Luteal phase consequences of low-dose gonadotropin-releasing hormone agonist therapy in nonluteal-supported in vitro fertilization cycles*

FERTILITY AND STERILITY Copyright

©

Vol. 64, No.3, September 1995

1995 American Society for Reproductive Medicine

Printed on acid-free paper in U. S. A.

Luteal phase consequences of low-dose gonadotropin-releasing hormone agonist therapy in nonluteal-supported in vitro fertilization cycles*

Cecil A. Long, M.D. t Victoria M. Sopelak, Ph.D. Stephen R. Lincoln, M.D. Bryan D. Cowan, M.D. Department of Obstetrics and Gynecology, University of Mississippi Medical Center, Jackson, Mississippi

Objective: To determine the follicular and luteal phase impact of low-dose GnRH agonist (GnRH-a) treatment during follicular stimulation for IVF. Design: A randomized prospective study compared patients receiving low-dose GnRH-a and hMG therapy to clomiphene citrate (CC) and hMG cycles. Setting: Patients were treated through a university-based IVF-ET program. Patients: Thirty-six patients underwent follicular stimulation with low-dose GnRH-a and hMG and were compared with 34 patients undergoing ovulation induction with CC and hMG. Results: Significantly shorter luteal phase length occurred with GnRH-a and hMG therapy; however, there was no statistically significant difference in luteal P levels. Follicular parameters were the same (peak E 2 , number offollicles, and number of oocytes), suggesting that folliculogenesis was not altered. There were no statistical differences in pregnancy rates. Conclusions: Sustained low-dose GnRH-a therapy during follicular stimulation does not have a clinical effect on luteal function. Fertil Steril 1995;64:573-6 Key Words: In vitro fertilization, ovulation induction, folliculogenesis, luteal phase, GnRH agonist

Regimens that use hMG for follicular stimulation during IVF are designed to produce multiple oocytes. Supraphysiologic steroid hormone concentrations are produced in response to hMG during both the follicular and luteal phase of the cycle. An unintended response during ovarian stimulation for IVF is a premature spontaneous LH surge. Premature LH surges occur in 10% to 50% of women undergoing ovulation induction for IVF and may alter the timing of oocyte retrieval, detrimentally affect the ovum, and accelerate the maturation of the endometrium (1-4). Because of these problems, many programs

Received April 28, 1994; revised and accepted March 16, 1995. * Supported in part by the Vicksburg Hospital Medical Foundation, Vicksburg, Mississippi. t Reprint requests: Cecil A. Long, M.D., Department ofObstetrics and Gynecology, University of Mississippi Medical Center, 2500 North State Street, Jackson, Mississippi 39216-4505 (FAX: 601-984-5301). Vol. 64, No.3, September 1995

do not perform follicular aspiration if a spontaneous LH surge occurs before hCG administration. To avoid unwanted LH surges, follicular stimulation regimens for IVF have been developed that incorporate GnRH agonists (GnRH-a). Gonadotropinreleasing hormone agonist therapy initiated in the luteal phase of the cycle immediately preceding follicular stimulation for IVF commonly is used to desensitize the pituitary to endogenous hypothalamic GnRH stimulation and prevent the LH surge (5). However, the amount of hMG used to stimulate the follicles often is increased with this therapy (6). In addition, premature luteal regression has been reported in women undergoing prolonged GnRH-a therapy during ovarian stimulation for IVF (4, 7, 8). Short-term "flare" treatments with GnRH-a and clomiphene citrate (CC) have been used (9, 10). The comparative benefits of GnRH-a "flare" versus CCinduced "flare" regimens have not been completely determined. In the present investigation we comLong et at. Luteal phase consequences of GnRH-a therapy

573

pared the clinical effects of low-dose leuprolide acetate (LA) combined with hMG to CC and hMG during follicular stimulation for NF. In a separate study, the hormonal response of patients treated with lowdose LA for follicular stimulation was determined. MATERIALS AND METHODS

Table 1 Comparison of Two IVF Ovulation Induction Protocols* LA and hMG* protocol I

CC and hMG protocol II

P value

36 18.2 ± 1.3 1,315 ± 130 1.0 11.0 ± 0.8 8 ± 0.5 4 ± 52 ± 3.3 13 ± 0.2

34 12.6 ± 0.4 1,461 ± 48 12.0 ± 0.9 8 ± 0.7 5 ± 0.5 64 ± 5.2 14 ± 0.2

<0.001 NS:j: NS NS NS NS <0.001

* Values are means ± SEM. t Conversion factor to SI unit, 3.671. :j: NS, not significant. § Conversion factor to SI unit, 3.180.

