Premature progesterone rise negatively correlated with live birth rate in IVF cycles with GnRH agonist: an analysis of 2,566 cycles

Transcription

1 Premature progesterone rise negatively correlated with live birth rate in IVF cycles with GnRH agonist: an analysis of 2,566 cycles Rui Huang, M.D., a,b Cong Fang, M.D., a Shuyi Xu, Ph.D., c Yanhong Yi, M.D., b and Xiaoyan Liang, M.D. a a Reproductive Medicine Center, Sixth Affiliated Hospital of Sun Yat-sen University; b Reproductive Medicine Center, Women and Children Hospital of Guangdong Province; and c Lingnan College of Sun Yat-sen University, Guangzhou, People's Republic of China Objective: To investigate the occurrence of premature progesterone rise (PPR) in GnRH agonist long or short protocol, address the relationship between circulating P levels and live birth rates, and explore the possible mechanism through which PPR affects clinical outcomes and the possible factors related to the occurrence of PPR. Design: Retrospective analysis. Setting: Reproductive medicine center of a public hospital. Patient(s): A total of 2,566 patients receiving in vitro fertilization/intracytoplasmic sperm injection treatment with GnRH agonist long or short protocol. Intervention(s): None. Main Outcome Measure(s): Live birth rates. Result(s): The corresponding incidence of PPR in long or short protocol was 22.86% (393/1,719) or 27.63% (234/847) with the cutoff value of 1.2 ng/ml or 2.0 ng/ml, respectively, being used to define PPR. Live birth rates decreased under the condition of PPR (40.65% vs % in long protocol; 30.18% vs % in short protocol). Logistic regression analysis showed that serum P level on the day of hcg administration was a strong predictor of live birth rate in both long and short protocols. Live birth rates in frozen embryo transfer cycles had no significant difference between groups with or without PPR (29.31% vs % in long protocol; 24.84% vs % in short protocol). Multivariate regression analysis showed that exogenous gonadotropin dose, the duration of stimulation, E 2 and LH levels on the day of hcg administration, the number of oocytes retrieved, and basal FSH level were all involved in PPR. Conclusion(s): In GnRH agonist cycles, PPR negatively correlated with live birth rate in fresh embryo transfer cycles, although no adverse impact on frozen embryo transfer was observed, implying that PPR may have deleterious effects on endometrial receptivity. (Fertil Steril Ò 2012;98: Ó2012 by American Society for Reproductive Medicine.) Key Words: Premature progesterone rise, live birth rate, in vitro fertilization, GnRH agonist Discuss: You can discuss this article with its authors and with other ASRM members at fertstertforum.com/huangr-premature-progesterone-rise-birth-rate-ivf-gnrh-agonist/ 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. Serum P level rise on the day of hcg administration in in vitro fertilization (IVF) cycles was not uncommon before the introduction of GnRH analogue. This situation, associated with poor oocyte quality, low fertilization rate, and adverse pregnancy outcome, is thought to be induced by inappropriate LH elevation (1) and therefore termed premature luteinization. It was thought that after the introduction of GnRH analogue in IVF Received February 28, 2012; revised May 20, 2012; accepted May 22, 2012; published online June 15, R.H. has nothing to disclose. C.F. has nothing to disclose. S.X. has nothing to disclose. Y.Y. has nothing to disclose. X.L. has nothing to disclose. R.H. and C.F. contributed equally to this work. Supported by the National Natural Science Foundation of China (grant no ). Reprint requests: Rui Huang, M.D., Reproductive Medicine Center, Sixth Affiliated Hospital of Sun Yat-sen University, Guangzhou, People's Republic of China ( drhuangr@hotmail.com). Fertility and Sterility Vol. 98, No. 3, September /$36.00 Copyright 2012 American Society for Reproductive Medicine, Published by Elsevier Inc. doi: /j.fertnstert cycles, P level rise would be rare because LH elevation would be effectively inhibited. However, the occurrence of subtle P rise on the day of hcg administration still reaches 5% 71% (2 5). This unexpected phenomenon deserves to be given more attention, and it may be proper to use the term premature progesterone rise (PPR) rather than premature luteinization in this circumstance. Nowadays, the influence of PPR on pregnancy outcomes remains controversial, although increasing numbers of papers have been published since 664 VOL. 98 NO. 3 / SEPTEMBER 2012

