In Vitro Fertilization

The concept of IVF as a treatment for infertility is straightforward: obtain eggs from the ovaries, mix them with sperm in a dish containing culture medium, and transfer the eggs back to the woman after fertilization has occurred. However, this technique took more than 100 years to develop.

In 1891, the first successful transfer of an embryo from one animal to another that resulted in birth was reported, using rabbits of two different strains. However, these embryos were obtained from eggs fertilized in vivo. Further progress toward the goal of IVF was slowed because of limited understanding of the maturation of eggs and sperm required to achieve fertilization and embryo development. In 1959, successful IVF was reported using rabbits.³

The first human birth to result from IVF was achieved in England in 1978.⁴ John and Lesley Brown had 9 years of infertility secondary to bilateral fallopian tube obstruction. Dr. Patrick Steptoe surgically retrieved a single mature oocyte from one of Lesley’s ovaries during a natural cycle. Dr. Robert G. Edwards combined John’s sperm with the oocyte in the laboratory and the resulting embryo was placed into Lesley’s uterus a few days later. On July 25, 1978, Louise Joy Brown was delivered by cesarean section at approximately 37 weeks’ gestation, and weighed 5 pounds 12 ounces.

Although this first human IVF birth employed a surgical procedure to retrieve the oocyte produced in a natural menstrual cycle, most IVF today is performed after ovarian stimulation so that multiple eggs can be retrieved transvaginally with a sonographically guided needle. Currently more than 99% of ART procedures in the United States employ IVF and transcervical embryo transfer.

Tubal Adhesions

Tubal adhesions account for 30% to 40% of cases of female infertility. Tubal damage or adnexal adhesion typically arise from salpingitis, appendicitis, endometriosis, or previous pelvic surgery. Laparoscopic surgery to remove adhesions and open tubes can yield pregnancy rates of more than 50% in women with relatively mild disease.¹¹ In women with severe tubal damage or extensive dense adhesions, surgery is unlikely to result in pregnancy. IVF is indicated in these patients and in women who have failed to conceive after infertility surgery.

Hydrosalpinges

Tubal surgery is also indicated in women with hydrosalpinges who are contemplating IVF. Hydrosalpinges are associated with decreased pregnancy and live birth rates after IVF.¹² Although the pathophysiology of this relationship is not completely understood, bilateral salpingectomy improves the success of subsequent IVF, especially in women with bilateral hydrosalpinges.

Tubal Ligation

Up to 25% of women who have undergone bilateral tubal sterilization will come to regret their decision and wish to have more children. Risk factors for tubal ligation regret include young age at time of sterilization, relationship with a new partner, loss of a child, and having been sterilized immediately postpartum or postabortion. Up to 5% will seek tubal anastomosis, which is a highly effective treatment option in properly selected patients.¹³ Success rates depend on the site of the anastomosis, the length of residual tube, the age of the patient, and the presence of other infertility causes.

In vitro fertilization is an effective alternative to surgery for women who have undergone a tubal ligation. This is particularly true for women without adequate residual tubal length, those who are poor surgical candidates, and those who would prefer to avoid surgery. IVF is also an option in women who have undergone tubal anastomosis, but who do not achieve pregnancy within 12 months.

The choice between IVF and tubal surgery is usually made based on relative success rates and the cost to the patients. If either IVF or surgery are covered completely or in part by insurance, the covered choice will usually see greater utilization.

Antisperm Antibodies

Antisperm antibodies are a relatively uncommon and difficult to treat cause of infertility. IVF with ICSI has been found to be an effective treatment for women with antisperm antibodies, even in patients with a higher density of such antibodies.²⁶ In one study, patients with antisperm antibodies had a 32% clinical pregnancy rate after IVF with ICSI.²⁷ Because antisperm antibodies are relatively uncommon, it does not appear that routine screening for antisperm antibodies before IVF is cost-effective.²⁸

Endometrial Polyps

Endometrial polyps, benign localized overgrowths of endometrial tissue of uncertain etiology, are present in up to 10% of asymptomatic premenopausal women over age 30.³⁴ It is assumed that endometrial polyps decrease fertility. For this reason, all large polyps are removed prior to IVF.

For smaller polyps that first appear during ovarian stimulation for IVF, optimal management remains to be determined. The problem is that removal requires interruption of the cycle and delay of IVF. A series of 49 women with endometrial polyps less than 2 cm in diameter who underwent IVF without removing the polyp found pregnancy and miscarriage rates were comparable to the general pregnancy rate of their IVF clinic population, although there was a trend toward a higher miscarriage rate.³⁵ At the present time, the decision to continue despite the polyp or stop the cycle and remove the polyp must be made on a case-by-case basis.

Hydrosalpinges

The presence of hydrosalpinges is well documented to decrease the success rate for IVF.³⁶ Whether pregnancy rates can be improved by salpingectomy remains less certain. A recent meta-analysis of three randomized, controlled trials indicated that the chance of live birth with IVF was doubled by pretreatment salpingectomy.¹² This effect seems to be most apparent in women with bilateral hydrosalpinges and with hydrosalpinges that are sonographically visible.³⁷ Drainage of hydrosalpinges at the time of egg retrieval can be performed, but it is uncertain if this improves IVF pregnancy rates. At present, many IVF programs offer patients with a sonographically visible hydrosalpinx the option of pretreatment salpingectomy.

