- Email: [email protected]
In vitro fertilization cycle outcome after coasting in gonadotropin-releasing hormone (GnRH) agonist versus GnRH antagonist protocols
In vitro fertilization cycle outcome after coasting in gonadotropin-releasing hormone (GnRH) agonist versus GnRH antagonist protocols Jacob Farhi, M.D., Avi Ben-Haroush, M.D., Yehezkel Lande, M.D., Onit Sapir, Ph.D., Haim Pinkas, M.D., and Benjamin Fisch, M.D., Ph.D. Infertility and IVF Unit, Department of Obstetrics and Gynecology, Helen Schneider Hospital for Women, Rabin Medical Center, Petach Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
Objectives: To compare the results of IVF cycles after coasting in patients treated with a GnRH antagonist or GnRH agonist protocol. Design: A retrospective case-control study. Setting: Infertility unit in a university-affiliated tertiary medical center. Patient(s): The study group included all women less than 38 years old attending the IVF unit from 2000 to 2006 in whom coasting was used. Data on E2 levels before and after coasting, duration of coasting, number of oocytes retrieved and fertilized, embryo quality, moderate-severe ovarian hyperstimulation syndrome (OHSS), and pregnancy were collected from the files and compared between GnRH agonist (n ¼ 329) and GnRH antagonist (n ¼ 45) cycles. Intervention(s): None. Main Outcome Measure(s): Number of retrieved oocytes and pregnancy rates. Result(s): There were no between-group differences in cycle parameters. In the antagonist group, there was no need for more than 2 days of coasting. There was a significant decrease in the number of retrieved oocytes even in short periods of coasting in the antagonist group but not in the agonist group. On the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. The OHSS rate after coasting was 4.6% in the agonist group and 4.4% in the antagonist group. Corresponding pregnancy rates after coasting were 27.4% and 24.4%. Conclusion(s): The same criteria for coasting can be applied in GnRH agonist as in GnRH antagonist cycles, with a similar IVF result. (Fertil Steril 2009;91:377–82. 2009 by American Society for Reproductive Medicine.) Key Words: Coasting, GnRH agonist, GnRH antagonist, IVF
Hyper-response to controlled ovarian hyperstimulation (COH) in IVF treatment can be managed by withholding gonadotropin stimulation. This strategy, called ‘‘coasting,’’ is currently considered the most popular means of reducing the risk of ovarian hyperstimulation syndrome (OHSS) without canceling the cycle (1–6). Nevertheless, evidence of its effectiveness in severe cases is still insufficient. The use of coasting is based on the assumption that the decline in FSH levels causes atresia of the small follicles, while at the same time larger follicles maintain their growth. Thus, mature oocytes can ultimately be collected from the still growing large follicles, while the risk of OHSS is reduced because of the atresia of small follicles (7). Extensive studies have been conducted on the effect of coasting in GnRH agonist protocols, providing strong grounds for its application and the prediction of clinical and laboratory performance as well as IVF outcome (8–10). Received September 3, 2007; revised and accepted November 28, 2007. The first two authors contributed equally to this study, and both are first authors. Reprint requests: J. Farhi, M.D., Infertility and IVF Unit, Department of Obstetrics and Gynecology, Helen Schneider Hospital for Women, Rabin Medical Center, Petach Tikva 49100, Israel (FAX: 972-3-5028107; E-mail: [email protected]).
0015-0282/09/$36.00 doi:10.1016/j.fertnstert.2007.11.086
On the basis of these results, the clinical practice of coasting has been extended to GnRH antagonist protocols. However, data on the effect on IVF outcome in this setting remain limited (11–13). The aim of the present study was to compare the IVF results after coasting in hyper-responders treated with a GnRH antagonist protocol versus a GnRH agonist protocol and to investigate whether the same criteria for coasting can be applied in these protocols. MATERIALS AND METHODS All IVF cycles from 2000 to 2006 in which gonadotropins were withheld for at least 1 day before oocyte retrieval due to hyper-response were retrospectively analyzed. Institutional Review Board approval was obtained for the study. Treatment Protocol The GnRH long protocol consisted of daily injections of Decapeptyl (Ferring, GmbH, Germany) 0.1 mg or a depot injection of Decapeptyl 3.75 mg at the early follicular or midluteal phase. Down-regulation was confirmed after 13–15 days (no ovarian cysts >18 mm, E2 <74 pmol/L) and was followed by gonadotropin stimulation. The GnRH short
Fertility and Sterility Vol. 91, No. 2, February 2009 Copyright ª2009 American Society for Reproductive Medicine, Published by Elsevier Inc.
377
protocol consisted of daily injections of Decapeptyl 0.1 mg starting on day 2 or 3 of menstruation, followed by gonadotropin stimulation from day 4 or 5 of menstruation. The antagonist protocol consisted of daily gonadotropin stimulation from day 3 or 4 of menstruation followed by daily injections of Cetrotide 0.25 mg (Serono, Switzerland) or Orgalutran 0.25 mg (N.V. Organon, the Netherlands) once the leading follicle reached 14 mm and until the day of hCG injection. In all treatment protocols, gonadotropin stimulation consisted of recombinant FSH (Gonal F, Serono, Postfach, Switzerland; or Puregon, N.V. Organon, Oss, The Netherlands), alone or in combination with urinary gonadotropins (Menogon, Ferring, Kiel, Germany).
Data Collection The following data were collected for analysis: baseline clinical parameters (patient age, day 3 FSH level, and cycle number); IVF cycle parameters (type of protocol, daily dose of gonadotropins, number of oocytes retrieved, number of mature oocytes, fertilization rate, embryo quality, and number of embryos transferred); IVF cycle outcome (pregnancy rate); and coasting parameters (E2 level on the day coasting was started, E2 level after coasting, day of hCG administration, and days of coasting). Findings were compared between GnRH agonist and GnRH antagonist coasting cycles. The main outcome measures were number of retrieved oocytes and pregnancy rate.
The choice of a GnRH agonist or antagonist protocol for COH was made by the treating physician on the basis of the patient characteristics or response during previous IVF cycles. In all protocols, the initial dose of gonadotropins was selected according to the individual response of the patient to previous stimulations, if available. During treatment, the ovarian response was monitored by vaginal ultrasound measurements of follicular growth and serum E2 level every 1–3 days, starting on the fifth or sixth day of stimulation, and the FSH and hMG dosages were adjusted accordingly.
