Pregnancy loss after frozen-embryo transfer a comparison of three protocols

Transcription

1 ORIGINAL ARTICLES: EARLY PREGNANCY Pregnancy loss after frozen-embryo transfer a comparison of three protocols Candido Tomas, M.D., Ph.D., a Birgit Alsbjerg, M.D., b Hannu Martikainen, M.D., Ph.D., c and Peter Humaidan, M.D., D.M.Sc. d a AVA Clinic, Fertility Center, Tampere, Finland; b Fertility Clinic, Skive Regional Hospital, Skive, Denmark; c Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Infertility, University of Oulu, Oulu, Finland; and d The Fertility Clinic, Department D, Odense University Hospital, Odense, Denmark Objective: To compare the reproductive outcome of three protocols for frozen ET treatment. Design: Retrospective follow-up study. Setting: Two public clinics and one private clinic. Patient(s): Four thousand four hundred seventy frozen ET between 2006 and Intervention(s): Thawing of embryos and ET. Main Outcome Measure(s): Pregnancy test rate, clinical pregnancy rate, and pregnancy loss rate. Result(s): The natural cycle followed by P (NC þ P) was used in 26% of, the natural cycle with hcg (NC þ hcg) in 10%, and the substituted cycle with estrogen and P (E þ P) in 64% of. The rate of transfers after thawing was similar in all groups (87.2%, 73.9%, and 87.2%, respectively). There was a significantly higher positive pregnancy test rate in the E þ P (34.3%) and NC þ hcg (35.5%) as compared with the NC þ P (26.7%). However, the clinical pregnancy rate was similar in all groups (27.7%, 29.1%, and 24.3%, respectively). Moreover, no differences were seen between groups regarding the live-birth rate (20.1%, 23.5%, and 20.7%, respectively). A logistic regression analysis showed that the type of protocol was the only predictor of pregnancy loss, while age, irregular, endometrial thickness, number, and quality of embryos transferred did not correlate to pregnancy loss. Conclusion(s): A higher positive pregnancy test rate was obtained in E þ P frozen ET in comparison with other protocols; however, due to an increased preclinical and clinical pregnancy loss, comparable clinical pregnancy, and delivery rates are reported for the three protocols. (Fertil Steril Ò 2012;98: Ó2012 by American Society for Reproductive Medicine.) Key Words: FET, frozen embryo transfer, substituted, pregnancy loss, luteal support Discuss: You can discuss this article with its authors and with other ASRM members at fertstertforum.com/tomasc-pregnancy-loss-frozen-embryo-transfer/ Use your smartphone to scan this QR code and connect to the discussion forum for this article now.* * Download a free QR code scanner by searching for QR scanner in your smartphone s app store or app marketplace. Cryopreservation of spare embryos after controlled ovarian stimulation has increased significantly during the last decade, coinciding with an increase in the elective single embryo transfer policy (e-set) and the improvement of freezing programs (1). As a result, an increase in the cumulative pregnancy rate per oocyte retrieval has been obtained (1 3), despite the fact that the pregnancy rate is lower after frozen ET (FET) compared with fresh ET (3, 4). In 2007, the proportion of FET treatments in relation to all treatments was 19.5% in Europe. However, this proportion was significantly higher in countries with a strong policy of e-set, like, for instance, Finland, where the proportion of FET was 42.8% during the same period (3). Received January 26, 2012; revised June 6, 2012; accepted July 3, 2012; published online July 27, C.T. reports honoraria for lectures from MSD (unrelated to this work). B.A. has nothing to disclose. H.M. has nothing to disclose. P.H. reports unrestricted research grants and honoraria for lectures from Merck Serono and MSD (unrelated to this work). Reprint requests: Candido Tomas, M.D., Ph.D., AVA Clinic, Fertility Center, Keskustori 1 A 9, Tampere, Finland ( ctomas@avaclinic.com). Fertility and Sterility Vol. 98, No. 5, November /$36.00 Copyright 2012 American Society for Reproductive Medicine, Published by Elsevier Inc. A variety of protocols may be used to prepare the endometrium for the transfer of frozen-thawed embryos. The most common protocols used in Scandinavia are the natural cycle, the substituted cycle, using estrogen followed by P to prime the endometrium, and the hcg-induced unstimulated cycle, hereafter referred to as the natural cycle with hcg (5, 6). A recent Cochrane review, comparing different protocols for FET, reported no significant differences in pregnancy rates between the protocols used, which prompted the authors to conclude that no evidence supports the use of one protocol over the other in women with regular spontaneous (7). VOL. 98 NO. 5 / NOVEMBER

