Before the advent of IVF, women with irreparable bilateral tubal obstruction were essentially sterile, and the prognosis for those with less severe distal disease was only fair. In the modern era of ART, surgical treatments are declining in importance and the prognosis for women with tubal factor infertility has improved dramatically. Approximately 9% of patients using ART have a primary diagnosis of tubal factor infertility.³ The relative advantages and disadvantages of surgery and IVF for the treatment of tubal factor infertility and the factors bearing on a choice between the two are discussed in depth in Chapter 27 and summarized here.

Reconstructive surgery remains a viable option for young women with mild distal tubal obstruction or peritubular adhesions (because postoperative live birth rates can exceed 50%),^4,5 and ⁶ but IVF is the treatment of choice for women with severe distal disease. Results achieved with surgery have varied, but success rates (10-35%) are generally lower than with IVF and the risk of ectopic pregnancy is higher (5-20%).^3,7,8,9 and ¹⁰ In 2007, the overall IVF live birth rate (per cycle start) for U.S. women with tubal factor infertility (all ages) was 30.7%.³ IVF is also the best treatment for women who remain infertile for more than a year after tubal surgery (the likelihood for success diminishes progressively with time after operation), for older women with significant distal tubal disease (cycle fecundity is low after distal tubal surgery and time is limited), and for women with recurrent distal tubal obstruction (repeated attempts to correct distal tubal occlusive disease are rarely successful).

Although not candidates for reconstructive surgery, women with severe distal tubal disease still can benefit from surgery before IVF. A substantial body of evidence indicates that communicating hydrosalpinges (proximal patency and distal occlusion) decrease the probability of both pregnancy and live birth after IVF by approximately one-half. The mechanism for the adverse effect of hydrosalpinges on IVF outcomes could involve mechanical interference with implantation or toxic effects on the embryo or endometrium.^11,12,13,14 and ¹⁵ A 2010 systematic review including five randomized trials involving 646 women observed that the odds of achieving an ongoing pregnancy were twice as great after laparoscopic salpingectomy for hydrosalpinges before IVF (OR=2.14, CI=1.23-3.73).¹⁶ Laparoscopic proximal occlusion of the tubes also increased the odds of clinical pregnancy, compared to no intervention (OR=4.66, CI=2.47-10.01), and neither surgical procedure was superior.¹⁶ Other treatments have been suggested, such as ultrasound-guided aspiration of hydrosalpinges after oocyte retrieval,¹⁷ but are unproven, and evidence suggests the fluid re-accumulates rapidly.¹⁸

Proximal tubal occlusion observed during hysterosalpingography (HSG) often is not real and results from “cornual spasm” or other technical pitfalls of the procedure (Chapter 27). Efforts to confirm the diagnosis are justified; otherwise many women may needlessly undergo IVF. Common methods include repeated HSG¹⁹ and laparoscopic “chromotubation.”^20,21 and ²² Fluoroscopic or hysteroscopic selective tubal cannulation both establish the diagnosis and provide the means for successful treatment.^{19,20,23,24,25,26} and ²⁷ Micro-surgical segmental resection and anastomosis is another proven treatment for true proximal tubal obstruction,^28,29,30 and ³¹ but requires uncommon technical expertise. IVF is the obvious alternative when cannulation is contraindicated (salpingitis isthmica nodosa) or technically unsuccessful, and when infertility persists for more than 6-12 months after the procedure.

Approximately 1 million U.S. women have an elective tubal sterilization procedure each year; up to 7% regret the decision and about 1% later request its reversal.^32,33 The most commonly cited reasons for regret include new relationships, changes in family planning goals, and death of a child. Regrets are more common in younger women, those who were unaware of the spectrum of contraceptive options, women whose decision for sterilization was influenced by a third-party (partner, other family member, friend, or physician), and in those sterilized postpartum or after an abortion.^34,35 Women 30 years old or younger are twice as likely as older women to express regret, 3.5 to 18 times more likely to request information about reversal of the procedure, and approximately eight times more likely to have a sterilization reversal or IVF.³⁶

Young women sterilized using rings or clips and women having no other infertility factors have the best surgical prognosis; success rates are lower for older women, those sterilized by cautery (particularly multiple-burn techniques), and women with other infertility factors.^{37,38,39,40,41,42,43} and ⁴⁴ Although conception rates are quite good (45-82%) after microsurgical tubal anastomosis in properly selected candidates, IVF is a legitimate alternative to surgery, particularly for older women, those with a poor surgical prognosis or preferring to avoid surgery, and women who desire only one additional pregnancy.

The association between endometriosis and infertility and the pathogenic mechanisms involved are considered at length in Chapter 29. In brief summary, 20-40% of infertile women have endometriosis and accumulated evidence indicates that fertility decreases with the severity of the disease. Endometriosis may cause infertility by distorting adnexal anatomy and interfering with ovum capture,⁴⁵ or possibly by impairing oocyte development, early embryogenesis, or endometrial receptivity.^46,47,48,49 and ⁵⁰ IVF should be expected to
overcome any anatomical obstacles, and although it would seem less likely to conquer functional disorders of oocyte, embryo, or endometrial development, outcomes in women with endometriosis suggest it can. Endometriosis is the primary diagnosis in approximately 5% of patients using ART.³

Treatment options for infertile women with advanced stages of endometriosis include conservative surgical treatment and IVF. For those with severe symptoms, surgery is the most logical treatment. Data from case series suggest that cumulative pregnancy rates 1-3 years after surgical treatment are approximately 50% for women with endometriomas,^51,52,53 and ⁵⁴ and about 30% for women with complete cul-de-sac obliteration.^51,55 Careful surgical technique is important because ovarian function can be compromised by excision of excessive tissue or damage to hilar vessels⁵⁶; the risk of ovarian failure after excision of bilateral ovarian endometriomas is approximately 2.5%.⁵⁷ After surgical treatment, the choice between expectant management, empirical treatment, and IVF should be based on age, the surgical results, and the severity of any other coexisting infertility factors.

Asymptomatic infertile women with advanced endometriosis, including those with ovarian endometriomas, can be treated surgically or proceed directly to IVF. There is no evidence to indicate that endometriomas have any important adverse effect on the response to ovarian stimulation or IVF outcomes.^{58,59,60,61,62,63,64} and ⁶⁵ Consequently, endometriomas can be left untreated before IVF. Aspiration of endometriomas before ovarian stimulation or at time of oocyte retrieval has been associated with an increased risk for developing an ovarian abscess,^66,67 and ⁶⁸ although the risk appears quite low.⁶⁹

Treatment options for asymptomatic women with known or suspected minimal or mild endometriosis and no other infertility factors include expectant management, surgical treatment, empiric treatment with clomiphene or exogenous gonadotropins and IUI, and IVF. In older women, those with other coexisting infertility factors, and women who have failed other forms of treatment, IVF is often the best overall choice.

