{"paper_id":"74381e7d-b51a-45ae-b950-c4696b674681","body_text":"During the follicular phase of the menstrual cycle, estrogen (E2) plays an essential\nrole in endometrial priming, as well as in the proliferation of uterine surface\nepithelium, glands, stroma, and blood vessels. However, the role of estrogen in\nendometrial preparation for embryo implantation during the luteal phase remains\nunclear, as some studies suggest that decreases in E2 during the luteal phase do not\nadversely affect the morphological developmental capacity of the endometrium ( Younis  et al ., 1994 ;  Pinheiro  et al ., 2017 ;  Ismail Madkour  et al ., 2016 ).\nSteroids secreted in supraphysiological levels during the early luteal phase inhibit\nLH production, thus engendering low E2 and progesterone (P) levels ( Hubayter & Muasher, 2008 ). Therefore, low\nE2 and P levels caused by a lack of luteal phase hormonal support in assisted\nreproduction technology cycles lead to decreased implantation and pregnancy rates\n(PR) ( Hutchinson-Williams  et al .,\n1989 ).\nDespite the necessity of luteal phase supplementation to improve  in\nvitro  fertilization (IVF) outcomes ( Nyboe Andersen  et al ., 2002 ), to our knowledge there is\nno consensus on the preferred type, dose, or timing of support. Although some\nstudies described benefits from E2 supplementation ( Lukaszuk  et al ., 2005 ;  Gorkemli  et al ., 2004 ), others failed to observe\npositive impacts on support by E2 ( Farhi  et\nal ., 2000 ;  Ghanem  et\nal ., 2009 ;  Smitz  et\nal ., 1993 ;  Lewin  et\nal ., 1994 ;  Tay & Lenton,\n2003 ). Meta-analyses including these studies showed that supplementing P\nwith E2 did not lead to better IVF outcomes ( Serna\n et al ., 2006 ;  Jee\n et al ., 2010 ). However, given the small size of\nthese studies, larger series are required to determine the importance of E2 in\nluteal phase support, along with the most efficient dose and route of\nadministration. To our knowledge, no publication has yet reported on the pregnancy\neffects of co-administering oral or transdermal E2 and progesterone in GnRH\nantagonist cycles.\nThis retrospective observational study compared the effects of three different luteal\nphase support protocols with estrogen on the outcomes of  in vitro \nfertilization-embryo transfer (IVF-ET) cycles of patients on a GnRH antagonist\nprotocol.\n\nThis study included 110 women undergoing intracytoplasmic sperm injection (ICSI)\nat private reproductive medicine center Brazilian Institute of Assisted\nReproduction (IBRRA) between March 2016 and February 2018. IVF-ET indications\nincluded tubal factor infertility, endometriosis, polycystic ovaries,\nnormozoospermia, and unexplained infertility.\nThe inclusion criteria were as follows: i) both ovaries present; ii) no current\nor past diseases affecting the ovaries or the secretion, clearance, or excretion\nof gonadotropins or sex steroids; iii) patients could not be on hormone therapy\nat the time of treatment; iv) adequate visualization of the ovaries on\ntransvaginal ultrasound examination; and v) small antral follicle (3-12 mm in\ndiameter) count between 1 and 32 in the two ovaries added. Informed consent was\nobtained from all patients. The Institutional Review Board and the IBRRA Ethics\nCommittee approved the study.\nOn oocyte pickup day, the patients were randomly assigned into one of three\ngroups: transdermal estrogen gel daily (Group 1: Oestrogel pump, Estradiol-\nBesins Pharmaceuticals, Belgium); oral estrogen daily (Group 2: Primogyna -\nestradiol valerate, Bayer Pharmaceuticals, Germany); or transdermal estrogen\npatches daily (Group 3: Estradott- Estradiol, Novartis Pharmaceuticals,\nSwitzerland) based on their application number. The researchers were blinded for\ntreatment allocation.\nOvarian stimulation was performed with recombinant FSH (Gonal-F; Merck-Serono\nPharmaceuticals, Italy), starting with a dose of 225-300IU on Day 2 of the\nmenstrual cycle. When needed, FSH doses were adjusted starting from the fourth\nday of stimulation based on ultrasound findings and E2 blood levels. A GnRH\nantagonist (Cetrotide; Merck-Serono Pharmaceuticals, Italy) was administered at\na dose of 250µg 0.5mL/day starting when the lead follicle reached 14-15mm\nin diameter, until the day of hCG injection.