Minimal stimulation achieves pregnancy rates comparable to human menopausal gonadotropins in the treatment of infertility*

Transcription

1 FERTILITY AND STERILITY Copyright :Q' 1996 American Society for Reproductive Medicine Printed on acid-free paper in U. S. A. Minimal stimulation achieves pregnancy rates comparable to human menopausal gonadotropins in the treatment of infertility* Peter Y. Lu, M.D.t Aeneid L. J. Chen, M.D.t Elizabeth J. Atkinson, M.S.:j: Susan H. Lee, M.D.t Lisa D. Erickson, M.D.t Steven J. Ory, M.D.t Mayo Clinic and Mayo Foundation, Rochester, Minnesota Objective: To examine the effectiveness of a novel clomiphene citrate (CC) and hmg combination protocol ("minimal stimulation") for controlled ovarian hyperstimulation. Minimal stimulation consists of administering 100 mg/d CC for 5 days followed by a single dose of 150 IU hmg. The results of this analysis are compared with those of an hmg-alone protocol. In vitro fertilization-embryo transfer and donor insemination patients are excluded from this analysis. Design: Retrospective review of minimal stimulation and hmg cycles from January 1, 1989 to December 31, Setting: Tertiary care center reproductive endocrinology and infertility clinic. Patients: Two hundred thirty-two women who underwent 549 treatment cycles. Main Outcome Measures: Clinical and multiple pregnancy rates (PRs) and medication costs. Results: Sixty-one women received 106 cycles of minimal stimulation and 183 received 443 cycles of hmg. Although subject groups were not assigned randomly, multivariate analysis detected no significant differences between the treatment groups. The total ampules of hmg required differed significantly (2.0 for minimal stimulation versus 16.8 :+: 8.5 [mean:+: SD] for hmg). Pregnancy rates and multiple gestation rates were similar. Medication expense of minimal stimulation is 21% that of the hmg protocol. Conclusions: Minimal stimulation is as effective as hmg in the population examined. The comparable PRs and decreased medication costs of minimal stimulation justifies further evaluation of its role in the treatment of infertility. Fertil Steril 1996; 65:583-7 Key Words: Clomiphene citrate, hmg, minimal stimulation, ovarian hyperstimulation, medication costs Indications for ovulation induction therapy have evolved to include a variety of nonovulatory-related causes of infertility such as endometriosis, male factor, corrected tubal factor, and idiopathic (unexplained) causes. The underlying precepts of this largely empirical approach are the [1] release of additional oocytes effectively creating more opportunities for conception; [2] the possible correction of subtle, less predictable causes of ovulatory dysfunction; Received June 15, 1995; revised and accepted August 30, * Presented at the 50th Annual Meeting of The American Fertility Society, San Antonio, Texas, November 7, t Section of Reproductive Endocrinology. :j: Section of Biostatistics. Reprint requests and present address: Steven J. Ory, M.D., 2825 North State Road 7, Suite 302, Margate, Florida (FAX: ). and [3] the possible overriding of factors causing subfertility (1). Although the precise mechanisms at work remain to be elucidated, the numerous reports of the success of controlled ovarian hyperstimulation (COR) in the treatment of nonovulatory forms of infertility attest to the effectiveness of this popular strategy (2-5). Clomiphene citrate (CC) generally is used as the initial agent for ovarian stimulation because of its relative economy and ease of administration. The medication and monitoring expenses per cycle of CC is generally much lower than ovarian stimulation with hmg. Women who remain anovulatory on CC or who have not conceived after several cycles are considered appropriate candidates for ovarian stimulation with hmg. Treatment regimens using a combination of CC and hmg have been devised to de- Lu et at. Assessment of minimal stimulation 583

2 crease costs by reducing the total amount of hmg necessary in an ovarian stimulation cycle, without sacrificing effectiveness (6-8). Corfman et al. (9) previously have described a novel ovarian stimulation protocol termed "minimal stimulation" (9). When used in our IVF-ET program, this protocol produced a clinical pregnancy rate (PR) comparable to pure hmg stimulation at a lower expense. The current study describes our preliminary experience with minimal stimulation in a non-ivf patient population over 3 years. For purposes of comparison, we have included the results of patients treated with a standard hmg ovarian stimulation protocol during the same time period. Patient Groups MATERIALS AND METHODS From January 1, 1989 to December 31, 1992 patients presenting to the Mayo Fertility Clinic for COR