Results of a 2006 systematic review including three randomized trials involving 165 infertile women with endometriosis of varying severity suggest that treatment with a GnRH agonist for 3-6 months before IVF can increase the odds of clinical pregnancy (OR=4.28, CI=2.0-9.15).⁷⁰ However, because prolonged treatment with a GnRH agonist also can decrease response to ovarian stimulation, most clinicians do not favor suppressive treatment before IVF.

Poor semen quality is the sole cause of infertility in approximately 20% of infertile couples and an important contributing factor in another 20-40% of couples with reproductive failure.^71,72 Many infertile men have disorders than can be corrected medically or surgically if properly diagnosed and treated, allowing them to achieve natural conception with their partners.⁷² In others, mild but important semen abnormalities can be overcome by IUI.

When treatment is not possible or fails, and insemination with donor sperm is not an acceptable option, IVF and ICSI, using sperm isolated from the ejaculate or extracted from the epididymis or testis, offers realistic hope for success. The evaluation and treatment of male factor infertility are the focus of Chapter 30. Discussion here is limited to the indications for ART.⁷²

The likelihood of male factor infertility is increased in men whose ejaculates consistently exhibit a sperm concentration under 15 million sperm/mL, less than 32% progressive motility, or fewer than 4% morphologically normal sperm (strict criteria, WHO III standard).⁷³
The overall odds of male infertility increase with the number of abnormal parameters in the subfertile range; the probability is two to three times higher when one is abnormal, five to seven times higher when two are abnormal, and approximately 16 times greater when all three parameters are abnormal.⁷⁴ Additional genetic evaluation is indicated for men with severe oligospermia (sperm concentration <5 million/mL) whose sperm may be used for ICSI (Chapter 30).

Medical or surgical treatment to improve or normalize poor semen quality is always the first and best option, when that is possible. When treatment is not feasible or proves unsuccessful, timely IUI can help to improve cycle fecundity in some couples with male factor infertility.

Best results with IUI are achieved when the number of total motile sperm in the insemination specimen exceeds a threshold of approximately 10 million^75,76 and ⁷⁷ and 14% or more of sperm have normal morphology (strict criteria; WHO III standard).⁷⁸ Higher counts do not further increase the likelihood for success^75,79 and IUI is seldom successful when fewer than 1 million total motile sperm are inseminated.^80,81 Success rates with IUI are best when 14% or more of sperm have normal morphology (strict criteria), intermediate with values between 4% and 14%, and generally quite poor when fewer than 4% of sperm are normal.⁷⁸ The likelihood of success with IUI also decreases with increasing female partner age and with coexisting infertility factors (ovulatory dysfunction, uterine and tubal factors).

When IUI is not possible, the prognosis for success with IUI is poor, or IUI proves unsuccessful and therapeutic donor insemination is rejected, IVF is the logical alternative. Approximately 18% of patients using ART have a primary diagnosis of male factor infertility.³

Conventional fertilization rates in IVF cycles are decreased when the total motile sperm count is less than 2-3 million (post-wash).⁸² Numerous studies have observed that conventional fertilization rates also are decreased when less than 4% of sperm are morphologically normal.^83,84,85,86 and ⁸⁷ Although severe teratospermia is widely accepted as an indication for assisted fertilization by ICSI, some observing no differences in fertilization, pregnancy, and live birth rates achieved with ICSI, compared with conventional fertilization, do not regard isolated teratospermia as an indication for ICSI.^88,89,90 and ⁹¹

For women with ovulatory disorders (hypogonadotropic hypogonadism, polycystic ovary syndrome, thyroid disorders, hyperprolactinemia), ovulation induction alone generally restores fertility (Chapter 31), but for some who require exogenous gonadotropins, ovulation induction proves difficult to achieve or consistently results in excessive ovarian stimulation and cycle cancellation for undue risk of ovarian hyperstimulation syndrome (OHSS) and high-order multiple gestation. For these difficult patients, IVF is an obvious treatment alternative, making their high sensitivity to gonadotropin stimulation an asset instead of a liability. Ovulatory dysfunction is the primary diagnosis in approximately 7% of patients using ART.³

The incidence of unexplained infertility ranges from 10% to as high as 30% among infertile populations, depending on diagnostic criteria.^92,93 and ⁹⁴ For women with unexplained infertility, treatment options (cycle fecundability in parentheses) include expectant management (2-4%),⁹⁵ IUI (2-4%),^96,97 empiric treatment with clomiphene (2-4%)^96,98 or exogenous
gonadotropins (5-7%),⁹⁹ combined treatment with IUI and clomiphene (5-10%)^100,101 and ¹⁰² or gonadotropins (7-10%),^99,100,103 and IVF (25-45%).^3,104 As might be expected, success rates with all forms of treatment decline progressively with increasing age of the female partner.

Among couples with unexplained infertility, IVF is the preferred treatment for some and the treatment of last resort for others. In either case, there is no question that IVF is the most effective treatment for couples with unexplained infertility. A higher incidence of fertilization failure has been observed in several, but not all, studies of IVF outcomes in couples with unexplained infertility,^105,106,107 and ¹⁰⁸ prompting many to recommend ICSI when IVF is planned. Approximately 12% of patients using ART have a diagnosis of unexplained infertility.³

IVF using oocytes from a known or anonymous young donor was first developed for women with premature ovarian failure or menopause.¹⁰⁹ Now, oocyte donation is most commonly performed in women over age 42, those with grossly abnormal ovarian reserve test results, and women whose IVF cycles consistently yield poor quality embryos (Chapter 27). Approximately 10% of patients using ART have a primary diagnosis of diminished ovarian reserve.³

Although less commonly encountered, there are a number of other legitimate indications for IVF and related ART procedures.

Fertility preservation is fast becoming a more common indication for ART. Women with cancer or other illnesses requiring treatments (chemotherapy, radiation therapy) that pose a serious threat to future fertility may be candidates for urgent IVF and cryopreservation of
embryos before treatment begins, if time and health allow.¹¹⁰ Oocyte cryopreservation is a viable option for women in similar circumstances having no male partner,¹¹¹ and is rapidly emerging as an option for young women at risk for premature ovarian failure,^112,113 healthy aging women, and others who anticipate delayed childbearing.^{111,112,114,115}

For women with normal ovaries but no functional uterus, due to a developmental anomaly (müllerian agenesis), advanced disease (multiple myomas, severe intrauterine adhesions), or a previous hysterectomy, and for women with medical conditions that preclude pregnancy due to serious health risks, gestational surrogacy offers the opportunity to have their own genetic offspring.^116,117

For couples at risk for transmitting a specific genetic disease or abnormality to their offspring, IVF with preimplantation genetic diagnosis (PGD) provides the means to identify and exclude affected embryos and thereby avoid that risk. PGD is applied most commonly in couples who carry autosomal recessive and sex-linked disorders or harbor a balanced chromosomal translocation.¹¹⁸ Women who carry a genetic disorder not amenable to diagnosis by PGD or who decline PGD may be candidates for oocyte donation. Preimplantation genetic screening (PGS) applies the same technology in couples having no known chromosomal or genetic abnormality in efforts to identify and exclude aneuploid embryos where the risk is increased, as in older women, those with a history of recurrent miscarriage, and in women with repeated unexplained IVF failure.¹¹⁸ Although the technical limitations of current methods for PGS have so far prevented the technology from improving live birth rates in at-risk couples, more sophisticated and reliable methods now emerging hold promise.¹¹⁹

The relationship between maternal age and fertility and the physiology of reproductive aging are discussed in detail in Chapters 27 and 28 and only briefly summarized again here, where the focus is on the relationship between maternal age and IVF outcomes.

