Minimal stimulation IVF with late follicular phase administration of the GnRH antagonist cetrorelix and concomitant substitution with recombinant FSH: a pilot study

Abstract

BACKGROUND: The use of the natural cycle for IVF offers the advantage of a patient-friendly and low-risk protocol. Its effectiveness is limited, but may be improved by using a GnRH antagonist to prevent untimely LH surges. METHODS: In this pilot study, minimal stimulation IVF with late follicular phase administration of the GnRH antagonist cetrorelix and simultaneous substitution with recombinant FSH was applied for a maximum of three cycles per patient. Main outcome measures were pregnancy rates per started cycle and cumulative pregnancy rates after three cycles. RESULTS: A total of 50 patients completed 119 cycles (2.4 per patient). Fifty-two embryo transfers resulted in 17 ongoing pregnancies [14.3% per started cycle; 32.7% per embryo transfer; 95% confidence interval (CI) 7.9–20.7% and 19.7–45.7%, respectively]. One dizygotic twin pregnancy occurred after transfer of two embryos, the other pregnancies were singletons. The cumulative ongoing pregnancy rate after three cycles was 34% (95% CI 20.6–47.4%). Live birth rate was 32% per patient (95% CI 18.8–45.2%). CONCLUSIONS: Pregnancy rates after IVF with minimal, late follicular phase stimulation are encouraging. Considering the low-risk and patient-friendly nature of this protocol, it may be a feasible alternative to IVF with ovarian hyperstimulation.

Introduction

Natural cycle IVF has received much attention in recent literature (Bassil et al., 1999; Reljič and Vlaisavljević, 1999; Janssens et al., 2000; Feldman et al., 2001; Ingerslev et al., 2001; Nargund et al., 2001; Ng et al., 2001; Omland et al., 2001; Reljič et al., 2001; Bauman et al., 2002; Lukassen et al., 2003; Morgia et al., 2004). The use of natural cycle IVF offers several advantages. Since no ovarian stimulation is performed, ovarian hyperstimulation syndrome (OHSS) does not occur, and compared with standard IVF treatment with controlled ovarian hyperstimulation (COH), natural cycle IVF is a cheap and patient-friendly procedure (Aboulghar et al., 1995; Daya et al., 1995; Olivennes and Frydman, 1998; Højgaard et al., 2001; Nargund et al., 2001).

The efficacy of natural cycle IVF, however, is limited. Reported ongoing pregnancy rates per started cycle are 7.2% on average, with a range of 0–18.8%, as reviewed in Pelinck et al. (2002). The lack of efficacy is related to high cancellation rates (28.9%), mainly caused by the occurrence of spontaneous LH surges (Pelinck et al., 2002).

The efficacy of natural cycle IVF can be improved by using a GnRH antagonist to prevent untimely LH surges and premature ovulations. In minimal stimulation IVF, a GnRH antagonist is started in the late follicular phase, after follicular dominance has developed. To substitute for the fall in gonadotrophins, HMG or FSH are administered together with the GnRH antagonist. Various protocols have been described with the GnRH antagonists Nal-Glu and cetrorelix at different doses, and substitution with HMG or urinary purified FSH (Meldrum et al., 1994; Paulson et al., 1994; Rongières-Bertrand et al., 1999). In these studies, a total of 54 cycles are described, resulting in seven ongoing pregnancies (13% per started cycle). In the largest of these three studies, only patients with severe male factor infertility were included and ICSI was applied. A total of 44 cycles in 33 patients led to seven clinical pregnancies, of which five were ongoing (Rongières-Bertrand et al., 1999).

So far, the use of recombinant FSH (rFSH) has not been described for the minimal stimulation protocol. Minimal stimulation IVF offers the same advantages as natural cycle IVF, being associated with zero risk of OHSS, since no ovarian hyperstimulation is performed, and being a cheap and patient-friendly protocol compared with COH IVF. Per cycle, minimal stimulation IVF is cheaper than COH IVF as considerably less hormonal medication is used, and laboratory procedures are less time-consuming as, in general, only one oocyte is obtained. The patient-friendliness of minimal stimulation IVF lies in the fact that hormonal medication is used for a few days only, and therefore few side-effects are experienced by patients, and that the oocyte retrieval is less painful than in COH IVF since, in general, only one follicle is aspirated. With minimal stimulation, usually no spare embryos are created, and therefore ethical dilemmas concerning them are avoided.