2 ORIGINAL ARTICLE: EARLY PREGNANCY In contrast, controversy exists as regards the results of studies comparing the reproductive outcome after stimulated versus natural cycle; thus, some investigators report no differences, while others report superiority of the natural cycle (7). The purpose of this large retrospective analysis was to evaluate and compare three different protocols for FET to identify predictive factors and influence the reproductive outcome. MATERIALS AND METHODS The study included 4,470 FET performed during a period of 5 years ( ) in three Scandinavian clinics: AVA Clinic, Fertility Center, Tampere, Finland; Department of Obstetrics and Gynecology of the University of Oulu, Finland; and the Fertility Clinic, Skive Regional Hospital, Skive, Denmark. The embryos were derived from either IVF or intracytoplasmic sperm injection treatment. The Endometrial Preparation Protocols The protocols used before the ET were the following: luteal P supplementation (NC þ P): this protocol was used in patients with regular menses (25 34 days). After a spontaneous menses a vaginal ultrasound examination was performed on the days day of the cycle. Based on the ultrasound examination, the patient was instructed to start morning urinary LH testing. Upon identification of a dominant follicle and an LH surge, the ET was scheduled for 3 5 days after the surge. Day 2 embryos were cultured overnight after thawing, and the ET was performed on the third day of embryo development. Vaginal luteal phase support commenced on the day of ET, using natural P 200 mg 3 daily, natural P 400 mg 2 daily, or Crinone 90 mg 2 daily during a 14-day period, after which P was discontinued, independent of the result of the pregnancy test. hcg for ovulation induction (NC þ hcg): this protocol was used in patients with regular menses (25 34 days). An ultrasound examination was performed 10 days after a spontaneous menses to detect the leading follicle. Once the dominant follicle had reached a size of mm, a bolus of 5,000 IU hcg IM (Pregnyl, NV Organon), was administered for induction of ovulation, and embryos were thawed and transferred 5 days after. No luteal support was given. cycle (E þ P): this protocol was used in patients with irregular (>34 days) as well as regular. The endometrial preparation started with E 2 2mg3/day, orally, from the first cycle day of a natural or P-induced menstrual cycle. A vaginal ultrasound was performed after 10 days. When the thickness of the endometrium was at least 7 mm, the ET was scheduled after overnight culturing of the embryos. Substitution with P vaginally 400 mg twice daily or Crinone 90 mg twice daily started 4 days before the day of the ET. The supplementation continued until a pregnancy test was performed. In case of a positive test, the patients were advised to continue the luteal support until the 10th gestational week. Pregnancy test and ultrasound examination: The pregnancy test was performed 14 days after ET. A vaginal ultrasound was performed during the sixth gestational week to confirm a clinical pregnancy, that is, presence of a gestational sac or heartbeat. A biochemical pregnancy was defined as a positive pregnancy test with the absence of a clinical pregnancy (preclinical); a miscarriage was considered to be a pregnancy loss until week 22 of gestation after previous confirmation of a clinical pregnancy. In this study, biochemical pregnancies (preclinical losses) and miscarriages (clinical losses) were considered to be pregnancy losses. The criteria for selection of embryos for transfer after thawing were similar for the three centers. Embryos were frozen according to the slow freezing protocol and thawed using a standard protocol as described elsewhere (8, 9). The transferred embryos had at