The average age of women using ART in the U.S. is 36 years.³ Maternal age is the one most important factor in determining the likelihood for success with IVF. Although IVF can overcome most causes of infertility in younger women, it cannot negate or reverse the age-related decrease in biologic fertility in older women, particularly those over the age of 40.¹²⁰ The success rates achieved with IVF, like natural fertility rates, decline as maternal age increases. The pattern reflects a progressive decline in response to ovarian stimulation, resulting in fewer oocytes and embryos, and a decreased embryo implantation rate, due to declining oocyte quality.^121,122,123 and ¹²⁴ In 2007, the percentage of cycles that resulted in a
live birth from fresh nondonor oocytes, by maternal age, was 39.6% for women under age 35, 30.5% for ages 35-37, 20.9% for ages 38-40 yr, 11.5% for ages 41-42, and 5.4% for ages 43-44 years.³ The pattern of decreasing success rates achieved with IVF parallels that associated with other, less complex, forms of treatment for infertility.¹²⁵

Evidence from numerous lines of investigation indicates that the age-dependent decrease in success rates achieved with IVF relates primarily to an increasing prevalence of aneuploidy in aging oocytes,^126,127 and ¹²⁸ which is reflected in the incidence of miscarriage in pregnancies achieved with ART: less than 14% for women under age 35, 19% at age 38, 28% at age 40, and nearly 60% for women over age 44.³ In a case series of IVF cycles involving women ages 45-49, 70/231 cycles (30%) were cancelled before oocyte retrieval and 34/161 retrievals (21%) resulted in a pregnancy, but only 5/34 pregnancies (15%) and 5/231 cycles (2%) resulted in a live birth.¹²⁹

The concept of ovarian reserve, generally defined as the size and quality of the remaining ovarian follicular pool, and the various methods for its measurement are discussed in detail in Chapter 27. In brief summary, the total number of oocytes in any given women is genetically determined and inexorably declines throughout life, from approximately 1-2 million at birth, to about 300,000 at puberty, 25,000 at age 40, and fewer than 1,000 at menopause.^128,130,131 and ¹³² The rate of follicular depletion is not constant, but increases gradually as the number of follicles remaining decreases.^133,134,135 and ¹³⁶ As the size of the remaining follicular pool decreases, circulating inhibin B levels (derived primarily from smaller antral follicles) decrease, resulting in lower levels of feedback inhibition and a progressive increase in serum follicle-stimulating hormone (FSH) levels, most noticeably during the early follicular phase.^{137,138,139,140,141,142,143,144} and ¹⁴⁵ Increasing inter-cycle FSH concentrations stimulate earlier follicular recruitment, resulting in advanced follicular development early in the cycle, an earlier rise in serum estradiol levels, a shorter follicular phase, and decreasing overall cycle length.^146,147 and ¹⁴⁸

The physiology of reproductive aging provides the foundation for all contemporary tests of ovarian reserve. In clinical practice, the basal early follicular phase (cycle day 2-4) FSH level is the most common test, but antimüllerian hormone (AMH) and antral follicle count are alternatives having significant potential advantages.

As basal FSH levels increase, peak estradiol levels during stimulation, the number of oocytes retrieved, and the probability for pregnancy or live birth decline steadily.^{149,150,151,152,153,154} and ¹⁵⁵

With current assays (using IRP 78/549), FSH levels greater than 10 IU/L (10-20 IU/L) have high specificity (80-100%) for predicting poor response to stimulation, but their sensitivity for identifying such women is generally low (10-30%) and decreases with the threshold value.¹⁵⁶ Although most women who are tested have a normal result, including those with a diminished ovarian reserve (DOR), the test is still useful because those with abnormal results are very likely to have DOR. In a 2008 study, an FSH concentration above 18 IU/L had 100% specificity for failure to achieve a live birth.¹⁵⁷

The basal serum estradiol concentration, by itself, has little value as an ovarian reserve test,^158,159,160 and ¹⁶¹ but can provide additional information that helps in the interpretation of the basal FSH level. An early elevation in serum estradiol reflects advanced follicular development and early selection of a dominant follicle (as classically observed in women with advanced reproductive aging), and will suppress FSH concentrations, thereby possibly masking an otherwise obviously high FSH level indicating DOR. When the basal FSH is normal and the estradiol concentration is elevated (>60-80 pg/mL), the likelihood of poor response to stimulation is increased and the chance for pregnancy is decreased.^162,163,164 and ¹⁶⁵ When both FSH and estradiol are elevated, ovarian response to stimulation is likely to be very poor.

Antimüllerian hormone (AMH) derives from preantral and small antral follicles. Levels are gonadotropin-independent and vary little within and between cycles.^166,167 and ¹⁶⁸ The number of small antral follicles correlates with the size of the residual follicular pool and AMH levels decline progressively with age, becoming undetectable near the menopause.^169,170,171 and ¹⁷²

Overall, lower AMH levels have been associated with poor response to ovarian stimulation and low oocyte yield, embryo quality, and pregnancy rates,^{173,174,175,176} and ¹⁷⁷ but studies correlating mean AMH levels with IVF outcomes have not yielded threshold values that can be applied confidently in clinical care.^{158,174,175,178} In the general IVF population, low AMH threshold values (0.2-0.7 ng/mL) have had 40-97% sensitivity, 78-92% speci-ficity, 22-88% positive predictive value (PPV) and 97-100% negative predictive value (NPV) for predicting poor response to stimulation (<3 follicles, or <2-4 oocytes), but
have proven neither sensitive nor specific for predicting pregnancy.^173,179,180 and ¹⁸¹ In women at low risk for DOR, values of 2.5-2.7 ng/mL have had 83% sensitivity, 82% specificity, 67-77% PPV, and 61-87% NPV for clinical pregnancy.^159,182 A study in women at high risk for DOR (involving older women, those with an elevated FSH, or history of poor response to stimulation) observed that an undetectable AMH had 76% sensitivity, 88% specificity, 68% PPV, and 92% NPV for three or fewer follicles.¹⁷⁴ A higher threshold value (1.25 ng/mL) had 85% sensitivity, 63% specificity, 41% PPV, and 57% NPV for cycle cancellation.¹⁶⁰

