{"paper_id":"c7e87c38-c592-4697-bff5-6ccd22783ee6","body_text":"Poor response to ovarian stimulation affects a significant proportion of infertile\ncouples seeking fertility advice. Although in the past few years a debate has arisen\nregarding the definition of poor ovarian response, the European Society of Human\nReproduction and Embryology (ESHRE) working group on Poor Ovarian Response\nDefinition recently developed new criteria to define patients who respond poorly to\novarian stimulation; the so called \"Bologna criteria\" ( Ferraretti  et al ., 2011 ). These criteria\nincorporate age, ovarian reserve tests (anti-Mullerian hormone-AMH-level or antral\nfollicle count - AFC) and ovarian response in previous IVF/ICSI cycles in the\ndefinition, and represent the first realistic attempt by the scientific community\n(ESHRE) to standardize the definition of poor ovarian response in a simple and\nreproductive manner.\nThe first studies published including women with poor ovarian response, according to\nthe Bologna criteria, have shown disappointingly low pregnancy rates, irrespectively\nof age. A recent observational study demonstrated a very poor prognosis for these\nwomen, given that live birth rates following treatment with natural cycle IVF was\n< 3% per patient, irrespective of age, and significantly lower when compared to\nwomen who did not fulfill the Bologna criteria ( Polyzos  et al ., 2012 ).\nA poor response to ovulation stimulation results in high cancellation rates of up to\n76% and extremely low pregnancy rates, from 3.2-14% ( Ulug  et al ., 2003 ;  Busnelli  et al ., 2015 ). Various strategies for poor\nresponders, including agonist and antagonist protocols have been attempted; however,\nat present, there is no definitive evidence that poor outcomes can be reversed by a\nspecific protocol ( Ubaldi  et al .,\n2014 ;  Ata & Seli, 2015 ).\nIt has been suggested that the buildup of androgens in the micro milieu of the\nprimate ovary, plays a critical role in early follicular development and granulosa\ncell proliferation, and increase the number of preantral and antral follicles ( Weil  et al ., 1999 ;  Hillier  et al ., 1997 ). In\naddition, increased intraovarian concentration of androgens seems to augment\nfollicle stimulating hormone (FSH) receptor expression in the granulosa cells ( Vendola  et al ., 1998 ;  1999 ).\nBased on the limited available evidence, transdermal testosterone pretreatment seems\nto increase clinical pregnancy and live birth rates in poor responders undergoing\novarian stimulation for IVF ( Ata & Seli,\n2015 ;  González-Comadran  et\nal ., 2012 ). However, there is insufficient data to support a\nbeneficial role of rLH, hCG, DHEA or letrozole administration in the probability of\npregnancy in poor responders undergoing ovarian stimulation for IVF ( Bosdou  et al ., 2012 ).\nTransdermal testosterone (TT) has been used at different doses and in different\nstimulation protocols ( Bosdou  et\nal ., 2016 ;  Kim  et\nal ., 2011 ;  Fàbregues\n et al ., 2009 ;  Massin\n et al ., 2006 ). However, it is difficult to establish\nits efficacy with sufficient evidence ( Polyzos\n et al ., 2018 ). This study compared luteal\nestradiol/GnRH antagonists protocol  versus  long GnRH agonists in\npoor responder patients according to the Bologna criteria, in which transdermal\ntestosterone has been used prior to the stimulation with gonadotropins.\n\nThis study was performed by a retrospective analysis of our database of women\nreferred to our center for IVF, and was conducted from January 2015 to May 2016\nin the Assisted Reproduction Unit of the Hospital Clinic in Barcelona (Spain).\nWe recruited 141 poor responder patients according to the Bologna criteria.\nAll the patients were in good health within normal limits of thyroid, kidney and\nhepatic laboratory results, and they had regular menstruation periods with\nduration of 21-35 days. None of them had taken any infertility medication in the\n3 months prior to the study.