MONITORING IN VITRO FERTILIZATION (IVF) CYCLES
E. Tawfik, A. Mastrorilli and A. Campana
Infertility and Gynecologic Endocrinology Clinic,
Department of Obstetrics and Gynecology,
University Cantonal Hospital, 1211 Geneva 14, Switzerland
The term " monitoring " means " close continuous observation ", so when we refer to monitoring an in vitro fertilization and embryo transfer (IVF-ET) cycle we mean close observation not only of a patient’s initial parameters and her own ovarian response to ovulation induction, but also events after completion of the therapy.
Why monitor the patient? Monitoring serves two purposes. On the one hand, it helps the physician to choose the most suitable protocol, or to modify the dose and/or the approach for the protocol being applied in an attempt to obtain the best possible outcome and avoid complications of therapy or of the procedure as a whole. On the other hand, monitoring our patients adds to the common pool of information which increases our knowledge and understanding of human reproduction.
In our opinion, monitoring IVF patients begins with the initial infertility work-up, and continues until after delivery. This chapter is concerned, however, with the time period of an IVF cycle which starts just before induction therapy and ends either by the establishment of a biochemical pregnancy, or failure of implantation. Before discussing different methods of monitoring, it is better to divide monitoring into three stages: before starting induction therapy, the period of induction and the period that follows completion of therapy.
Before induction therapy
During this period, one has to think about which protocol should be offered to the patient. This depends on many factors such as the patient’s endocrine profile and general health, her age and financial situation, and the physician’s previous experience.
Before prescribing clomiphene citrate (CC), the physician should be sure that the follicle stimulating hormone (FSH) is not abnormally elevated and that the patient is not hypoestrogenic. One should rule out disorders of the pituitary, adrenal and thyroid which require specific treatment. Liver function evaluation should precede CC therapy if history and physical examination findings suggest liver disease. Ultrasonography (US) should be done just before starting the therapy to exclude ovarian cysts.
If gonadotropin therapy is chosen, it is of utmost importance to exclude ovarian incompetence because this type of treatment is very expensive and is not free of complications. Abnormally high serum levels of gonadotropins with low estrogen levels indicate ovarian failure which precludes induction of ovulation except in a few special cases. Non gynecological endocrine problems should be treated before starting the therapy. Hypogonadotropic function with galactorrhea requires evaluation for an intracranial lesion. It is important to know that hyperprolactinemia has no adverse effect on ovarian response to exogenous gonadotropin therapy (8). Ultrasonography should be done to exclude the presence of ovarian cysts and/or polycystic ovary disease (PCO) which require special care.
Gonadotropin releasing hormone analogue (GnRHa) combined with exogenous gonadotropin therapy
This approach is especially effective for women who either show no response to exogenous gonadotropins, or who develop premature spontaneous luteinizing hormone (LH) and progesterone rise. Indeed the major effect appears to be the prevention of premature luteinization which is a major reason for decreased success with other therapies. Patients with significant estrogen and gonadotropin levels, especially anovulatory women with PCO do not respond well to gonadotropins but the response can improve after GnRHa desensitization. Some protocols call for GnRHa use during the luteal phase of the preceding IVF cycle, others advocate its use during the follicular phase concomitantly with human menopausal gonadotropin (hMG) and/or pure FSH. At present there is no agreement as to which protocol is best, and the information available is rather conflicting.
If it is decided to prescribe the long term GnRHa protocol, the patient should be monitored for the criteria of pituitary and ovarian suppression. Complete suppression is verified by the onset of menstruation associated with a serum LH <2 IU/L, estradiol (E2) <50 pg/ml (16) and better <30 pg/ml (31) and by the absence of any ovarian follicles >10 mm in diameter. If all these criteria are not met on day 12, GnRHa should be continued and the patient assessed at weekly intervals until suppression is complete, then induction can be started.
Monitoring ovarian response to induction therapy
Monitoring ovarian response to induction therapy depends mainly on the biophysical parameters of follicular growth, and hormonal parameters, principally E2 levels.