The antral follicle count (AFC) is the total number of antral follicles measuring 2-10 mm in both ovaries during the early follicular phase and is a useful measure of ovarian reserve because it quantifies the number of follicles at the stage of development that responds to FSH stimulation.^{183,184,185,186} and ¹⁸⁷

Several studies have observed a relationship between the AFC and response to ovarian stimulation in IVF cycles. In the general IVF population, including women at low and high risk for DOR, an AFC threshold value of three to four follicles has high specificity (73-100%) for predicting poor response to ovarian stimulation and failure to conceive (64-100%), but relatively low sensitivity for both endpoints (9-73% for poor response, 8-33% for failure to conceive).^{160,188,189,190,191,192,193} and ¹⁹⁴ The PPV and NPV of AFC have varied widely in studies. A low AFC has high specificity for predicting poor response to ovarian stimulation and treatment failure, making it a useful test, but low sensitivity limits its overall clinical utility.

In summary, none of the ovarian reserve tests currently in use is an accurate predictor of pregnancy in IVF cycles, unless extreme abnormal threshold values are applied, which results in very low sensitivity for identifying women having a poor prognosis.¹⁵⁷ The tests are adequate for predicting poor response, which does have prognostic value, although not as much in young women as in older women.^195,196 and ¹⁹⁷ Although ovarian reserve tests have become a routine element of pre-treatment evaluation for couples planning IVF, it can be argued that routine testing has limited clinical utility in the large majority of patients and can be misleading, especially in women at low risk for having a diminished ovarian reserve.¹⁹⁸

Ovarian reserve tests always should be interpreted with caution. Rigid application of test results risks inappropriate recommendations for treatment, or for no treatment, and both must be avoided. An abnormal test result does not preclude the possibility of pregnancy. Except perhaps when grossly abnormal, test results should not be used to deny treatment, but only to obtain prognostic information that may help to guide the choice of treatment and best use of available resources. Although the probability of pregnancy may be low, many with abnormal test results will achieve pregnancy if afforded the chance. Ultimately, regardless of the prognosis, the success rate for any individual woman will be zero or 100%.

Although the average overall IVF live birth rate per cycle is approximately 29% for all women in the U.S., success rates vary, to some extent, with the cause of infertility. In 2007, the success rates for women with tubal factor infertility, ovulatory dysfunction, endometriosis, male factor, and unexplained infertility were above average, and those for women with multiple infertility factors, a uterine factor, and diminished ovarian reserve were below average. Whereas these data are useful, it is important to note that criteria for the different diagnoses are not standardized and likely vary among treatment centers.

Women with a previous live birth are more likely to succeed with IVF than nulliparous women. In all age categories, success rates for women having one or more previous live births are modestly higher (2-4%) than for women with no previous live births.³ A previous unsuccessful IVF cycle does not decrease the likelihood for success in subsequent cycles until approximately the fourth IVF cycle.¹⁹⁹ A history of an earlier unsuccessful pregnancy also has no effect on success rates.³

As discussed earlier in the section focused on indications for ART, there is substantial evidence indicating that hydrosalpinges adversely affect IVF outcomes, and that salpingectomy or proximal tubal occlusion before IVF increases the likelihood for achieving a live birth by 2-fold.¹⁶ A study evaluating the cost-effectiveness of preliminary salpingectomy concluded the procedure decreases the average cost per live birth, compared to no treatment.²⁰⁰ Laparoscopic salpingectomy before IVF is generally recommended for women with hydrosalpinges.

The effect of uterine myomas on IVF outcomes depends on their location. Submucosal myomas significantly decrease the likelihood for success, subserosal myomas have no significant impact, and the effect of intramural myomas is unclear. Overall, studies examining the effect of submucosal myomas on IVF outcomes indicate they decrease clinical pregnancy rates and delivery rates by approximately 70%,^{201,202,203,204,205,206} and ²⁰⁷ and increase risk for miscarriage by more than 3-fold.^206,207 A 2009 systematic review of studies examining outcomes after submucosal myomectomy concluded that clinical pregnancy rates achieved with IVF were 2-fold higher after surgery than in women with submucous myomas in situ, and comparable to those observed in women without myomas.²⁰⁷ Results of individual studies examining the effects of intramural myomas on IVF outcome are inconsistent, with some observing an adverse effect,^{208,209,210,211} and ²¹² and others not.^{206,213,214,215,216,217} and ²¹⁸ Although systematic reviews have concluded that intramural myomas have significant negative impact on implantation rates and live birth rates,^204,205,219 there is no compelling evidence that their removal improves outcomes.²²⁰

The first birth resulting from IVF derived from a single oocyte collected in a natural ovulatory cycle.² Compared to stimulated IVF cycles, natural cycle IVF offers a number of attractive advantages. Natural cycle IVF involves only monitoring the spontaneous cycle and retrieving a single oocyte before the midcycle LH surge occurs. It is physically less demanding, requires little or no medication, decreases costs by 75-80%,^234,235 and all but eliminates risks for multiple pregnancy and ovarian hyperstimulation syndrome (OHSS). The chief disadvantages of natural cycle IVF are high cancellation rates due to premature LH surges and ovulation, and the comparatively low success rate, which is approximately 7%.²³⁶

When oocyte retrieval is based on detection of the midcycle rise in LH, careful and frequent monitoring is required and procedures are difficult to schedule efficiently. Alternatively, exogenous human chorionic gonadotropin (hCG) can be administered when the lead follicle reaches a size consistent with maturity, thereby better defining the optimum time for oocyte retrieval.²³⁵ Adjuvant treatment with a GnRH antagonist also can be used to prevent a premature LH surge, but requires “add-back” treatment with exogenous FSH, and success rates are still quite low, ranging up to 14% per cycle in non-randomized trials.^237,238,239 and ²⁴⁰ In one large cohort study involving 844 treatment cycles in 350 good prognosis patients, the cancellation rate was 13%, the pregnancy rate was 8% per cycle and the cumulative pregnancy rate after three “modified natural IVF cycles” was 21%.²⁴¹ In a cohort of infertile couples with male factor infertility, success rates in modified natural cycles have reached as high as 13% per cycle, with a cumulative pregnancy rate of 44% after six treatment cycles.²⁴²

Clomiphene citrate was the first method of ovarian stimulation used in IVF,^243,244 but now has been almost entirely replaced by more effective stimulation regimens using human menopausal gonadotropins (hMG) or FSH, in combination with a GnRH agonist or antagonist.²⁴⁵

Clomiphene (100 mg daily) usually is administered for 5-8 days, beginning on cycle day 3, and induces development of two or more follicles in most normally ovulating women,^246,247 and ²⁴⁸ although egg yields (1-3) are only slightly greater than in unstimulated cycles and substantially lower than in cycles stimulated with exogenous gonadotropins.^248,249 and ²⁵⁰ Cycle cancellation rates are somewhat lower than in natural cycles and the numbers of oocytes retrieved, embryos transferred, and pregnancy rates are greater. As in natural cycles, exogenous hCG is administered when the lead follicle reaches mature size and a GnRH antagonist can be used to prevent a premature endogenous LH surge.