\nOvulation was induced by a subcutaneous (SC) injection of 250 mcg of recombinant\nhCG (Ovidrel, Merck-Serono Pharmaceuticals, Italy) when three follicles of at\nleast 18 mm in diameter were observed on ultrasound examination. Oocyte pickup\nwas performed 34 to 36 hours after hCG injection. Intracytoplasmic sperm\ninjection (ICSI) was performed in all metaphase II oocytes. All patients\nunderwent embryo transfer with ultrasound guidance on Day 3.\nSupplementation with estrogen (transdermal gel, oral, or transdermal patches,\naccording to randomization group) and intravaginal P 600mg once a day\n(Utrogestan, progesterone micronized, Besins Pharmaceuticals, Belgium) were\nadministered to all patients on the day of oocyte retrieval. The three different\nadministration routines of estrogen provided each the equivalent to 4 mg of\nestradiol daily. Blood samples were drawn on the day of hCG administration and\non beta hCG measurement day (two weeks after ET), to measure E2 and P levels.\nEstrogen administration and intravaginal P were continued until pregnancy was\nruled out by a negative serum beta-hCG test performed on day 14 after ET or\nuntil the twelfth week of pregnancy for pregnant patients. Clinical pregnancies\nwere detected with the confirmation of a fetal heartbeat on transvaginal\nultrasound examination. No drug-related side effects were reported in our\nstudy.\nICSI was routinely performed in all fertilization procedures. Fertilization was\nevinced when two pronuclei were observed. Embryos were cultured until the day of\ntransfer (Day 3) in IVF Global® media (Life Global, Canada) supplemented\nwith 10% synthetic serum substitute (SSS) and graded based on the Veeck scoring\nsystem ( Veeck, 1996)  before transfer. The\nsame embryologist performed all embryology procedures and embryo assessments in\nthis study. All women received one or two embryos categorized as I and/or II.\nThe definition over the number of embryos transferred was based on the\nguidelines of the Brazilian Federal Council of Medicine (FCM).\nEmbryo transfers were performed three days after oocyte retrieval. The patients\nwere instructed to have a full bladder to provide for an acoustic window to\nvisualize the uterus in preparation for the ultrasound-guided embryo transfer.\nEach patient was placed in the lithotomy position without anesthesia or\nsedation. The embryo transfers were performed with a Wallace Classic Soft Embryo\nTransfer Catheter, and abdominal ultrasound was performed using a 5 MHz probe\n(GE Logiq 400 Pro Series, General Electric Company, Pewaukee, WI).\nSerum AMH levels were measured using a second-generation enzyme-linked\nimmunosorbent assay. Intra- and inter-assay coefficients of variation (CV) were\n<6% and <10% respectively, with a lower detection limit of 0.13ng/mL and\nlinearity up to 21ng/mL for AMH. E2 and P levels were determined by\nelectrochemiluminescence immunoassay (Elecsys and Cobas e analyzers; Roche\nDiagnostics GmbH, Mannheim, Germany). The results were determined via a curve\nspecifically generated for the instrument by two-point calibration and based on\nthe provided master curve. Sensitivity was 5pg/mL, and the linear interval of\nthe test was 5 to 4,300 pg/mL for estrogen. E2 levels were determined with\nintra-assay and inter-assay coefficients of variation of <3.3% and <4.9%,\nrespectively. Sensitivity was 0.21 ng/mL, and the linear interval of the test\nwas 0.21 to 60ng/mL for P. P levels were assayed with intra-assay and\ninter-assay coefficients of variation of <8% and <9.1%, respectively.\nTransvaginal ultrasound to assess the baseline antral follicle count was\nperformed on Day 3 of the menstrual cycle. Follicles with a mean diameter of\n3-12mm (mean of two orthogonal diameters) from both ovaries were considered. To\noptimize the reliability of ovarian follicular assessment, the ultrasound\nscanner was equipped with a tissue harmonic imaging system, which allowed for\nimproved image resolution and adequate recognition of follicular borders.