2 Fertility and Sterility the 1990s. Some showed that PPR was associated with decreased pregnancy rate (6 8), some presented data against the negative effect of PPR (9 11), and some even reported positive correlation between PPR and pregnancy rate (12). In 2007, Venetis et al. (13) published the first systematic review and meta-analysis, including 12 studies, and concluded that PPR was not associated with the probability of pregnancy. But in 2010, Bosch et al. (14) conducted a retrospective study recruiting the largest number of patients (>4,000), showing that PPR (P level >1.5 ng/ml), in both GnRH agonist and GnRH antagonist cycles, was associated with decreased ongoing pregnancy rate. This discrepancy highlights the necessity to reevaluate the effect of PPR on clinical outcomes. In addition, the underlying mechanism through which PPR influences clinical outcomes is elusive. Some showed that the premature of endometrium, a detrimental factor to the endometrial receptivity, was the reason for low pregnancy rate related to PPR (15), whereas some documented that the compromised quality of oocytes might also be a cause (16, 17). As for the pathogenesis and etiology of PPR, which are also poorly understood, several possible mechanisms were suggested, including: 1) the accumulated P production from multiple follicles being developed during controlled ovarian stimulation (COS), each one producing a normal amount of P (8, 9); 2) the overdose of exogenous gonadotropin promoting the conversion of P from cholesterol by acting on granulosa cells (14); 3) premature luteinization, the consequence of subtle LH rise which occurs even when GnRH analogue used (18); and 4) poor ovarian response with increased LH sensitivity (19). It is clinically relevant to address these questions, because if PPR is harmful to clinical outcomes, either triggering hcg in advance or freezing all embryos should be recommended to avoid the adverse consequences. The present retrospective analysis aimed to quantify the incidence of PPR in a large cohort of patients (n ¼ 2,566) undergoing COS with GnRH agonist long or short protocol, to investigate the relationship between serum P level and the probability of live birth, and to explore the possible mechanism through which PPR affects live birth rate and the possible factor(s) related to the occurrence of PPR. MATERIALS AND METHODS This study is a noninterventional retrospective analysis of all IVF/intracytoplasmic sperm injection (ICSI) cycles in women who started their first IVF/ICSI cycles with GnRH agonist treatment in the Reproductive Medicine Center of the Women and Children Hospital of Guangdong Province from January 1, 2002, to December 30, During this period, a total of 3,034 patients initiated their first IVF/ICSI embryo transfer (ET) cycles, of which 468 cycles were excluded either because no embryo formed or because all embryos were frozen owing to ovarian hyperstimulation syndrome or other personal reasons. No fresh ET was canceled in the case of P level rise. COS was performed in patients using either long protocol (n ¼ 1,719) or short protocol (n ¼ 847). From 2002 to 2004, short protocol was prevalent (650/733, 88.7%) except for polycystic ovary syndrome patients or some potential high responders, who received long protocol, whereas later on, long protocol was dominant (1,636/1,833, 89.3%), with short protocol being used only in those potential poor responders or patients refusing long protocol for personal reasons. To collect the clinical data of frozen ET, patients were followed through December No additional ethics approval was needed, because the study was retrospective, our center is licensed and supervised by the Ministry of Public Health of China, and all of the data used in the study were collected from the patients undergoing routine and standard IVF treatment in our center without any additional intervention. Controlled Ovarian Stimulation Protocol Briefly, in long protocol, mg or 1.3 mg triptorelin depot (Ipsen Pharma Biothech) were administered for downregulation, and recombinant FSH (rfsh: Puregon, NV Organon; or Gonal-F, Serono) were administered for ovarian stimulation. For short protocol, 0.1 mg triptorelin (Ipsen Pharma Biothech) was administered in cycle day 2 and rfsh was given from day 3. Follicle growth monitoring included serum E 2, P, and LH level measurements and vaginal ultrasound investigation. When two or more follicles reached the size of R18 mm, 10,000 IU hcg (Profasi; Serono) was injected for the oocyte retrieval 35 hours later. Embryo Culture, Transfer, Freezing and Thawing After routine IVF/ICSI procedure, embryos were cultured in Quinnes sequential culture media. No more than three embryos were transferred on day 3; mg P in oil was given for luteal support. Surplus embryos with no less than four cells and no more than 25% fragmentation were frozen. For frozen ET, no more than 3 frozen-thawed survival embryos were transferred in a natural or hormonal replacement cycle. Hormone Assays Sera were obtained at 8 9 a.m. at each time point for FSH, LH, E 2, and P measurements and analyzed by radioimmunoassay (Tianjin Depu Diagnostic Product Co.). The intra- and interassay coefficients of variation, respectively, were 5.8% and 9.1% for LH, 5.6% and 7.8% for FSH, 8.2% and 9.3% for E 2, and 7.0% and 9.6% for P. The sensitivity for progesterone was 0.1 ng/ml, and the range of measurement was ng/ml. Outcome Variables The primary outcome variable was live birth rate, and the secondary outcome variables were clinical pregnancy rate and implantation rate. Clinical pregnancy was defined as the presence of heartbeat of fetus on ultrasound examination at 7 weeks' pregnancy. Implantation rate was defined as the number of gestational sacs seen on the ultrasound divided by the total number of embryos transferred. Statistical Analysis Patients were divided into six distinct groups according to their serum P levels on the day of hcg administration: VOL. 98 NO. 3 / SEPTEMBER

3 ORIGINAL ARTICLE: ASSISTED REPRODUCTION %0.9, , , , , or >2.00 ng/ml. To determine the cutoff value of PPR, trend analysis was performed with the use of the Bosch method (14). Briefly, the odds ratio (OR) of live birth rate for each P interval, compared with the preceding interval, was calculated. Logistic regression analysis was used to assess 14 parameters possibly related to live birth rate, in which P level, age, and the number of high-quality embryos transferred were categorized. Factors related to PPR were assessed using a multivariate analysis. Stata 11.0 was applied to data analysis. Data were expressed as mean SD unless stated otherwise. Chi-square test was used to analyze categoric data, and independentsample Student t test was used for continuous variables. A P value of <.05 was considered to be statistically significant. RESULTS General Information The baseline and cycle characteristics are listed in Table 1. The mean P level on the day of hcg administration in long or short protocol was or ng/ml, respectively, and the corresponding live birth rate 38% or 28%, respectively. Because of the different characteristics of patients treated with long or short protocol, the following analyses were conducted separately. The Effect of P Levels on Live Birth Rates Figure 1 depicts the relationship between serum P levels and live birth rates. Figure 1A shows that live birth rate in all patients decreased when P level was >1.2 ng/ml; the same trend was observed in long protocol (Fig. 1B), whereas in short protocol the decreased live birth rate appeared only under the condition of P level >2.0 ng/ml (Fig. 1C). The OR (95% confidence interval [CI]) for live birth rate for each group compared with the preceding one is listed in Table 2, showing the difference of relative change in OR between intervals, confirming the nonlinear relationship between live birth rates and intervals of serum P level. Furthermore, statistically significant difference was observed between the and ng/ml intervals (P¼.0405), and the ng/ml and >2.0 ng/ml intervals (P¼.0090) in the overall study group. In long protocol, the difference was statistically significant only between the and ng/ml intervals (P¼.0316), and in short protocol, the difference was statistically significant only between the ng/ml and >2.0 ng/ml intervals (P¼.0156). These data indicate that a serum P level of 1.2 ng/ml or 2.0 ng/ ml may represent the cutoff value to define PPR where there is a negative impact of P on live birth rate in long or short protocol, respectively. The corresponding incidence of PPR