Human Reproduction vol.7 no.9 pp. 1235-1239, 1992

A randomized trial between GIFT and ovarian stimulation for the treatment of unexplained infertility and failed artificial insemination by donor

Hendrikus V.Hogerzeil1, Jan C.M.Spiekerman, Jan W.A.de Vries and Gerard de Schepper2 Section of Reproductive Endocrinology and Fertility, Department of Obstetrics and Gynaecology, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam and 2Institute of Human Genetics, University of Amsterdam, Amsterdam, The Netherlands 'To whom correspondence should be addressed

This study was designed to investigate the surplus effect of gamete intra-FaDopian transfer (GIFT) over ovarian stimulation alone, in patients with unexplained infertility. A total of 50 patients with unexplained infertility of at least 3 years duration, or unexplained failure of artificial insemination by donor (AID) for at least 12 cycles, meeting strict inclusion criteria, were randomly selected for either two GIFT cycles or two ovarian stimulation cycles. Ovarian stimulation was combined with timed intercourse, or timed cervical donor insemination. In 38 completed GIFT cycles, five clinkal pregnancies (13.2% per cycle) occurred and in 44 ovarian stimulation cycles four clinical pregnancies occurred (9.1% per cycle). Five remaining GIFT cycles were converted into in-vitro fertilization leading to two pregnancies. Of the 50 patients suffering from unexplained infertility, the 23 who did not have AID gave rise to four pregnancies out of 39 cycles (10.3%); from the remaining 27 patients who underwent AID, seven pregnancies were achieved out of 48 cycles (14.6%). No statistical differences between GIFT and ovarian stimulation treatment were found. Therefore, the GIFT success rates can be explained at least in part, if not fully, by the effect of ovarian stimulation alone. Consequently, ovarian stimulation should be considered in unexplained infertility before more elaborate forms of assisted reproduction are used. Key words: GIFT/ovarian stimulation/unexplained infertility/ artificial insemination by donor

Introduction Many forms of assisted reproduction have been recommended for the treatment of unexplained infertility, of which gamete intraFallopian transfer (GIFT), in particular, has been suggested as a successful method. In 100 consecutive GIFT attempts Braeckmans et al. (1987) reported 24 clinical pregnancies. In selected series, 40% (Craft et al., 1988) and even >50% (Formigli et al., 1990) pregnancy rates have been achieved with GIFT. In controlled studies, however, GIFT pregnancy rates are often indistinguishable from those achieved with in-vitro fertiliza© Oxford University Press

tion (TVF) and other methods (Leeton et al., 1987; Yovich et al., 1988b; Tanbo et al., 1990). Invariably, assisted reproduction treatment includes ovarian stimulation and subsequent shortening of the distance which the gametes have to cover before they can meet. It is likely that the claimed success rates are, at least in part, due to the stimulation itself, because more oocytes become available for fertilization per cycle (Corsan and Kemmann, 1991). Indeed, pregnancy rates in GIFT seem to correlate with the number of oocytes transferred (Craft et al., 1988; Yovich et al., 1988a). Furthermore, success rates depend on the duration of infertility. Unexplained infertility of short duration has a higher potential for spontaneous cure (Hull et al., 1985) and presumably a better response to assisted reproduction techniques. The present study was designed to determine the added effect of GIFT over ovarian stimulation alone on the pregnancy rate per cycle, in patients with unexplained infertility. For this, all patients at the infertility clinic meeting strict inclusion criteria for unexplained infertility of S 3 years' duration were selected for the study. Also all patients that had been unsuccessfully treated for at least 12 ovulatory cycles with artificial insemination by donor (AID) using cryopreserved semen of fertile donors, and who met similar selection criteria for unexplained infertility, were allowed to participate. Materials and methods Patients Unexplained infertility was defined as primary or secondary infertility of at least 3 years' duration, or unsuccessful treatment for a minimum of 12 ovulatory donor insemination cycles, when all of the following criteria were met: (i) regular menstrual cycle (between 25 and 35 days) and biphasic basal body temperature curve; (ii) hormonal plasma levels measured at least once and in the normal range: prolactin 9 nmol/1, to diminish the risk of large multiple pregnancies. Otherwise, intercourse or cervical insemination were scheduled 34 h after the administration of HCG. At donor insemination, the contents of one thawed cryopreserved semen straw of 0.25 ml, containing a minimum of 2.5 x 106 progressive motile spermatozoa, were inseminated intracervically. In the GIFT cycles, laparoscopic follicle aspiration followed 36 h after HCG injection. A sperm cell suspension was prepared 2 h before follicle aspiration either through a Percoll gradient or by the swim-up method, and diluted in Ham's F10 medium supplemented with 10% patients' serum. No more than three oocytes and approximately 100 000 motile spermatozoa were transferred, in a volume of 40 /tl, into one Fallopian tube with 1236

