Department of Gynaecology and Reproductive Medicine, University Hospital, University of Western Ontario, London, Ontario, Canada

Transcription

1 FERTILITY AND STERILITY Copyright 1992 The American Fertility Society Printed on acid-free paper in U.S.A. Randomized, prospective comparison of luteal leuprolide acetate and gonadotropins versus clomiphene citrate and gonadotropins in 408 first cycles of in vitro fertilization*t Ian S. Tummon, M.D.:j: Susan A. J. Daniel, Ph.D. Brian R. Kaplan, M.B., Ch.B. Jeffrey A. Nisker, M.D. A. Albert Yuzpe, M.D. Department of Gynaecology and Reproductive Medicine, University Hospital, University of Western Ontario, London, Ontario, Canada Objective: To compare luteal phase leuprolide acetate (LA) initiated pituitary down regulation followed by human menopausal gonadotropins (hmg) versus clomiphene citrate (CC) and hmg for follicular recruitment and oocyte maturation before in vitro fertilization (IVF). Design: Randomized, prospective comparison in first cycles of IVF. Setting: University Hospital, a tertiary referral center offering assisted reproductive technologies. Participants: Participants were couples undergoing their first ever cycle of IVF and consenting to participation in the trial Results: Luteal phase initiated LA/hMG was associated with a lower probability of cycle cancellation, improved folliculogenesis, and a higher probability of embryo transfer (ET) compared with CC/hMG alone. Implantation rates were not different. Conclusion: A higher rate of ET with LA/hMG suggests that gonadotropin-releasing hormone agonist for the induction of folliculogenesis before IVF may be appropriate. Fertil Steril 1992;58:563-8 Key Words: Randomized, in vitro fertilization, luteal, leuprolide acetate, clomiphene citrate, cancellation, folliculogenesis, embryo transfer, pregnancy Considerable enthusiasm has developed for the use of gonadotropin-releasing hormone agonist (GnRH-a) as adjuncts to controlled ovarian stimulation in assisted reproductive technologies. Justification for their use remains disappointingly empirical and anecdotal (1). The rationale for cotreatment with GnRH-a includes development of Received March 2, 1992; revised and accepted May 20, * Presented in part at the 47th Annual Meeting of the Society of Obstetricians and Gynaecologists of Canada, Toronto, Ontario, Canada, June 11 to 15, t Supported by a grant from Abbott Laboratories Limited, Montreal, Quebec and Takeda Abbott Pharmaceuticals, North Chicago, Illinois. :j: Reprint requests: Ian S. Tummon, M.D., Department of Gynaecology and Reproductive Medicine, University Hospital, 339 Windermere Road, London, Ontario, Canada N6A 5A5. a more uniform cohort of follicles, suppression of tonic increases of luteinizing hormone (LH), and prevention of premature LH surges. Reports comparing clomiphene citrate (CC)/human menopausal gonadotropins (hmg) with GnRH-a/hMG confirm a reduction in the incidence of LH surges (2, 3). Less clear-cut are differences in folliculogenesis and fertilization as well as implantation rates between the two groups. Compared with CC/hMG, GnRH-a/ hmg improved folliculogenesis for the poor responder (2), but no improvement was demonstrated for the unselected first cycle patient (3). Meta-analysis of comparisons of ovarian stimulation protocols is handicapped by variations in specific protocols, and conclusions need to be drawn with caution (4). The present study is a prospective, randomized comparison of luteal phase leuprolide acetate (LA) / Tummon et al. Luteal LA versus CC in IVF 563

2 hmg and CC/hMG during first cycles of in vitro fertilization (IVF). Subject Selection MATERIALS AND METHODS Between September 1989 and June 1991, 508 couples undergoing their first ever IVF cycle voluntarily consented to participate in this study in the Department of Gynaecology and Reproductive Medicine at University Hospital. Two hundred seventy-five women were randomized to receive CC (Serophene; Serono Canada, Mississauga, Ontario, Canada)/hMG, whereas 233 women were randomized to treatment with LA (Lupron; Abbott Laboratories, Montreal, Quebec, Canada) before hmg stimulation. Those initiating cycles were 229 women in the CC/hMG group and 179 in the LA/hMG group. Thus 17% of the CC/hMG group did not begin treatment in contrast to 23% of the LA/hMG group (P = 0.08). Couples not