In this pilot study we performed IVF with minimal stimulation. The GnRH antagonist cetrorelix was administered in the late follicular phase and rFSH was used for substitution. The goal of this study was to make an estimation of pregnancy rate per started cycle and of cumulative pregnancy rates after three cycles. This is the first study in which a minimal stimulation protocol is described using rFSH for substitution.

Materials and methods

Study protocol

This study protocol was reviewed and approved by the ethics committee of the Academic Hospital of Groningen, The Netherlands. Inclusion criteria for this study were: female patient age 18–36 years, first IVF treatment ever or first IVF treatment after a pregnancy, the presence of a regular and proven ovulatory menstrual cycle with a length of 26–35 days and body mass index (BMI) of 18–28 kg/m². Indications for IVF were tubal pathology, unexplained subfertility, male factor, endometriosis, cervical factor or failed donor inseminations.

Patients with male factor or unexplained subfertility had undergone treatment with intrauterine insemination for three to six cycles before starting IVF treatment, as is standard protocol in our centre. Patients were excluded from the study if an endometriosis cyst was seen on ultrasound. Patients requiring ICSI were not included in this study.

Patients were offered a maximum of three free treatment cycles. Treatments were performed in three consecutive menstrual cycles, unless patients otherwise requested. Patients who decided not to participate in this study underwent COH IVF treatment according to our standard protocol.

Ultrasound monitoring was started on cycle day 3 or 8, and repeated daily or every other day, according to the size of the lead follicle. Follicle diameter was measured in three perpendicular planes, and the mean value was taken. When a lead follicle with a mean diameter of at least 14 mm was observed, daily injections of cetrorelix 0.25 mg (Cetrotide^®; Serono, The Hague, The Netherlands) together with 150 IU rFSH (75 IU FSH per ampoule; Gonal-F^®; Serono, Benelux BV, The Netherlands) were started. Cetrorelix was continued up to and including the day of ovulation triggering; rFSH was continued up to the day of ovulation triggering. Patients were instructed to have their injections in the evening and at the same time daily, to ensure a 24-h interval between injections.

Serum concentrations of LH and estradiol (E₂) were assessed on the days that ultrasound was performed. Blood samples were taken in the morning, so serum concentrations reflected levels 12–16 h after administration of the medication.

Ovulation triggering was achieved by subcutaneous injection of 10 000 IU of HCG (Pregnyl^®; Organon, Oss, The Netherlands) when a follicle with a diameter of at least 18 mm was observed and plasma E₂ levels were ≥0.8 nmol/l (equivalent to 218 pg/ml; Immulite 2000) or ≥1.06 nmol/l (equivalent to 288 pg/ml; Architect i-2000). In case a follicle with a diameter of 18 mm was observed together with an E₂ level of <0.8 nmol/l (218 pg/ml; Immulite 2000) or <1.06 nmol/l (288 pg/ml; Architect i-2000), ovulation triggering was postponed for 1 day. If an LH rise was noticed, cycles were not cancelled since we hypothesized that cetrorelix administered on the day of ovulation triggering should be capable of blunting the LH surge, thus allowing for planned oocyte retrieval. Transvaginal ultrasound-guided follicle aspiration was performed 34 h after ovulation triggering. A single lumen aspiration needle (MDT^®; Hilvarenbeek, The Netherlands) was used. No flushing of the follicle was performed. The aspiration pressure was −50 to −100 mmHg, according to standard procedure. Analgesia (fentanyl 2 mg intravenously) was only given on patient request.

Oocytes were inseminated 2–5 h after oocyte retrieval with 100 000 motile spermatozoa prepared by centrifugation for 15 min at 300 g over a 45/90% gradient of Suprasperm^® (Medicult a/s, Jylling, Denmark), followed by washing and a swim-up procedure in culture medium. Oocytes were cultured in human tubal fluid medium (Cambrex Bio Science, Verviers, Belgium), supplemented with 10% plasma solution (CLB, Amsterdam, The Netherlands). Fertilization was assessed 17–20 h after insemination. Embryo transfer was performed 72–76 h after oocyte retrieval, using a TDT^® (Prodimed, Neuilly-en-Thielle, France) or Wallace^® (SIMS Portex Ltd, Hythe, UK) catheter. For luteal support, HCG 1500 IU was given 5, 8 and 11 days after oocyte retrieval. Since no conclusive data are available on the need for luteal phase support in minimal stimulation IVF, luteal support was administered in all cycles where embryo transfer was performed.