in long or short protocol was 22.86% (393/1719) or 27.63% (234/847). The clinical outcomes of patients with or without PPR were compared, showing that the clinical pregnancy rate (OR 0.74, P¼.004 in long protocol; OR 0.84, P¼.002 in short protocol) and implantation rate (OR 0.73, P¼.000 in long protocol; OR 0.71, P¼.005 in short protocol) of patients with PPR decreased significantly, and the live birth rate decreased significantly in long protocol (40.65% vs %, OR 0.73; P¼.004) whereas in short protocol, the live birth rate of patients with PPR demonstrated a tendency of declining (30.18% vs %, OR 0.78; P¼.083). Considering potential factors that might influence live birth rate, we also conducted logistic regression analysis in long or short protocol separately, and the result confirmed TABLE 1 Baseline and cycle characteristics of the patients with long or short protocol. Characteristic Long protocol (n [ 1,719) Short protocol (n [ 847) P value Age (y) a Duration of infertility (y) b Basal FSH level (IU/L) a Antral follicle count c BMI (kg/m 2 ) a Duration of stimulation (d) c Total dose of rfsh administered (IU) 2, , , b LH level on D hcg (IU/L) b E 2 level on D hcg (pg/ml) 2, , , , b P level on D hcg (ng/ml) c No. of oocytes retrieved c No. of embryos transferred d No. of embryos frozen c Clinical pregnancy rate 712/1,719 (41.42%) 263/847 (31.05%).000 d Live birth rate 656/1,719 (38.16%) 240/847 (28.33%).000 d Implantation rate 1,070/3,713 (28.82%) 390/2,157 (18.08%).000 d Spontaneous miscarriage rate 51/712 (7.16%) 18/263 (6.84%).495 d Note: BMI ¼ body mass index; DhCG ¼ day of hcg administration. t test with unequal variance. Median Pearson chi-square test. Kruskal-Wallis equality-of-populations rank test (chi-square). One-sided Fisher exact chi-square test. 666 VOL. 98 NO. 3 / SEPTEMBER 2012

4 Fertility and Sterility FIGURE 1 in long protocol were categorized according to their P levels in fresh cycles as group without PPR (P level %1.2 ng/ml) or group with PPR (P level >1.2 ng/ml), whose live birth rate was 197/672 (29.31%) or 72/284 (25.35%), respectively, with no statistical significance observed (P¼.213). In patients receiving short protocol, 2.0 ng/ml was regarded as the cutoff value for grouping. The live birth rate in the 2 groups was 80/ 322 (24.84%) or 39/161 (24.22%), respectively (P¼.881). Multivariate Analysis of Factors Involved in PPR Multivariate regression analysis on the nine possible factors related to PPR showed that in both protocols, the basal FSH level, the duration of stimulation, the total dose of rfsh administered, E 2 level on the day of hcg administration, the number of oocytes retrieved, and LH level on the day of hcg administration were all associated with PPR (Table 3). The relationship between serum P levels and live birth rates in (A) all patients and in patients treated with (B) long protocol or (C) short protocol. that P level did have a negative effect on live birth rate in both protocols (Supplemental Tables 1 and 2, available online at Effect of PPR on Live Birth Rate in Frozen ET Cycles To explore the mechanism of the effect of PPR on live birth rate, the outcomes of 956 frozen ET cycles (670 patients) in long protocol and the outcomes of 483 frozen ET cycles (324 patients) in short protocol were analyzed. The patients DISCUSSION At present, knowledge remains uncertain about whether and how PPR influences clinical outcomes, although increasing literature can be reviewed (20), which may result from such reasons as the distinct definition of PPR, the different statistic methods used, the different characteristics of patients recruited, and the retrospective nature of studies analyzing the relationship between PPR and clinical outcomes. Controversy remains regarding the relationship between PPR and clinical outcomes since Schoolcraft et al. first reported in 1991 that premature luteinization was associated with lower pregnancy rate during pituitary suppression with GnRH agonist (2). In 2007, Venetis et al. (13), based on the first meta-analysis, denied the negative effect of PPR. But that paper was criticized for the heterogeneity of studies included, such as the different threshold of P level to define PPR. Actually, in most studies, the methods to determine the cutoff value (varying from 0.9 to 2.0 ng/ml) defining PPR were arbitrarily chosen (6, 8, 12), which