beginning treatment did so for a wide range of personal reasons. The study design was approved by the Review Board for Health Sciences Research Involving Human Subjects of the University of Western Ontario. Randomization was performed by simple randomization using a sequence of randomized numbers. The minimum sample size estimate, based on 80% power and 5% level of significance and assuming a 15% difference in pregnancy rates (PRs), was 270 couples. Couples were referred for the usual IVF indications, including tubal disease (n = 188), endometriosis (n = 54), idiopathic infertility (n = 49), male factor (22), or multifactorial abnormalities (95) (5). In vitro fertilization was not offered if <10 5 motile spermatozoa were obtained after sperm processing (6). In Vitro Fertilization Patients were randomized to receive ovarian stimulation with one of two protocols. The first protocol consisted of a combination of CC 100 mg orally during cycle days 5 to 9 of the treatment cycle and hmg (Pergonal; Serono Canada) one ampule per day intramuscularly (1M), beginning on cycle day 6. In the second protocol, subjects received hmg alone after prior pituitary deregulation from the midluteal phase of the previous cycle using LA, 0.5 mg subcutaneously (SC) beginning day 22 of the menstrual cycle preceding IVF stimulation and continuing through the entire period of hmg therapy. The initial dose ofhmg in the LA/hMG group was adjusted according to body weight: <52 kg one ampule per day; 52 to 76 kg one and a half ampules per day; > 76 to 88 kg two ampules per day; >88 kg three ampules per day. According to individual estradiol (E2) response, the daily dosage ofhmg was modified. The laboratory aspects of IVF were performed as previously described (6). Up to five embryos were transferred at the time of embryo transfer (ET). Follicular Phase Monitoring and Luteal Support Daily monitoring of serum E2 and follicular diameters as determined by vaginal ultrasound (US) were performed as previously described (7). Final oocyte maturation was induced using 5,000 IU 1M of human chorionic gonadotropin (hcg, Profasi; Serono Canada). Luteal support with oral micronized progesterone 200 mg four times daily was administered in both treatment groups (8). Cancellations Reasons for cycle cancellations were considered in four categories: [1] elevated endogenous LH (>25 U/L for 3 consecutive days); [2] spontaneous LH surges (>50% increase over baseline; [3] abnormal E2 response (levels increased < 1 during any 72 hours of stimulation or >1 drop on the day of hcg administration or E2 exceeding 12,000 pmoljl); and [4] inappropriate follicular growth «2 follicles> 1.6 cm at hcg administration). Pregnancy Pregnancy was defined as a positive serum {3-hCG value obtained 18 days after oocyte retrieval. An active fetal heart within an intrauterine gestational sac as determined by vaginal US examination 40 days after oocyte retrieval was considered positive evidence for implantation. Data Analysis Demographic characteristics and mean cycle outcomes were compared between treatment groups using Student's t-tests. Diagnostic categories, categories of reasons for cancellation, and treatment outcome between groups were compared using X2 tests. Comparison of cancellation, retrieval, transfer, and pregnancy rates (PRs) between treatments were examined using calculations of relative risks (RRs). Binomial regression was used to estimate RRs adjusted simultaneously for various covariates that in- 564 Tummon et a1. Luteal LA versus CC in IVF Fertility and Sterility

3 cluded age (in 4 categories: <25, 25 to 30, 30 to 35, and >35 years), duration of infertility (in 3 categories: <3, 3 to 5, and >5 years), and diagnosis (tubal, endometriosis, idiopathic, multifactorial). In all analyses, P < 0.05 was considered significant. Results are presented as means ± SD. Down Regulation Phase RESULTS In the LA/hMG group, ovarian suppression (serum E2 < 73 pmoljl) was achieved at 0.5 mg in 133 (74%) in 12.4 ± 2.9 days. In the remaining 46 patients, administration was increased to 1 mg/d. Appropriate suppression was achieved during the next 10 days in 35 subjects. The mean time to suppression in the subjects requiring an increase in LA was 19.7 ± 4.5 days. Eleven subjects (6%) remained unsuppressed, and treatment was discontinued. Stimulation Phase At hcg