the inner catheter of a TDT embryo transfer catheter (Prodimed, Neuilly-en-Thelle, France). The catheter was inserted into the tube through a curved metal outer catheter with which the ampulla was cannulated (Cook Europe, Eindhoven, The Netherlands). If transfer failed, an FVF procedure followed. The same Ham's F10 medium supplemented with 10% patients' serum was used for embryo culture and transfer. Embryo transfer (in 30 /tl) followed on the 2nd or 3rd day. Supernumerary oocytes were also fertilized in vitro and the embryos, if possible, cryopreserved for replacement in a natural cycle. The luteal phase of GIFT or IVF cycles was supported by HCG, 5000 IU on the day of follicle aspiration, and if no signs of imminent ovarian hyperstimulation syndrome were present, again 3 and 6 days later. In ovarian stimulation cycles, luteal support was only given if buserelin was used in the follicular phase. Statistical analysis was by Fisher's exact test, chi-square test and Mest where appropriate.

Results The 50 patients (Table I) had a total of 87 treatment cycles leading to 11 clinical pregnancies, with a mean success rate per cycle of 12.6%. An additional pregnancy was established in one of 11 cryo-transfer cycles, resulting from supernumerary oocytes of the GIFT cycles. This pregnancy was not included in the statistical analysis. In 43 initial GIFT cycles seven pregnancies occurred (16.3%), and in 44 ovarian-stimulation cycles four pregnancies occurred (9.1 %), but in five of the GIFT cycles an opportunity for IVF was taken, leading to two pregnancies. Therefore, in 38 completed GIFT cycles, five pregnancies were established (13.2%). During the study period, the IVF programme also had a clinical pregnancy rate of 13.3% per puncture. No significant difference in the mean success rate per cycle was found between GIFT and ovarian stimulation, counting either all 43 initial GIFT cycles or 38 completed cycles. By random selection, 26 patients had been treated with GIFT, and 24 with ovarian stimulation. The mean age of the 26 GIFTtreated patients was 33.5 ± 3.8 years and of the 24 ovarian stimulation patients 31.8 ± 4.5 years; these were not significantly different. In the first GIFT cycles (n = 26) four (15.4%) pregnancies occurred; in three of these cycles a change to IVF was made, leading to one pregnancy. Five patients then ceased or chose a different form of treatment. In the remaining second cycles (n = 17), IVF was used twice and three (17.6%) patients became pregnant, one of whom had undergone IVF. In the first ovarian stimulation cycles {n = 24), two (8.3%) pregnancies were established. Two patients did not continue treatment after

Table I. Number and category of patients in each treatment group

GIFT Ovarian stimulation Total

Unexplained infertility

Failed AID

TotaJ

14 9 23

12 15 27

26 24 50

GIFT = gamete intra-Fallopian transfer. AID = artificial insemination by donor.

GIFT or ovarian stimulation in unexplained infertility

the first cycle. In the remaining second ovarian stimulation cycles (n = 20), again two (10%) patients became pregnant, bringing the total to four pregnancies out of 44 ovarian stimulation cycles (9.1%) (Table IT). Since the patients with unexplained failure of AID possibly represented a more fertile subgroup than the patients with 3 years of unexplained infertility and normal partner, the results of these subgroups were then compared. From the 23 unexplained infertility patients, 14 were treated by GIFT and nine by ovarian stimulation. From the 27 patients with unexplained failure of AID, 12 had GIFT and 15 ovarian stimulation. There was no significant difference in age between the two groups, the mean age of the unexplained infertility patients being 31.5 ± 3.8 years, and the patients with unexplained failure of AID 33.6 ± 4.4 years. The 23 patients with unexplained infertility had 39 treatment cycles and four pregnancies (10.3%). The 27 patients with failed AID had an overall total of 48 treatment cycles, from which seven pregnancies resulted (14.6%). The difference in pregnancy rate was not significant.