undoubtedly undermines the reliability of any conclusion. In some other studies, receiver operating characteristic curve analysis was used to determine the cutoff value (5, 11). Nevertheless, that is not a proper method, because our data showed that the relationship between P levels and live birth rates is not linear, which is consistent with Bosch et al.'s report (14). To remedy the defects in the above-mentioned methods, trend analysis, introduced by Bosch et al. (14), was applied in our research. We found a negative relationship between PPR and live birth rate, with the cutoff value being 1.2 ng/ml in long protocol and 2.0 ng/ml in short protocol; the former is different from the cutoff value of 1.5 ng/ml in Bosch et al.'s research (14). This difference may be attributed to the different outcomes measured, because ongoing pregnancy rate was chosen to be analyzed in Bosch et al.'s study (14) whereas live birth rate was chosen in our study. In addition, our patients were much younger (31.1 vs years) and the methods of progesterone measurement were different, which might also make a difference. Among the research analyzing the relationship of PPR and clinical outcomes in GnRH agonist long protocol, some studies failed to observe negative correlation, which probably can be attributed to either the VOL. 98 NO. 3 / SEPTEMBER

5 ORIGINAL ARTICLE: ASSISTED REPRODUCTION TABLE 2 Live birth rates according to serum P levels. P level (ng/ml) Overall Long protocol Short protocol OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value % ( ) ( ) ( ) ( ).0405 a ( ).0316 a ( ) ( ) ( ) ( ) ( ) ( ) ( ).3002 > ( ).0090 a ( ) ( ).0156 a Note: CI ¼ confidence interval; OR ¼ odds ratio. a P<.05. limited number of patients recruited or a relatively low threshold used to define PPR (5, 11, 21). As far as we know, the present study included the largest number of patients (n ¼ 1,719) treated with long protocol, strengthening the reliability of our conclusion. The occurrence of PPR and its impact on clinical outcomes in short protocol have been poorly understood, with only one paper published before now. Our data showed that: 1) P level was higher in short protocol than in long protocol, which might be the consequence of relatively high LH level resulting from the flare-up effect of GnRH agonist (Table 1), adding further evidence that LH might also be involved in the occurrence of PPR; and 2) both clinical pregnancy rate (P¼.002) and live birth rate (P¼.083) fell when P level increased to >2 ng/ml. In contrast to our results, Martinez et al. (5) did not observe the higher P level and the negative correlation between PPR and clinical pregnancy rate in short protocol. This discrepancy may result from three reasons: 1) Only poor responders were recruited in Martinez et al.'s study, whereas in our study, both normal and poor responders were included and the former were predominant; 2) the number of patients involved in Martinez et al.'s study (n ¼ 159) was somewhat limited; and 3) the cutoff value to define PPR in Martinez et al.'s study was 0.9 ng/ml, much lower than that in our study (2.0 ng/ml). Because the literature comparing the occurrence of PPR in long or short protocol is not conclusive, we try to explain the reasons for the distinct cutoff values determined in the present study as follows: 1) The characteristics of patients in long or short protocol were different from each other, more poor responders being included in short protocol; 2) the cycle properties of the two protocols are different, with the LH level in the follicle phase being much higher in short protocol; 3) the number of patients included in short protocol was relatively few, which might have introduced a bias. Retrospective study is the method commonly used to analyze the correlation between serum P levels and clinical outcomes. However, it should be noted that the weakness in the nature of retrospective study makes it difficult to exclude interference from various confounding factors; nor can prospective studies (14, 22) exclude such interference, because serum P level in vivo cannot be subjectively intervened. For this reason, we performed logistic regression analysis covering 13 confounding factors, and the results confirmed the conclusion that PPR was a negative predictor of live birth rate, which was consistent with Shapiro et al.'s study (23) indicating