Monitoring follicular growth
Sonography can depict developing follicles, beginning at the time they measure between 3 and 5 mm. As follicles spontaneously reach maturity in the natural cycle their inner dimensions range from 17 to 25 mm (9). Within the same individual however, the size of a mature follicle is relatively constant. Intrafollicular echoes may be observed within mature follicles probably arising from clusters of granulosa cells that shear off the wall near the time of ovulation. After ovulation, the follicular wall becomes irregular. The fresh corpus luteum usually appears as an echogenic structure with a small hypoechoic center. Patients undergoing ovulation induction are usually examined every other day beginning at day 10, but those undergoing IVF–ET are examined earlier, usually starting between day 5 and 8 of their cycles, and daily thereafter.
In CC-treated cycles, each follicle seems to develop at an individual rate, and at times may be accelerated or slowed down. Therefore the largest follicle on a given day may not be the same one that is the largest two days later, and it may not even be the one that is most mature. Furthermore, correlation of E2 and follicle size is poor and the maximum preovulatory diameter can range from 19 to 24 mm. However, the largest diameter in these cases estimated by Fossum et al. (12) ranged between 22 and 31 mm.
In hMG-treated patients, there seem to be two distinct patterns of follicular development (35). In amenorrheic women with dormant ovaries, a small number of large follicles develops. The growth rate and E2 production are linear, correlate well and are of equal predictive value. A high pregnancy rate is achieved in this group. In contrast, stimulation of patients with estrogenic activity requires less hMG and usually results in the rapid recruitment of many follicles with different growth rates and E2 secretory capacity. The rate at which E2 increases is exponential, increasing the risk of hyperstimulation. The growth rate and functional maturity are asynchronous. In this group of women, both E2 and sonographic follicular monitoring are essential.
The biophysical indicators that correlate best with the day of LH surge (12) have been found to be the follicular volume in spontaneous cycles (range: 3.4-5.6 ml), the cross-sectional area in GnRH stimulated cycles (range 1074-1382 mm2) and the largest diameter in CC-treated cycles (range: 22-31 mm). Because no significant difference was seen in the correlation among the various biophysical variables and the mid cycle LH peak however, it could be concluded that in women ovulating spontaneously, or in those induced to ovulate with CC or GnRH, any available biophysical index will have the same predictive value. In contrast, correlation analysis in cycles treated with hMG indicates that both the follicular diameter and E2 are required for optimal timing of human chorionic gonadotropin (hCG) administration.
Sonographic delineation of follicle size is crucial because hCG is best administered once follicles reach 15 to 18 mm in size even in non-IVF cycles when ovulation is allowed to occur, as the LH surge is less frequent when hMG is used for stimulation. For IVF, follicles are typically aspirated when they reach 15 to 18 mm in average diameter and when the E2 level is approximately 400 pg/ml per large follicle (20). Another sonographic sign of mature follicles is the presence of low level intrafollicular echoes, as mentioned earlier. When follicles >15 mm are aspirated, oocytes are at all stages of maturity (23). Therefore one can rely on follicular diameters alone if the patient’s previous cycles and her E2 response are known.
There is no difference in E2 production between follicles measuring 14 mm and those that are smaller, nor between follicles measuring 17 mm and those which are larger (32). The authors devised an equation to determine expected serum E2 levels depending on number and size of follicles in both ovaries. Thus the serum E2 level on the day of hCG injection is:
E2 = 291 pg/ml + 180 (x) + 64 (y) + 18.7 (z)
where x, y and z represent follicles measuring >17 mm, 15 to 16 mm and <14 mm respectively.
Whereas the sonographic finding of enlarged ovaries with multiple immature follicles may suggest the possibility of hyperstimulation, extremely high levels of E2 (over 3000 pg/ml) can be a more accurate predictor of this syndrome (2). On sonography, patients with ovarian hyperstimulation syndrome (OHSS) usually have enlarged ovaries (over 10 cm) that may contain several hypoechoic areas. The hypoechoic areas may correspond to atretic follicles, or to regions of hemorrhage within the ovary.
Sonography of the endometrium
In the late proliferative or periovulatory phase of endometrial development, a multilayered endometrium can be distinguished (11). The inner hypoechoic area probably represents edema in the compact layer of the endometrium. The endometrium would have the configuration of a theta with respect to the hypoechoic area as imaged in the semi-axial or semi-coronal plane. Patients who achieve pregnancy more frequently have a multilayered periovulatory endometrium than those in the group who do not conceive (5).
In the secretory phase, the thickness of the endometrium increases to between 8 and 16 mm and is echoic, probably due to the increased mucus and glycogen within the glands.