\nThe use of agonists or antagonists depended on the criterion of the specialist\nthat indicated the treatment; however, the pattern of androgenization was\nsimilar in both groups of patients. All patients were treated with transdermal\ntestosterone (TT) preceding ovarian stimulation with gonadotropins, but in one\ngroup we used luteal estradiol valerate and the GnRH antagonist protocol (Group\n1); whereas in the second group (Group 2) we used the long GnRH agonist protocol\n( Fig. 1 ). The study was approved by our\nInstitutional Review Board and informed consent was obtained from all individual\nparticipants included in the study (HB-15-EL-RS-C).\nFigure 1 Schematic representation protocols.\nSchematic representation protocols.\nStudy parameters, including days of stimulation, dose of gonadotropin\nadministered, peak E2 level on the day of human chorionic gonadotropin (hCG)\nadministration, number of oocytes retrieved , number of embryos and high quality\nembryos were evaluated. Pregnancy outcomes, including clinical and ongoing\npregnancy rates were also analyzed.\nIn no cycle we performed preimplantational diagnosis\nAll patients included in the study performed the same pattern with transdermal\ntestosterone (TT). Testosterone therapy was commenced on the first day of the\nnext menstrual cycle in Group 1, whereas in Group 2 testosterone began on the\nday when pituitary-ovarian suppression was confirmed. The therapy with\ntestosterone was continued for 5 days.\nTransdermal testosterone treatment was carried out using a daily single patch\nwith a 2.5 mg/day nominal delivery rate of testosterone (Testopatch, Pierre\nFabre Iberica SA, Barcelona, Spain) which was applied on the thigh at night and\nremoved always at 09:00h in the morning.\nThis transdermal delivery system maintains stable testosterone levels within\nnarrow ranges with little within - and between - subject variation, providing a\nhighly controllable way of delivering testosterone reliably, and the hormonal\ndose administered can be modified according to the duration of patch application\n( Buckler  et al .,\n1998 ;  De Sanctis  et al .,\n1998 ;  Mazer, 2000 ). We chose\nto use testosterone 20 mg/kg per day for 5 days on the basis of previous\nexperimental studies in primates ( Vendola\n et al ., 1998 ;  1999 ).\nThus, in each patient, the patch was applied at night at a time aimed to leave it\nin place for a predetermined number of hours in order to provide the desired\ndaily dose of testosterone (e.g. in a woman weighing 60kg and needing\n1200mg/day, the patch was used for 12h [0.1mg/h delivery rate 12h. 1.2mg or\n1200mg] and thus applied at 21:00h). Testosterone therapy was performed\naccording to a routinely used protocol ( Balasch\n et al ., 2006 ;  Fàbregues  et al ., 2009 ).\nIn 53 patients (Group 1), estradiol priming (4mg of oral estradiol valerate\n(E 2 ) (Progynova; Bayer, Spain)) was initiated on luteal day 21th\nand stopped in the first day of the next menstrual cycle. After TT therapy,\nrecombinant FSH (Gonal-F, Merck S.A., Madrid, Spain.) was initiated at an\ninitial dose of 300IU/day together with 75IU HMG (Menopur, Ferring SA, Madrid,\nSpain). The gonadotropin dose was adjusted according to serum E2 levels and\nserial ultrasound monitoring. The GnRH antagonist (Cetrotide, Merck S.A.,\nMadrid, Spain) was administered at a dose of 250µg/0.5ml/day when the\nleading follicle reached 14-15mm in its maximum diameter. GnRH administration\ncontinued until the day of hCG injection.\nIn 88 patients (Group 2), pituitary suppression was achieved by subcutaneous\nadministration of leuprolide acetate (Procrin; Abbott Laboratories, Madrid,\nSpain). This treatment was started in the mid-luteal phase of the previous cycle\nand given 1 mg daily, then reduced to 0.5mg after ovarian arrest, when serum\nestradiol (E2) concentration declined to < 50pg/ml and a vaginal ultrasound\nscan showed an absence of 10mm-diameter follicles. Transdermal testosterone was\nadministered during 5 days and gonadotropin ovarian stimulation was started the\nday following the last testosterone patch application. On Days 1 to 4 of ovarian\nstimulation, 300IU per day of r-hFSH (Gonal-F, Merck S.A., Madrid, Spain)\ntogether with 75IU HMG (Menopur, Ferring S.A., Madrid, Spain) were administered.\nOn day 5 onward, the gonadotropin dose was administered on an individual basis\naccording to ovarian response.