that low preovulatory serum P was a dominant predictor of clinical pregnancy. It is clinically relevant to answer how PPR affects live birth rate: whether it is due to compromised oocyte quality, TABLE 3 Multivariate regression analysis of factors related to premature P rise. Variable Long protocol Short protocol Regression coefficient Standard error P value Regression coefficient Standard error P value Age (y) Basal FSH level (IU/L) a a Basal E 2 level (pg/ml) BMI (kg/m 2 ) Duration of stimulation (d) a a Total dose of rfsh administered (IU) e a e a LH level on D hcg (IU/L) a a E 2 level on D hcg (pg/ml) 8.40e e a e a No. of oocytes retrieved a a Note: Abbreviations as in Table 1. a P< VOL. 98 NO. 3 / SEPTEMBER 2012

6 Fertility and Sterility decreased endometrial receptivity, or both. If the former is the main reason, triggering hcg in advance would be the first choice to avoid the adverse effects of PPR on the quality of oocytes. If decreased endometrial receptivity plays a key role, freezing embryos would be a better choice, because triggering hcg in advance would compromise the overall quality of the oocyte cohort when most of the trailing follicles are still small. To address this question, we analyzed the outcomes of frozen ET cycles in long protocol (n ¼ 956) and short protocol (n ¼ 483), and the results showed that live birth rates were not significantly different between the groups with or without PPR, implying that PPR may negatively affect live birth rate through decreased endometrial receptivity. This result was consistent with Shapiro et al.'s (24) and Lahoud et al.'s (25) studies which indicated that embryo cryopreservation could rescue cycles with PPR. Additionally, based on a donor oocyte IVF model, Melo et al. (26) found that the pregnancy rates of recipients were not influenced by the P levels of donors, suggesting that increased P level did not deteriorate the quality of oocytes. van Vaerenberg et al. (27) observed that the gene expression profile of endometrium with P level >1.5 ng/ml was obviously different from that with normal P level, and Labarta et al. (28) found that this different gene expression profile was maintained until implantation occurred, providing evidence that decreased endometrial receptivity may be a cause of decreased live birth rate. Therefore, in the case of P rise on the day of hcg administration, the most appropriate choice to avoid the negative effect is to cancel fresh ET and freeze all embryos rather than triggering hcg in advance. Several methods have been applied to studying the underlying mechanism of the occurrence of PPR. In some research, the authors just compared multiple variables between groups with or without PPR and the factors showing difference were regarded to be the reason of PPR (8, 29, 30). Such kind of comparison is somewhat simple and weak to distinguish causal from noncausal correlation, nor can it reflect the interaction among variables. Some studies used logistic regression to analyze which factors were involved in PPR (14), but P level must be categorized into two groups in that method, which may magnify or minify the influence of some factors, because P level is a continuous variable. Based on these pioneer studies, four possible explanations of the occurrence of PPR were documented, i.e., multiple follicle growth (8, 9), overdose of exogenous gonadotropin (14), premature lutineization (18), and poor ovarian response (19). The former three partly explain the reasons for PPR from limited facets, and the last one is not that convincing, because we found that women with lower basal FSH level were inclined to have PPR. Actually, according to the present study using multivariate regression analysis, excess growth of multiple follicles, overdose of exogenous gonadotropin, and LH rise all contributed to the occurrence of PPR, consistent with Hugues et al.'s conclusion (18). Furthermore, the dose of gonadotropin administered and patients' ovarian response, including the number of oocytes retrieved and the E 2 level on the day of hcg administration, were influenced by patients' baseline characteristics, such as basal FSH, E 2 level, body mass index, and age. Therefore, these baseline characteristics were also included in our multivariate regression analysis, making our conclusion more reliable and comprehensive. In summary, our analysis on the