Statistical Analysis Data were managed and analyzed with the SPSS statistical package, version 15, for Windows (SPSS, Chicago). A logistic multivariate regression analysis (with enter and forward likelihood ratio analysis), analysis of variance with post hoc Tukey’s multiple comparison tests, the KruskalWallis test, the Mann-Whitney test, and the c2-test were used, as appropriate. P<.05 was considered statistically significant.
Management During Coasting Coasting was applied in all COH cycles, regardless of the type of protocol, when the serum E2 concentration measured 10,000–12,000 pmol/mL in the absence of a leading follicle of 14 mm or more in diameter. During the coasting period, gonadotropins were withheld while administration of the GnRH agonist or antagonist was continued. In all coasting cycles, hCG (Chorigon, Teva, Israel; or Pregnyl, N.V. Organon, 10,000 IU; or Ovitrelle, Serono, Italy, 250 mg) was administered when E2 concentration levels dropped at least 15% relative to the peak levels of E2 in the presence of at least two leading follicles measuring 17 mm or more. Cycles were canceled if coasting lasted more than 6 days.
RESULTS During the study period, 7866 IVF cycles were performed in our IVF unit; a GnRH antagonist protocol was used in 3107 cycles (39.5%), and a GnRH agonist protocol was used in the remainder (4759 cycles). Coasting was applied in 374 cycles: in 45 after stimulation with a GnRH antagonist protocol (antagonist group, 1.5% of all GnRH antagonist cycles) and in 329 after stimulation with a GnRH agonist protocol (agonist group; 6.9% of all GnRH agonist cycles). The difference in the use of coasting between the two protocols was statistically significant (P<.001, odds ratio [OR] ¼ 5.06; 95% confidence interval [CI], 3.7–6.8).
Oocytes were fertilized by standard insemination or intracytoplasmic sperm injection, depending on sperm parameters or previous IVF performance. Embryos were transferred on day 2 or 3. The luteal phase was supported by daily vaginal P (Utrogestan, Besins International Laboratories, France, 600 mg/day; or Endometrin, Ferring, 200 mg/day). A serum pregnancy test was performed 12 days after ET. Clinical pregnancy was defined as the presence of an intrauterine gestational sac on ultrasound scan 4 weeks after oocyte retrieval. Embryo Quality Embryos were graded by gross morphological appearance under light microscopy at 48 or 72 hours after oocyte collection, using the system of Staessen et al. (14): grade 1, even and homogeneous blastomeres without fragmentation; grade 2, even and homogeneous blastomeres with <20% fragmentation; grade 3, uneven and nonhomogeneous blastomeres with 20%–50% fragmentation; grade 4, uneven and nonhomogeneous blastomeres with >50% fragmentation. 378
Farhi et al.
Coasting in agonist vs. antagonist cycles
There was no significant difference between the antagonist and agonist groups in patient age (30.5 4.8 vs. 33.1 6.2 years), baseline FSH level (5.7 1.9 vs. 6.1 1.8 IU/L), or number of previous IVF attempts before the coasting cycle (3.8 3.1 vs. 4.4 4.0). The characteristics of the GnRH antagonist and GnRH agonist coasting cycles are shown in Table 1. The mean total number of gonadotropin ampoules used before coasting was similar in the agonist and antagonist groups, as was the duration of coasting (in days). On the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. However, there was no between-group difference in the mean number of oocytes retrieved, fertilization rate, or rate of high-quality embryos. The mean number of transferred embryos was also similar in the two groups. The rate of OHSS after coasting (defined as moderate-severe OHSS needing hospitalization) was 4.6% in the agonist group and 4.4% in the antagonist group. The corresponding pregnancy rates were 27.4% and 24.4% (Table 1). Vol. 91, No. 2, February 2009
TABLE 1 Characteristics of the GnRH antagonist and GnRH agonist coasted cycles.
No. of ampoules (units) Coasting days Coasting of more than 2 days, n (%) E2 level at initiation of coasting (pmol/L) E2 level at end of coasting (pmol/L) (hCG day) E2 ratio (peak/nadir) (pmol/L) Endometrial width, mm No. of retrieved oocytes Fertilization rate, (%) No. of high-quality embryos available for transfera No. of transferred embryos Hospitalization for moderatesevere OHSS, n (%) Pregnancy rate, n (%)
Agonist group (n [ 329 cycles)
Antagonist group (n [ 45 cycles)
P
27.7 12.5 (2078 939) 1.3 0.7 22 (6.7)
28.3 14.7 (2129 1106) 1.2 0.3 0
.742 .080 .074
11,447 4507 12,741 5573
10,750 4196 10,140 4916
.328 .030
1.07 0.9 10.0 2.1 13.6 6.5 60.0 24.8 1.7 2.2
1.3 1.2 9.6 2.1 14.1 7.9 62.5 28.3 1.2 1.2
.142 .259 .630 .500 .314
2.2 1.0 15 (4.6)
2.1 1.1 2 (4.4)
.500 .972
90 (27.4)
11 (24.4)
.672
Note: Values given as mean SD unless otherwise specified. a Day 2 embryos: grade A or AB, 4 cells; day 3 embryos: grade A or AB, 6–8 cells. Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
None of the antagonist cycles required more than 2 days of coasting, whereas 5.8% of the agonist cycles required 3 or more days of coasting (Table 2). Longer coasting was associated with a significant decrease in the number of retrieved oocytes in the antagonist group (2 days of coasting compared with 1 day of coasting) but not in the agonist group (1–4 days of coasting). In addition, the mean number of retrieved oocytes after 1 day of coasting was similar in the agonist and antagonist groups (13.6 6.5 and 15.4 7.7, respectively; P¼.526), but after 2 days of coasting the number of retrieved oocytes was significantly lower in the antagonist group versus the agonist group (7.0 5.1 and 13.9 6.6, respectively; P¼.01). The association between duration of coasting and E2 level at the end of the coasting period (hCG day) is presented in Table 3. There was no significant difference in E2 levels between the pregnancy and nonpregnancy cycles; the E2 levels at initiation of coasting were 10,801 3810 versus 11,571 4682 (pmol/L), respectively (P¼.140), and the E2 levels at end of coasting period on hCG day were 12,723 4238 versus 12,351 6004 (pmol/L), respectively (P¼.656). A multivariate logistic regression analysis was performed to identify factors predicting outcome (pregnancy/no pregnancy) in the coasting cycles (Table 4). We found that the treatment protocol (antagonist vs. agonist coasting cycle) was not a predictor of a successful cycle and that it did not influence laboratory parameters or outcome. Lower fertilization rate was the only factor associated with a lower success rate after coasting. Fertility and Sterility