In one study (31), vaginal sonograms were performed during the late proliferative phase of natural and gonadotropin-stimulated cycles. The endometrium was classified into two grades according to the findings. Grade I was characterized by homogenous echogenicity of the endometrium, while grade II was characterized by an outer peripheral layer of dense echogenicity surrounding a central sonolucent area (halo pattern). Grades I and II were subclassified into group A (>9 mm thick) and group B (<9 mm). Grade IIA was optimal, as it was associated with a clinical pregnancy rate per embryo transfer of 33% while the other three groups were poor as they were associated with a rate of 7% only. Women aged 41-45 years experienced a 25% incidence of poor sonographic grades compared to only 5% incidence in women <40 years.
Monitoring E2 levels
Ovarian response can be identified early in the cycle. Hodgen (15) defined non responders as those whose E2 levels did not reach 300 pg/ml by day 8 of stimulation. Slow responders were those whose E2 levels were <300 pg/ml by day 5, but >300 pg/ml by day 8 of stimulation. Fast responders had their E2 levels >300 pg/ml by day 5 of stimulation. However because E2 levels can be augmented to comparable levels by increasing the dose of gonadotropins, a correlation between E2 levels and gonadotropin dose is needed. Ibrahim and co-workers (16) defined poor response in desensitization protocols as the need for 4 or more ampoules of hMG/day to induce ovulation.
The dose of gonadotropin should not be changed as long as serial E2 levels rise between 50 and 100% every other day (32). Dirnfeld et al. (6), showed that very slow or very rapid estrogen growth rates (EGRs), calculated from the 4 days preceding oocyte aspiration in CC/hMG stimulated cycles, were associated with a reduced pregnancy rate. EGRs of 0.31 to 0.41 were associated with optimal pregnancy rates. EGR is calculated by the formula:
EGR = e-B -1
where B is the slope of the least square line corresponding to the semilogarithmic plot of E2 values versus time and e = 2.718.
Using GnRHa and gonadotropin in a desensitization protocol, the ovarian response was evaluated in terms of E2 levels on the day of hCG injection, and 36 hours later at egg retrieval (23). Low responders, medium responders and high responders were those with E2 levels of <800 pg/ml, 800-1500 pg/ml and >1500 pg/ml respectively on the day of hCG injection or <400 pg/ml, 400-1000 pg/ml and >1000 pg/ml respectively at egg retrieval. There were no differences between the three groups in respect to development of mature oocytes and rapidly cleaving embryos. The pregnancy rate in the low responding group, however, was significantly lower than in the other two groups, despite replacement of an equivalent number of oocytes and cleaving embryos. Thus it seems that the receptivity of the endometrium depends at least partially on adequate E2 levels. It also seems that E2 levels do not directly correlate with oocyte maturity and embryonic growth.
An upper limit of estradiol of 3800 pg/ml for anovulatory women (with polycystic ovaries) and 2400 for women with hypothalamic amenorrhea produces a risk of severe hyperstimulation of 5% in pregnant cycles and 1% in non conceptional cycles (14).
Paltieli and colleagues (28) found that in hMG cycles in which ovulation was triggered by using hCG injections, at least 80% of pregnancies were achieved when the E2 rise (active phase) was 6±1 days, whereas only 15% of pregnancies were achieved when the active phase was >7 days. They attributed the high incidence of early abortion, when the active phase was >7 days, to be an expression of oocyte overexposure to hMG prior to hCG injection. Such overexposure may result in postmature oocytes and end in early abortion. The same group of investigators noted also that in good outcome cycles, E2 continued to rise until hCG was administered, but in nonpregnant cycles, E2 plateaued on the day before hCG administration, which suggests that luteinization or atresia of the more advanced follicles had commenced spontaneously.
Monitoring special situations
Although adequate follicular development occurs with CC and hMG combination regimen, it is thought that one problem with that regimen is premature luteinization (13). In general, it is believed that the rise in serum progesterone occurs 12 hours before or on the day of the onset of a spontaneous LH surge in a natural cycle, or in a controlled ovarian hyperstimulation for IVF-ET program (36). Fleming and Coutts (10) defined the criteria for premature luteinization to be: serum progesterone >1.5 ng/ml associated with a rise in serum LH concentration before maturation of the developing follicles, together with a decline or plateauing of the serum E2 concentration despite continued hMG administration. However, there were reports that a significant rise in serum progesterone occurs in advance of the onset of the LH surge in regimens using a combination of CC and hMG (30).