GIFT and ovarian stimulation between unexplained infertility patients and patients with unexplained failure of AID. As for the outcome of the 11 clinical pregnancies, two ended in a spontaneous abortion, six were ongoing singleton, and three were twin pregnancies. All three twin pregnancies were in failed AID patients, of which two had occurred after ovarian stimulation. The 12th pregnancy, which resulted from transfer of two cryopreserved embryos, proceeded to ongoing singleton. All ongoing pregnancies were delivered and are well (Table HI). Discussion In this study, we found no significant difference in the pregnancy rate between GIFT and ovarian stimulation for the treatment of unexplained infertility. Apparently, the intra-Fallopian transfer of gametes did not add significantly to the already accomplished effect of ovarian stimulation alone, which happens to be an obligatory preamble in the follicular phase of all GIFT programmes. We know of no other prospective study that has compared these two components of GIFT within the same setting. The pregnancy rate for GIFT (13.3%) in this study does not seem to equal results that are commonly reported. For instance, the success rate for GIFT in a large multicentred survey in the USA has risen from a 25% clinical pregnancy rate in 1968 retrievals in 1987 (Medical Research International, National IVF-ET Registry, 1989), to 30% in 3652 retrievals in 1989 (Medical Research International IVF-ET Registry, 1991). Two possible explanations for this difference might be the definition of unexplained infertility, and the GIFT method. Concerning the definition, at the onset of the trial we expected no difficulty in

The results for the two subgroups were then considered according to type of treatment. The 23 patients with unexplained infertility had 24 GIFT cycles, of which three converted to IVF (no pregnancies); the 21 completed GIFT cycles led to two pregnancies (9.5%), while 15 ovarian stimulation cycles also led to two pregnancies (13.3%). The 27 failed AID patients had 19 initial GIFT cycles and five pregnancies (26.3%), but two converted to IVF (with two clinical pregnancies); thus 17 completed GIFT cycles led to three pregnancies (17.3%), while 29 ovarian stimulation cycles resulted in two pregnancies (6.9%). Statistical analysis showed no difference in the success rate for

Table n . All 87 treatment cycles in the two treatment groups, per category of patient

GIFT GIFT-rVF* GIFT GIFT-IVF*

1st 1st 2nd 2nd

Total Ovarian stimulation

1st 2nd

Total

Unexplained infertility

Total

Failed AID b

All

(Pregn.)"

All

(Pregn.)

11 2 10 1

(1)

12 1 5 1

SSSS

Cycle

23(3) 3(1) 15(2) 2(1)

24

(2)

19

(5)

43(7)

9 6

(1) (1)

15 14

(1) (1)

24(2) 20(2)

15

(2)

29

(2)

44 (4)

(1)

•GIFT cycles in which a change to IVF was made. ^Cycles leading to pregnancy are shown in parentheses.

Table ID. Outcome of 12 clinical pregnancies Total

Spontaneous abortion Singleton ongoing Twin ongoing Total

2 7* 3 12

GIFT

Ovarian stimulation

Unexplained infertility

Failed AID

1 6" 1 8

•One pregnancy after transfer of cryopreserved embryos from a previous GIFT cycle.

1237

H.V.HogeraU ct al.