574

Treatment

Cycles

Recoveries

Transfers

Pregnant

LA and hMG CC and hMG

36 34

30 28

29 26

5* 5t

* Three delivered (two twins, one singleton), two miscarried. t Four delivered (all singleton), one ectopic.

Seventy consecutive women (25 to 45 years old) undergoing their (first) initial treatment cycle in the university medical center NF -ET program were randomized to one of two ovulation induction protocols. All participants consented to participate in the study through an approved (Institutional Review Board) protocol. Patients randomized to protocol I (n = 36) received 0.25 mg SC LA (Lupron; TAP Pharmaceuticals, Deerfield, IL) with 150 IV hMG beginning on day 2 of the menstrual cycle. On days 2 to 6 of the menstrual cycle, another 34 patients took 50 mg CC and received 150 IV of hMG (Pergonal; Serono Laboratories, Randolph, MA) 1M on a daily basis beginning on cycle day 3 (protocol II). Estradiol and follicular sonographic monitoring began on day 7 for both groups. Human chorionic gonadotropin (10,000 IV) was administered when three or more follicles measured ~15 mm (largest diameter) and the circulating E2 was >200 pg/mL per follicle. Transvaginal ultrasound-directed follicular aspiration was performed 34 hours after hCG administration. Luteal phase P concentrations were obtained 6 days after oocyte retrieval. No luteal support was administered to any patient. A clinical pregnancy was confirmed if rising hCG concentrations were observed and an intrauterine gestation or tubal pregnancy was confirmed. Previous and separate to this randomized study, 14 patients treated with protocol I underwent daily hormone monitoring beginning on cycle day 2 to determine the follicular phase responses of the pituitary and ovary. Serum LH, FSH, and E2 were mea-

Cycles HMG (75 IV ampules) Peak E2 (pg/mLlt Follicles Oocytes Embryos Progesterone (ng/mL)§ Luteal length (d)

Table 2 Pregnancy Rates Mter Two Different Ovulation Induction Regimens for IVF

Long et at. Luteal phase consequences of GnRH-a therapy

sured by RIA upon entry to the cycle before the administration of any drug. Subsequently, 0.25 mg LA was administered and LH, FSH, and E2 measurements were repeated 2 hours later. Thereafter, these hormones were measured 2 hours after (10:00 A.M.) each daily dose of LA (8:00 A.M.) until hCG was administered. Human menopausal gonadotropin was administered each day after hormone measurements were obtained (4:00 P.M.). Statistical analysis was performed using Student's t-test for comparison of numeric data and X 2 for comparison of nominal data. Analysis of variance of multiple means was performed for LH and FSH data, and post hoc testing with Dunnett's method was performed against baseline controls. A P value of <0.05 was considered significant. RESULTS

Table 1 contains the average hMG requirements, preovulatory E2 concentration, midluteal phase P concentration, number of follicles aspirated, number of oocytes recovered, number of embryos produced, and length of the luteal phase in each group of patients. Protocol I patients had significantly shorter luteal phase lengths (P < 0.001) and required more hMG (P < 0.001), however, there was no significant difference in P levels (P = 0.088) and no other significant hormone or gamete differences were observed between the two groups. There were six cancellations in each group of patients. Protocol I patients had four cancellations because of poor response, one for enlarged ovarian cyst, and one for hyperstimulation. Patients randomized to protocol II had three patients cancel because of poor response, two because of a premature LH surge, and one because of an enlarged ovarian cyst. Table 2 contains the number of cycles, recoveries, transfers, and clinical pregnancies achieved in each group. Of 10 pregnancies, there were two spontaneous abortions (protocol 1) and one ectopic pregnancy (protocol II). There were no significant differences between the number of oocytes recovered embryos transferred, or pregnancy rates (PRs) between the groups (X2; P = 0.712). The pregnancy rate per cycle in the Protocol I patients was 13.8% and Protocol II patients had a 14.7% PR per cycle. Fertility and Sterility

10

• FSH

:J E

D

8

LH

....... =>

E 6

'-'

c: '0.. 0

....0

40

0

20

~

.,c:

"0

C>

0

5

10

15

Day of Stimulation

Figure 1 Gonadotropin levels after daily injection of GnRH-a (0.25 mg LA). Follicle-stimulating hormone and LH significantly increase and remain elevated for approximately 24 hours. Luteinizing hormone steadily declines to reach baseline level on approximately day 10 of induction cycle. A secondary rise of FSH is seen on day 7 and persists throughout the follicular phase ofthe cycle. P < 0.05 compared with time 0 measurements.