least 50% intact blastomeres after thawing and cleaved after overnight culturing. All transfers were performed with an endometrial thickness of at least 6 mm. The number of embryos transferred in each case depended on the number of previous treatments, the number of embryos frozen in the same straw, and the quality of available embryos. During the study period, there were no differences throughout the years concerning the use of the various protocols and their outcomes. Institutional Review Board approval was not required for the present study owing to its retrospective nature and the fact that the study data completely excluded the identification of subjects. All patients had given written authorization at the time of treatment for the future use of their clinical data. Statistics The clinical and embryological data were collected from the IVF database program Babe 3.3 (Cleodora, Medical Software Lda.) and analyzed using the software package JMP 8.0 for Macintosh (SAS Institute Inc.). A comparison of the treatment outcomes among the three protocols was performed using the c 2 -test. Univariate and multivariate logistic regression analysis were performed to identify predictive factors of pregnancy loss. RESULTS A total of 4,470 consecutive FET were included in this retrospective analysis. The natural cycle protocol with luteal P supplementation (NC þ P) was used in 26%, the natural cycle protocol with hcg for ovulation induction (NC þ hcg) in 10%, and the substituted cycle protocol (E þ P) in 64% of the. The demographic data of each group are shown in Table 1. At the time of the FET cycle, the patients in the E þ P group were significantly older and had significantly more irregular (cycle >34 days) than the patients in the two other groups (Table 1). Compared with the NC þ P group, the patients in the E þ P group had significantly more primary infertility (P<.0002) and a significantly thicker endometrium at the 1166 VOL. 98 NO. 5 / NOVEMBER 2012

3 Fertility and Sterility TABLE 1 Demographic data of patients with FET in natural with P, natural with hcg, and substituted. Treatments, n (%) 1,168 (26) 444 (10) 2,858 (64) Age (y), mean SD <.0001 BMI (kg/m 2 ), mean SD NS Irregular cycling patients, % <.0001 Duration of infertility (y), mean SD a NS Primary infertility, % 56 a 69 <.0002 Endometrial thickness at ET (mm), mean SD a <.0001 Embryos transferred, mean SD <.0001 Note: NS ¼ no significant differences. Not available. time of the ET (P<.0001; Table 1). Pregnant patients had a significantly thicker endometrium than nonpregnant patients ( mm vs mm, respectively; P<.0001). In the E þ P, the endometrium was significantly thicker in pregnant than in nonpregnant patients ( mm vs mm, respectively; P¼.01). Similarly, in the NC þ P group, pregnant patients had a thicker endometrium than nonpregnant patients ( mm vs mm, respectively; P¼.02; data not shown). The clinical outcome is shown in Table 2. Patients in the NC þ P had ET after thawing in 87.2% of the cases, in the NC þ hcg protocol in 73.9% of the cases, and in the E þ P in 87.2% of the cases (NS). A significantly higher positive pregnancy test rate was seen in the E þ P (34.3%) and NC þ hcg (35.5%) as compared with in the NC þ P (26.7%; P<.0001). However, the clinical pregnancy rate per transfer was similar in E þ P, NC þ hcg. and NC þ P groups (27.7%, 29.1%, and 24.3%, respectively; NS). Moreover, live-birth rates per ET were similar in the three groups (20.1%, 23.5%, and 20.7%, respectively; NS; Table 2). In the E þ P, both the preclinical pregnancy losses (19.1% vs. 8.8%; P<.001) and clinical pregnancy losses (22.4% vs. 13.6%, respectively; P<.005) were significantly higher than in the NC þ P. The preclinical and clinical pregnancy loss rate after a positive pregnancy test for the three protocols is shown in Table 3. A significantly higher pregnancy loss (P<.0001) was seen in the E þ P (41.5%) as compared with in the NC þ hcg (33.6%) and NC þ P (22.4 %). A logistic regression model was used to identify factors that significantly correlate to the pregnancy loss after a positive pregnancy test in FET. Only the type of protocol used correlated to the pregnancy loss, independently of other factors. In contrast, the age of the patient, the