Sequential treatment with clomiphene (100 mg daily for 5 days) and modest doses of exogenous gonadotropins (150-225 IU daily beginning on the last day of clomiphene treatment or the day after) stimulates multifollicular development more effectively than treatment with
clomiphene alone.^251,252 and ²⁵³ Drug costs and monitoring requirements are moderately higher, but still substantially less than in standard stimulation regimens involving higher dose gonadotropin treatment after down-regulation with a long-acting GnRH agonist (described below).^254,255 In one comparative trial, higher cancellation rates and lower pregnancy rates were observed in sequential clomiphene/gonadotropin cycles.²⁵⁵ In another, the sequential stimulation regimen yielded fewer oocytes and embryos, but pregnancy rates were similar and the risks of OHSS were lower.²⁵⁴ Adding a GnRH antagonist to the treatment regimen can prevent premature LH surges and improve outcomes, but also increases costs. In a randomized trial, sequential clomiphene/gonadotropin stimulation and GnRH antagonist treatment yielded a pregnancy rate comparable to that achieved with a more aggressive standard treatment protocol,²⁵⁶ confirming the results of two earlier retrospective studies,^{257, 258} but contrasting with those of another observing lower pregnancy rates.²⁵⁹

The introduction of long-acting GnRH agonists in the late 1980s revolutionized the approach to ovarian stimulation in ART by providing the means to suppress endogenous pituitary gonadotropin secretion and thereby prevent a premature LH surge during exogenous gonadotropin stimulation. Adjuvant treatment with a GnRH agonist eliminated the need for frequent serum LH measurements and assuaged fears of premature luteinization which previously had required cancellation of approximately 20% of all IVF cycles before oocyte retrieval.^260,261 and ²⁶² Because fewer than 2% of cycles are complicated by a premature LH surge after down-regulation with a GnRH agonist,²⁶³ stimulation could continue until follicles were larger and more mature. Numerous clinical trials subsequently demonstrated that egg yields and pregnancy rates were significantly higher than in cycles stimulated with exogenous gonadotropins alone.^264,265 Moreover, GnRH agonist treatment offered the welcome additional advantage of scheduling flexibility, allowing programs to coordinate cycle starts for groups of women simply by varying the duration of GnRH agonist suppression. Not surprisingly, the “long protocol” quickly became the preferred ovarian stimulation regimen for all forms of ART. Its only disadvantages are that GnRH agonist treatment sometimes blunts the response to gonadotropin stimulation and increases the dose and duration of gondotropin therapy required to stimulate follicular development. The combined costs of the additional gonadotropins and the agonist itself increase the total cost of treatment substantially. Nevertheless, because GnRH agonists have more advantages than disadvantages, the long protocol became and has remained the standard ovarian stimulation regimen in IVF cycles.

In the typical cycle, GnRH agonist treatment begins during the midluteal phase, approximately 1 week after ovulation, at a time when endogenous gonadotropin levels are at or near their nadir and the acute release of stored pituitary gonadotropins in response to the agonist, known as the “flare” effect, is least likely to stimulate a new wave of follicular development.^266,267 GnRH agonist treatment can be scheduled to begin on cycle day 21 (assuming a normal cycle of approximately 28 days duration), but most prefer to first con-firm that ovulation has occurred by measuring the serum progesterone concentration. In women who do not cycle predictably, oral contraceptives (OC) can be used to control the onset of menses, starting GnRH agonist treatment 1 week before their discontinuation. In the U.S., leuprolide acetate (administered by s.c. injection) is the most commonly used GnRH agonist. In Europe and elsewhere, buserelin acetate (administered by s.c. injection or intranasal spray) and triptorelin (administered subcutaneously) are more common²⁶⁸; all work equally well. For leuprolide, the usual treatment regimen begins with 1.0 mg daily for approximately 10 days or until onset of menses or gonadotropin stimulation, decreasing to
0.5 mg daily thereafter until hCG is administered. A single dose of a longer-acting depot form of GnRH agonist (leuprolide, goserelin) offers greater convenience, but evidence indicates the total dose and duration of gonadotropin stimulation required are increased significantly when depot forms of the agonists are used.²⁶⁹

Gonadotropin stimulation begins after confirming that effective pituitary down-regulation has been achieved (serum estradiol level <30-40 pg/mL, no follicles >10 mm in diameter). Some women require longer durations of treatment to achieve suppression or may develop an ovarian cyst.²⁶⁰ The significance of an ovarian cyst has been controversial. Whereas some investigators have observed that baseline cysts are associated with a poorer response to gonadotropin stimulation, decreased numbers of oocytes and embryos, and lower overall IVF success rates,^270,271 and ²⁷² others have not.^{273,274,275,276} and ²⁷⁷ Overall, the weight of available evidence suggests that women who develop cysts or require longer durations of GnRH agonist treatment to achieve suppression are more likely to respond poorly to gonadotropin stimulation and less likely to achieve pregnancy. Cyst aspiration immediately before stimulation does not appear to adversely affect response²⁷⁸ and may even improve response in the aspirated ovary,²⁷⁹ but probably is not warranted in women with a normal contralateral ovary.

The initial dose of exogenous gonadotropins must be tailored to the needs of the individual woman. Typical starting doses range between 150 and 300 IU of urinary FSH (uFSH), recombinant FSH (rFSH), or urinary menotropins (hMG) daily, depending on age, the results of ovarian reserve testing, and the response observed in any previous stimulation cycles. Either a “step-up” (beginning with a low dose, increasing as necessary based on response) or a “step-down” (beginning with a higher dose, decreasing as necessary based on response) can be used, but the latter approach is generally preferred. All contemporary gonadotropin preparations, including hCG, can be administered subcutaneously.