\nThe criteria for hCG administration (250mg s.c.Ovitrelle, Serono S.A.) were the\npresence of two or more follicles >18 mm in diameter, with >4 follicles\nmeasuring >14 mm in association with a consistent rise in serum E2\nconcentration. The cycle was cancelled when there were less than 3 follicles\nwith diameter >14 mm after 8-9 days of gonadotropin therapy, or after 4-5\nadditional treatment days without attaining, or the imminent prospect of\nattaining, the criteria for hCG administration.\nOocyte aspiration was performed with vaginal ultrassonography 35-36h after hCG\nadministration. Embryo grading was recorded according to published criteria\n( Veeck, 1999 ); embryos graded 1 or 2\nwere considered of high quality. In both groups, embryo transfer was performed\nin the cleavage stage (day 3). The luteal phase was supported with vaginal\nmicronized progesterone (600mg/day given at 8h intervals) starting on the day\nfollowing oocyte aspiration and continuing either up to menstruation or, if the\npatients became pregnant, for at least the first 3 weeks of pregnancy.\nPregnancy was diagnosed by a positive serum β-hCG test 12 days after ET.\nClinical pregnancy was defined by observation of a fetal heartbeat using\ntransvaginal ultrasonography at 5-6 weeks gestation.\nAll statistical analyses were performed using the SPSS version 23.0 software\n(Chicago, IL, USA). We used a t-test to compare the mean values between two\ndifferent stimulation protocols.\nDifferences in outcome rates were analyzed using an χ 2  test or\nFisher's exact test.  p <0.05 was considered statistically\nsignificant.\n\nTable 1  depicts the baseline characteristics\nof the patients enrolled in the two different stimulation protocols. The groups were\nsimilar with respect to age, body mass index (BMI), duration of infertility, antral\nfollicle count, AMH levels and basal FSH and estradiol.\nComparison of patient characteristics for cycles using Luteal\nE 2 /TT/GnRH antagonist vs. TT/GnRH agonist protocol\nValues are mean ± DE unless specified otherwise\nIncluding cancelled cycles and cycles with ≤3 oocytes\ncollected\nThere were no reported major side effects after testosterone therapy and two\nprotocols were well-tolerated by all patients.\nTable 2  shows the stimulation parameters in\nboth groups studied. The number of cancelled cycles due to inadequate response was\nsimilar 13.6%  vs.  15.1%. The number of follicles and estradiol\nlevels on hCG day were not significantly different. However, the total gonadotropin\ndose used was significantly higher (2709±123IU  vs. \n2258±13;  p =0.023) in group 2 compared to group 1. In\naddition the mean length of stimulation was significantly higher (9.5±0.2\n vs.  7.9±0.3 days;  p =0.001) in group 2,\nwhen compared to group 1.\nOvarian stimulation characteristics in Groups 1 and 2\nValues are mean ± DE unless specified otherwise\nWhen comparing ovum retrieval and IVF outcomes in groups 1 and 2, there were no\nsignificant differences. However, there was a trend towards a slight improvement in\nthe implantation rate 27.3%  vs.  19%, pregnancy rate per oocyte\nretrieval (37.8%  vs.  31.6%) and per embryo transfer (38.6%\n vs.  34.3%) in group 1 as compared with group 2 ( Table 3 ).\nOvum retrieval and IVF/ICSI outcome in groups 1 and 2\nValues are mean ± unless specified otherwise\n\nThis is the first study comparing different GnRH analogues protocols in poor\nresponder patients according to the Bologna criteria in which TT has been used. The\npotential stimulating role of androgens on folliculogenesis has been suggested by a\nnumber of basic research studies ( Weil  et\nal ., 1999 ;  Vendola\n et al ., 1999 ;  Hillier  et al ., 1997 ), and illustrated by some\npathophysiological conditions ( Norman, 2002 ;\n Pigny  et al ., 2003 ) and\nclinical models ( Nagels  et al .,\n2015 ;  Grynberg  et al .,\n2010 ;  Futterweit & Deligdisch,\n1986 ). Transdermal testosterone has been shown in previous small RCTs to\nincrease the reproductive outcomes of IVF/ICSI patients ( González-Comadran  et al ., 2012 ). In\nmost of these studies, transdermal testosterone in relatively high doses was\nadministered before ovarian stimulation with a duration varying from 5 to 21 days\n( Bosdou  et al ., 2016 ;\n Kim  et al ., 2011 ;  Fàbregues  et al ., 2009 ;\n Massin  et al .,\n2006 ).