In 1992, Mio and colleagues (24) defined " subtle progesterone rise " as a fluctuation in the serum progesterone concentration of between 1 and 2 ng/ml from day 7 of the cycle until 24 hours before the hCG administration, or the onset of the LH surge. This is not coupled with a significant increase in the serum LH concentration, defined as an increase of <100% of the mean value of the 2 preceding days, nor with a decline or plateauing of the serum E2 concentration during ovarian stimulation. It was reported that subtle progesterone rise occurred in 31.7% of patients and 20.5% of cycles in which CC and hMG were used to induce ovulation.
A significantly higher serum E2 concentration and a greater number of developed and collected oocytes were observed in those cycles with subtle progesterone rise. The rate of mature oocyte formation and fertilization were significantly lower, however, and the rate of embryos with cytoplasmic fragments was significantly higher in those cycles. A low pregnancy rate which did not progress to ongoing pregnancy was also observed in those cycles (24). Granulosa cells may develop LH receptors as a result of high concentration of serum E2 induced by exogenous FSH (3). An excessive sensitivity of granulosa cells to LH might induce untimely progesterone production even for a low concentration of serum LH. The low fertilization rate may be due to the direct determining effect of the subtle progesterone rise on oocyte quality and maturity. A low pregnancy rate after ET may be related to the poor quality of embryos as well as to the direct effect of progesterone on the receptivity of the endometrium due to asynchrony between it and the embryos.
It was concluded that elevated FSH value (cut limit of 26 IU/L) following CC administration predicts a poor response to further stimulation by gonadotropins—with an accuracy of 92.3%—and should result in cancellation of the cycle (34). It could be that those patients have diminished ovarian reserves and consequently a poor prognosis for future IVF based on the findings of Navot et al. (27) and Muasher et al. (26). A reverse effect of CC on ovarian steroid synthesis however cannot be excluded (18,21). These patients may benefit from a combination of GnRHa and gonadotropin therapy (1,30).
Gonadotropin only protocols
It has been demonstrated that basal LH levels decline, and LH surges are often absent in gonadotropin-treated cycles of humans (17) and animals (29). Suppression of LH secretion during stimulated cycles (in which serum E2 is often elevated beyond normal mid cycle levels) is considered to be due to inhibition of the pituitary by ovarian factors, or a direct effect of exogenous gonadotropins (29). In some cases however, increase in LH is often observed during gonadotropin therapy (22).
Mizunuma et al. (25) defined premature LH release as an LH release that exceeds pretreatment values for at least two consecutive days before hCG administration. They could recognize three types of premature LH release. The sustained type, defined as an LH release that occurred during the treatment and lasted until hCG administration. The transient type, so-called when LH release occurred during treatment, but returned to normal levels before LH administration. If LH was released only on the day of hCG administration, it was called the onset type. Those cycles with premature LH release were accompanied by increased FSH levels, and a high incidence of ovarian hyperstimulation. They concluded that ovarian hyperstimulation can be reduced by modulating the dose of FSH and the intervals of administration in cycles showing premature LH release when it occurs early in the cycle and is discovered early.
Timing of hCG administration
The doses of hCG administered are in the range 2000-10,000 IU. The regimen used by many clinics (37) calls for the administration of hCG on the sixth day of a sustained increase in serum estradiol levels. Patients who fail to achieve adequate follicular development after 6-8 days of ovarian stimulation do not receive hCG, and the treatment cycle is cancelled.
In other centres, hCG is administered when the serum estradiol level reaches 200-300 pg/ml per follicle >17-18 mm in diameter. Patients with poor follicular development or with only one developing follicle are not given hCG. It is inadvisable to give hCG to patients in whom the serum estradiol level is seen to increase rapidly (i.e. doubling in 24 hours) in order to minimize the risk of the OHSS.
Just prior to hCG injection, a serum LH can be drawn and compared to values earlier in the cycle. This helps to identify women who have initiated a premature LH surge (LH value 2.5 times baseline). However, without frequent sampling of LH (every 3 hours), the onset of the surge cannot be identified with precision (33). LH sampling is not required in patients who are treated with GnRHa. If a spontaneous LH surge occurs in a stimulated cycle, some centres cancel the treatment cycle, whereas others give hCG if there is a satisfactory estradiol response and adequate follicular growth has taken place (37). In these cases, it is necessary to adjust the timing of oocyte recovery.