Numerous clinical trials and meta-analyses have compared outcomes in cycles stimulated with uFSH, rFSH, or hMG, with or without GnRH agonist pretreatment, concluding that there is no compelling evidence to indicate the superiority of one gonadotropin preparation over others.^280,281,282 and ²⁸³ However, a 2008 systematic review including seven trials comparing outcomes in cycles stimulated with rFSH or hMG, involving 2,159 patients, observed a significant increase in live birth rate with hMG (RR=1.18, CI=1.02-1.38); the pooled risk difference for live birth was 4%.²⁸⁴

Recombinant DNA technology has been used to develop a new longer-acting form of rFSH. Corifollitropin alpha is the product of a chimeric gene containing the sequences of the FSH-β subunit and the C-terminal peptide of the hCG-β subunit, which bears four O-linked glycosylation sites, and has a half-life three times longer than standard rFSH (95 vs. 32 hours).^285,286 A single dose (100 μg for women <60 kg, 150 μg for those >60 kg) can induce and sustain multi-follicular growth for a week in women receiving ovarian stimulation for IVF. Corifollitropin has shown considerable promise in phase II trials, currently is being evaluated for safety and efficacy in large phase III trials, and is the first step
towards a new generation of recombinant gonadotropins.²⁸⁷ The first live birth resulting from treatment with corifollitropin was reported in 2003.²⁸⁸

The low levels of LH secretion remaining after down-regulation with a GnRH agonist are sufficient to support normal follicular development in most women stimulated with uFSH or rFSH alone,²⁸⁹ because only about 1% of LH receptors must be occupied to sustain normal levels of steroidogenesis.²⁹⁰ However, in some women treated only with FSH, LH levels are markedly suppressed (<1 IU/L) and may be inadequate.^291,292 In such cycles, the dose and duration of gonadotropin stimulation required is higher, peak estradiol levels are lower, and the numbers of oocytes and embryos may be reduced.^293,294 Extremely low LH levels also may adversely affect fertilization, implantation, and pregnancy rates,^{295,296,297,298} and ²⁹⁹ and have been associated with a higher incidence of biochemical pregnancy and early pregnancy loss.^300,301 A 2007 systematic review and meta-analysis including 11 trials comparing stimulation with rFSH alone or in combination with recombinant LH (rLH) after GnRH agonist-induced down-regulation in IVF and ICSI cycles observed no significant differences in clinical or ongoing pregnancy rates.³⁰² However, in three trials including only poor responders, the pregnancy rate was higher in those receiving combined stimulation with rFSH and rLH.³⁰² In sum, the evidence indicates there may be a subgroup of women who could benefit from supplemental rLH or hMG during ovarian stimulation. In the absence of any reliable method for identifying such women, and in light of recent evidence suggesting that use of hMG may increase live birth rates,²⁸⁴ many clinicians favor combined stimulation with FSH and hMG over stimulation with FSH alone.

The response to stimulation is monitored with serial measurements of serum estradiol and transvaginal ultrasonography. The first estradiol level usually is obtained after 3-5 days of stimulation to determine whether the chosen dose of gonadotropins requires adjustment. Thereafter, serum estradiol concentrations and sonography are obtained every 1-3 days, based on the quality of the response and the need to evaluate the impact of any further adjustments in the dose of gonadotropin treatment. In general, stimulation continues until at least two follicles measure 17-18 mm in mean diameter, when others typically measure 14-16 mm and the serum estradiol concentration reflects the overall size and maturity of the cohort. Most women require a total of 7-12 days of stimulation. It is important to emphasize that these parameters only approximate the goals of stimulation. In clinical practice, follicle measurements vary among observers and estradiol assays vary in their performance characteristics. Ultimately, each program must empirically establish its own thresholds, based on its own experience.

The endometrium is monitored during stimulation by measuring the endometrial thickness or “stripe” (the sum thickness of the two layers, measured in the mid-sagittal plane). Numerous studies have examined the prognostic value of endometrial thickness and pattern in ART cycles, but the issue remains unsettled. Many have suggested that results are best when endometrial thickness measures 8-9 mm or greater or appears “trilaminar,” and poor when the endometrium is less than 6-7 mm in thickness or appears homogeneous on the day of hCG administration.^{303,304,305,306,307} and ³⁰⁸ However, numerous others have failed to observe any clear correlation between endometrial thickness or appearance and outcomes.^{309,310,311,312,313} and ³¹⁴ Some have suggested that excessive endometrial growth (>14 mm) also is a poor prognostic indicator,^305,315 but that too has been refuted.^316,317 Overall, although measurements of endometrial growth are routine, their utility remains unclear. Consequently, changes in stimulation regimens and cycle cancellations based on endometrial thickness or appearance alone are difficult to justify.³¹⁸

When the cohort of ovarian follicles reaches maturity, hCG (5,000-10,000 IU) is administered to stimulate the final stages of follicular development. The equivalent dose of the recombinant form of hCG now available is 250 μg.^319,320 A 2005 systematic review including seven trials comparing recombinant and urinary hCG observed no differences in clinical outcomes.³²¹ The predictive value of the serum progesterone concentration
on the day of hCG administration has been debated vigorously, with some arguing that pregnancy rates were substantially lower when levels exceeded 0.9-1.0 ng/mL,^{322,323,324,325,326} and ³²⁷ and others refuting the contention.^{328,329,330,331} and ³³² It is now clear that mildly increased progesterone levels are relatively common in women who respond well to gonadotropin stimulation and are a poor prognostic indicator only in poor responders.³³³ Whereas it may be tempting to delay hCG administration in poor responders to afford smaller follicles the opportunity to further mature, the strategy is not likely to succeed and may be detrimental.

Approximately 7-18% of stimulation cycles are cancelled before oocyte retrieval, most for lack of adequate response, and some for excessive response.³ When the ovaries become grossly enlarged, containing large numbers of follicles of all sizes, and serum estradiol concentrations are markedly elevated (>5,000 pg/mL), the risk for OHSS increases substantially.^334,335 and ³³⁶ Management options in “high responders” include all of the following:

Canceling the cycle and starting anew using a more conservative stimulation regimen may ultimately decrease overall costs and maximize the chances for success.³³⁷ The prognosis for high responders in subsequent cycles is generally very good. Dual suppression with both an OC (one pill daily for 21 days or more) and a GnRH agonist (leuprolide 1.0 mg s.c. daily, beginning 1 week before discontinuation of contraceptive treatment) can attenuate the response to subsequent lower dose gonadotropin stimulation.³³⁸ Coasting allows larger follicles to continue growing but withdraws support from small and intermediate-sized follicles.^339,340 Although approximately 20-30% of coasted cycles are ultimately cancelled, the strategy can help to reduce the risks for developing severe OHSS and avoid cancellation.^339,341 Proceeding to oocyte retrieval and fertilization and freezing all embryos can salvage the cycle but avoid the greater risks of serious or prolonged OHSS observed in conception cycles.^342,343 Delaying transfer until after symptoms abate and freezing all embryos when they persist is another option.³⁴⁴

The challenges presented by “poor responders” are far greater. Poor responders include women who develop few follicles (<3-5) despite high doses of gonadotropin stimulation or have relatively low peak estradiol levels (<500-1,000 pg.mL); there are no consensus criteria defining a poor responder. The prognosis is relatively poor for such women^345,346 and ³⁴⁷ and the important decision centers on whether to attempt stimulation again using a different or more aggressive treatment regimen.³⁴⁸ Some of the more commonly employed options include the following:

Higher doses of gonadotropin stimulation may generate a somewhat more vigorous follicular response, but doses greater than 450 IU daily generally have little or no additional benefit.^349,350,351 and ³⁵² Decreasing the dose or discontinuing GnRH agonist treatment early
or completely may help to improve the quality of response.^{353,354,355,356} and ³⁵⁷ A standard or microdose GnRH agonist “flare protocol” (described below) may stimulate an improved response in some poor responders.^358,359 and ³⁶⁰ Stimulation regimens employing a GnRH antagonist instead of a long-acting agonist eliminate any suppressive effects of the agonist altogether.³⁶¹ Other strategies have included efforts to increase androgen concentrations by treatment with dehydroepiandrosterone (DHEA)³⁶² or an aromatase inhibitor,³⁶³ and the addition of growth hormone to the stimulation regimen.³⁶⁴ A 2009 systematic review and meta- analysis of randomized trials comparing different stimulation regimens in poor responders found insufficient evidence to support the routine use of any particular intervention.³⁶⁵ A 2010 systematic review including 10 trials involving eight different comparison groups reached the same conclusion.³⁴⁵

The “short” or “flare” protocol is an alternative stimulation regimen designed to exploit both the brief initial agonistic phase of response to a GnRH agonist and the suppression that results from longer-term treatment.^359,366 In a typical standard short protocol, leuprolide acetate (1.0 mg daily) is administered on cycle days 2-4, continuing thereafter at a reduced dose (0.5 mg daily), and gonadotropin stimulation (225-450 IU daily) begins on cycle day 3. Later adjustments in the dose of gonadotropin stimulation, if needed, are based on response and indications for hCG administration are the same as in the long protocol (described above).

An early meta-analysis including seven clinical trials comparing the short and long GnRH agonist treatment regimens determined that the two protocols yielded similar cancellation and pregnancy rates.²⁶⁴ A 2000 systematic review including 22 trials concluded that pregnancy rates achieved with the long protocol were superior to those using the flare regimen (OR=1.27, CI=1.04-1.56) overall,²⁶⁵ but the analysis did not control for diagnosis and other prognostic factors and results may not apply to all women, or to poor responders in particular. Whereas some have observed improved follicular response and lower cycle cancellation rates in poor responders treated with a flare protocol, pregnancy and live birth rates remained low.^367,368 Decreased scheduling flexibility is a distinct disadvantage of the flare protocol, unless the onset of menses is controlled by preliminary treatment with an OC. The regimen also can result in a significant increase in serum progesterone and androgen levels, presumably resulting from late corpus luteum rescue,^369,370 which may adversely affect oocyte quality and fertilization and pregnancy rates.³⁷¹

The “OC microdose GnRH agonist flare” stimulation regimen is a variation of the standard short protocol involving 14-21 days of preliminary ovarian suppression with an OC (one pill daily), followed by microdose leuprolide treatment (40 μg twice daily) beginning 3 days after discontinuation of OC treatment, and high-dose gonadotropin stimulation (300-450 IU daily) starting on day 3 of leuprolide therapy. Indications for later gonadotropin dose adjustments and hCG administration are the same as in other stimulation regimens. Its primary advantage over the standard short protocol is that it does not induce any increases in serum progesterone or androgen concentrations,³⁶⁰ possibly because the doses of GnRH agonist administered are much lower, but likely also because preliminary OC treatment all but eliminates the possibility there may be a corpus luteum left to respond.^372,373 The OC-microdose GnRH agonist flare protocol may be useful in previous poor responders, in whom it can stimulate increased endogenous FSH release and may yield lower cancellation rates and higher peak serum estradiol levels, transfer rates, and pregnancy rates.^360,374,375

The introduction of GnRH antagonists into clinical practice provided another option for ovarian stimulation in ART. In contrast to the long-acting agonists, which first stimulate and later inhibit pituitary gonadotropin secretion by desensitizing gonadotropes to GnRH via receptor down-regulation, the antagonists block the GnRH receptor in a dose- dependent competitive fashion and have no similar flare effect^376,377; gonadotropin suppression is almost immediate.

GnRH antagonists offer several potential advantages over agonists. First, the duration of treatment for an antagonist is substantially shorter than for an agonist. Since its only purpose is to prevent a premature endogenous LH surge and its effects are immediate, antagonist treatment can be postponed until later in follicular development (after 5-6 days of gonadotropin stimulation), after estradiol levels are already elevated, thereby eliminating the estrogen deficiency symptoms that may emerge in women treated with an agonist.³⁷⁸ Second, because any suppressive effects that agonists may exert on the ovarian response to gonadotropin stimulation also are eliminated, the total dose and duration of gonadotropin stimulation required is decreased.^378,379 For the same reason, GnRH antagonist stimulation protocols may benefit women who are poor responders when treated with a standard long protocol.^378,380 Third, by eliminating the flare effect of agonists, GnRH antagonists avoid the risk of stimulating development of a follicular cyst. Finally, the risk of severe OHSS associated with use of antagonists also appears lower than with agonists.^381,382 and ³⁸³

GnRH antagonists have some potential disadvantages. When administered in small daily doses, strict compliance with the prescribed treatment regimen is essential.³⁷⁸ Antagonists suppress endogenous gonadotropin secretion more completely than agonists. Whereas the low levels of LH observed during agonist treatment are usually sufficient to support normal follicular steroidogenesis during stimulation with uFSH or rFSH, the even lower concentrations in women treated with an antagonist may not be. Indeed, serum estradiol levels may plateau or fall when antagonist treatment begins.^299,378,384 Although follicular growth appears unaffected, most prefer to add or substitute a low dose of hMG (75 IU) at the same time if it was not already part of the stimulation regimen. Evidence also suggests that pregnancy rates in antagonist treatment cycles may be modestly lower than in cycles using agonists in the long protocol.³⁸⁵

The two GnRH antagonists available for clinical use, ganirelix and cetrorelix, are equally potent and effective. For both, the minimum effective dose to prevent a premature LH surge is 0.25 mg daily, administered subcutaneously.^299,386 Either can be administered in a series of small daily doses (0.25 mg). The treatment protocol may be fixed and begin after 5-6 days of gonadotropin stimulation,^299,386,387 or tailored to the response of the individual, starting
treatment when the lead follicle reaches approximately 13-14 mm in diameter.^388,389 The individualized treatment regimen generally requires fewer total doses and may yield better overall results.³⁸⁸ Alternatively, a single larger dose of cetrorelix (3.0 mg) will effectively prevent an LH surge for 96 hours. If given on day 6-7 of stimulation, the interval of effective suppression will encompass the day of hCG administration in most women (75-90%); the remainder may receive additional daily doses (0.25 mg) as needed, ending on the day of hCG treatment.^390,391 and ³⁹² The single dose antagonist treatment regimen also can be withheld until the lead follicle reaches 13-14 mm in diameter.³⁹³