\nSeveral previous studies have shown that testosterone may indeed have a role during\nthe later stages of follicular growth by increasing follicle-stimulating hormone\nreceptor messenger RNA in preovulatory follicles, and by stimulating oocyte\nmaturation. However, most of the published experiments indicate that testosterone\nmainly acts during the earlier stages of folliculogenesis by playing a role in\nfollicle activation and growth ( Walters,\n2015 ).\nIn this study we chose to use TT for 5 days on the basis of studies in primates and\nalso available reports from previous clinical studies ( Fàbregues  et al ., 2009 ;  2013 ;  Balasch\n et al ., 2006 ). Studies suggest that IGF-I appears to\nmediate or facilitate the effect of TT on early follicle development, and also\nimproves oocyte and embryo quality ( Meldrum\n et al ., 2013 ). IGF-I stimulation by testosterone may\nexplain the unusually high implantation rates reported in some studies with\ntreatments aimed at increasing the exposure of any kind of testosterone to ovarian\nfollicles in poor responders ( Bosdou  et\nal ., 2012 ;  Kim  et\nal ., 2011 ).\nRegardless of the dose and duration of the treatment with TT, it has been used both\nin long GnRH agonist ( Bosdou  et al .,\n2016 ;  Walters, 2015 ;  Fàbregues  et al ., 2009 )\nand short GnRH antagonist protocols ( Doan  et\nal ., 2017 ;  Kim  et\nal ., 2011 ;  Massin  et\nal ., 2006 ), but these protocols have never been compared in\nthis context before. Several studies suggested that there was no significant\ndifference on the number of oocytes retrieved, mature oocytes and pregnancy rates in\nboth GnRH antagonist and GnRH agonist protocols in poor responders ( Pandian  et al.,  2010 ;  Devesa  et al.,  2010 ). However,\n Pu  et al . (2011) \ndemonstrated that the stimulation duration was significantly lower with the GnRH\nantagonist protocol. The results of our study coincide with that provided in the\nliterature in the sense that the use of TT in a GnRH antagonist protocol could be a\nuseful option in these patients, shortening the duration of stimulation and the\nquantity of gonadotropins used.\nSeveral studies suggested that luteal estradiol could improve the results in poor\nresponders, shortening GnRH antagonist stimulation cycles ( Chang  et al.,  2012 ) , \ndecreasing cancellation cycles ( Reynolds  et\nal ., 2013 ), and improving FSH effects in granulosa cells\n( Ireland & Richards, 1978 ;  Wang & Greenwald, 1993 ). In our study it\nhas not been possible to evaluate the luteal estradiol efficacy, because we did not\nhave a control group in which we used the antagonist protocol without previous\nestradiol. However, taking into account what is suggested in the literature, this\ncould be a valid treatment option that should be analyzed in subsequent randomized\nstudies.\nThe main limitation of this study was its retrospective design and small sample size.\nHowever, the poor responder population according to the Bologna criteria represents\nonly a 5 to 10% of patients in most assisted reproduction clinics, which creates\nlogistic problems when performing a prospective study with sufficient power.\nAlthough the patients were not randomized, the two populations had similar baseline\ncharacteristics, which made possible to compare IVF outcomes between the groups.\nAdjuvant therapy with TT can be used with similar efficacy with both GnRH agonist and\nGnRH antagonist protocols in poor responders. More studies are needed to analyze\nwhether luteal estradiol can improve the response profile when TT is applied in GnRH\nantagonist protocol in these patients.\n\nAlthough there are controversial aspects regarding androgenic therapy in\nlow-responders, it seems that it can be a valid option as adjuvant therapy to\ngonadotropins. Its efficacy is not significantly different when different GnRH\nanalogues are used; however, short regimes with antagonists with previous estradiol\nin the luteal phase facilitate ovarian stimulation by shortening the days of\ntreatment and gonadotropin use.","source_license":"CC-BY-4.0","license_restricted":false}