Clinical pregnancy was defined as the ultrasound visualization of a intrauterine gestational sac. Ongoing pregnancy was defined as the presence of a intrauterine gestational sac with fetal heart beat at 12 weeks amenorrhoea.

Serum assays

Serum LH was measured using immunoassay station AutoDELFIA^® (EEG Wallac, Turku, Finland), with inter-assay coefficients of variation (lower, middle and upper range of concentrations) of 5.4%, 5.7% and 6%.

E₂ concentrations were measured on the Immulite 2000^® station (Diagnostic Products Corporation, Los Angeles, CA, USA) for the first 94 cycles of the study. In the other 25 cycles in this study, the Architect-i2000^® (Abbott Diagnostics, Chicago, IL, USA) was used for E₂ measurements. The interassay coefficients of variation (lower, middle and upper range of concentrations) were 17.7%, 8% and 9.3% for Immulite 2000 and 7.9%, 3.6% and 2.8% for Architect i-2000. E₂ values measured with Architect-i2000=values measured with Immulite 2000 × 1.20+0.1 nmol/l.

Data analysis

The percentage and 95% confidence interval (CI) of succesful oocyte retrievals was calculated per attempt, the 2PN fertilization rate was calculated per inseminated oocyte, and the number of embryo transfers and clinical and ongoing pregnancy rates were calculated per started cycle. The percentage and 95% CIs of cumulative ongoing pregnancy and live birth rate after three cycles was calculated per patient. The group of cycles where premature ovulation occurred (n=5) was compared with the group where oocyte retrieval was performed as planned (n=104), in order to investigate whether patient or cycle characteristics were different between these groups. Patient and cycle characteristics were compared using Student's t-test and the Mann–Whitney U-test where appropriate. A P-value of <0.05 was considered significant.

Results

Of 51 consecutive patients asked, 50 agreed to participate in this study. Patient characteristics are shown in Table I. The median age of the patients was 32 years (range 26–36). Indications for IVF are shown in Table I. Cycle characteristics are shown in Table II. Nine patients completed one cycle, 13 patients completed two cycles and 28 patients completed three cycles, for a total of 119 treatment cycles (2.4 cycles per patient).

The median number of days of cetrorelix administration was 4 (range 1–11). Follicle size at which ovulation was triggered was 19 mm (range 16–25) (Table II).

Ten out of 119 started cycles were cancelled before the start of medication, because of lack of follicular development, premature luteinization or personal reasons of the patient. Five planned oocyte retrievals were cancelled (4.2% of started cycles). In these cases, premature ovulation had occurred at the time of planned oocyte retrieval and no follicle was present. In two of these, two co-dominant follicles were present at the time of ovulation triggering.

Patient and cycle characteristics of the group where premature ovulation occurred (n=5) and of the group where oocyte retrieval was performed as planned (n=104) are shown in Table III. Female patient age, BMI and number of days of cetrorelix administration were not different between groups. Mean LH levels at ovulation triggering were higher in cancelled oocyte retrievals compared with cycles where oocyte retrieval was performed (13.7±3.3 versus 4.8±4.1 IU/l; P=0.001) (Table III). Fourteen LH surges occurred (LH value on the day of ovulation triggering >10 IU/l). In all these cycles, ovulation was triggered and oocyte retrieval was planned according to protocol. In four of them (LH value at ovulation triggering 13–16 IU/l), premature ovulation had occurred at the time of oocyte retrieval. In the other 10 cases (LH value 11–28 IU/l), ovulation had not yet taken place and oocyte retrieval was successful in seven cases. In one cycle, premature ovulation had occurred at the time of oocyte retrieval, despite the absence of a detected LH surge (LH value at ovulation triggering 8.3 IU/l).