in conjunction with timed intercourse or IVI were offered the option of a stimulation protocol combining CC and hmg or the conventional hmg protocol. The infertility diagnoses of these patients included endometriosis, corrected tubal disease, ovulatory dysfunction, male factor, and/or unexplained infertility. For purposes of this study, patients undergoing ovarian hyperstimulation for IVF ET or donor semen insemination were excluded from this analysis. The majority of these women had failed prior CC therapy and were candidates for more aggressive ovarian stimulation. The final choice of stimulation protocol was made by the treating physician in conjunction with the patient's wishes and pertinent clinical information but assignment was not randomized. Patients on the minimal stimulation protocol who responded with fewer than two dominant follicles (>20 mm in diameter), had a thin endometrium «8 mm), or were unsuccessful in achieving pregnancy after three or four cycles were progressed to the hmg protocol. Serum quantitative,b-subunit of hcg level was determined if menstruation did not occur within 17 days after hcg administration. If the result was >5 miv/ml (conversion factor to SI unit, 1.00), the test was repeated 48 hours later. A clinical pregnancy was defined as visualization of the gestational sac by transvaginal ultrasound (VS) performed 4 weeks after hcg administration. Significant ovarian hyperstimulation was defined as the development of signs, symptoms, and laboratory data necessitating hospitalization. The histories of all patients were abstracted for pertinent data and submitted for statistical analysis. 584 Lu et al. Assessment of minimal stimulation Ovarian Stimulation Protocols Minimal Stimulation The protocol for minimal stimulation is 100 mg/d CC orally, administered during cycle days 3 through 7. On cycle day 9, a single injection of 150 IV 1M hmg (Pergonal; Serono Laboratories, Randolph, MA) was given (9). On cycle day 12, transvaginal ultrasonography was performed to assess follicle development. Administration of 10,000 IV 1M hcg (Profasi; Serono Laboratories) was timed with the appearance of a lead mean follicle diameter of 20 mm. In general, one or two sonograms were performed and peak follicle diameter was recorded or projected assuming a 1 to 2 mm increase in mean follicle diameter of the lead follicle per day. Patients received luteal intravaginal P support until the following menses or positive pregnancy test. Human Menopausal Gonadotropin Protocol Patients receiving hmg were given daily doses ranging from 75 to 600 V, usually initiating therapy with 150 IU/d unless the patient had exhibited unusual sensitivity or tolerance to gonadotropins. This was individualized and was determined by the treating physician based on available clinical data such as the patient's previous response to ovarian stimulation therapy (CC, minimal stimulation, and/or previous hmg). Management during the cycle also was individualized; adjustments to hmg dosage and the timing of hcg administration were determined by transvaginal VS and/or serum E2 results. Luteal phase intravaginal P support was administered until menstruation or pregnancy confirmation. Statistical Analysis Statistical analysis was performed by comparing the treatment groups according to the Wilcoxon Rank Sum method for continuous variables and the X 2 test for categorical variables, including pregnancy. In these analyses, cycle was considered the experimental unit, in part because women switched treatments. Statistical analysis was performed by comparing the treatment groups according to the Wilcoxon Rank Sum method for continuous variables and the X2 test for categorical variables, including pregnancy. In these analyses, cycle was considered the experimental unit, in part because women switched treatments. To adjust for differences between the patients in the regimen groups, the Anderson-Gill model was used to analyze the influence of regimen on pregnancy. In this model, woman was considered the experimental unit, regimen was treated as a time-dependent covariate, and the model endpoint was the time to pregnancy. A step- Fertility and Sterility