DISCUSSION Despite the many studies conducted on coasting in GnRH agonist cycles, there is still a paucity of data for GnRH antagonist protocols. The use of coasting in GnRH antagonist protocols was first described in case reports that showed no compromise of IVF outcome (11, 12). A similar conclusion was reached in the first cohort study of coasting in 29 GnRH antagonist cycles, which was reported by Bahceci et al. (13) in 2006. The present study compared the results of IVF cycles after coasting in 45 patients treated with a GnRH antagonist protocol and 329 patients treated with a GnRH agonist protocol. This comparison was feasible, since we used the same criteria for the application of coasting in hyper-responders in both protocols. The key findings of the study were as follows: [1] In the antagonist group, there was no need for more than 2 days of coasting, whereas in the agonist group, 5.8% of the cycles required 3 or more days of coasting. [2] There was a significant decrease in the number of retrieved oocytes with longer coasting in the antagonist group but not in the agonist group. [3] On the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. [4] There was no between-group difference in mean number of oocytes retrieved, fertilization rate, or rate of high-quality embryos. [5] The OHSS and pregnancy rates after coasting were similar in the agonist and antagonist groups. The difference between the protocols in the duration of coasting until the E2 levels dropped sufficiently for hCG 379
TABLE 2 Association of duration of coasting with number of retrieved oocytes and pregnancy rate. All cycles (n [ 374) Agonist group (n [ 329 cycles) Antagonist group (n [ 45 cycles) Duration of coasting, days 1 2 3 R4 P
Cycles 284 68 14 8
Pregnancies Cycles (%) (oocytes/cycle) 77 (27.1) 17 (25) 4 (28.6) 3 (37.5) .895c
Pregnancies (%)
Cycles (oocytes/cycle)
Pregnancies (%)
68 (27.6) 15 (24.6) 4 (28.6) 3 (37.5) .880c
38 (15.4 7.7) 7 (7.0 5.1) — — .007b
9 (23.7) 2 (28.6) — — .782c
246 (13.6 6.5) 61 (13.9 6.6) 14 (12.4 7.9) 8 (11.5 4.1) .624a
Note: Number of retrieved oocytes given as mean SD. a Kruskal-Wallis test. b Mann-Whitney test. c 2 c -test. Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
administration was not statistically significant. In both groups, the vast majority of patients needed up to 2 days of coasting (Table 2). The fact that fewer oocytes were retrieved after 2 days of coasting in the antagonist group only may indicate a higher rate of atresia of oocytes in coasted antagonist versus coasted agonist cycles. It seems that atresia induced by coasting affects the number of retrieved oocytes (Table 2) but is not concomitantly reflected in the E2 levels. This may be explained by different effects of duration of coasting on oocyte survival and steroidogenesis. Previous studies noted that the duration of coasting in long GnRH agonist protocols had no deleterious effect on IVF outcome, provided that coasting did not last for more than 4 days (8, 9, 15–17). In GnRH agonist cycles, cycle cancellation should be discussed in cases of a longer duration (9). In GnRH antagonist protocols, prolonged coasting was not needed (Table 2), and Bahceci et al. (13) reported similar pregnancy rates after prolonged coasting in GnRH antagonist protocols. The association of E2 levels before and after coasting with cycle outcome has been evaluated in GnRH agonist cycles. No correlation was found between pregnancy or implantation
rates and either the magnitude of the E2 drop (10) or the level to which E2 decreased during coasting (8, 17, 18). There are no studies of this factor in GnRH antagonist protocols. In our study, on the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. Despite the greater magnitude of the drop, however, the difference in E2 levels from the agonist cycles was not statistically significant (Table 1). In agreement with earlier findings on the effect of E2 levels in GnRH agonist coasted cycles, neither univariate nor multivariate analysis (Table 4) in the present study yielded a correlation of pregnancy rate with E2 dynamics in GnRH antagonist coasting cycles. Regarding the correlation between the duration of coasting and number of oocytes retrieved, Mansour et al. (9) found that in a GnRH agonist protocol, both the number of oocytes retrieved and the number of metaphase II oocytes were significantly higher in cycles in which coasting lasted less than 3 days compared with cycles with coasting of 4 days and more. In our study, we found that the effect of the duration of coasting on the number of oocytes retrieved is much higher in the antagonist group. A significant decrease in the
TABLE 3 Association of duration of coasting with E2 level (pmol/L) at the end of coasting period on hCG day. Duration of coasting, days
Agonist group (n [ 329 cycles)
Antagonist group (n [ 45 cycles)
11,472 3592 12,542 5683 22,489 7984 18,358 8217 < .001
9963 4504 10,672 6463 — — .767
1 2 3 R4 P Note: Values given as mean SD. Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
380
Farhi et al.