finding at least 100 participants and 160 treatment cycles, since - 4 0 0 new female patients register each year at the infertility clinic. Yet, when scrutinizing eligible patients for the strict criteria of unexplained infertility, this diagnosis became a rare event. Even though 30% of suitable patients did not participate, the prevalence of unexplained infertility (excluding failed AID) in the out-patient population came to a surprisingly low 5%. In contrast, Hull et al. (1985) found a prevalence of 9.5% of 3 or more years' duration, but allowed minor abnormalities such as mild oligozoospermia, and minor adhesions to be included. In general, the definitions of unexplained infertility are highly variable, and sometimes not even mentioned. For instance, Trounson et al. (1980), Templeton and Penney (1982), and Audibert et al. (1989) demand a minimum period of only 2 years' unexplained infertility; Tanbo et al. (1990) accept 10 x lO^ml spermatozoa as the lower limit for sperm counts; the US (National) IVF-ET Registry (Medical Research International 1988, 1991) singles out unexplained infertility as an indication for GIFT, but does not give a definition. This makes results hard to compare, and we believe that our study population with a minimum of 3 years' infertility and (apart from all other parameters) absence in both partners of auto-immune spermatozoal antibodies, represents a less favourable selection, which may be reflected in the GIFT results. One is tempted to conclude that the more infertility is truly unexplained, the less effective are GIFT and ovarian stimulation. Concerning the GIFT method, we had > 3 years' experience with laparoscopic follicle aspiration for IVF but relatively little experience with GIFT when the study started, having previously initiated a total of three GIFT pregnancies. This may be reflected in the GIFT results of this study and probably accounts for the necessity to turn to IVF in some cases after follicle aspiration because of difficulty in cannulating the Fallopian tube. Of more importance, however, was the maximum number of oocytes transferred to the tube, which was limited to three by the study protocol. This necessarily restricted the results but also reduced the risk of multiple pregnancy. We had a clinical pregnancy rate per GIFT/rVF cycle of 16.3% and per completed GIFT cycle of 13.3%. Craft et al. (1988) found a 13.9% clinical pregnancy rate when one to two oocytes were transferred, and 24.7% if three to four oocytes were replaced. Yeeetal. (1989) had 18.8% success rate if two oocytes were replaced and 23.2% for three oocytes. Yovich et al. (1988a) had 16.6% pregnancy rate with two oocytes replaced and 27.3% with three oocytes. All authors found higher results if more than three oocytes were transferred. On average therefore, one might expect a 23 % - 2 7 % pregnancy rate, when replacing three oocytes. However, none of the above authors specified the duration of unexplained fertility in their patients, and only one specified the nature of the infertility per patient, which again makes it difficult to compare their results to those of the present study. The power of this study is expressed by the 95% confidence interval of the observed difference (4.1 %) between GIFT (13.2%) and ovarian stimulation (9.1 %) pregnancy rates, which runs from - 9 . 6 % to 17.8%. The possibility of a type II error, i.e. 'GIFT results do not significantly differ from ovarian stimulation in the treatment of unexplained infertility, whereas in trum they do', is present within these limits (Detsky and Sackett, 1985). 1238

Despite the shorter duration of female unexplained infertility, we have included in the study population patients with failed AID, who otherwise fulfilled the unexplained infertility criteria. Indeed, failed AID patients had been unsuccessfully exposed to donor semen for a minimum of only 12 ovulatory cycles, as opposed to the minimal 3 years exposure to their partner's semen in unexplained infertility patients, and could therefore be expected to have better pregnancy rates. However, since these patients were stratified before randomization, equal numbers of failed AID patients could be expected in both treatment categories and therefore would make equal contributions to the pregnancy rates, thus maintaining the validity of the comparison between the two treatment modalities. The effective cumulative success rate of our AID programme is - 5 5 % (Hogerzeil etal., 1988), after 12 well-timed ovulatory insemination cycles. By that time some 25% of the initial patients have dropped out for various reasons, and 20% are not pregnant, numbering yearly - 15 women. Unlike unexplained infertility patients, these failed ADD patients have very little remaining chance of pregnancy, as our AID programme usually has to finish after 12 insemination cycles. The Federation CECOS (1989) estimates the mean success rate per cycle between the 12th and 24th donor insemination cycle at 4.5%, and therefore recommends assisted reproduction, i.e. IVF, as the treatment of choice from the 18th insemination cycle onwards. Others have successfully treated failed AID with GIFT, sometimes even before trying cervical insemination (Formigli et al., 1990), or after failure of nine to 24 cycles (Cefalu et al., 1988). For that reason, it seemed appropriate to offer the failed AID patients a further chance by an assisted reproductive procedure in this trial. The overall pregnancy rate per cycle for failed AID patients (14.6%) was, as expected, slightly higher than for unexplained infertility patients (10.3%), and more than three times the mean success rate per cycle of 4.5% in the CECOS study; the results confirm their recommendation to make use of assisted reproduction in these cases. We found a 26.3% pregnancy rate per GIFT/rVF cycle in the failed AID group, and 6.9% per ovarian stimulation cycle. Thus, here we do see a tendency towards a better performance for GIFT/TVF (P < 0.08). This again suggests that female infertility of short duration is readily treated by GIFT (but may not need treatment at all, if time for natural fecundation is available), whereas long standing unexplained infertility does not respond to GIFT all that easily. On the other hand, ovarian stimulation seemed to give rather poor results in the failed AID subgroup. This may be due to the relative paucity of available sperm cells, since the contents of only one thawed cryopreserved straw (2.5 x 106 motile spermatozoa) were used for intracervical insemination, as opposed to the full ejaculate in timed intercourse for unexplained infertility patients. Pregnancy rates in AID cycles, using cryopreserved semen, tend to drop behind those where a full fresh ejaculate is used (Richter et al., 1984). The mean success rate per cycle in the first six AID cycles of the Federation CECOS (1989) study is only 10.3%, using intracervical insemination of one thawed cryopreserved straw, as opposed to an estimated 25 % in natural cycles of fertile couples. Only when high doses (40 x 106) of thawed cryopreserved motile spermatozoa are used (Bordson et al., 1986) are results for fresh and frozen spermatozoa