administration, serum E2 and number of follicles > 1.5 cm in greater diameter were greater in the LA/hMG group, as shown in Table 1. Comparable General Characteristics of the Groups The age of the female partner, duration of infertility, and other general characteristics were not different between groups (Table 1). There was a single principal diagnosis in 181 couples (79%) in the CC/ hmg group and 132 couples (73%) in the LA/hMG. The distribution of diagnoses was not different between groups (x 2 = 2.27, P = 0.81). Risk of Cancellation Compared with CC /hmg, the risk of cancellation with LA/hMG was 46% that of the risk of cancellation with CC/hMG. Because the 95% confidence interval (CI) for the RR does not contain 1, the risk of cancellation for LA/hMG was lower than that for CC/hMG. Examination of the reasons for cancellation revealed differences in the categories responsible for cycle cancellation. In the LA/hMG group, there were no cancellations for tonic increases in LH or premature LH surges. Eighteen of the 179 cycles (10.1%) were canceled because of abnormal E2 or follicular responses. In contrast, 31.6% of CC/ hmg cycles were canceled. Ten CC/hMG cycles (4.4 %) were canceled because of tonic elevation of LH, whereas 42 (18.4%) were canceled because of a premature rise in LH, and 18 (7.8%) were canceled because of abnormal E2 or follicular responses (x 2 = 32.0, P = 0.001). With CC/hMG, the overall cancellation rate was 30.8%. Oocyte Recovery, Fertilization, ET, and PRs A summary of outcomes of treatment groups is illustrated in Figure 1. Retrieval rates with LA/hMG (89.9%) were higher (P < 0.001) per cycle initiated Table 1 Demographic Characteristics, Responses, and Treatment Outcomes* LA/hMG CC/hMG Probability Age (y) Weight (kg) Gravidity Parity No. of ovaries No. of patent tubes Menstrual cycle (d) Infertility (y) Stimulation (d) HMG (no. of ampules) Serum E2 on day of hcg (pmol/l) Follicles> 1.5 cm on day of hcg Oocytes/ cycle Oocytes/retrieval Zygotes/cycle Zygotes/retrieval Embryos/cycle Embryos/retrieval 31.6 ± ± ± ± ± ± ± ± ± ± 9.4 6,327 ± 2, ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3.0 4,689 ± 2, ± ± ± ± ± ± ± <0.002 * Values are means ± SD. Tummon et al. Luteal LA versus CC in IVF 565

4 ~.OOOl... ~r Ill!,,","ouO</",,,, 0-0, I/) w ~ " CC/hMG 0-0 ~ 0 INITIATION RETRIEVAL TRANSFER PREGNANCY Figure 1 Outcomes of treatment groups. than with CC/hMG (68.4%). Similarly, transfer rates with LA/hMG (57.5%) were higher (P < 0.001) per cycle initiated than with CC/hMG (45.4%). Pregnancy rates were 10.7% and 9.2% in the LAI hmg and CC/hMG groups, respectively, per stimulation initiated (P = 0.89). Retrieval, transfer, and PRs were not different when only uncanceled cycles were considered. Similarly, both crude and adjusted PRs of pregnancy were not different. (Table 2) In the CC/hMG group, the overall fertilization rate was 36.3% of oocytes compared with 41.1 % in the LA/hMG group (P = 0.049). Fertilization of one or more oocytes was obtained in 71.5% of CC/hMG cycles compared with 73.5% in the LA/hMG group, a proportion not different between groups (P = 0.66). The number of oocytes was correlated with serum E2 on the morning of hcg administration in both the CC/hMG and LA/hMG groups, respectively, (r = 0.64, P = and r = 0.53, P = 0.001). Overall, serum E2 on the day of hcg administration was higher with LA/hMG. Controlling for serum E2 the number of oocytes was compared between groups. In the serum E2 stratum between 7,000 to 7,999 pmol/l, the number of oocytes in the LA/hMG group was 2.33 ± 2.32, higher than the CC/hMG group, 1.78 ± 1.66 (P = 0.04). Controlling for the number of embryos transferred, there was no difference in the PR between stimulation protocols. For transfer of more than three embryos, pregnancy occurred in 12 of 43 (27.9%) LA/hMG cycles and 9 of 19 (47.4%) CCI hmg cycles (x 2 = 1.44, P = 0.22). For transfer of three embryos, pregnancies occurred in 4 of 26 (15.4%) LA/hMG cycles and 4 of 22 (18.2%) CCI hmg cycles (x 2 = 0.16, P = 0.90). If two embryos were transferred, pregnancies occurred in 1 of 15 (6.7%) and 4 of 26 (15.4%) LA/hMG and CC/hMG cycles, respectively, (X 2 = 0.20, P = 0.87). For a single embryo transferred, results were pregnancies in 1 of 19 (5.3%) and 4 of 34 (11.8%) LA/hMG and CCI hmg cycles, respectively, (x 2 = 0.030, P = 0.86). In