\nIntra-analysis CV for follicular and ovarian measurements were <5%, and the\nlower limit of detection was 0.1mm. In an effort to evaluate the bulk of\ngranulosa cells in both ovaries, we calculated the mean follicle diameter\n(cumulative follicle diameter divided by the number of follicles measuring\n3-12mm in diameter from both ovaries) and the largest follicle diameter.\nThis study aimed to evaluate whether the dose and mode of administration of\nestrogen affected the levels of estrogen on beta hCG measurement day and\npregnancy rates, and which markers of ovarian reserve correlated with pregnancy\nrates. A secondary objective was to assess whether the levels of progesterone on\nbeta hCG measurement day correlated with pregnancy rates.\nStatistical power calculation revealed that at least 30 patients were required in\neach arm of the study to attain significance in clinical pregnancy rates, the\nmain endpoint analyzed in this randomized study. It was calculated for a\ndifference of 25% in clinical pregnancy rates, as observed in the pilot\nstudy.\nThe level of significance (α) was 0.05 with a power of 0.95. The analysis\nof the number of clinical pregnancies showed that we had enough numbers to reach\nthe required level of statistical power. Thus, enrollment was discontinued and\nthe analysis of results commenced.\nData sets were analyzed on SPSS for Windows release 15.0 (SPSS, Inc., Chicago,\nIL). Continuous data were expressed as mean values ± SD. Data following a\nnormal distribution were analyzed with one-way ANOVA, whereas the Kruskal-Wallis\ntest was used for the remaining data. Categorical data were analyzed with\nPearson’s c 2  test. If statistical difference was found, the groups\nwere compared by the c 2  test with the Spearman correction. E2, P, and\nE2/P rates for ongoing pregnancies in all groups were analyzed with the\nMann-Whitney U test. Significance was set at 5%.\n\nOur retrospective study included 110 patients. Six patients were not present on\nthe day of beta HCG measurement and were thus excluded. The final study\npopulation was 104. Group 1 had 32 patients, Group 2 had 33 patients, and Group\n3 had 39 patients. Patient characteristics are described in  Table 1 . There was no statistically\nsignificant difference between groups for age, body mass index (BMI), Day 3 FSH\nand Anti-Müllerian hormone (AMH) levels, antral follicle count (AFC),\nlength of stimulation, total dose of gonadotropin, or peak E2 and P levels on\nhCG injection day ( Table 1 ).\nPatient and cycle characteristics for the three treatment groups\nNote:  p <0.05 was considered to be statistically\nsignificant. Data are expressed as mean values ± SD\nNo significant difference was found in the number of oocytes retrieved, number of\nembryos I + II, number of embryos transferred, implantation rates, clinical PR,\nmiscarriage rates, multiple-pregnancy rates, E2 and P levels on beta hCG\nmeasurement day ( Table 2 ) ( Figure 1 ).\nIn vitro  fertilization-embryo transfer cycle\ncharacteristics of the three treatment groups\nNote:  p <0.05 was considered to be statistically\nsignificant. Data are expressed as mean values ± SD or as\nproportions and absolute numbers.\nΔ – Mean of Variation of hormone profile\nFigure 1 Level of estradiol (E2) and progesterone (P) on beta hCG measurement\nday of the three treatment groups ( p >0.05)\nLevel of estradiol (E2) and progesterone (P) on beta hCG measurement\nday of the three treatment groups ( p >0.05)\nThere was no significant difference in the levels of E2 and progesterone on hCG\ninjection day and ß hCG measurement day in the three groups\n( p >0.05). The E2/P ratio on beta HCG measurement day was\ncomparable between the three groups, showing that the mode of administration of\nestrogen in the luteal phase did not lead to different effects on hormonal\nprofiles ( Table 3 ).