GIFT or ovarian stimulation in unexplained infertility comparable. Apparently, if using ovarian stimulation in failed ADD patients, a quantity of motile spermatozoa similar to that of a fresh ejaculate should be used. The success of ovarian stimulation is primarily due to the higher number of available oocytes and possibly the solving of occult ovulation disturbances. The method is simpler and cheaper than procedures that require oocyte collection. However, higher numbers of oocytes introduce higher risks of multiple pregnancy, and the management of the follicular phase in ovulation induction is not always easy. In this study, two of the four pregnancies induced by ovarian stimulation were twins, although HCG was withheld if more than four follicles with a minimum diameter of 17 mm were seen; conversely, only one twin pregnancy resulted out of seven GIFT/IVF pregnancies. The number of multiple pregnancies in The Netherlands has risen steeply in recent years, of which 50% are attributed to ovulation induction in non-invasive assisted reproductive techniques. Therefore in future it may be wise to use adapted ovarian stimulation regimes to prevent unwanted multiple pregnancies. Although the power of this study is limited by its sample size, we have demonstrated that ovarian stimulation with timed intercourse is a successful alternative to GIFT in the treatment of unexplained infertility. It is clear that the success of assisted reproductive methods in this area which include tubal oocyte transfer is at least in part, if not fully, caused by ovarian stimulation leading to higher numbers of available oocytes. Since GIFT is an elaborate procedure, routinely requiring full anaesthesia, we advocate ovarian stimulation as precursor treatment in unexplained infertility, although care should be taken to avoid multiple pregnancies. In failed AID, however, ovarian stimulation seems to be less useful if cervical insemination of only one thawed cryopreserved semen straw is used, probably because of low concentration. As the results of GIFT in our hands did not surpass those of FVF, we now prefer FVF, both for the additional information on the fertilization process and the out-patient character of the procedure.

Acknowledgements This study was made possible through a grant by Organon Nederland B.V. Oss, The Netherlands. References Audibert.F., Hedon.B., Amal.F., Humeau.C, Badoc.E., Virenque,V., Boulot.P., Mares.P., Laffargue.F. and VialaJ.L. (1989) Results of FVF attempts in patients with unexplained infertility. Hum. Reprod., 4, 766-771. Braeckmans.P., Devroey.P., Camus.M., Khan.I., Staessen.C, Van Waesberghe,L., Wisanto.A. and Van Steirteghem.A.C. (1987) Gamete intra-Fallopian transfer: evaluation of 100 consecutive attempts. Hum. Reprod., 2, 201-205. Bordson.B.L., Ricci.E., Dickey,R.P., Dunaway.H., Taylor.S.N. and Curole,D.N. (1986) Comparison of fecundabflity with fresh and frozen semen in therapeutic donor insemination. FertiL Steril., 46, 466—469. Cefalu,E., Cittadini.E., Balmaceda^I.P., Guastella.G., Ord.T., Rojas.F. and Asch.R.H. (1988) Successful gamete intrafallopian transfer following failed artificial insemination by donor: evidence for a defect in gamete transport. Fertil. Steril., 50, 279-282.