Age

It has long been known that reproductive capacity declines with increasing age.³⁸ The age-dependent decline in female fertility can be partly attributed to the fact that women have a finite and nonreplenishable number of germ cells. The peak number of germ cells occurs at midgestation during fetal life and declines continuously thereafter, with an accelerated loss of oocytes between ages 37 and 38.³⁹ Diminished ovarian reserve is a term used to indicate decline in reproductive capacity associated with ovarian follicular depletion and diminished oocyte quality.

The success of IVF declines with age in a similar fashion (Fig. 38-1). For women undergoing IVF, diminished ovarian reserve is associated with poor ovarian response to gonadotropins, cycle cancellation, and lower chances of conception. Despite the known correlation with chronologic age and ovarian reserve, there exists a tremendous amount of variability in patients. As a result, multiple markers of ovarian reserve have been sought to supplement age as a predictor of ovarian response to stimulation in IVF. Timely recognition of reduced ovarian reserve is important for counseling patients prior to IVF.

Figure 38-1 Live birth rates per cycle start by age of the woman, for cycles performed in the United States in the year 2003

(From Centers for Disease Control and Prevention: 2003 Assisted Reproductive Technology Success Rates. Available at: http://www.cdc.gov/reproductivehealth/ART/index.htm. Accessed 17 September 2005.

Basal FSH

Measurement of serum follicle-stimulating hormone (FSH) in the early follicular phase of the menstrual cycle (day 3) is widely employed to assess the potential responsiveness of the ovaries to stimulation and can predict to some degree subsequent IVF pregnancy rates. This measurement of “basal” FSH is relatively inexpensive and easily obtained. However, the ideal way to use basal FSH levels for counseling prior to IVF remains controversial.

A direct correlation of the basal FSH and IVF outcome measurements was found in a study of 441 patients undergoing 758 consecutive IVF cycles.⁴⁰ Patients with basal FSH levels greater than 25 mIU/mL had only a 3.6% ongoing pregnancy rate, whereas those with basal FSH levels less than 15 mIU/mL had a 17% pregnancy rate. They also noted that fewer follicles were aspirated, fewer oocytes were obtained, and fewer embryos were available for transfer in the high FSH group compared to the low FSH group.

A recent meta-analysis of 21 studies indicated that basal FSH levels were moderately predictive for poor response but poorly predictive for pregnancy. In women over age 40, neither the basal nor the stimulated (clomiphene citrate challenge test [CCCT]) FSH level were able to predict pregnancy rates with IVF. However, no patient with a basal FSH level greater than 11.1 mIU/mL or a CCCT FSH greater than 13.5 mIU/mL (i.e., day 10 serum FSH) carried a pregnancy past 20 weeks.

Many IVF programs use a cut-off value for the basal FSH to determine when to cancel an IVF cycle. A study of 230 consecutive IVF cycles using GnRH antagonists found that the 97.5th percentile predictive value of basal FSH resulting in pregnancy was 10 mIU/mL. Patients should be counseled that these data suggest that pregnancy with IVF is unlikely in patients with basal FSH levels above this cutoff.

However, it is important to note that FSH has a wide fluctuation both between cycles in the same patients as well as between particular hormonal assays. It is therefore critical for each program to evaluate its own normative data for this assay as well as consider repetition of borderline hormonal values in patients.

Day 3 Estradiol

Serum estradiol levels on day 3 of the menstrual cycle have also been found to be predictive of subsequent IVF pregnancy rates. One study found that the ongoing pregnancy rates for patients with day 3 estradiol levels less than 30 pg/mL were significantly higher than for patients with estradiol levels between 31 and 75 pg/mL.⁴¹ No pregnancies occurred in patients with day 3 estradiol levels greater than 75 pg/mL. Another study found that the 97.5th percentile predictive value for day 3 estradiol for pregnancy was 56 pg/mL.

Basal elevations of estradiol are an independent marker of poor ovarian response to stimulation even in cycles without a rise in FSH. This is presumably because high circulating levels of serum estradiol suppress FSH levels. This hypothesis was confirmed in a study of 225 patients, where no pregnancies occurred after IVF with day 3 estradiol greater than 100 pg/mL, despite FSH levels less than 15 mIU/mL in all patients.⁴² In a study of 2476 IVF patients who had normal day 3 FSH levels, patients with day 3 estradiol levels either less than 20 pg/mL or greater than 80 pg/mL had an increased cancellation rate.⁴³ However, estradiol levels are no longer predictive of IVF pregnancy rates once the patients had more than three maturing follicles.