Coasting in agonist vs. antagonist cycles
Vol. 91, No. 2, February 2009
TABLE 4 Multivariate regression analysis of all coasted cycles: predictors for pregnancy. Variable Age (years) Cycle number Previous pregnancy? (yes or no) No. of ampoules Coasting days E2 level before coasting (mmol/L) E2 level after coasting (mmol/L) Endometrial width (mm) No. of retrieved oocytes Fertilization rate (%) No. of transferred embryos Treatment protocol (agonist vs. antagonist)
P .776 .887 .545 .077 .583 .079 .104 .184 .741 .034 .877 .873
Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
number of retrieved oocytes was noted between 1 and 2 days of coasting, whereas in the agonist group, the number of oocytes retrieved was not changed even after 4 days of coasting (Table 2). The overall fertilization rate in our study after coasting was similar between the agonist and antagonist groups and similar to the expected rate in coasting cycles reported in earlier studies for the GnRH agonist protocol (9) and the GnRH antagonist protocol (13) cycles. The proportion of high-quality embryos was also similar in our GnRH antagonist and agonist groups; thus it seems that the protocol does not affect embryo quality after coasting. In our study, we compared embryo quality only in coasted cycles. We did not include a comparison with noncoasted cycles, although previous studies of embryo quality after coasting have suggested that coasting causes a reduction in oocyte quality in both GnRH agonist protocols (17) and GnRH antagonist protocols (13). Moreover, there are data suggesting that a profound drop in E2 level (<1000 pg/mL) in GnRH agonist protocols may be related to low oocyte quality (16). Pregnancy rates were similar in the GnRH antagonist and GnRH agonist groups in our study; therefore, it seems that the type of protocol has no effect on pregnancy rates after coasting. Previous studies, assessing each protocol separately, reported normal implantation and pregnancy rates for both protocols, as long as coasting lasted for up to 4 days (9, 10, 13). Overall, the need for coasting in our study population was significantly lower in GnRH antagonist protocols (1.5%) than in GnRH agonist protocols (6.9%, P<.001; OR ¼ 5.06; 95% CI, 3.7–6.8). However, the choice of protocols for selected populations with different ovarian response probably precludes independent analysis. Be that as it may, once Fertility and Sterility
hyper-response had developed and coasting was applied, the OHSS rate in the agonist and antagonist groups was similar (4.6% and 4.4%, respectively). One limitation of the current study is the relatively small number of coasted antagonist cycles. In a sample sizes analysis, 3645 subjects in each group are needed to achieve 80% power to detect a difference between 27.3% and 24.4% (the actual pregnancy rates in the agonist and antagonist groups, respectively) with a significance level of .05. Since coasting is needed in only 1.5% of the antagonist cycles, such numbers are difficult to achieve; moreover, the difference between these pregnancy rates is clinically insignificant. In conclusion, the same criteria for coasting can be applied in GnRH antagonist and GnRH agonist protocols, with a similar laboratory performance and cycle outcome.
REFERENCES 1. Sher G, Zouves C, Feinman M, Maassarani G. Prolonged coasting: an effective method for preventing severe ovarian hyperstimulation syndrome in patients undergoing in-vitro fertilization. Hum Reprod 1995;10:3107–9. 2. Benadiva CA, Davis O, Klingman I, Moomjy M, Liu HC, Rosenwaks Z. Withholding gonadotropin administration is an effective alternative for the prevention of ovarian hyperstimulation syndrome. Fertil Steril 1997;67:724–7. 3. Delvigne A, Rozenberg S. Preventive attitude of physicians to avoid OHSS in IVF patients. Hum Reprod 2001;16:2491–5. 4. D’Angelo A, Amson N. ‘‘Coasting’’ (withholding gonadotrophins) for preventing ovarian hyperstimulation syndrome. Cochrane Database Syst Rev 2002;3:CD002811. 5. Chen D, Burmeister L, Goldschlag D, Rosenwaks Z. Ovarian hyperstimulation syndrome (strategies for prevention). Reprod Biomed Online 2003;7:43–9. 6. Garcia-Velasco JA, Isaza V, Quea G, Pellicer A. Coasting for the prevention of ovarian hyperstimulation syndrome: much ado about nothing? Fertil Steril 2006;85:547–54. 7. Garcia-Velasco JA, Zuniqa A, Pacheco A, Gomez R, Simon C, Remohi J, et al. Coasting acts through downregulation of VEGF gene expression and protein secretion. Hum Reprod 2004;19:1530–8. 8. Ulug U, Ben-Shlomo I, Bahceci M. Predictors of success during the coasting period in high-responder patients undergoing controlled ovarian hyperstimulation for assisted reproduction. Fertil Steril 2004;82: 338–42. 9. Mansour R, Aboulghar M, Serour G, Amin Y, Abou-Setta AM. Criteria of a successful coasting protocol for the prevention of severe ovarian hyperstimulation syndrome. Hum Reprod 2005;20:3167–72. 10. Kovacs P, Matyas S, Kaali SG. Effect of coasting on cycle outcome during in vitro fertilization/intracytoplasmic sperm injection cycles in hyper-responders. Fertil Steril 2006;85:913–7. 11. Delvigne A, Carlier C, Rozenberg S. Is coasting effective for preventing ovarian hyperstimulation syndrome in patients receiving a gonadotropinreleasing hormone antagonist during an in vitro fertilization cycle? Fertil Steril 2001;76:844–6. 12. Fatemi HM, Platteau P, Albano C, Van Steirteghem A, Devroey P. Rescue IVF and coasting with the use of a GnRH antagonist after ovulation induction. Reprod Biomed Online 2002;5:273–5. 13. Bahceci M, Ulug U, Tosun S, Erden HF, Bayazit N. Impact of coasting in patients undergoing controlled ovarian stimulation with the gonadotropin-releasing hormone antagonist cetrorelix. Fertil Steril 2006;85: 1523–5. 14. Staessen C, Camus M, Bollen N, Devroey P, Van-Steirteghem AC. The relationship between embryo quality and the occurrence of multiple pregnancies. Fertil Steril 1992;3:626–30.
381
15. Tortoriello DV, McGovern PG, Colon JM, Skurnick JH, Lipetz K, Santoro N. ‘‘Coasting’’ does not adversely affect cycle outcome in a subset of highly responsive in vitro fertilization patients. Fertil Steril 1998;69:454–60. 16. Isaza V, Garcia-Velasco JA, Aragones M, Remhi J, Simon C, Pellicer A. Oocyte and embryo quality after coasting: the experience from oocyte donation. Hum Reprod 2002;17:1777–82.
382
Farhi et al.
Coasting in agonist vs. antagonist cycles
17. Delvigne A, Kostyla K, Murillo D, Van Hoeck J, Rozenbers S. Oocyte quality and IVF outcome after coasting to prevent ovarian hyperstimulation syndrome. Int J Fertil Womens Med 2003;48:25–31. 18. Levinsohn-Tavor O, Friedler S, Schachter M, Raziel A, Strassburger D, Ron-El R. Coasting—what is the best formula? Hum Reprod 2003;18: 937–40.