Corsan.G.H. and Kemman.E. (1991) The role of superovulation with menotropins in ovulatory fertility: a review. Fertil. Steril., 55, 468-477. Craft.I., Al-Shawaf,T., Lewis.P., Serhal,P., Simons,E., Ah-Moye,M., Fiamanya.W., Robertson.D., Shrivastav.P. and Brindsen.P. (1988) Analysis of 1071 GIFT procedures—the case for a flexible approach to treatment. Lancet, 1, 1094-1097. Detsky.M.S. and Sackett.D.L. (1985) When was a 'negative' clinical trial big enough? Arch. Intern. Med., 145, 709-712. Federation CECOS, Le Lannou.D. and Lansac.J. (1989) Artificial procreation with frozen donor semen: experience of the French Federation CECOS. Hum. Reprod., 4, 757-761. Formigli.L., Coglitore.M.T., Roccio.C, Beloti.G., Stangalini.A. and Formigli.G. (1990) One-hundred-and-six gamete intra-Fallopian transfer procedures with donor semen. Hum. Reprod., 5, 549—552. Hogerzeil.H.V., Hamerlynck.J.V.Th.H., Amstel,v.,N., Nagelkerke.N.J.D. and Lammes.F.B. (1988) Results of artificial insemination at home by the partner with cryoprcserved donor semen: a randomized study. Fertil. Steril., 49, 1030-1035. Hull.M.G.R., Glazener,C.M.A., Kelly,N.J., Conway.D.I., Foster.P.A., Hinton.R.A., Coulson.C, Lambert.P.A., Watt.E.M. and Desai.K.M. (1985) Population study of causes, treatment, and outcome of fertility. Br. Med. J., 291, 1693-1697. Kurzrok,R. and Miller.E.G. (1932) Biochemical studies of human semen. HI. Factors affecting migration of sperm through the cervix. Am. J. Obstet. Gynecol., 24, 19-23. LeetonJ., Rogers.P., Caro.C, Healy.D. and Yates,C. (1987) A controlled study between the use of gamete intrafallopian transfer (GIFT) and in vitro fertilization and embryo transfer in the management of idiopathic and male infertility. Fertil. Steril., 48, 605—607. Medical Research International and the Society of Assisted Reproductive Technology, The American Fertility Association (1989) In vitro fertilization/embryo transfer in the United States: 1987 results from the National IVF-ET Registry. Fertil. Steril., 51, 13-19. Medical Research International, Society of Assisted Reproductive Technology, and The American Fertility Society (1991) In vitro fertilization-embryo transfer (TVF-ET) in the United States: 1989 results from the IVF-ET Registry. Fertil. Steril., 55, 14-23. Richter,M.A., Haning.R.V. and Shapiro.S.S. (1984) Artificial donor insemination: fresh versus frozen semen; the patient as her own control. Fertil. Steril., 41, 277-280. Tanbo.T., Olav.P.O. and Abyholm.T. (1990) Assisted fertilization in infertile women with patent Fallopian tubes. A comparison of in-vhro fertilization, gamete intra-Fallopian transfer and tubal embryo stage transfer. Hum. Reprod., 5, 266-270. Templeton^A.A. and Penney,G.C. (1982) The incidence, characteristics, and prognosis of patients whose infertility is unexplained. Fertil. Steril., 37, 175-182. Trounson.A.O., LeetonJ.F., Wood.C, WebbJ. and Kovacs.G. (1980) The investigation of idiopathic infertility by in vitro fertilization. FertiL Steril., 34, 431-438. Yee.B., Rosen.G.F., Chacon.R.R., Soubra.S. and Stone.S.C. (1989) Gamete intrafallopian transfer: the effect of the number of eggs used and the depth of gamete placement on pregnancy initiation. Fertil. Steril., 52, 639-644. Yovich.J.L., Matson.P.L., Blackledge.D.G., Turoer.S.R., Richardson.P.A., YovichJ.M. and Edirisinghe.W.R. (1988a) The treatment of normospermic infertility by gamete intrafallopian transfer (GIFT). Br. J. Obstet. Gynaecol., 95, 361-366. Yovich,J.L., YovichJ.M. and Edirisinghe.W.R. (1988b) The relative chance of pregnancy following tubal or uterine transfer procedures. Fertil. Steril., 49, 858-864. Received on February 14, 1992; accepted on July 1, 1992

1239