As a general rule, hyperstimulation is associated with the presence of many follicles. It is advisable that hCG not be administered if there are more than 3-4 follicles of 14 mm or more in diameter (33). Mild hyperstimulation has been associated with an increased number of intermediate size follicles and severe hyperstimulation with an increase in small follicles (2). A large number (11 or more) of small follicles should also preclude hCG administration.
Check and colleagues (4) used hCG to trigger ovulation in their patients in whom ovulation was induced by hMG. The timing of injection of hCG was influenced by the serum progesterone level as follows: if the serum progesterone was >1.8 ng/ml, then 10,000 units of hCG would be given as long as there was at least one dominant follicle with serum estradiol >200 pg/ml, even if multiple follicles were present and the serum estradiol was <200 pg/ml/mature follicle.
Monitoring after completion of the induction therapy
Some programmes measure the estrogen level on the day following hCG, and if there is a marked drop in the value at that time, retrieval is cancelled because that pattern is associated with a poor chance of pregnancy (33). Oocyte retrieval is performed approximately 35 hours after the hCG injection. Nowadays, ultrasonically-guided retrieval has replaced laparoscopic oocyte retrieval to a large degree. During the process of oocyte pick-up, one should be sure of aspirating all the follicles, to decrease the risk of hyperstimulation syndrome, especially when the ovaries are enlarged and/or in the presence of many follicles (7). After completion of the procedure, the ovarian pedicles should be thoroughly observed for torsion, especially in the presence of enlarged ovaries, as the repeated manipulations in the course of the procedure to reach the follicles located at different parts of the ovary may directly lead to iatrogenic torsion especially when the ovaries are enlarged.
Another kind of monitoring is carried out in the laboratory after oocyte retrieval, but this is again beyond the scope of this paper. The patient should also be prepared during the laboratory work for the next step, which is the embryo transfer. For the purpose of analytical studies, it is mandatory to document the number of transferred embryos, their stage of development, whether the transfer was easy, difficult or very difficult, whether mucus or blood entered into the catheter and so on.
After embryo transfer, the patient should inform the physician if any symptoms of OHSS develop. It is a common practice to supplement the luteal phase with progesterone, although limited studies indicate that this may not be necessary (19).
The patient has to be followed up for the occurrence of pregnancy. The diagnosis of a biochemical pregnancy depends on the measurement of ß-subunit hCG levels in the patient’s serum about 2 weeks after oocyte retrieval. A value >25 mIU/ml is diagnostic and is confirmed by a rising titre 3 days later. When pregnancy is diagnosed, it may be necessary to support it by exogenous hCG administration until 12 weeks gestation. The diagnosis of a clinical pregnancy is made when one or more gestational sacs can be identified by ultrasound image 4 to 6 weeks after oocyte retrieval. Embryonic viability is diagnosed when the heart beats can be detected on the screen.
This is not the whole story. Pregnant patients still need close follow-up and special care in relation to an expensive precious pregnancy until after labour and delivery.
- Belaisch-Allart, J., Testart, J., and Frydman, R. (1989): Hum. Reprod., 4:33-34.
- Blankstein, J., Shalev, J., Saadon, T., Kukia, E.E., Rabinovici, J., Pariente, C., Lunenfeld, B., Serr, D.M., and Mashiach, S. (1987): Fertil. Steril., 47:597-602.
- Channing, C.P., Kammerman, S. (1974): Biol. Reprod.,10:179-198.
- Check, J.H., Adelson, H.G., Stern, J., and Lauer, C. (1992): Int. J. Fertil., 37:103-105.
- deCrespigny, L., Cooper, D., and McKenna, M. (1988): J. Ultrasound Med., 7:7-10.
- Dirnfeld, M., Lejeune, B., Camus, M., Vekemans, M., and Leroy, F. (1985): Fertil. Steril., 43:379-384.
- Fakih, H., and Bello, S. (1992): Fertil. Steril., 58:829-832.
- Farine, D., Dor, J., Lupovici, N., Lunenfeld, B., and Mashiach, S. (1985): Obstet. Gynecol., 65:658-660.