Vol. 91, No. 2, February 2009
Objectives: To compare the results of IVF cycles after coasting in patients treated with a GnRH antagonist or GnRH agonist protocol. Design: A retrospective case-control study. Setting: Infertility unit in a university-affiliated tertiary medical center. Patient(s): The study group included all women less than 38 years old attending the IVF unit from 2000 to 2006 in whom coasting was used. Data on E2 levels before and after coasting, duration of coasting, number of oocytes retrieved and fertilized, embryo quality, moderate-severe ovarian hyperstimulation syndrome (OHSS), and pregnancy were collected from the files and compared between GnRH agonist (n ¼ 329) and GnRH antagonist (n ¼ 45) cycles. Intervention(s): None. Main Outcome Measure(s): Number of retrieved oocytes and pregnancy rates. Result(s): There were no between-group differences in cycle parameters. In the antagonist group, there was no need for more than 2 days of coasting. There was a significant decrease in the number of retrieved oocytes even in short periods of coasting in the antagonist group but not in the agonist group. On the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. The OHSS rate after coasting was 4.6% in the agonist group and 4.4% in the antagonist group. Corresponding pregnancy rates after coasting were 27.4% and 24.4%. Conclusion(s): The same criteria for coasting can be applied in GnRH agonist as in GnRH antagonist cycles, with a similar IVF result. (Fertil Steril 2009;91:377–82. 2009 by American Society for Reproductive Medicine.) Key Words: Coasting, GnRH agonist, GnRH antagonist, IVF
Hyper-response to controlled ovarian hyperstimulation (COH) in IVF treatment can be managed by withholding gonadotropin stimulation. This strategy, called ‘‘coasting,’’ is currently considered the most popular means of reducing the risk of ovarian hyperstimulation syndrome (OHSS) without canceling the cycle (1–6). Nevertheless, evidence of its effectiveness in severe cases is still insufficient. The use of coasting is based on the assumption that the decline in FSH levels causes atresia of the small follicles, while at the same time larger follicles maintain their growth. Thus, mature oocytes can ultimately be collected from the still growing large follicles, while the risk of OHSS is reduced because of the atresia of small follicles (7). Extensive studies have been conducted on the effect of coasting in GnRH agonist protocols, providing strong grounds for its application and the prediction of clinical and laboratory performance as well as IVF outcome (8–10). Received September 3, 2007; revised and accepted November 28, 2007. The first two authors contributed equally to this study, and both are first authors. Reprint requests: J. Farhi, M.D., Infertility and IVF Unit, Department of Obstetrics and Gynecology, Helen Schneider Hospital for Women, Rabin Medical Center, Petach Tikva 49100, Israel (FAX: 972-3-5028107; E-mail: [email protected]).
0015-0282/09/$36.00 doi:10.1016/j.fertnstert.2007.11.086
On the basis of these results, the clinical practice of coasting has been extended to GnRH antagonist protocols. However, data on the effect on IVF outcome in this setting remain limited (11–13). The aim of the present study was to compare the IVF results after coasting in hyper-responders treated with a GnRH antagonist protocol versus a GnRH agonist protocol and to investigate whether the same criteria for coasting can be applied in these protocols. MATERIALS AND METHODS All IVF cycles from 2000 to 2006 in which gonadotropins were withheld for at least 1 day before oocyte retrieval due to hyper-response were retrospectively analyzed. Institutional Review Board approval was obtained for the study. Treatment Protocol The GnRH long protocol consisted of daily injections of Decapeptyl (Ferring, GmbH, Germany) 0.1 mg or a depot injection of Decapeptyl 3.75 mg at the early follicular or midluteal phase. Down-regulation was confirmed after 13–15 days (no ovarian cysts >18 mm, E2 <74 pmol/L) and was followed by gonadotropin stimulation. The GnRH short
Fertility and Sterility Vol. 91, No. 2, February 2009 Copyright ª2009 American Society for Reproductive Medicine, Published by Elsevier Inc.
377
protocol consisted of daily injections of Decapeptyl 0.1 mg starting on day 2 or 3 of menstruation, followed by gonadotropin stimulation from day 4 or 5 of menstruation. The antagonist protocol consisted of daily gonadotropin stimulation from day 3 or 4 of menstruation followed by daily injections of Cetrotide 0.25 mg (Serono, Switzerland) or Orgalutran 0.25 mg (N.V. Organon, the Netherlands) once the leading follicle reached 14 mm and until the day of hCG injection. In all treatment protocols, gonadotropin stimulation consisted of recombinant FSH (Gonal F, Serono, Postfach, Switzerland; or Puregon, N.V. Organon, Oss, The Netherlands), alone or in combination with urinary gonadotropins (Menogon, Ferring, Kiel, Germany).
Data Collection The following data were collected for analysis: baseline clinical parameters (patient age, day 3 FSH level, and cycle number); IVF cycle parameters (type of protocol, daily dose of gonadotropins, number of oocytes retrieved, number of mature oocytes, fertilization rate, embryo quality, and number of embryos transferred); IVF cycle outcome (pregnancy rate); and coasting parameters (E2 level on the day coasting was started, E2 level after coasting, day of hCG administration, and days of coasting). Findings were compared between GnRH agonist and GnRH antagonist coasting cycles. The main outcome measures were number of retrieved oocytes and pregnancy rate.
The choice of a GnRH agonist or antagonist protocol for COH was made by the treating physician on the basis of the patient characteristics or response during previous IVF cycles. In all protocols, the initial dose of gonadotropins was selected according to the individual response of the patient to previous stimulations, if available. During treatment, the ovarian response was monitored by vaginal ultrasound measurements of follicular growth and serum E2 level every 1–3 days, starting on the fifth or sixth day of stimulation, and the FSH and hMG dosages were adjusted accordingly.