Mean E₂ levels at ovulation triggering were not different in cancelled oocyte retrievals compared with cycles where oocyte retrieval was performed (0.93±0.17 nmol/l, 253±46 pg/ml versus 0.91±0.50 nmol/l, 248±136 pg/ml; P=0.95). The follicle size at ovulation triggering was smaller in cycles where oocyte retrieval was cancelled than in cycles where oocyte retrieval was performed (18±2.1 versus 19.6±1.5 mm; P=0.02) (Table III).

Planned oocyte retrieval was cancelled because of premature ovulation in five different patients, who completed a further seven cycles. In these cycles, oocyte retrieval was planned at a 1–2 mm smaller follicle size than in the preceding cycle, in order to avoid another cancellation. None of these cycles were cancelled, and six out of seven oocyte retrievals were successful.

Eighty out of 104 oocyte retrievals were successful (76.9%; Table II). In eight of these, two co-dominant follicles were aspirated. In one case, five co-dominant follicles were present and in one case six. One, two and three oocytes were obtained in 73, six and one case, respectively. In 60 out of a total of 88 oocytes, 2PN fertilization occurred (normal fertilization rate 68.2%).

Fifty-two embryo transfers were performed (43.7% per started cycle), leading to 19 clinical pregnancies (16% per started cycle; 36.5% per embryo transfer). In 49 and three cases, respectively, one and two embryos were transferred. Eighteen pregnancies were singleton, and one pregnancy was a dichorionic twin pregnancy after transfer of two embryos. Two miscarriages occurred. The ongoing pregnancy rate was thus 14.3% per started cycle and 32.7% per embryo transfer. The cumulative ongoing pregnancy rate after three cycles was 34% per patient. One singleton pregnancy ended with immature birth at gestational age of 18 weeks. The other pregnancies ended with live births, giving a live birth rate of 32% per patient (Table II).

Results according to indication are shown in Table IV. For male factor subfertility, the fertilization rate per inseminated oocyte was 46.7%. For the other indications, fertilization rate ranged from 75% to 100%. Owing to the low fertilization rate in male factor subfertility, the number of embryo transfers per started cycle was only 25%, whereas for the other indications, the number of embryo transfers per started cycle ranged from 45.7% to 100%.

Discussion

In this study, minimal stimulation IVF with late follicular phase administration of the GnRH antagonist cetrorelix and concomitant substitution with rFSH is described. This is the first study in which the use of rFSH is described for this purpose.

We hypothesized that during daily administration of cetrorelix 0.25 mg, substitution with gonadotrophins is necessary since follicular developmental arrest is expected in some cases if cetrorelix is administered without any substitution (Duijkers et al., 1998).

E₂ production in the follicle is both LH- and FSH-dependent. Androgens, which are needed as a substrate for FSH-dependent production of E₂ in the granulosa cells, are produced in the theca cells in response to LH. In two studies where rFSH was administered after downregulation with GnRH agonists, normal follicular development was demonstrated in the presence of very low or even undetectable LH levels (Ben-Chetrit et al., 1996; Sullivan et al., 1999). We therefore hypothesized that residual LH levels after administration of cetrorelix 0.25 mg daily would suffice for androgen production and maintenance of a physiological E₂ level, provided that exogenous FSH was administered for substitution.

In this study, no developmental arrest of the dominant follicle was seen after start of medication. It seems therefore that substitution with 150 IU of rFSH after administration of 0.25 mg of cetrorelix is effective in this protocol.

The overall premature ovulation rate (4.2% per started cycle) in our study is in accordance with the ovulation rate in an earlier study, where HMG was used for substitution after administration of a single dose of 0.5 or 1 mg of cetrorelix in the late follicular phase. In this study, four out of 44 planned oocyte retrievals were cancelled because of premature ovulation (Rongières-Bertrand et al., 1999). In one of these cycles, an early LH surge was detected after inadvert omission of cetrorelix injection. In the other three, no LH surge was detected but at the time of oocyte retrieval no follicle was seen.