3 % Cycle Number Figure 1 Cumulative PRs of minimal stimulation and hmg treatment groups expressed as percentages. Pregnancy rates are based on the cycles of women who initially started each treatment. Data do not include the subsequent treatment cycles after switching from one protocol to the other. No significant difference between treatments. f':" minimal stimulation; D, hmg. wise model selection process was used to select significant variable from among the following: prior pregnancy, primary infertility, anovulation, endo-. metriosis, unexplained infertility, tubal disease, male factor, age, and method of insemination. After those variables were selected, a regimen variable was included in the model. Interactions with the regimen also were investigated. Throughout the analyses, P < 0.05 defined the level of statistical significance. Life-table analysis was applied to create cumulative PRs by treatment. The plots only include data while on the first method treatment (Fig. 1). Throughout the analyses, P < 0.05 defined the level of statistical significance. RESULTS A total of 232 women completed 549 ovarian stimulation cycles and were included in the analysis. Sixty-one women completed 106 cycles of minimal stimulation and 183 women completed 443 hmg cycles. The mean age (:::,::SD) at the time of treatment in the minimal stimulation cycles was lower than that of the hmg cycles (31.9 :::':: 4.2 and 33.0 :::':: 4.9 years, respectively, P = 0.034). A significantly greater percentage of hmg cycles were associated with lui than in minimal stimulation cycles (66.1% and 55.7%, respectively, P = 0.043). The infertility characteristics and diagnoses of the groups are shown in Table 1. There was a significantly greater percentage of patients with a diagnosis of endometriosis in the hmg group while the minimal stimulation group contained a higher incidence of unexplained infertility. The total ampules of hmg used in the two treatments was significantly different (2.0 Table 1 Infertility Characteristics and Diagnoses of Minimal Stimulation and hmg Treatment Cycles Minimal stimulation hmg Duration of infertility 34.5::': ::': 35.7t (mo)* Primary infertility t (%) Ovulatory t Dysfunction (%) Endometriosis (%) Male factor (%) t Tubal disease (%) t Unexplained (%) * Values are means::': SD. t Not significantly different, Wilcoxon Rank Sum Test. t Not significantly different, X 2 test. Significantly different, < II Significantly different, for minimal stimulation and 16.8 :::':: 8.5 for hmg). The number of days to hcg in each protocol was defined as the first day of medication administration (CC or hmg) up to, and including, the day of hcg injection. The mean (:::'::SD) number of days to hcg for minimal stimulation was 10.2 :::':: 1.4 versus 8.6 :::':: 2.5 days for hmg (P < 0.001). Mter adjusting for differences in age, use of lui, and infertility diagnoses in a multivariate model, no differences between the two groups were detected. Table 2 summarizes the pregnancy outcome data of each treatment. There was no significant difference between the cumulative PRs of each treatment over five cycles. (Fig. 1). Table 3 shows the results of the subgroup of patients who were not successful on minimal stimulation and proceeded to further hmg cycles. Three of these women (33%) achieved clinical pregnancies during their first hmg cycle. The patient charge at our hospital pharmacy for 100 mg/d CC for 5 days is approximately $ The cost for 150 IU hmg is approximately $ and 10,000 IU hcg costs approximately $ Thus, based on these figures and the mean amount ofmedication used, the mean medication expenses of minimal stimulation and hmg stimulation are $ and $1,215.36, respectively. Table 2 Pregnancy Data of Minimal Stimulation and hmg Cycles* Clinical pregnancy (%) Miscarriage (%) Singleton (%) Hyperstimulation (%) * Clinical pregnancy rate per cycle. t Not significantly different. t Not significantly different, Minimal stimulation hmg 20.lt 33.0t 78.0t 2.5t Lu et al. Assessment of minimal stimulation 585

4 Table 3 Pregnancy Data of Women who Failed Minimal Stimulation Therapy and Subsequently Received hmg No. of minimal stimulation No.ofhMG Clinical Patient cycles cycles pregnancy No No Yes Yes No No Yes No No Total DISCUSSION Protocols for COH that require a lower dosages of hmg are attractive from both a financial and patient comfort perspective. Patients who fail to ovulate or do not conceive with CC therapy and are candidates for more aggressive ovarian stimulation therapy may benefit from a CC and hmg regimen before hmg-alone therapy. Kistner (6) first reported the successful use ofhmg followed by CC to induce ovulation in women with endogenous estrogenic activity who had responded poorly to CC alone. The success of sequential administration of these drugs for ovulation induction has been reported subsequently as well CC (7, 10). Jarrell et al. (8) reported that not only did sequential CC and hmg decrease by up to 50% the total hmg required to achieve ovulation, but also established the effectiveness of their protocol by observing that there was no difference in PR when compared with hmg therapy. The precise dosages and schedule ofcc and hmg differed widely in these studies and all used several days of hmg administration with varying techniques of