Nonprotocol patients that previously underwent daily LH and FSH monitoring during LA injection revealed a significant increase (P < 0.001) in both LH and FSH concentrations from baseline values (Fig. 1). Luteinizing hormone levels remained elevated (P < 0.05) for approximately 5 days (day 7 of cycle) and then approached baseline values thereafter. Follicle-stimulating hormone concentrations only were increased (P < 0.05) 2 and 24 hours after initiating follicular stimulation. However, after returning to a level comparable to control, a secondary rise in FSH began and persisted for the remainder of follicular stimulation beginning on day 7. DISCUSSION

"Flare" protocols incorporating GnRH-a or CC have been used extensively in IVF induction regimens. Gonadotropin-releasing hormone agonist with hMG for controlled ovarian hyperstimulation (COH) attempts to incorporate the transient early phase rise ofFSH into follicular stimulation, avoid the deleterious LH effects on the follicle during the early phase, and desensitize the pituitary to an unwanted LH surge (10, 11). By achieving these goals early in the cycle, endogenous FSH can augment follicular stimulation, but the developing follicle avoids LH effects because follicular LH receptors have not been expressed yet (7). After initial stimulation with GnRH-a, gonadotropin production is suppressed and the unwanted LH surge is prevented. Previous reports have demonstrated that 0.5 to 1.0 mg of LA will stimulate and then desensitize the pituitary and prevent an unwanted LH surge when Vol. 64, No.3, September 1995

administered during the follicular phase concomitantly wit'h hMG (10, 11). Our study shows that daily doses of only 0.25 mg of LA also will achieve these effects. Our treatment protocol uses less GnRH-a than most reports and fulfills the strategy of premature LH surge avoidance by suppressing endogenous pituitary gonadotropin release by day 5 of drug administration (day 7 of cycle). Despite the appealing features of gonadotropin alteration, it is the ability of GnRH-a and hMG follicular stimulation protocols to achieve multiple oocyte recruitment that attracts our clinical interest. When the number of follicles, peak E 2 , and number of 00cytes derived from protocol I were compared with non-GnRH-a protocol II, no differences were observed. However, no LH surges occurred in protocol I, but one third of cycles cancelled in protocol II were cancelled because of detectable LH surges. Thus, GnRH-a and hMG does not alter ovarian follicular events when compared with CC and hMG but significantly reduces the occurrence of endogenous LH surges. The luteal effects of GnRH-a treatment during controlled follicular stimulation for IVF have not been established clearly. Combined high-dose LA (0.5 to 1.0 mg) with hMG does not affect P production during the early luteal phase (12, 13). In our study there was no significant difference in midluteal P concentrations between protocol I cycles and protocol II cycles. However, low-dose GnRH-a may have attenuated the life span of the corpus luteum, evidenced by shorter luteal phases observed in women treated with LA and hMG than women treated with CC and hMG (P < 0.001). Follicular phase responses probably do not account for the luteal phase differences because the follicular phase events were similar in both groups. These data suggest that even lowdose GnRH-a therapy during follicular stimulation may shorten the luteal phase. This may be similar to the immediate decline in corpus luteum P production seen in monkeys that received a potent GnRH antagonist in the late luteal phase (14). Therefore, our data and others (11, 13, 14) suggest that alteration of pituitary LH secretion affects midcorpus luteum and late corpus luteum function in GnRH-a and hMG-stimulated cycles. Although luteal phase duration was shorter in protocol I patients, our evidence suggests that these differences are unimportant clinically. First, the average serum P concentration observed in each follicular stimulation group is much higher than the P concentration observed during spontaneous ovulations and conception. Second, as Table 2 demonstrates, the conception rates (clinical pregnancy per cycle) in the LA and hMG group versus the CC and hMG group were similar (13.8% and 14.7%, respecLong et al. Luteal phase consequences of GnRH-a therapy