Statistical Analysis Data were managed and analyzed with the SPSS statistical package, version 15, for Windows (SPSS, Chicago). A logistic multivariate regression analysis (with enter and forward likelihood ratio analysis), analysis of variance with post hoc Tukey’s multiple comparison tests, the KruskalWallis test, the Mann-Whitney test, and the c2-test were used, as appropriate. P<.05 was considered statistically significant.
Management During Coasting Coasting was applied in all COH cycles, regardless of the type of protocol, when the serum E2 concentration measured 10,000–12,000 pmol/mL in the absence of a leading follicle of 14 mm or more in diameter. During the coasting period, gonadotropins were withheld while administration of the GnRH agonist or antagonist was continued. In all coasting cycles, hCG (Chorigon, Teva, Israel; or Pregnyl, N.V. Organon, 10,000 IU; or Ovitrelle, Serono, Italy, 250 mg) was administered when E2 concentration levels dropped at least 15% relative to the peak levels of E2 in the presence of at least two leading follicles measuring 17 mm or more. Cycles were canceled if coasting lasted more than 6 days.
RESULTS During the study period, 7866 IVF cycles were performed in our IVF unit; a GnRH antagonist protocol was used in 3107 cycles (39.5%), and a GnRH agonist protocol was used in the remainder (4759 cycles). Coasting was applied in 374 cycles: in 45 after stimulation with a GnRH antagonist protocol (antagonist group, 1.5% of all GnRH antagonist cycles) and in 329 after stimulation with a GnRH agonist protocol (agonist group; 6.9% of all GnRH agonist cycles). The difference in the use of coasting between the two protocols was statistically significant (P<.001, odds ratio [OR] ¼ 5.06; 95% confidence interval [CI], 3.7–6.8).
Oocytes were fertilized by standard insemination or intracytoplasmic sperm injection, depending on sperm parameters or previous IVF performance. Embryos were transferred on day 2 or 3. The luteal phase was supported by daily vaginal P (Utrogestan, Besins International Laboratories, France, 600 mg/day; or Endometrin, Ferring, 200 mg/day). A serum pregnancy test was performed 12 days after ET. Clinical pregnancy was defined as the presence of an intrauterine gestational sac on ultrasound scan 4 weeks after oocyte retrieval. Embryo Quality Embryos were graded by gross morphological appearance under light microscopy at 48 or 72 hours after oocyte collection, using the system of Staessen et al. (14): grade 1, even and homogeneous blastomeres without fragmentation; grade 2, even and homogeneous blastomeres with <20% fragmentation; grade 3, uneven and nonhomogeneous blastomeres with 20%–50% fragmentation; grade 4, uneven and nonhomogeneous blastomeres with >50% fragmentation. 378
Farhi et al.
Coasting in agonist vs. antagonist cycles
There was no significant difference between the antagonist and agonist groups in patient age (30.5 4.8 vs. 33.1 6.2 years), baseline FSH level (5.7 1.9 vs. 6.1 1.8 IU/L), or number of previous IVF attempts before the coasting cycle (3.8 3.1 vs. 4.4 4.0). The characteristics of the GnRH antagonist and GnRH agonist coasting cycles are shown in Table 1. The mean total number of gonadotropin ampoules used before coasting was similar in the agonist and antagonist groups, as was the duration of coasting (in days). On the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. However, there was no between-group difference in the mean number of oocytes retrieved, fertilization rate, or rate of high-quality embryos. The mean number of transferred embryos was also similar in the two groups. The rate of OHSS after coasting (defined as moderate-severe OHSS needing hospitalization) was 4.6% in the agonist group and 4.4% in the antagonist group. The corresponding pregnancy rates were 27.4% and 24.4% (Table 1). Vol. 91, No. 2, February 2009
TABLE 1 Characteristics of the GnRH antagonist and GnRH agonist coasted cycles.
No. of ampoules (units) Coasting days Coasting of more than 2 days, n (%) E2 level at initiation of coasting (pmol/L) E2 level at end of coasting (pmol/L) (hCG day) E2 ratio (peak/nadir) (pmol/L) Endometrial width, mm No. of retrieved oocytes Fertilization rate, (%) No. of high-quality embryos available for transfera No. of transferred embryos Hospitalization for moderatesevere OHSS, n (%) Pregnancy rate, n (%)
Agonist group (n [ 329 cycles)
Antagonist group (n [ 45 cycles)
P
27.7 12.5 (2078 939) 1.3 0.7 22 (6.7)
28.3 14.7 (2129 1106) 1.2 0.3 0
.742 .080 .074
11,447 4507 12,741 5573
10,750 4196 10,140 4916
.328 .030
1.07 0.9 10.0 2.1 13.6 6.5 60.0 24.8 1.7 2.2
1.3 1.2 9.6 2.1 14.1 7.9 62.5 28.3 1.2 1.2
.142 .259 .630 .500 .314
2.2 1.0 15 (4.6)
2.1 1.1 2 (4.4)
.500 .972
90 (27.4)
11 (24.4)
.672
Note: Values given as mean SD unless otherwise specified. a Day 2 embryos: grade A or AB, 4 cells; day 3 embryos: grade A or AB, 6–8 cells. Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
None of the antagonist cycles required more than 2 days of coasting, whereas 5.8% of the agonist cycles required 3 or more days of coasting (Table 2). Longer coasting was associated with a significant decrease in the number of retrieved oocytes in the antagonist group (2 days of coasting compared with 1 day of coasting) but not in the agonist group (1–4 days of coasting). In addition, the mean number of retrieved oocytes after 1 day of coasting was similar in the agonist and antagonist groups (13.6 6.5 and 15.4 7.7, respectively; P¼.526), but after 2 days of coasting the number of retrieved oocytes was significantly lower in the antagonist group versus the agonist group (7.0 5.1 and 13.9 6.6, respectively; P¼.01). The association between duration of coasting and E2 level at the end of the coasting period (hCG day) is presented in Table 3. There was no significant difference in E2 levels between the pregnancy and nonpregnancy cycles; the E2 levels at initiation of coasting were 10,801 3810 versus 11,571 4682 (pmol/L), respectively (P¼.140), and the E2 levels at end of coasting period on hCG day were 12,723 4238 versus 12,351 6004 (pmol/L), respectively (P¼.656). A multivariate logistic regression analysis was performed to identify factors predicting outcome (pregnancy/no pregnancy) in the coasting cycles (Table 4). We found that the treatment protocol (antagonist vs. agonist coasting cycle) was not a predictor of a successful cycle and that it did not influence laboratory parameters or outcome. Lower fertilization rate was the only factor associated with a lower success rate after coasting. Fertility and Sterility