monitoring serum estrogen. In light of the significant financial burden for couples in need of assisted reproductive technologies, it becomes more worthwhile to compare the effectiveness of superovulation regimens that may have significant cost differences (11, 12). We have examined our experience with minimal stimulation and hmg cycles over 3 years. Based on our practice standards, the majority of the patients included in this study had previous unsuccessful CC cycles. Minimal stimulation often was offered as a "next step" after CC failures and the reduced costs and monitoring of this protocol was appealing to patients. Though the characteristics that were examined between the groups indicated that they were similar, we do not purport to represent them as equivalent. The hmg cycles did have significantly more patients with endometriosis, whereas unexplained infertility was seen more often in the mini- 586 Lu et al. Assessment of minimal stimulation mal stimulation group. Also, there was an increased use ofiui in the hmg cycles. However, an extensive multivariate analysis did not detect any significant impact of these differences on outcomes. A more ideal control for this analysis could have been achieved by randomization of patients by stimulation protocol and method of insemination. However, given the absence of such controls, it seems reasonable to make the present comparisons. Based on the present sample size, this analysis has an 80% power to detect a 13% difference between the PRs of the minimal stimulation and hmg protocols. We found no significant differences or trends toward differences between the PRs and singleton PRs ofthe two treatments. Multivariate model analysis detected no differences after adjusting for age, use ofiui, or infertility diagnosis. Within each treatment group, PRs did not differ significantly between infertility diagnoses. Similarly, over the course of five cycles, no significant differences existed between their cumulative PRs (Fig. 1). The 33% miscarriage rate observed with hmg is surprising but consistent with previous studies, whereas the finding of a 4.55% miscarriage rate with minimal stimulation is less than the reported miscarriage rates CC or hmg (13). The reasons for these differences are not apparent from this study but selection bias may have contributed to some of the observed differences. Some authors have expressed concern regarding the antiestrogenic properties CC. There may be defective estrogen-driven stimulation of the endometrium, resulting in a suboptimal environment for implantation (14,15). The results ofthis analysis result do not confirm a negative impact ofcc on early pregnancy. In our practice, patients who responded poorly to minimal stimulation (fewer than two dominant follicles) or do not conceive after three to four treatment cycles are considered for hmg therapy. Examination of endometrial thickness is also performed at the time of transvaginal US. In light of reports of a potential correlation between endometrial thickness and conception along with concerns of a decrease in endometrial thickness during CC therapy, it is our practice also to consider discontinuation of minimal stimulation in the presence of a thin endometrial stripe (16,17). In this study, there were nine women who completed a total of 20 minimal stimulation cycles without conceiving and subsequently proceeded to hmg therapy. This subgroup underwent a cumulative total of 13 additional hmg cycles, which resulted in three clinical pregnancies. These conceptions all occurred in the first hmg cycle after failed minimal stimulation. This suggests that hmg therapy still may be effective in women who are unsuccessful with minimal stimulation. Patients failing hmg therapy are advised to pursue IVF-ET. Fertility and Sterility

5 The mean medication expense ofhmg stimulation was approximately five times the cost of minimal stimulation. There also are differences in the cost of monitoring between the two protocols. Monitoring hmg stimulation usually requires several transvaginal USs and serum E2 determinations. The only monitoring required during minimal stimulation is a transvaginal US on cycle day 12. No serum E2 determinations were obtained. If the lead follicle was not at least 20 mm in diameter on the day of US, the timing of hcg administration was projected by assuming a follicular growth rate of approximately 1 to 2 mm per day (18). This usually resulted in hcg being given within 3 days (between days 13 to 15 of the cycle). Rarely, a second US was performed if the lead follicle diameter was < 14 mm in order to properly time the ovulation inducing dose of hcg. Patients without follicular development of 2: 16 mm by cycle day 15 were considered treatment