contrast, the PR was related to the number of embryos transferred X2 = 6.5, P = 0.08 for the LAI hmg group and X2 = 12.2, P = for the CCI hmg group. Per transfer, PRs were not different between groups: 17.5% and 19.2% in the LA/hMG and CC/hMG groups, respectively, (x 2 = 0.01, P = 0.97). There was no difference in implantation rates per embryo between groups: 19 of 228 (8.3%) transferred embryos in the CC/hMG group and (6.4%) in the LA/hMG group (P = 0.36). DISCUSSION This study, a randomized, prospective comparison of 408 first IVF cycles, clearly demonstrates a reduced risk of cancellation with luteal phase initiated LA/hMG compared with CC/hMG. Adjusted for demographic covariates relating to fecundity, the RR of cancellation with LA/hMG is 44% that of CCI hmg. This result confirms earlier reports of a reduction in cancellation rates for IVF cycles using GnRH-a/hMG versus CC/hMG (2, 3). In one report, no difference was found; however, a small sample size limited statistical power (9). In the present study, the difference in cancellation rates between treatment groups is because of abnormalities of LH associated with CC/hMG. With LA/hMG, 10% of cycles were canceled, all because of abnormalities in E2 response or folliculogenesis, whereas with CCI hmg, 31 % of cycles were canceled, 23% because of LH abnormalities and 8% because of abnormalities in E2 response or folliculogenesis. This finding is similar to that of Kubik and co-workers (3) who compared CC/hMG with luteal LA/hMG in 60 unselected women undergoing their first cycle of IVF or gamete intrafallopian transfer. In contrast, they Table 2 Crude and Adjusted RRs: LA/hMG versus CC/hMG Crude RR cancellation/cycle Crude RR retrieval/cycle Crude RR transfer/cycle Crude RR pregnancy/cycle Adjusted RR cancellation/cycle Adjusted RR % retrieval/cycle Adjusted RR transfer/cycle Adjusted RR pregnancy/cycle RR % CI 0.31, , , , , , , , Tummon et a1. Luteal LA versus CC in IVF Fertility and Sterility

5 demonstrated no improvement in folliculogenesis, with similar numbers of mature oocytes recovered using both protocols, whereas we have demonstrated an increase in oocyte numbers with LA/hMG. This apparent discrepancy may relate to Kubik's protocol using a uniform definition of follicle maturity (3). In the present study, overall oocyte numbers are higher using GnRH -a. Even controlling for E2 response and comparing oocyte numbers in the E2 range from 7,000 to 7,999 pmoljl, oocyte numbers were higher with GnRH -a. The proportion of oocytes that fertilized was higher with LA/hMG, compared with CC/hMG. Thus it appears that LA/hMG treatment is associated with either improved oocyte quality or more mature oocytes. Likely as a consequence of more fertilized oocytes, the present study also demonstrates a significantly greater chance of achieving ET with GnRH-a. Adjusted for demographic covariates relating to fecundity, the PR of having an ET per cycle initiated with LA/hMG is 128% that of CC/hMG. Because the CI of this calculation does not include 1, it is concluded that this difference is significant. Both treatments resulted in similar implantation rates, but it is not known whether any changes in conceptus quality and uterine receptivity might be induced by the different treatment protocols. Clomiphene citrate, but not GnRH-a, has been linked with a disordered endometrial response involving a direct effect ofthe drug on the endometrium (10). In the present study, PRs were directly related to the number of embryos transferred. Controlling for the number of embryos transferred, there was no difference in PRs between groups. It is concluded that use of LA/hMG offers the distinct advantage of fewer canceled cycles resulting in a higher chance of achieving ET. The final decision about which superovulation protocol to use will depend on many factors. Use of luteal LA/hMG is associated with extended ovarian stimulation and increased hmg requirements. Drug costs for GnRH -a as well as gonadotropins are higher with LA/hMG compared with CC/hMG. With LA/hMG treatment, the average cost for hmg was more than $1,000 greater than in the CC/hMG group. With this formulation of LA/hMG, there is need for a SC injection, and fear of SC administration may have influenced some women not to use GnRH-a. Additionally, use of GnRH-a requires the cost of drugs for luteal