DISCUSSION Despite the many studies conducted on coasting in GnRH agonist cycles, there is still a paucity of data for GnRH antagonist protocols. The use of coasting in GnRH antagonist protocols was first described in case reports that showed no compromise of IVF outcome (11, 12). A similar conclusion was reached in the first cohort study of coasting in 29 GnRH antagonist cycles, which was reported by Bahceci et al. (13) in 2006. The present study compared the results of IVF cycles after coasting in 45 patients treated with a GnRH antagonist protocol and 329 patients treated with a GnRH agonist protocol. This comparison was feasible, since we used the same criteria for the application of coasting in hyper-responders in both protocols. The key findings of the study were as follows: [1] In the antagonist group, there was no need for more than 2 days of coasting, whereas in the agonist group, 5.8% of the cycles required 3 or more days of coasting. [2] There was a significant decrease in the number of retrieved oocytes with longer coasting in the antagonist group but not in the agonist group. [3] On the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. [4] There was no between-group difference in mean number of oocytes retrieved, fertilization rate, or rate of high-quality embryos. [5] The OHSS and pregnancy rates after coasting were similar in the agonist and antagonist groups. The difference between the protocols in the duration of coasting until the E2 levels dropped sufficiently for hCG 379
TABLE 2 Association of duration of coasting with number of retrieved oocytes and pregnancy rate. All cycles (n [ 374) Agonist group (n [ 329 cycles) Antagonist group (n [ 45 cycles) Duration of coasting, days 1 2 3 R4 P
Cycles 284 68 14 8
Pregnancies Cycles (%) (oocytes/cycle) 77 (27.1) 17 (25) 4 (28.6) 3 (37.5) .895c
Pregnancies (%)
Cycles (oocytes/cycle)
Pregnancies (%)
68 (27.6) 15 (24.6) 4 (28.6) 3 (37.5) .880c
38 (15.4 7.7) 7 (7.0 5.1) — — .007b
9 (23.7) 2 (28.6) — — .782c
246 (13.6 6.5) 61 (13.9 6.6) 14 (12.4 7.9) 8 (11.5 4.1) .624a
Note: Number of retrieved oocytes given as mean SD. a Kruskal-Wallis test. b Mann-Whitney test. c 2 c -test. Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
administration was not statistically significant. In both groups, the vast majority of patients needed up to 2 days of coasting (Table 2). The fact that fewer oocytes were retrieved after 2 days of coasting in the antagonist group only may indicate a higher rate of atresia of oocytes in coasted antagonist versus coasted agonist cycles. It seems that atresia induced by coasting affects the number of retrieved oocytes (Table 2) but is not concomitantly reflected in the E2 levels. This may be explained by different effects of duration of coasting on oocyte survival and steroidogenesis. Previous studies noted that the duration of coasting in long GnRH agonist protocols had no deleterious effect on IVF outcome, provided that coasting did not last for more than 4 days (8, 9, 15–17). In GnRH agonist cycles, cycle cancellation should be discussed in cases of a longer duration (9). In GnRH antagonist protocols, prolonged coasting was not needed (Table 2), and Bahceci et al. (13) reported similar pregnancy rates after prolonged coasting in GnRH antagonist protocols. The association of E2 levels before and after coasting with cycle outcome has been evaluated in GnRH agonist cycles. No correlation was found between pregnancy or implantation
rates and either the magnitude of the E2 drop (10) or the level to which E2 decreased during coasting (8, 17, 18). There are no studies of this factor in GnRH antagonist protocols. In our study, on the day of hCG administration, E2 levels dropped to a lower level in the antagonist cycles. Despite the greater magnitude of the drop, however, the difference in E2 levels from the agonist cycles was not statistically significant (Table 1). In agreement with earlier findings on the effect of E2 levels in GnRH agonist coasted cycles, neither univariate nor multivariate analysis (Table 4) in the present study yielded a correlation of pregnancy rate with E2 dynamics in GnRH antagonist coasting cycles. Regarding the correlation between the duration of coasting and number of oocytes retrieved, Mansour et al. (9) found that in a GnRH agonist protocol, both the number of oocytes retrieved and the number of metaphase II oocytes were significantly higher in cycles in which coasting lasted less than 3 days compared with cycles with coasting of 4 days and more. In our study, we found that the effect of the duration of coasting on the number of oocytes retrieved is much higher in the antagonist group. A significant decrease in the
TABLE 3 Association of duration of coasting with E2 level (pmol/L) at the end of coasting period on hCG day. Duration of coasting, days
Agonist group (n [ 329 cycles)
Antagonist group (n [ 45 cycles)
11,472 3592 12,542 5683 22,489 7984 18,358 8217 < .001
9963 4504 10,672 6463 — — .767
1 2 3 R4 P Note: Values given as mean SD. Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
380
Farhi et al.
Coasting in agonist vs. antagonist cycles
Vol. 91, No. 2, February 2009
TABLE 4 Multivariate regression analysis of all coasted cycles: predictors for pregnancy. Variable Age (years) Cycle number Previous pregnancy? (yes or no) No. of ampoules Coasting days E2 level before coasting (mmol/L) E2 level after coasting (mmol/L) Endometrial width (mm) No. of retrieved oocytes Fertilization rate (%) No. of transferred embryos Treatment protocol (agonist vs. antagonist)
P .776 .887 .545 .077 .583 .079 .104 .184 .741 .034 .877 .873
Farhi. Coasting in agonist vs. antagonist cycles. Fertil Steril 2009.
