a considerable amount of other urinary protein and still required intramuscular administration. Further purification using monoclonal antibodies specific for FSH yielded a preparation containing less than 0.1 IU LH and less than 5% unidentified protein. The even more highly purified products now in use contain less than 0.001 IU LH, very low levels of urinary protein, and can be administered subcutaneously.

Just over 15 years ago, the in vitro production of recombinant human FSH was achieved through genetic engineering. Briefly summarized, the process involves introduction of the genes encoding the a- and b-FSH subunits into the genome of a Chinese hamster ovary cell line, which then synthesizes and secretes a glycosylated bioactive dimeric FSH that is finally purified by immunochromatography using a specific anti-FSH monoclonal antibody. Recombinant FSH preparations contain less acidic FSH isoforms that have a shorter half-life than those derived from human urine but stimulate estrogen secretion as or even more efficiently.²³³ The advantages of recombinant FSH preparations include the absence of urinary protein, more consistent supply, and less batch-to-batch variation in biologic activity. The two recombinant FSH preparations currently available are marketed
as follictropin alpha and follitropin beta. They are both structurally identical to native FSH and contain 1 alpha and 1 beta glycoprotein chain, but the post-translational glycosylation process and purification procedures for the two are different.²³⁴ Despite the subtle differences in structure, they are functionally the same. The biological activity of all FSH preparations, including recombinant formulations, is ultimately confirmed using the classic Steelman-Pohley ovarian bioassay.²³⁵

Most recently, recombinant technology has been used to create a new chimeric gene containing the coding sequences of the FSH β-subunit and the C-terminal peptide of the hCG β-subunit (containing additional glycosylation sites). Co-expression of the α-subunit and the chimeric FSH β-subunit produces a new molecule called corifollitropin alpha, which has a prolonged half-life and enhanced in vivo bioactivity compared with wild-type FSH. Early studies in women suppressed by treatment with a long-acting GnRH agonist have confirmed the extended half-life of the compound and clinical trials have demonstrated that corifollitropin alpha can induce and sustain multifollicular growth for a week in women receiving ovarian stimulation for IVF. Corifollitropin alpha provides the means for simpler and more convenient treatment compared with conventional treatment regimens involving daily injections with FSH and GnRH agonist. Although the new recombinant gonadotropin is likely to find applications in assisted reproductive technologies where the goal is to induce multi-follicular development, it is not well suited for ovulation induction where unifollicular development is the objective.^236,237