Inhibin B

Elevation of FSH occurs in part as a result of diminished inhibin secretion by developing follicles. It follows that inhibin levels may also predict IVF success. Inhibin is a heterodimer that is secreted in two forms: inhibin A and inhibin B. In one IVF study, women with a day 3 serum inhibin B concentration less than 45 pg/mL had lower estradiol levels, fewer oocytes retrieved, a higher cycle cancellation rate, and a lower clinical pregnancy rate than women whose inhibin B concentration was at least 45 pg/mL.⁴⁴ However, another study of 120 women undergoing IVF found that inhibin B levels did not improve the prediction of IVF success over using patient age and basal FSH level.

Antimüllerian Hormone

Antimüllerian hormone is produced by developing granulosa cells. A recent study of 130 women undergoing their first IVF cycle found that low antimüullerian hormone levels were associated with fewer eggs retrieved and increased chance of cycle cancellation.⁴⁵ According to another study, basal serum antimüllerian hormone level may be a better predictor of IVF outcome than FSH, estradiol, or inhibin levels.⁴⁶

Antral Follicle Count

The number of small follicles visible by transvaginal ultrasonography early in the follicular phase is another means used to predict ovarian response in IVF. A prospective IVF study of 130 women younger than age 45 indicated that the number of antral follicles on cycle day 3 provided better prognostic information regarding poor ovarian response during hormone stimulation for IVF than age or basal FSH, estradiol, or inhibin B levels. Another IVF study of 120 women also found that a single antral follicle count was predictive of poor ovarian response, although the predictive accuracy could be increased by repeating the count on a subsequent cycle and using the higher of the two counts.⁴⁷ The exact number of antral follicles that predicts outcome is still uncertain.

Clomiphene Citrate Challenge Test

The CCCT decribed by Navot and colleagues is a method to dynamically elucidate the ovarian response.⁴⁸ The test is performed by measuring basal FSH on day 3 (day 2 is acceptable), administering clomiphene citrate (100 mg, days 5 to 9), and then remeasuring a FSH on day 10 (days 9 and 11 are also acceptable). An abnormal test was originally defined as an FSH level after clomiphene citrate more than 2 standard deviations from the basal level. Many clinicians define an abnormal CCCT as an FSH value greater than 12 mIU/mL on either cycle day 3 or 10.

The CCCT has been shown to have a sensitivity of 43% and a specificity of 76%, using IVF cycle cancellation as an endpoint in a study of 198 women.⁴⁹ Positive and negative predictive values were 37% and 80%, respectively. The estradiol levels during ovarian stimulation, the number of retrieved oocytes, and the rate of transfer cycles were significantly lower in patients with an abnormal CCCT. Forty-three percent of the abnormal test results were abnormal only on their elevation of day 10 or 11 FSH and not on their basal FSH level. However, the rate of pregnancies per started cycle did not show a statistically significant difference, which was attributed to the low numbers of patients.

A recent meta-analysis of a total of 1352 patients from 12 studies on basal FSH and 7 studies on CCCT found that basal FSH had a sensitivity of 6.6% and a specificity of 99.6% for identifying inability to achieve pregnancy in an IVF cycle, whereas CCCT sensitivity was 25.9% and specificity was 98.1%.⁵⁰ This study suggests that basal FSH and CCCT are similar in the ability to predict a clinical pregnancy and that, although a normal test is not helpful, an abnormal test is highly predictive that pregnancy will not occur with IVF. Based on this, the authors recommended that basal FSH be used rather than CCCT because of its simplicity and lower cost.

Combining Screening Markers

It is apparent that there are clear limitations to each of the proposed markers for ovarian response to stimulation. Therefore, attempts have been made to combine some of these markers to improve the predictive ability compared to a single marker alone. When basal FSH and estradiol are combined for patients undergoing IVF cycles, one study reported no pregnancies in patients with basal FSH levels greater than 17 mIU/mL and basal estradiol level greater than 45 pg/mL.⁴¹ Another study of 74 IVF patients with basal FSH levels less than 15 mIU/mL examined the ratio of basal FSH to LH.⁵¹ This study found that an elevated FSH-to-LH ratio greater than 3.6 correlated with a lower day 8 estradiol, lower peak estradiol, and fewer follicles greater than 15 mm in size

A randomized study of 110 patients evaluated multiple markers for ovarian reserve, including the exogenous FSH ovarian reserve test. This test involves the administration of 300 IU of rFSH (Gonal-F) on cycle day 3 with serum measurements of FSH, estradiol, and inhibin B before and 24 hours after administration. They concluded that the exogenous FSH ovarian reserve test was better than the CCCT at predicting the number of large follicles resulting from controlled ovarian hyperstimulation.⁵²

Conclusion

At present, basal FSH screening remains the screening tool of choice in many IVF centers in the United States. Issues examining different assay standards, variation by laboratory, and cycle-to-cycle variability of any biomarker of ovarian reserve will have to be addressed. Perhaps the most important aspect of ovarian reserve testing is the way in which this information is used to counsel the infertile couple. None of these tests have absolute sensitivity or specificity; however, the counseling of patients with abnormal tests to consider other options of treatment such as oocyte donation or adoption remains reasonable. Moreover, none of the data presented here indicates the ability of a single or even multiple tests to accurately exclude patients from treatment. The ultimate decision about the treatment course planned should be an individualized process between the infertile woman or couple and the physician.