number of retrieved oocytes was noted between 1 and 2 days of coasting, whereas in the agonist group, the number of oocytes retrieved was not changed even after 4 days of coasting (Table 2). The overall fertilization rate in our study after coasting was similar between the agonist and antagonist groups and similar to the expected rate in coasting cycles reported in earlier studies for the GnRH agonist protocol (9) and the GnRH antagonist protocol (13) cycles. The proportion of high-quality embryos was also similar in our GnRH antagonist and agonist groups; thus it seems that the protocol does not affect embryo quality after coasting. In our study, we compared embryo quality only in coasted cycles. We did not include a comparison with noncoasted cycles, although previous studies of embryo quality after coasting have suggested that coasting causes a reduction in oocyte quality in both GnRH agonist protocols (17) and GnRH antagonist protocols (13). Moreover, there are data suggesting that a profound drop in E2 level (<1000 pg/mL) in GnRH agonist protocols may be related to low oocyte quality (16). Pregnancy rates were similar in the GnRH antagonist and GnRH agonist groups in our study; therefore, it seems that the type of protocol has no effect on pregnancy rates after coasting. Previous studies, assessing each protocol separately, reported normal implantation and pregnancy rates for both protocols, as long as coasting lasted for up to 4 days (9, 10, 13). Overall, the need for coasting in our study population was significantly lower in GnRH antagonist protocols (1.5%) than in GnRH agonist protocols (6.9%, P<.001; OR ¼ 5.06; 95% CI, 3.7–6.8). However, the choice of protocols for selected populations with different ovarian response probably precludes independent analysis. Be that as it may, once Fertility and Sterility
hyper-response had developed and coasting was applied, the OHSS rate in the agonist and antagonist groups was similar (4.6% and 4.4%, respectively). One limitation of the current study is the relatively small number of coasted antagonist cycles. In a sample sizes analysis, 3645 subjects in each group are needed to achieve 80% power to detect a difference between 27.3% and 24.4% (the actual pregnancy rates in the agonist and antagonist groups, respectively) with a significance level of .05. Since coasting is needed in only 1.5% of the antagonist cycles, such numbers are difficult to achieve; moreover, the difference between these pregnancy rates is clinically insignificant. In conclusion, the same criteria for coasting can be applied in GnRH antagonist and GnRH agonist protocols, with a similar laboratory performance and cycle outcome.
REFERENCES 1. Sher G, Zouves C, Feinman M, Maassarani G. Prolonged coasting: an effective method for preventing severe ovarian hyperstimulation syndrome in patients undergoing in-vitro fertilization. Hum Reprod 1995;10:3107–9. 2. Benadiva CA, Davis O, Klingman I, Moomjy M, Liu HC, Rosenwaks Z. Withholding gonadotropin administration is an effective alternative for the prevention of ovarian hyperstimulation syndrome. Fertil Steril 1997;67:724–7. 3. Delvigne A, Rozenberg S. Preventive attitude of physicians to avoid OHSS in IVF patients. Hum Reprod 2001;16:2491–5. 4. D’Angelo A, Amson N. ‘‘Coasting’’ (withholding gonadotrophins) for preventing ovarian hyperstimulation syndrome. Cochrane Database Syst Rev 2002;3:CD002811. 5. Chen D, Burmeister L, Goldschlag D, Rosenwaks Z. Ovarian hyperstimulation syndrome (strategies for prevention). Reprod Biomed Online 2003;7:43–9. 6. Garcia-Velasco JA, Isaza V, Quea G, Pellicer A. Coasting for the prevention of ovarian hyperstimulation syndrome: much ado about nothing? Fertil Steril 2006;85:547–54. 7. Garcia-Velasco JA, Zuniqa A, Pacheco A, Gomez R, Simon C, Remohi J, et al. Coasting acts through downregulation of VEGF gene expression and protein secretion. Hum Reprod 2004;19:1530–8. 8. Ulug U, Ben-Shlomo I, Bahceci M. Predictors of success during the coasting period in high-responder patients undergoing controlled ovarian hyperstimulation for assisted reproduction. Fertil Steril 2004;82: 338–42. 9. Mansour R, Aboulghar M, Serour G, Amin Y, Abou-Setta AM. Criteria of a successful coasting protocol for the prevention of severe ovarian hyperstimulation syndrome. Hum Reprod 2005;20:3167–72. 10. Kovacs P, Matyas S, Kaali SG. Effect of coasting on cycle outcome during in vitro fertilization/intracytoplasmic sperm injection cycles in hyper-responders. Fertil Steril 2006;85:913–7. 11. Delvigne A, Carlier C, Rozenberg S. Is coasting effective for preventing ovarian hyperstimulation syndrome in patients receiving a gonadotropinreleasing hormone antagonist during an in vitro fertilization cycle? Fertil Steril 2001;76:844–6. 12. Fatemi HM, Platteau P, Albano C, Van Steirteghem A, Devroey P. Rescue IVF and coasting with the use of a GnRH antagonist after ovulation induction. Reprod Biomed Online 2002;5:273–5. 13. Bahceci M, Ulug U, Tosun S, Erden HF, Bayazit N. Impact of coasting in patients undergoing controlled ovarian stimulation with the gonadotropin-releasing hormone antagonist cetrorelix. Fertil Steril 2006;85: 1523–5. 14. Staessen C, Camus M, Bollen N, Devroey P, Van-Steirteghem AC. The relationship between embryo quality and the occurrence of multiple pregnancies. Fertil Steril 1992;3:626–30.
381
15. Tortoriello DV, McGovern PG, Colon JM, Skurnick JH, Lipetz K, Santoro N. ‘‘Coasting’’ does not adversely affect cycle outcome in a subset of highly responsive in vitro fertilization patients. Fertil Steril 1998;69:454–60. 16. Isaza V, Garcia-Velasco JA, Aragones M, Remhi J, Simon C, Pellicer A. Oocyte and embryo quality after coasting: the experience from oocyte donation. Hum Reprod 2002;17:1777–82.
382
Farhi et al.
Coasting in agonist vs. antagonist cycles
17. Delvigne A, Kostyla K, Murillo D, Van Hoeck J, Rozenbers S. Oocyte quality and IVF outcome after coasting to prevent ovarian hyperstimulation syndrome. Int J Fertil Womens Med 2003;48:25–31. 18. Levinsohn-Tavor O, Friedler S, Schachter M, Raziel A, Strassburger D, Ron-El R. Coasting—what is the best formula? Hum Reprod 2003;18: 937–40.
Vol. 91, No. 2, February 2009