outcomes of 2,566 patients undergoing their first IVF/ICSI cycles treated with GnRH agonist long or short protocol showed that PPR negatively correlated with live birth rate in fresh ET cycles, whereas no adverse impact on frozen ET was observed, implying that PPR may have deleterious effects on endometrial receptivity. REFERENCES 1. Hamori M, Stuckensen JA, Rumpf D, Kniewald T, Kniewald A, Kurz CS. Premature luteinization of follicles during ovarian stimulation for in-vitro fertilization. Hum Reprod 1987;2: Schoolcraft W, Sinton E, Schlenker T, Huynh D, Hamilton F, Meldrum DR. Lower pregnancy rate with premature luteinization during pituitary suppression with leuprolide acetate. Fertil Steril 1991;55: Silverberg KM, Burns WN, Olive DL, Riehl RM, Schenken RS. Serum progesterone levels predict success of in vitro fertilization/embryo transfer in patients stimulated with leuprolide acetate and human menopausal gonadotropins. J Clin Endocrinol Metab 1991;73: Givens CR, Schriock ED, Dandekar PV, Martin MC. Elevated serum progesterone levels on the day of human chorionic gonadotropin administration do not predict outcome in assisted reproduction cycles. Fertil Steril 1994;62: Martinez F, Coroleu B, Clua E, Tur R, Buxaderas R, Parera N, et al. Serum progeseterone concentrations on the day of hcg administration can not predict pregnancy in assisted reproduction cycles. Reprod Biomed Online 2004;8: Bosch E, Valencia I, Escudero E, Crespo J, Simon C, Remohi J, et al. Premature luteinization during gonadotropin-releasing hormone antagonist cycles and its relationship with in vitro fertilization outcome. Fertil Steril 2003;80: Lee FK, Lai TH, Lin TK, Horng SG, Chen SC. Relationship of progesterone/estradiol ratio on day of hcg administration and pregnancy outcomes in high responders undergoing in vitro fertilization. Fertil Steril 2009;92: Kiliçdag EB, Haydardedeoglu B, Cok T, Hacivelioglu SO, Bagis T. Premature progesterone elevation impairs implantation and live birth rates in GnRH-agonist IVF/ICSI cycles. Arch Gynecol Obstet 2010;281: Miller KF, Behnke EJ, Arciaga RL, Goldberg JM, Chin NW, Awadalla SG. The significance of elevated progesterone at the time of human chorionic gonadotropin may be related to luteal support. J Assist Reprod Genet 1996; 13: Lai TH, Lee FK, Lin TK, Horng SG, Chen SC, Chen YH, et al. An increased serum progesterone-to-estradiol ratio on the day of human chorionic gonadotropin administration does not have a negative impact on clinical pregnancy rate in women with normal ovarian reserve treated with a long gonadotropin releasing hormone agonist protocol. Fertil Steril 2009;92: Saleh HA, Omran MS, Draz M. Dose subtle progesterone rise on the day of HCG affect pregnancy rate in long agonist ICSI cycles? J Assist Reprod Genet 2009;26: Doldi N, Marsiglio E, Destefani A, Gessi A, Merati G, Ferrari A. Elevated serum progesterone on the day of HCG administration in IVF is associated with a higher pregnancy rate in polycystic ovary syndrome. Hum Reprod 1999;14: Venetis CA, Kolibianakis EM, Papanikolaou E, Bontis J, Devroey P, Tarlatzis BC. Is progesterone elevation on the day of human chorionic gonadotrophin administration associated with the probability of pregnancy in in vitro feritlization? A systematic review and meta-analysis. Hum Reprod Update 2007;13: VOL. 98 NO. 3 / SEPTEMBER

7 ORIGINAL ARTICLE: ASSISTED REPRODUCTION 14. Bosch E, Labarta E, Crespo J, Simon C, Remohi J, Jenkins J, et al. Circulating progesterone levels and ongoing pregnancy rates in controlled ovarian stimulation cycles for in vitro fertilization: analysis of over 4000 cycles. Hum Reprod 2010;25: Papanikolaou EG, Kolibianakis EM, Pozzobon C, Tank P, Tournaye H, Bourgain C, et al. Progesterone rise on the day of human chorionic gonadotropin administration impairs pregnancy outcome in day 3 single-embryo transfer, while has no effect on day 5 single blastocyst transfer. Fertil Steril 2009;91: Long X, Pen C, Lu G. Isolation and identification of genes differentially expressed in premature luteinization granulosa cell during controlled ovarian hyperstimulation. Fertil Steril 2009;92: Yovel I, Yaron Y, Amit A, Peyser MR, David MP, Kogosowski A, et al. High progesterone levels adversely affect embryo quality and pregnancy rates in in vitro fertilization and oocyte donation programs. Fertil Steril 1995;64: Hugues JN, Masse-Laroche