presence of irregular, the thickness of the endometrium, the number of embryos transferred, and the number of top-quality embryos transferred did not correlate to pregnancy loss (Table 4). DISCUSSION This analysis of more than 4,000 revealed that the chance of achieving a positive pregnancy test was significantly higher in NC þ hcg and E þ P than in NC þ P. However, owing to a significantly higher preclinical as well as clinical pregnancy loss in the E þ P and the NC þ hcg compared with in the NCþP, the delivery rates per ET were similar for all three protocols. These differences are seen in both preclinical and clinical pregnancy losses. The present findings are in line with the results of a large retrospective study including 1,677 patients that also reported a higher pregnancy rate in the E þ P cycle in comparison with NC þ hcg ; however, no significant differences in live-birth rate was seen owing to a higher pregnancy loss in the substituted cycle (10). The reason for the higher pregnancy loss rate seen in E þ P remains unclear but could be related to the higher number of patients with polycystic ovarian syndrome in this group. However, Veleva et al. found that hormonal TABLE 2 Treatment outcome after FET in natural with P, natural with hcg, and substituted. Started, n 1, ,858 ET, n (%) 1,019 (87.2) 327 (73.9) 2,492 (87.2) NS Positive pregnancy test/et, n (%) 272 (26.7) 116 (35.5) 854 (34.3).0001 Clinical pregnancy/et, n (%) 248 (24.3) 95 (29.1) 691 (27.7) NS Deliveries/ET, n (%) 211 (20.7) 77 (23.5) 500 (20.1) NS Deliveries/started cycle, % NS Note: NS ¼ no significant differences. VOL. 98 NO. 5 / NOVEMBER

4 ORIGINAL ARTICLE: EARLY PREGNANCY TABLE 3 Pregnancy loss after FET in natural with P, natural with hcg, and substituted. ET 1, ,492 Pregnancy test/et, n (%) 272 (26.7) 116 (35.5) 854 (34.3) <.0001 Clinical pregnancy/et, n (%) 248 (24.3) 95 (29.1) 691 (27.7) NS Deliveries/ET, n (%) 211 (20.7) 77 (23.5) 500 (20.1) NS Preclinical pregnancy loss, n (%) 24 (8.8) 21 (18.1) 163 (19.1) <.0001 Clinical pregnancy loss, n (%) 37 (13.6) 18 (15.5) 191 (22.4) <.005 Total pregnancy loss, n (%) 61 (22.4) 39 (33.6) 354 (41.5) <.0001 substitution per se for FET was associated with a 1.7-fold higher miscarriage rate compared with fresh assisted reproductive technology (ART) (11). Moreover, an increased pregnancy loss in FET compared with other ART treatments has been reported in studies in which the FET protocol was not taken into account (12, 13). High and low body mass index (BMI) has been associated with increased miscarriage after FET (11, 14, 15). In the present study, however, the BMI was similar in all groups. Another possible explanation for the higher pregnancy loss in E þ P can be an excessive estrogen environment or a suboptimal ratio between P and E 2. In the study of Morozov et al., the level of E 2 before transfer was higher in the substitution in comparison with the natural cycle (16). On the other hand, some studies indicate that neither supraphysiological nor low serum E 2 levels are related to the success of treatment (17). The logistic regression analysis of data from the present study revealed that the protocol used for FET was the only factor influencing pregnancy loss, independently of other factors. Thus, irregular, the age of the woman at the time of ET, and the number of top-quality embryos transferred do not seem to play an independent role for pregnancy loss after FET. The finding that age is not an independent factor in the logistic regression might be due to the fact that the embryos transferred in this study were high-quality embryos collected in women with a mean age below 35 years. Thus, one would expect a similar aneuploidy rate between the groups. Finally, when discussing early pregnancy loss, apart from the protocol used, one also needs to consider the role of the embryo. Thus, although a previous study has shown that freezing may increase aneuploidy and polyploidy in frozen mouse embryos (18), in the present study changes should be similar in all groups and would not explain the difference in pregnancy loss. It is known from previous studies that the window of embryo implantation starts after 48 hours of initiation of P and remains open for approximately 4 days (19). All