failures and were advised to proceed with hmg therapy. From our clinical experience, we found that follicle growth is variable between patients and that information gained from transvaginal US in the first cycle of minimal stimulation may be useful in determining the optimal day to perform US in subsequent cycles. Because cycle monitoring data was not included in this analysis, we are unable to report their associated costs. However, if the costs of monitoring were taken into account, they would amplify further differences between protocols. Thus, we conclude from this analysis that the minimal stimulation appears to be an effective protocol for COH in the treatment of infertility. The clinical PRs were comparable to a standard hmg regimen in our patient population. In light of the differences in medication costs, minimal stimulation may be a reasonable therapeutic option before consideration of pure hmg treatment. Although we acknowledge the limitations of this retrospective, nonrandomized analysis, these findings suggest that, on the basis of cost and efficacy, a prospective assessment of minimal stimulation would be worthwhile. Acknowledgment. The authors acknowledge the advice of Peter C. O'Brien, Ph.D., Section of Biostatistics, Mayo Clinic, Rochester, Minnesota, in the statistical analysis of this data. REFERENCES 1. Stovall DW, Guzick DS. Current management of unexplained infertility. Curr Opin Obstet Gynecol 1993;5: WeIner S, DeCherney AB, Polan ML. Human menopausal gonadotropins: a justifiable therapy in ovulatory women with long-standing idiopathic infertility. Am J Obstet Gynecol 1988; 158: Glazener CM, Coulson C, Lambert PA, Watt EM, Hinton RA, Kelly NG, et al. Clomiphene treatment for women with unexplained infertility: placebo-controlled study of hormonal responses and conception rates. Gynecol EndocrinoI1990;4: Fedele L, Bianchi S, Marchini M, Villa L, Brioschi D, Parazzini F. Superovulation with human menopausal gonadotropins in the treatment of infertility associated with minimal or mild endometriosis: a controlled randomized study. Fertil Steril 1992;58: Nulsen JC, Walsh S, Dumez S, Metzger DA. A randomized and longitudinal study of human menopausal gonadotropin with intrauterine insemination in the treatment of infertility. Obstet GynecoI1993;82: Kistner RW. Use of clomiphene citrate, human chorionic gonadotropin, and human menopausal gonadotropin for induction of ovulation in the human female. Fertil Steril 1966; 17: March CM, Tredway DR, Mishell DR Jr. Effect of clomiphene citrate upon amount and duration of human menopausal gonadotropin therapy. Am J Obstet Gynecol1976; 125: Jarrell J, McInnes R, Cooke R, Arronet G. Observations on the combination of clomiphene citrate-human menopausal gonadotropin-human chorionic gonadotropin in the management of anovulation. Fertil Steril 1981;35: Corfman RS, Milad MP, Bellavance TL, Ory SJ, Erickson LD, Ball GD. A novel ovarian stimulation protocol for use with the assisted reproductive technologies. Fertil Steri11993; 60: Rabau E, Mashiach S, Serr DM, Melmed H. Induction of ovulation by combined administration of human menopausal gonadotropin and clomiphene citrate. Obstet Gynecol 1968; 31: Diamond MP, Hill GA, Webster BW, Herbert CM, Rogers BJ, Osteen KG, et al. Comparison of human menopausal gonadotropin, clomiphene citrate and combined human menopausal gonadotropin-clomiphene citrate stimulation protocols for in vitro fertilization. Fertil SteriI1986;46: Yuen BH, Pride SM, Rowe TC, Moon YS, McComb PF, Poland BJ, et al. Comparison of the outcome of ovulation induction therapy in an in vitro fertilization program employing a low dose and individually adjusted high dose schedule of human menopausal gonadotropins. Am J Obstet Gynecol 1985; 151: Shoham Z, Zosmer A, Insler V. Early miscarriage and fetal malformations after induction of ovulation (by clomiphene citrate and/or human menotropins), in vitro fertilization, and gamete intrafallopian transfer. Fertil Steril 1991;55: Garcia J, Jones GS, Wentz AC. The use of clomiphene citrate. Fertil Steril1977;28: Nelson LM, Hershlag A, Kurl RS, Hall JL, Stillman RJ. Clomiphene citrate directly impairs endometrial receptivity in the mouse. Fertil Steril 1990; 53: Randall JM, Templeton A. Transvaginal sonographic assessment of follicular and endometrial growth in spontaneous and clomiphene citrate cycles. Fertil Steril 1991;56: Gonen Y, Casper RF, Jacobson W, Blankier J. Endometrial thickness and growth during ovarian stimulation: a possible predictor of implantation in in vitro fertilization. Fertil Steril 1989; 52: Gougeon A. Dynamics of follicular growth in the human: a model from preliminary results. Hum Reprod 1986;1:81-7. Lu et al. Assessment of minimal stimulation 587