support, although for many centers this step is routine. There are little data whereby the optimum dose of GnRH-a has been ascertained (11). Successful results have been reported with a step-down dosage of LA (0.5 to 0.25 mg SC) initiated in the follicular phase (12), but most reports employing LA/hMG have adopted a higher stepdown regimen (1.0 to 0.5 mg SC) (13). From a GnRH-a cost perspective, the 0.5-mg dosage of LA employed in the present study is advantageous because in 74% of cycles only a single vial of LA was required. The use of CC/hMG offers advantages of simplicity and reduced pharmaceutical costs. If no abnormalities in LH occur, the cycle has similar results to LA/hMG. Use of GnRH-a/hMG will be more attractive when cryopreservation is effective and greater advantage can be taken of increased folliculogenesis. On the other hand, if there is no underlying tubal disease, it may be possible to successfully convert CC/hMG IVF cycles to intrauterine insemination (14). The major disadvantages of CC/hMG are the unpredictability of LH abnormalities and the financial and emotional distress resulting from cycle cancellation. Acknowledgments. The authors are grateful for the assistance of Jeanne Adam, M.Sc., Senior Biostatistician, Abbott Laboratories, Montreal, Quebec, with statistical analysis, Catherine B. Benson, R.N., University Hospital, with data collection and James E. Brown, M.D., University Hospital, University of Western Ontario, London, Ontario, for pharmacology expertise. REFERENCES 1. Maroulis GB, Emery M, Verkauf BS, Saphier A, Bernhisel M, Yeko MR. Prospective randomized study of human menotropin versus a follicular and a luteal phase gonadotropinreleasing hormone analog-human menotropin stimulation protocols for in vitro fertilization. Fertil Steril1991;55: MacLachlan V, Besanko M, O'Shea F, Wade H, Wood C, Trounsoncouples A, et al. A controlled study of luteinizing hormone-releasing hormone agonist (buserelin) for the induction of folliculogenesis before in vitro fertilization. N Engl J Med 1989;320: Kubik CG, Guzick DS, Berga SL, Zelenik AJ. Randomized, prospective trial of leuprolide acetate and conventional superovulation in first cycles of in vitro fertilization and gamete intrafallopian transfer. Fertil Steril 1990;54: Meldrum DR. Editorial response. J In Vitro Fert Embryo Transf 1990;7: Yuzpe AA, Brown SE, Casper RF, Nisker JA, Graves G. Rates and outcomes of pregnancies achieved in the first 4 years of an in-vitro fertilization program. Can Med Assoc J 1989;140: Tummon IS, Yuzpe AA, Daniel SAJ, Deutsch A. Total acrosin activity correlates with fertility potential after fertilization in vitro. Fertil Steril 1991;56: Tummon et al. Luteal LA versus CC in IVF 567

6 7. Yuzpe AA, Brown SE, Casper RF, Nisker J, Graves G, Shatford L. Transvaginal, ultrasound-guided oocyte retrieval for in vitro fertilization. J Reprod Med 1989;34: Colwell KA, Tummon IS. Oral micronized progesterone elevates serum progesterone after in vitro fertilization: a randomized controlled trial. J Reprod Med 1991;36: Ferrier A, Rasweiler JJ, Bedford JM, Prey K, Berkeley AS. Evaluation of leuprolide acetate and gonadotropins versus clomiphene citrate and gonadotropins for in vitro fertilization or gamete intrafallopian transfer. Fertil SteriI1990;54: Wood PL, IfHand CA, Allen E, Bentick B, Burton P, Shaw RW, et al. Serum levels of pregnancy-associated alpha 2- globulin (alpha 2-PEG), a glycosylate beta-lactoglobulin homologue, in successful and unsuccessful assisted conception. Hum Reprod 1989;4: Polson DW, MacLachlan V, Krapez JA, Wood C, Healy DL. A controlled study of gonadotropin-releasing hormone agonist (buserelin acetate) for folliculogenesis in routine in vitro fertilization patients. Fertil Steril1991;56: Seifer DB, Thornton KL, DeCherney AH, Lavy G. Early pituitary desensitization and ovarian suppression with leuprolide acetate is associated with in vitro fertilization-embryo transfer success. Fertil Steril 1991;56: Meldrum DR, Wisot A, Hamilton F, Gutlay AL, Kempton W, Huynh D. Routine pituitary suppression with leuprolide before ovarian stimulation for oocyte retrieval. Fertil Steril 1989;51: Curole DN, Dickey RP, Taylor SN, Rye PH, Olar TT. Pregnancies in cancelled gamete intrafallopian transfer cycles. Fertil Steril1989;51: Tummon et al. Luteal LA versus CC in IVF Fertility and Sterility