A common variation of the antagonist stimulation regimen uses preliminary treatment with an OC to control the onset of menses, typically ending approximately 5 days before the scheduled start, which also may help to synchronize the follicular cohort before stimulation begins. Another variation advocated for poor responders uses micronized estradiol (2 mg twice daily, administered orally, beginning on day 21 of the preceding cycle) to suppress FSH during the late luteal phase for the same purpose, ending on the day before gonadotroins stimulation begins,^393,394 or continuing through the first 3 days of gonadotropin stimulation.³⁴⁷ The improved follicular dynamics observed are similar to those achieved by down-regulation with a GnRH agonist in the long protocol. The rebound increase in endogenous FSH levels that follows the discontinuation of estradiol treatment also may synergize with exogenous gonadotropins to promote multifollicular development.^395,396

Results of a number of early trials comparing a fixed antagonist treatment protocol to the standard long protocol suggested that the two stimulation regimens yielded similar pregnancy rates.^{379,391,397,398} However, a 2006 systematic review and meta-analysis including 27 trials comparing different antagonist stimulation protocols with the long GnRH agonist protocol observed a significantly lower clinical pregnancy rate (OR=0.84, CI=0.72-0.97) and ongoing pregnancy/live birth rate (OR=0.82, CI=0.69-0.98). Overall, the total dose and duration of gonadotropin stimulation required, peak serum estradiol levels, and the number of follicles and oocytes were lower in antagonist cycles.

The explanation for the modestly lower pregnancy rates observed in antagonist treatment cycles is not clear. It is possible, but unlikely, that GnRH antagonists may have adverse effects on oocytes, embryos, or the endometrium.^399,400 It is far more likely that early results reflected inexperience and improved with time and further refinements in the treatment regimen like those described above. Many of the advantages originally envisioned for GnRH antagonists already have been realized. Whether antagonists ultimately will replace agonists and become the standard ovarian stimulation regimen in ART cycles remains to be seen, but their place in the therapeutic arsenal already is firmly established.

Women with polycystic ovary syndrome (PCOS) characteristically exhibit high tonic LH secretion and are predisposed to premature LH surges when treated with standard ovulation induction regimens. Women with PCOS also are at increased risk for developing OHSS when aggressively stimulated with exogenous gonadotropins. Whereas both GnRH agonists and antagonists can suppress elevated circulating LH concentrations, the smaller
follicular cohorts observed in antagonist cycles may help to reduce the risk of OHSS in women with PCOS who tend to be high responders. The use of antagonists, rather than agonists, provides the opportunity to use an agonist instead of hCG to induce final oocyte maturation, thereby possibly further decreasing the risk of OHSS.⁴⁰¹ Whereas a single bolus injection of an agonist (leuprolide 0.5 mg, triptorelin 0.2 mg) triggers a physiologic LH surge that lasts less than 24 hours, hCG levels remain elevated for several days and stimulate markedly higher estradiol and progesterone concentrations.⁴⁰²

The antagonist treatment regimens currently in use have potential disadvantages for women with PCOS. Their tonically elevated LH levels will remain high until antagonist treatment begins. Consequently, LH levels may rise prematurely, particularly if antagonist treatment is withheld until the lead follicle reaches 14 mm or more. Moreover, evidence indicates that increased LH exposure during early follicular development may be detrimental and predispose to lower pregnancy rates.^403,404,405 and ⁴⁰⁶ In theory, pretreatment with an OC might prove quite useful by suppressing LH and androgen levels before stimulation begins, decreasing exposure during early follicular development and the risk of rising LH levels before antagonist treatment starts. Preliminary OC suppression and later antagonist treatment may help to limit the follicular response to gonadotropin stimulation while preserving the option to use an agonist to trigger final oocyte maturation. These considerations simply serve to illustrate that GnRH antagonists are not a panacea and are not necessarily the best choice even for women with PCOS.

Antagonist stimulation protocols are advocated for poor responders, primarily because they avoid the suppressive effects that agonists can have on follicular response and can prevent the premature LH surges observed commonly in women stimulated with gonadotropins alone.⁴⁰⁷ However, evidence is insufficient to indicate they yield results consistently better than other stimulation regimens.^345,365

Up to 20-30% of retrieved oocytes may be immature at the time of retrieval, reflecting the varying size and maturity of follicles in the cohort at the time hCG is administered. An accurate assessment of oocyte maturity is important to the timing of fertilization, even more so when ICSI is to be performed.

Like the LH surge in natural cycles, hCG triggers the resumption of meiosis in primary ooyctes previously arrested at prophase I of the first meiotic division. Oocyte maturity
generally can be judged by the expansion of the cumulus mass, radiance of the corona cells, the size and cohesiveness of granulosa cells, and the shape and color of the oocyte. When the cumulus mass is removed, as it is in preparation for ICSI, the oocyte can be further evaluated according to the presence or absence of the first polar body and germinal vesicle (nuclear membrane).

A mature (metaphase II) oocyte has extruded the first polar body and is in the resting phase of meiosis II. The cumulus cells are typically expanded and luteinized and the corona radiata exhibits a sunburst pattern. A metaphase I oocyte of intermediate maturity has no polar body and denser cumulus cells, but the germinal vesicle and nucleolus have faded. Metaphase I oocytes require additional time in culture before fertilization and must be examined periodically to document extrusion of the first polar body. A prophase I oocyte is grossly immature and exhibits a compact corona containing relatively few cumulus cells and a prominent germinal vesicle and nucleolus; dissolution of the germinal vesicle signals the resumption of meiosis I.

In the past, men with non-obstructive azoospermia were considered sterile and untreatable by any means other than the use of donor sperm. However, testis biopsy specimens in such men often demonstrate sperm,⁴⁷² suggesting low level production of sperm that do not survive epididymal transit to reach the ejaculate.⁴⁷³ Whereas conventional wisdom was that sperm must traverse the male reproductive tract to acquire the ability to fertilize an oocyte, success with ICSI using epididymal or testicular sperm has demonstrated otherwise. Even grossly immature sperm (round spermatid nuclear injection; ROSNI) have been used to achieve fertilization, albeit with limited success.⁴⁷⁴

It is important to emphasize again that genetic evaluation and counseling are indicated for men with severe seminal abnormalities before their sperm are used for ICSI. Men with congenital bilateral absence of the vas deferens (CBAVD) or less severe forms of vasal aplasia, and their female partners, should be screened for cystic fibrosis gene mutations
before any attempts at pregnancy via ART to determine the risk for transmitting cystic fibrosis or CBAVD to offspring.^475,476 and ⁴⁷⁷ Men with non-obstructive azoospermia or severe oligospermia (less than 5 million/mL) should be offered karyotyping and screening for Y chromosome microdeletions.⁴⁷⁷