575

tively). Conclusions surrounding PRs are limited because of the small number of cycles. Lastly, endogenous hCG production by the implanted conceptus stimulates ("rescues") the corpus luteum, and this rescue occurs before the termination of the luteal phase in either group (15). In conclusion, our protocol of low-dose GnRH-a treatment during COH cycles for IVF offers several clinical advantages. Elevated concentrations of LH in the early follicular phase (days 2 to 4) as a result of flare do not have an adverse effect on follicular development. Down-regulation of pituitary responsiveness to GnRH-a occurs in a timely fashion (day 7 of cycle) and unwanted LH surges are prevented. Although the luteal length in patients treated with GnRH-a is shorter compared with patients treated with CC, the PRs of both groups are equivalent and the average P is greater than that observed in unstimulated cycles and conceptions. A shorter luteal phase in GnRH-a-treated women probably is the result of impaired gonadotropin stimulation of the corpus luteum in nonconception cycles. Although this event abbreviates the luteal phase, it is not associated with pregnancy failure. This indicates that there is no adverse clinical impact of low-dose GnRH-a on the function of the corpus luteum and that supplemental luteal phase support is not required with this protocol.

4.

5.

6.

7.

8. 9.

10.

11.

12.

13. REFERENCES 1. Mausher SJ, Garcia JE, Rosenwaks Z. The combination of follicle-stimulating hormone and human menopausal gonadotropin for the induction of multiple follicular maturation for in vitro fertilization. Fertil Steril 1985;44:62-9. 2. Fishel SB, Edwards RG, Purdy JM. Analysis of 25 infertile patients treated consecutively by in vitro fertilization at Bourn Hall. Fertil Steril 1984;42:191-7. 3. Dodds WG, Awadalla SG, Hisson C, Roh SI, Friedman CI, Kim MH. Atypical luteinizing hormone rise and associated

576

Long et at. Luteal phase consequences of GnRH-a therapy

14.

15.

fertilization failure in non-male factor in vitro fertilization patients. Obstet Gynecol 1989;73:191-5. Palermo R, Amodeo G, Navot D, Rosenwaks Z, Cittadini E. Concomitant gonadotropin-releasing hormone agonist and menotropin treatment for the synchronized induction ofmultiple follicles. Fertil Steril 1989;49:290-5. Fleming R, Coutts JRT. Induction of multiple follicular growth in normally menstruating women with endogenous gonadotropin suppression. Fertil Steril 1986;45:226-30. Insler Y, Potashnik G, Lunenfeld E, Meizher I, Levy J. The combined suppression/stimulation therapy in IVF-ET programs: expectations and fact. Hum Reprod 1989;4:47-51. Feichtinger W, Kemeter P, Szalay S, Beck A, Janisch H. Could aspiration of the Graffian follicle cause luteal phase deficiency? Fertil Steril 1982;37:205-8. Filicori M, Butler JP, Crawley WF Jr. Neuroendocrine control of the human corpus luteum. J Clin Invest 1984; 73:1638-42. Frydman R, Belaisch-Allart J, Parneix I, Forman R, Hazout A, Testart J. Comparison between flare up and down regulation effects of luteinizing hormone-releasing hormone agonists in an in vitro fertilization program. Fertil Steril 1988; 50:471-5. Macnamee MC, Howles CM, Edwards RG, Taylor PJ, Elder KT. Short-term luteinizing hormone-releasing hormone agonist treatment: prospective trial of a novel ovarian stimulation regimen for in vitro fertilization. Fertil Steril 1989; 52:264-9. Braendle W, Lindner C, Lichtenberg Y, Goepel E, Bettendorf G. Endocrine profiles and luteal function during GnRH-analoguelhMG therapy. Hum Reprod 1989;4:121-6. Smitz J, Devroey P, Camus M, Deschacht J, Khan I, Stessen C, et al. The luteal phase and early pregnancy after combined GnRH-agonistIhMG treatment for super ovulation in IVF or GIFT. Hum Reprod 1988;3:585-91. Belaisch-Allart J, DeMarzan J. Effect ofluteal phase supplementation in an IVF programme after ovarian stimulation by LHRH analogues: multicenter analysis. Contracept Fertil Sex (Paris) 1988; 16:654-9. Collins RL, Sopelak VM, Williams RF, Hodgen GD. Prevention of gonadotropin releasing hormone antagonist induced luteal regression by concurrent exogenous pulsatile gonadotropin administration in monkeys. Fertil Steril 1986;46: 945-53. Fritz MA, Hess DL, Patton PE. Influence of corpus luteum age on steroidogenic response to exogenous human chorionic gonadotropin in normal cycling women. Am J Obstet Gynecol 1992; 167:709-16.

Fertility and Sterility