E, Reboul-Marty J, Boîko O, Meynant C, Cedrin- Durnerin I. Impact of endogenous luteinizing hormone serum levels on progesterone elevation on the day of human chorionic gonadotropin administration. Fertil Steril 2011;96: Younis JS, Matilsky M, Radin O, Ben-Ami M. Increased progesterone/estradiol ratio in the late follicular phase could be related to low ovarian reserve in in vitro fertilization embryo transfer cycles with a long gonadotropinreleasing hormone agonist. Fertil Steril 2001;76: Elnashar AM. Progesterone rise on the day of HCG administration (premature luteinization) in IVF: an overdue update. J Assist Reprod Genet 2010; 27: Urman B, Alatas C, Aksoy S, Mercan R, Isiklar A, Balaban B. Elevated serum progesterone level on the day of human chorionic gonadotropin administration does not adversely affect implantation rates after intracytoplasmic sperm injection and embryo transfer. Fertil Steril 1999;72: Papanikolaou EG, Zepyridis L, Kolibianakis E, Grimpizis G, Pantos G, Bili E, et al. Comparing GnRH-agonist with antagonist IVF cycles, is the clinical outcome affected by the incidence of progesterone elevation. A RCT. Hum Reprod 2009;24(Suppl 1):i Shapiro BS, Daneshmand ST, Garner FC, Aguirre M, Thomas S. Large blastocyst diameter, early blastulation, and low preovulatory serum progesterone are dominant predictors of clinical pregnancy in fresh autologous cycles. Fertil Steril 2008;90: Shapiro BS, Daneshmand ST, Garner FC, Aguirre M, Hudson C, Thomas S. Embryo cryopreservation rescues cycles with premature luteinization. Fertil Steril 2010;93: Lahoud R, Kwik M, Ryan J, Al-Jefout M, Foley J, Illingworth P. Elevated progesterone in GnRH agonist down regulated in vitro fertilisation (IVFICSI) cycles reduces live birth rates but not embryo quality. Arch Gynecol Obstet 2012;285: Melo MAB, Meseguer M, Garrido N, Bosch E, Pellicer A, Remohí J. The significance of premature luteinization in an oocyte-donation programme. Hum Reprod 2006;21: van Vaerenbergh I, Fatemi HM, Blockeel C, van Lommel L, In't Veld P, Schuit F, et al. Progesterone rise on HCG day in GnRH antagonist/rfsh stimulated cycles affects endometrial gene expression. Reprod Biomed Online 2011;22: Labarta E, Martinez-Conejero JA, Alama P, Horcajadas JA, Pellicer A, Simon C, et al. Endometrial receptivity is affected in women with high circulating progesterone levels at the end of the follicular phase: a functional genomics analysis. Hum Reprod 2011;26: Segal S, Glatstein I, McShane P, Hotamisligil S, Ezcurra D, Carson R. Premature luteinization and in vitro ferilization outcome in gonadotropin/gonadotropinreleasing hormone antagonist cycles in women with polycystic ovary syndrome. Fertil Steril 2009;91: Seow KM, Lin YH, Hsieh BC, Huang LW, Huang SC, Chen CY, et al. Characteristics of progesterone changes in women with subtle progesterone rise in recombinant follicle-stimulating hormone and gonadotropinreleasing hormone antagonist cycle. Gynecol Obstet Invest 2010;70: VOL. 98 NO. 3 / SEPTEMBER 2012

8 Fertility and Sterility SUPPLEMENTAL TABLE 1 Logistic regression analyses on the predictors of live birth in long protocol. Variable Regression coefficient Standard error P value OR 95% CI P level (ng/ml) a a > a Age (y) a a > Basal FSH level (IU/L) Basal E 2 level (pg/ml) a BMI (kg/m 2 ) Duration of infertility (y) Antral follicle count Duration of stimulation (d) Total dose of rfsh administered (IU) e LH level on D hcg (IU/L) a E 2 level on D hcg (pg/ml) 5.24e e No. of oocytes retrieved No. of high-quality embryos transferred a a a Endometrial thickness (mm) a Note: BMI ¼ body mass index; CI ¼ confidence interval; DhCG ¼ day of hcg administration; OR ¼ odds ratio. a P<.05. VOL. 98 NO. 3 / SEPTEMBER e1

9 ORIGINAL ARTICLE: ASSISTED REPRODUCTION SUPPLEMENTAL TABLE 2 Logistic regression analyses on the predictors of live birth in short protocol. Variable Regression coefficient Standand error P value OR 95% CI P level (ng/ml) > a Age (y) > Basal FSH level (IU/L) Basal E 2 level (pg/ml) BMI (kg/m 2 ) Duration of infertility (y) a Antral follicle count Duration of stimulation (d) Total dose of Gn administered (IU) LH level on D hcg (IU/L) E 2 level on D hcg (pg/ml) 5.02e e No. of oocytes retrieved No. of high-quality embryos transferred a a a Endometrial thickness (mm) a Note: Abbreviations as in Supplemental Table 1. P< e2 VOL. 98 NO. 3 / SEPTEMBER 2012