The suppressive effect of cetrorelix is dose-dependent, 0.25 mg daily being the lowest effective dose (Albano et al., 1997; Duijkers et al., 1998). However, in individual cases, LH surges and subsequent ovulations have been demonstrated during daily administration of 0.25 mg of cetrorelix (Duijkers et al., 1998; Ragni et al., 2001; Al-Inany and Aboulghar, 2002). In our study, serum LH levels on the day of ovulation triggering in the cycles that were cancelled because of premature ovulation were rather high, and in four out of five cancelled oocyte retrievals a LH surge (LH >10 IU/l) was detected. Medication was administered in the evening, while blood was drawn 12–16 h before. It is therefore possible that a LH surge occurred without detection. However, in 10 cycles, LH value was >10 IU/l on the day of ovulation triggering, but ovulation had not occurred at the time of oocyte retrieval. Apparently, in some cases, cetrorelix 0.25 mg is indeed capable of blunting the LH surge enough to allow for planned oocyte retrieval.

The overall premature ovulation rate (4.2%) we found in this study is low, and probably lower than the ovulation rate with natural cycle IVF without use of GnRH antagonists, which is ∼16.6% (Pelinck et al., 2002). Further research should be performed comparing natural cycle IVF with minimal stimulation IVF to clarify whether the benefit of a lowered cancellation rate with a minimal stimulation protocol outweighs the inconvenience of treatment with GnRH antagonist and gonadotrophins. On the other hand, an increase in GnRH antagonist dose might further reduce ovulation rates and thus lead to better results in minimal stimulation IVF. In this study, we chose to use 150 IU rFSH as substitution, and this seems to be effective in the minimal stimulation protocol. However, we do not know whether a lower dose of gonadotrophins or the use of LH-containing preparations will be equally effective. This also should be a subject of further research.

In the present study, medication was started at a follicle size of 14 mm, assuming that follicular dominance had developed, in order to avoid multiple follicular recruitment. In 97 cycles, a single dominant follicle developed and subsequently, no more than one oocyte and one embryo for transfer was obtained. In 10 cycles, two co-dominant follicles were present when medication was started and both follicles continued to grow during administration of cetrorelix and rFSH. In two cycles, five and six co-dominant follicles developed, whereas only one was present when medication was started. It seems therefore that in most but not all cases, multiple follicle development is prevented by starting medication no earlier than at a follicular size of 14 mm. Out of 52 embryo transfers, 49 were of one single embryo.

In the present study, the overall ongoing pregnancy rate was 32.7% per embryo transfer. The ongoing implantation rate was 30.9% per transferred embryo. This implantation rate seems to be comparable to implantation rates of embryos obtained after COH IVF. This is surprising since, other than in COH IVF, in our protocol no selection of the best-quality embryo was possible since in most cases only one was obtained. It may be that the oocyte from the dominant follicle represents the best-quality oocyte in a cohort of oocytes, leading to an embryo with good implantation potential. An alternative explanation for the surprisingly good implantation rates found in our study is that the endometrium is more receptive after minimal stimulation than in stimulated IVF. Some authors claim a disturbed endometrial receptivity after ovarian stimulation for IVF, related to supraphysiological levels of E₂ (Fossum et al., 1989; Simón et al., 1998; Basir et al., 2001; Bourgain and Devroey, 2003). Physiological levels of steroid hormones as present in minimal stimulation cycles may thus be associated with better endometrial receptivity compared with endometrial recepticity after COH IVF.

In this study, the patient population is probably somewhat biased towards a good chance for pregnancy, the maximum patient age at inclusion being 36 years. The overall ongoing pregnancy rate was 14.3% per started cycle, which seems to be higher than pregnancy rates obtained after natural cycle IVF without the use of a GnRH antagonist (Pelinck et al., 2002).

The pregnancy rate in the present study seems comparable to those found in an earlier study, where a single injection of cetrorelix 0.5 or 1 mg was given together with HMG (Rongières-Bertrand et al., 1999). In that study, where in all cases ICSI was applied, seven clinical pregnancies, of which five ongoing, were obtained in 44 cycles (15.9% clinical and 11.4% ongoing pregnancy rate per started cycle) (Rongières-Bertrand et al., 1999).