\nComparison of hormone profile variance in the three treatment groups\nNote:  p <0.05 was considered to be statistically\nsignificant\nIn terms of ovarian reserve markers, a significant correlation was observed\nbetween testing positive for clinical pregnancy and AMH levels (r=0.66 ,\np <0.0001) ( Figure 2 ). In\nrelation to the hormonal profile, positive pregnancy tests were significantly\nassociated with E2 levels on beta hCG measurement day (r=0.77\n p <0.0001) ( Figure 3 ),\nregardless of the estrogen protocol chosen for luteal phase support. Concerning\nthe variables significantly correlated with positive pregnancy tests, the median\nE2 level on beta hCG measurement day was 903.65±127.85pg/mL and the\nmedian level of AMH on Day 3 was 4.43±3.14ng/mL, regardless of the\nestrogen protocol chosen for luteal phase support. Thus, the mode of\nadministration of estrogen in the luteal phase did not affect the outcome of ART\ntreatment.\nFigure 2 Spearman’ s correlation of positive clinical pregnancy test and AMH\nlevel ( p < 0.0001)\nSpearman’ s correlation of positive clinical pregnancy test and AMH\nlevel ( p < 0.0001)\nFigure 3 Spearman’ s correlation for positive clinical pregnancy tests and E2\nlevels on beta hCG measurement day\n( p <0.0001)\nSpearman’ s correlation for positive clinical pregnancy tests and E2\nlevels on beta hCG measurement day\n( p <0.0001)\n\nBased on the findings from a cohort of 104 patients, our retrospective study showed\nthat in IVF-ET protocols including cycles with a GnRH antagonist, the use of oral\nmedication, transdermal patches, or transdermal gel during the luteal phase did not\nsignificantly affect pregnancy rates. To our knowledge, our study was the first to\ncompare the effects on IVF-ET cycle outcomes of three different luteal phase support\nprotocols with estrogen in patients given a GnRH antagonist.\nEmbryo implantation is a complex dynamic process that involves structural and\nmorphological changes to the embryo and the endometrium. Adequate levels of estrogen\nand P may be essential for optimal endometrial maturation before embryo\nimplantation, as a lack of synchrony between the stages of embryo development and\nendometrial maturation may result in implantation failure. The idea that E2 might be\nused throughout the luteal phase in IVF cycles emerged when  Smitz  et al . (1988)  showed that serum E2\nconcentrations dropped at the end of the luteal phase. Cycles using GnRH agonists\nand antagonists have been associated with poor luteal phase hormonal production.\nAlthough the role of P supplementation in the luteal phase of down-regulated cycles\nis well established, there have been only a few attempts to clarify the benefits of\nadding E2 therapy in these cycles. The use of E2 during the luteal phase, including\nits role in the preparation of the endometrium for implantation, remains rather\ncontroversial.\nStewart  et al . (1993)  were\nthe first to identify a significant difference in serum E2 levels between conception\nand non-conception cycles in fertile women undergoing donor insemination. This\ndifference was noted as early as Day 6 after the LH surge. Similarly, a rise in\nluteal E2 on Day 6 in conception cycles compared with non-conception cycles was\nfound in a group of 32 women trying to conceive spontaneously ( Baird  et al ., 1997 ). Other studies have\nreported an association between elevated and steadily increasing serum E2 levels in\nthe luteal phase of IVF-ET cycles and higher PRs ( Emperaire  et al ., 1984 ). Subsequently,  Sharara & McClamrock (1999)  revealed that\nthe magnitude of the decline in serum E2 concentrations, measured by the ratio of\npeak E2 (on the day of hCG administration) to midluteal E2 (10 days after hCG\nadministration), was predictive of IVF success. A sharp decline in midluteal E2,\ndefined as a peak E2 to midluteal E2 ratio greater than 5, resulted in significantly\nlower implantation and OP rates. All of the above data raised the issue around a\npotential positive correlation between elevated E2 levels in the luteal phase and\nconception, in addition to the need to elucidate the relationship between estradiol\non the day of beta hCG measurement and pregnancy, as seen in our study.  