ETs in this study were performed within the expected window of implantation. However, it cannot be excluded that differences in preclinical and clinical pregnancy losses found between the protocols might be correlated to a slightly different timing and/or endometrial receptivity. In line with this, Wilcox et al. showed that the risk of early pregnancy loss increased significantly with a later implantation within the window (20). The advantages of a hormonal substitution regimen with estrogen and P in FET treatment are the simplicity and flexibility in terms of timing of transfer, which make it the protocol of choice for women with irregular. A thicker endometrium at the time of transfer was associated with higher rates of positive pregnancy tests in the present study, a finding that was also reported in previous studies (21, 22). The difference in endometrial thickness found in the present study (approximately 1 mm) cannot be considered TABLE 4 Likelihood of pregnancy loss in a logistic regression model including the protocol and other demographic factors. Pregnancy loss, univariate c 2 -test, multivariate Age Irregular in the diagnosis, %: Yes 42.8 No 35.3 Endometrial thickness at ET Number of embryos transferred, % One 36.9 Two 34.7 Protocol, % < <.0001 luteal P cycle VOL. 98 NO. 5 / NOVEMBER 2012

5 Fertility and Sterility clinically significant. Furthermore, the endometrial thickness was not correlated to early pregnancy loss, as seen in the logistic regression analysis. However, the higher preclinical and clinical pregnancy losses suggest that the placentation process might be suboptimal under certain conditions. So far, there is a lack of evidence regarding the most optimal protocol for FET. In a Cochrane review including seven randomized trials, natural cycle, natural cycle with hcg, and substituted cycle with or without GnRH-analog down-regulation, the investigators concluded that no protocol seemed to be superior to the other in terms of the reproductive outcome (7). In the present study, the group of patients with NC þ hcg had similar pregnancy rates as patients in the NC þ P group (29.1% vs. 24.3%, respectively; NS). This is in contrast to a recent controlled randomized study, in which the ongoing pregnancy rate in natural was shown to be significantly higher (31.1% vs. 14.3%, respectively; P¼.025) than in the NC þ hcg (23). Moreover, significantly more embryos were transferred in the NC þ hcg compared with in the NCþP (1.7 vs. 1.5), which may hide a possible negative impact of hcg. During recent years, Scandinavian countries have adopted a strong SET policy, and this policy is applied to FET as well. The latest report from the National Institute for Health and Welfare from Finland (2011) reporting treatments from 2009 shows that the rate of SET in frozen is similar to the rate of fresh SET transfers (62% vs. 66%, respectively). However, the present study reports the results of treatments over a 5-year period, for which reason the recent strong SET policy in FET is not yet clearly evident. The role of luteal support with P in FET is still unclear. Fatemi et al. found no effect on ongoing pregnancy rate or pregnancy loss by adding vaginal P in the hcg-induced natural (23) and no difference in pregnancy loss. On the other hand, in a recent prospective randomized controlled study, the live-birth rate was significantly higher (30% vs. 20%) in women receiving vaginal P for luteal phase support after FET in natural compared with those who did not receive P (24). Thus, luteal support may increase the success rate after FET in natural. A drawback of the present study is the difference in size of the three groups analyzed. This is obviously owing to the retrospective nature of the study, which reflects daily practice rather than a predetermined scientific hypothesis. In conclusion, this 5-year survey of 4,470 FET performed in three clinics shows that, although a higher number of positive pregnancy tests are seen in E þ P, there is a significantly higher early pregnancy loss in these, both preclinical and clinical, as compared with the two other treatment protocols analyzed. Thus, the delivery rate per ET is similar between protocols. Future research should focus on an optimization of