- Fleischer, A.C., Daniell, J.F., Rodier, J., Lindsay, A.M., and James, A.E. (1981): J. Clin. Ultrasound, 9:275-280.
- Fleming, R., and Coutts, J.R.T. (1986): Fertil. Steril., 45:226-230.
- Forrest, T.S., Elyadereni, M.K., Muilenburg, M.I., Bewtra, C., Koble, W.T., and Sullivan, P. (1988): Radiology, 167:233-237.
- Fossum, G.T., Vermesh, M., and Kletzky, O.A. (1990):Obstet. Gynecol., 75:407-411.
- Hamori, M., Stuckensen, J.A., Rumpf, D., Kniewald, T., Kniewald, A., and Kurz, C.S. (1987): Hum. Reprod., 2:639-643.
- Haning, R.V. Jr., Boehnlein, L.M., Carlson, I.H., Kuzma, D.L., and Zweibel, W.J. (1984): Fertil. Steril., 42:882-889.
- Hodgen, G.D. (1989): Hum. Reprod., 4:37-46.
- Ibrahim, Z.H., Matson, P.L., Puck, P., and Lieberman, B.A. (1991): Fertil. Steril., 55:202-204.
- Kamrava, M.M., Seibel, M.M., Berger, M.J., Thompson, I., and Taymor, M.L. (1982): Fertil. Steril., 37:520-523.
- Laufer, N., Reich, R., Braw, R., Shenker, J.G., and Tsafriri, A. (1982): Biol. Reprod., 27:463-470.
- Leeton, J., Trounson, A., and Jessup, D. (1985): J. In Vitro Fert. Embryo Transf., 2:166-169.
- Marrs, R.P., Vargyas, J.M., and March, C.M. (1983): Am. J. Obstet. Gynecol., 145:417-421.
- Marut, E.L., and Hodgen, G.D. (1982): Fertil. Steril., 38:100-104.
- McFaul, P.B., Traub, A.I., and Thompson, W. (1989): Acta Eur. Fertil., 20:157-161.
- Mettler, L., and Tavmergen, E.N. (1989): Hum. Reprod., 4:59-64.
- Mio, Y., Sekijima, A., Iwabe, T., Onohara, Y., Harada, T., and Terakawa, N. (1992): Fertil. Steril., 58:159-166.
- Mizunuma, H., Andoh, K., Yamada, K., Takagi, T., Kamijo, T., and Ibuki, Y. (1992): Fertil. Steril., 58:46-50.
- Muasher, S., Oehninger, S., Simonetti, S., Matta, J., Ellis, L.M., Liu, H.C., Jones, G.S., and Rosenwaks, Z. (1988): Fertil. Steril., 50:298-307.
- Navot, D., Rosenwaks, Z., and Margalioth, E.J. (1987): Lancet, 2:645-647.
- Paltieli, Y., Tal, J., Porat, N., Tesler, B., Abramovici, D., and Sharf, M. (1991): Int. J. Fertil., 36:94-98.
- Schenken, R.S., and Hodgen, G.D. (1983): J. Clin. Endocrinol. Metab., 57:50-55.
- Serafini, P., Stone, B., Kerin, J., Batzofin, J., Quinn, P., and Marrs, R.P. (1988): Fertil. Steril., 49:86-89.
- Sher, G., Herbert, C., Maassarani, G., and Jacobs, M.H. (1991): Hum. Reprod., 6:232-237.
- Silverberg, K.M., Olive, D.L., Burns, W.N., Johnson, J.V., Groff, T.R., and Schenken, R.S. (1991): Fertil. Steril., 56:296-300.
- Speroff, L, Glass, R.H., and Kase, N.G. (1989): Clinical Gynecologic Endocrinology and Infertility, 4th ed. Williams & Wilkins, Baltimore.
- Tanbo, T., Dale, P.O., Abyholm, T., and Stokke, K.T. (1989): Hum. Reprod., 4:647-650.
- Tarlatizis, B.C., Laufer, N., and DeCherney, A.H. (1984): J. In Vitro Fert. Embryo Transf., 1:226-232.
- Trounson, A.O., and Calabrese, R. (1984): J . Clin. Endocrinol . Metab., 59:1075-1080.
- WHO, editor (1992): Recent advances in medically assisted conception. Geneva.