Cancellation rates and number of unsuccessful oocyte retrievals were rather high, leading to a disappointingly low number of embryo transfers per cycle (43.7%). Implantation rates were high, leading to a 14.3% ongoing pregnancy rate per started cycle. Compared with pregnancy rates after COH IVF, this ongoing pregnancy rate per cycle is low. However, minimal stimulation IVF is easily repeated in consecutive cycles, and the duration of one cycle of minimal stimulation IVF is about one-third of a COH IVF treatment cycle, downregulation preceding the actual treatment cycle and a resting cycle afterwards included. Therefore, in terms of effectiveness, it seems logical to compare cumulative pregnancy rates after three cycles of minimal stimulation IVF with those after one cycle of COH IVF.

Future research should clarify which patients will benefit most from minimal stimulation IVF. From the present study, it seems that this protocol is effective for all indications studied, albeit possibly less so for male factor subfertility, where, owing to a low fertilization rate, the number of embryo tranfers per started cycle seems lower than for the other indications. Of course, no firm conclusions can be drawn owing to the small numbers involved. In analogy with natural cycle IVF, this protocol could be a valuable alternative for poor responders (Lindheim et al., 1997; Bassil et al., 1999; Feldman et al., 2001; Morgia et al., 2004). Since with ovarian stimulation, these patients will have only few oocytes, it seems feasible to apply a minimal stimulation IVF protocol, where with less costs and less time per cycle, comparable results could be obtained. In addition, patients at risk for OHSS or with a history of OHSS could also benefit from this protocol.

Minimal stimulation IVF has the advantage of being a patient-friendly protocol, since medication is administered for a few days only and therefore few side-effects are experienced. Oocyte retrieval is less painful than in COH IVF, since usually only one follicle is aspirated. Also, minimal stimulation IVF is associated with zero risk of OHSS and therefore is a low-risk alternative to COH IVF.

As yet there are no cost-effectiveness analyses available on minimal stimulation IVF, but per cycle, minimal stimulation IVF will be far cheaper than COH IVF, since considerably less hormonal medication is used.

We consider the very low multiple pregnancy rate in minimal stimulation IVF an advantage. Clearly, multiple pregnancies are a consequence of embryo transfer policy and not of ovarian stimulation, and so can be prevented after COH IVF by performing elective single embryo transfer (SET) in cases where at least one embryo of good quality is available (Gerris and Van Royen, 2000; Ozturk et al., 2001). The addition of pregnancies after transfer of frozen–thawed spare embryos raises the delivery rate per oocyte retrieval even further (Tiitinen et al., 2001). With judicious application of elective SET, the multiple pregnancy rate is reduced to 7.5–21% without a drop in overall pregnancy rates (Gerris et al., 2002; De Sutter et al., 2003; Tiitinen et al., 2003). However, this strategy still requires ovarian hyperstimulation, leading to high costs, patient discomfort and risk for OHSS, and is only possible if patients are willing to undergo elective SET, which is often not the case (Gerris et al., 2004; Murray et al., 2004).

Of course, it is not known how willing patients will be to undergo minimal stimulation IVF, since the high inclusion rate in the present study is probably, at least in part, caused by the fact that treatment cycles were offered for free.

A randomized controlled study comparing minimal stimulation IVF with COH IVF seems warranted. In such a study, focus should not be on success rates per cycle, but rather on pregnancy rates per time spent by the patient; so for instance, three or four cycles of minimal stimulation IVF can be compared with one cycle of COH IVF, frozen embryo transfers included. In such a study, elective SET could be applied in the COH IVF cycles as well as in those cycles of minimal stimulation IVF where two or more embryos are obtained. Concerning cost-effectiveness, costs per live birth should be calculated, including costs of pregnancy and delivery and costs related to OHSS. Also, quality of life of patients should be taken into account in such a study.

In conclusion, pregnancy rates after minimal stimulation IVF with use of the GnRH antagonist cetrorelix in the late follicular phase are encouraging. For substitution, rFSH alone seems to be sufficient. A randomized controlled trial comparing minimal stimulation IVF with current standard treatment protocols is warranted, to clarify whether minimal stimulation IVF is a feasible alternative to COH IVF.

This study was supported by a grant from the Academic Hospital Groningen, Groningen, The Netherlands. These data were presented at the 18th Annual Meeting of ESHRE, July 1–3, 2002, Vienna, Austria.

References