the hormonal substitution used for FET. REFERENCES 1. McLernon DJ, Harrild K, Bergh C, Davies MJ, de Neubourg D, Dumoulin JC, et al. Clinical effectiveness of elective single versus double embryo transfer: meta-analysis of individual patient data from randomised trials. Br Med J 2010;341:c Veleva Z, Karinen P, Tomas C, Tapanainen JS, Martikainen H. Elective single embryo transfer with cryopreservation improves the outcome and diminishes the costs of IVF/ICSI. Hum Reprod 2009;24: de Mouzon J, Goossens V, Bhattacharya S, Castilla JA, Ferraretti AP, Korsak V, et al. Assisted reproductive technology in Europe, 2007: results generated from European registers by ESHRE. Hum Reprod 2012;27: Andersen AN, Goossens V, Ferraretti AP, Bhattacharya S, Felberbaum R, de Mouzon J, et al. Assisted reproductive technology in Europe, 2004: results generated from European registers by ESHRE. Hum Reprod 2008;23: Weissman A, Horowitz E, Ravhon A, Steinfeld Z, Mutzafi R, Golan A, et al. Spontaneous ovulation versus HCG triggering for timing natural-cycle frozen-thawed embryo transfer: a randomized study. Reprod Biomed Online 2011;23: Weissman A, Levin D, Ravhon A, Eran H, Golan A, Levran D. What is the preferred method for timing natural cycle frozen-thawed embryo transfer? Reprod Biomed Online 2009;19: Ghobara T, Vandekerckhove P. Cycle regimens for frozen-thawed embryo transfer. Cochrane Database Syst Rev 2008;1:CD Lassalle B, Testart J, Renard JP. Human embryo features that influence the success of cryopreservation with the use of 1,2 propanediol. Fertil Steril 1985;44: Elder K, Dale B. In-vitro fertilization. In: Elder K, Dale B, ed. In-vitro fertilization. Cambridge: Cambridge University Press, 2010: Givens CR, Markun LC, Ryan IP, Chenette PE, Herbert CM, Schriock ED. Outcomes of natural versus programmed for 1677 frozen-thawed embryo transfers. Reprod Biomed Online 2009;19: Veleva Z, Tiitinen A, Vilska S, Hyden-Granskog C, Tomas C, Martikainen H, et al. High and low BMI increase the risk of miscarriage after IVF/ICSI and FET. Hum Reprod 2008;23: Brandes M, Verzijden JC, Hamilton CJ, de Weys NP, de Bruin JP, Bots RS, et al. Is the fertility treatment itself a risk factor for early pregnancy loss? Reprod Biomed Online 2011;22: Farr SL, Schieve LA, Jamieson DJ. Pregnancy loss among pregnancies conceived through assisted reproductive technology, United States, Am J Epidemiol 2007;165: Helgstrand S, Andersen AM. Maternal underweight and the risk of spontaneous abortion. Acta Obstet Gynecol Scand 2005;84: Lashen H, Fear K, Sturdee DW. Obesity is associated with increased risk of first trimester and recurrent miscarriage: matched case-control study. Hum Reprod 2004;19: Morozov V, Ruman J, Kenigsberg D, Moodie G, Brenner S. cryo-thaw transfer may improve pregnancy outcome. J Assist Reprod Genet 2007;24: Soares SR, Troncoso C, Bosch E, Serra V, Simon C, Remohi J, et al. Age and uterine receptiveness: predicting the outcome of oocyte donation. J Clin Endocrinol Metab 2005;90: Mozdarani H, Moradi SZ. Effect of vitrification on viability and chromosome abnormalities in 8-cell mouse embryos at various storage durations. Biol Res 2007;40: Prapas Y, Prapas N, Jones EE, Duleba AJ, Olive DL, Chatziparasidou A, et al. The window for embryo transfer in oocyte donation depends on the duration of progesterone therapy. Hum Reprod 1998;13: Wilcox AJ, Baird DD, Weinberg CR. Time of implantation of the conceptus and loss of pregnancy. N Eng J Med 1999;340: El-Toukhy T, Coomarasamy A, Khairy M, Sunkara K, Seed P, Khalaf Y, et al. The relationship between endometrial thickness and outcome of medicated frozen embryo replacement. Fertil Steril 2008;89: Dessolle L, Darai E, Cornet D, Rouzier R, Coutant C, Mandelbaum J, et al. Determinants of pregnancy rate in the donor oocyte model: a multivariate analysis. Hum Reprod 2009;24: Fatemi HM, Kyrou D, Bourgain C, Van den Abbeel E, Griesinger G, Devroey P. Cryopreserved-thawed human embryo transfer: spontaneous natural cycle is superior to human chorionic gonadotropin-induced natural cycle. Fertil Steril 2010;94: Bjuresten K, Landgren BM, Hovatta O, Stavreus-Evers A. Luteal phase progesterone increases live birth rate after frozen embryo transfer. Fertil Steril 2011;95: VOL. 98 NO. 5 / NOVEMBER