Elgindy  et al . (2010)  found the\nlowest E2 levels on Days 7, 10, and 13 in the P-only group during the luteal phase,\nand further showed that the decreases on days 7 and 10 were the highest. A\ndose-finding RCT ( Lukaszuk  et al .,\n2005 ) reported that the best implantation and pregnancy rates were\nrecorded in the group given 6 mg E2 supplementation compared with 2 mg or no E2\nsupplementation. Regarding the route of administration, most of the previous studies\nused the oral route ( Lukaszuk  et\nal ., 2005 ;  Farhi  et\nal ., 2000 ;  Smitz  et\nal ., 1993 ;  Lewin  et\nal ., 1994 ;  Fatemi  et\nal ., 2006 ;  2007 )\nIn our study, the three groups did not differ in regards to maximum E2 levels on the\nday of hCG administration. Addition of E2 (oral or transdermal) did not\nsignificantly change the endocrine profile of the luteal phase. Although  Lewin  et al . (1994)  reported\nno significant difference in luteal E2 levels upon oral supplementation with 2 mg E2\nvalerate,  Farhi  et al . (2000) \nreported significantly higher E2 levels in non-conception cycles on Days 11, 14, and\n16 after hCG administration upon supplementation with 4 mg E2 valerate.  Fatemi  et al . (2007)  observed\nthat the addition of 4 mg E2 valerate to P for luteal phase support in antagonist\ncycles did not affect the E2 level significantly until Day 10 after hCG\nadministration, when it was associated with significantly higher E2 levels. However,\nthe effect of higher levels of E2 on the endocrine profile could not be ruled out.\nMorphologic studies have demonstrated that the endometrium is sensitive to decreases\nin steroid levels and subnormal midluteal E2 concentrations (24). During the luteal\nphase, estrogen has a modulatory effect on the secretory endometrial P receptor\nconcentration and may serve to replenish and maintain a requisite level of P\nreceptors to mediate and complete the P response ( Fritz  et al ., 1987 ;  Goldstein  et al ., 1982 )\nThere is no consensus regarding the optimal dose and duration of E2 administration\nduring the luteal phase. A well-conducted randomized trial ( Hutchinson-Williams  et al ., 1989 ) looked into\nthe effects of different E2 supplementation doses on IRs and PRs using the long\nGnRH-a protocol. All patients received P4 vaginally (600 mg/day) and were randomly\nallocated to daily doses of 0, 2, or 6mg of E2. Significantly higher IRs and PRs\nwere recorded in patients given low-dose E2 supplementation compared to those who\ndid not. The subgroup meta-analyses on different doses of E2 suggested similar\ntrends toward favorable outcomes in the group given a combination of E2 and P4, but\nthe number of studies was very limited, precluding the extraction of clear\nconclusions regarding optimal E2 doses. These discrepancies may be attributed to the\ndifferent methodological designs across studies. Further studies are required to\ndetermine the role of luteal E2 supplementation in IVF and investigate its optimal\nregimen (dose and route).\nAlthough our study showed a significant association between testing positive for\nclinical pregnancy and AMH levels, ovarian reserve markers reportedly have some\npredictive power in the realm of assisted reproductive technology (ART) treatments.\nHowever, there is consensus that they provide only general approximations of\nstimulation quantity (e.g., the number of oocytes retrieved in ART treatment\ncycles). The main limitations of these tests include their poor sensitivity and, in\nmost cases, their dependency on cycle stage. Furthermore, once a woman tests\nabnormal, poor prognosis is assigned to her ART treatment possibilities ( Scheffer  et al ., 2017 )\nIn our study, no significant difference on pregnancy rates was observed between the\nadministration of oral estrogen, transdermal estrogen patches, or transdermal\nestrogen gel as luteal phase support in IVF-ET GnRH antagonist protocols. Further\nresearch in this area is warranted to confirm and advance these findings.\n\nIn IVF-ET cycles with a GnRH antagonist, no significant difference was observed on\npregnancy rates when patients were given oral E2, transdermal E2 patches, or\ntransdermal E2 gel during the luteal phase. Clinical pregnancy rates correlated with\nAMH and E2 levels on beta hCG measurement day.","source_license":"CC-BY-4.0","license_restricted":false}