{"paper_id":"bc6da553-e0db-40d3-b010-97642b87cd4b","body_text":"One in six couples worldwide will experience at least one infertility problem during\ntheir reproductive years ( ESHRE, 2018 ) and the\nmajority may benefit from assisted reproduction technology (ART) treatments. Between\n5.6% and 35.1% of women will exhibit poor ovarian response (POR) ( Oudendijk  et al. , 2012 ).  Zhang  et al.  (2016)  estimated\nthat 5-18% of all  in vitro  fertilization (IVF) cycles are canceled\nfor poor ovarian response.\nThe European Society of Human Reproduction and Embryology (ESHRE) standardized the\ncriteria for poor ovarian response in the Bologna criteria. At least two out of\nthree following criteria are required to define poor ovarian response during IVF: 1)\nMaternal age > 40 or any other risk factor for poor ovarian response; 2) prior\npoor ovarian response (≤ 3 oocytes with a conventional stimulation protocol);\n3) abnormal ovarian reserve test (antral follicle count [AFC] less than 5-7\nfollicles or anti-Müllerian hormone [AMH] less than 0.5- 1.1ng/ml). The ESHRE\nalso indicated that history of two episodes of poor ovarian response after a maximum\nstimulation protocol is enough to define poor ovarian response ( Qin  et al.,  2017 ).\nA variety of methods have been applied to improve ovarian response, including\nincreased gonadotropin dosage, modulation with gonadotropin-releasing hormone\n(GnRH), flare-up regimes, adjunctive human growth hormone therapy, minimal ovarian\nstimulation with clomiphene citrate, and unstimulated or natural cycle IVF. However,\nthe outcomes of these treatments have been less than satisfactory ( Wiser  et al.,  2010 ).\nDehydroepiandrosterone (DHEA) is an endogenous steroid produced in the zona\nreticularis of the adrenal cortex and by ovarian theca cells. In the ovary, it\npromotes follicular development and granulosa cell proliferation by increasing\nintraovarian androgen concentrations. DHEA also enhances the level of follicular\ninsulin-like growth factor-1 (IGF-1), which promotes folliculogenesis by enhancing\nthe effect of gonadotropin and reducing follicular regression ( Wiser  et al.,  2010 ).\nDespite the wider use of DHEA in poor responders, views among clinicians vary\nconsiderably. The aim of this study was to review the potential benefits of DHEA\nsupplementation for poor responders undergoing ART treatment.\n\nSearches were carried out on PubMed and the Cochrane Library for relevant literature\non the efficacy of DHEA at improving the ovarian response of women with poor ovarian\nresponse and women with premature ovarian aging after failed IVF. In all cases, DHEA\nsupplementation was administered before ovarian stimulation in IVF cycles.\nThe following keywords were used in the searches: \"DHEA\"; \"dehydroepiandrosterone\";\n\"poor ovarian responder\"; \"low response\"; \"diminished ovarian reserve\"; \"IVF\"; and\n\"ICSI\".\nRandomized controlled trials (RCT), meta-analyses, systematic reviews, retrospective\nand prospective controlled studies were eligible for inclusion.\nOur manual and automatic searches yielded a total of 38 publications. The full texts\nof the articles were retrieved, and 25 met the inclusion criteria. The selected\narticles were designed to compare whether pre-treatment with DHEA improved the IVF\noutcomes of patients with poor ovarian response.\nIn order to facilitate the understanding of the effects of DHEA supplementation on\nthe response to IVF, we separated the results into topics as follows: effect of DHEA\non ovarian reserve markers; effect of DHEA on response to ART treatment; effect of\nDHEA on oocyte quality; effect of DHEA on pregnancy/live birth rates; and other\ntherapies for poor responders.\n\nThe meta-analysis published by  Zhang  et\nal.  (2016)  investigated the AMH levels and the AFC of\nthe same patients before and after DHEA supplementation. Only one of the\nanalyzed studies was an RCT. The authors of the included articles did not report\nthe administered doses of DHEA. The study population featured women with\ndiminished ovarian reserve undergoing ovarian stimulation in IVF protocols. Six\nself-control studies indicated that DHEA treatment increased the AFC, with no\nsignificant heterogeneity observed (95% CI 0.14-0.66,  p =0.002,\nI 2 =24%). Two articles described significant increases in AMH\nlevels for all age groups after DHEA supplementation (95% CI 1.27-1.6,\n p <0.0001, I 2 =0%).\nConversely, the randomized double blind placebo-controlled pilot study by  Yeung  et al.  (2014)  did\nnot detect statistically significant differences in the median AFC of\nindividuals given DHEA compared with placebo (3.5 [1.75-4.25]\n vs.  4 [3-4];  p =0.436). This study included\nindividuals with ages ≤ 40 years and subfertility lasting for more than a\nyear, with poor ovarian response defined as an AFC < 5. Patients with a\nhistory of ovarian cystectomy or oophorectomy, individuals previously submitted\nto cytotoxic chemotherapy or pelvic irradiation, and subjects with a history of\ntaking testosterone supplementation were excluded. The patients were given 25mg\nof DHEA three times a day, starting at least 12 weeks before the scheduled IVF\ntreatment.\nSimilarly to the AFC, serum follicle stimulating hormone (FSH), AMH, and serum\ntestosterone (cutoff at 1.0ng/ml) levels were not statistically different\nbetween the groups throughout the study period ( Yeung  et al. , 2014 ). This same study found that\nwomen with higher follicular DHEA levels (cutoff at 180µg/dL) had a\nstatistically higher number of good-quality embryos (1 [0-2]\n vs.  0 [0-0.25];  p =0.013). Therefore, DHEA\nsupplementation may have improved the ovarian environment in which follicular\nmaturation takes place, leading to decreased aneuploidy. However, the underlying\nmechanism is still unknown.\nWiser  et al.  (2010) \ncarried out a randomized prospective controlled study to evaluate the effects of\nDHEA supplementation on IVF. The case group received 75 mg of DHEA once a day at\nleast six weeks prior to the IVF cycle and during treatment. Patients with prior\npoor response to ovarian stimulation were included in the study. Patients over\nthe age of 42 and patients given DHEA at any time before the start of the study\nwere excluded. The mean peak estradiol levels of the DHEA and control groups on\nthe day of chorionic gonadotropin hormone (HCG) administration were not\nstatistically different (732pg/ml and 917pg/ml, respectively;\n p =0.2).\nTherefore, according to the literature, the effect of DHEA supplementation on\novarian reserve markers is still controversial. More trials are needed to\nunderstand the potential effects of DHEA on ovarian reserve markers and\npregnancy hormones.\nA meta-analysis by  Li  et al. \n(2015)  concluded that DHEA did not increase the number of oocytes\nretrieved among poor responders submitted to IVF (95% CI 1.43 to 0.96;\n p =0.70). Participants in the case group received 75mg of\nDHEA daily for at least 12 weeks.\nZhang  et al.  (2016) \npublished a meta-analysis and found three trials including 69 patients in which\nwomen aged less than 36 years had significant increases in the number of oocytes\nafter treatment with DHEA (95% CI 2.15-2.61,  p <0.0001,\nI 2 =1%). In a study included in this meta-analysis,  Jirge  et al . (2014) \nreported that the mean of number of retrieved oocytes from controls was 2.09,\n versus  4.45 in the DHEA group (RR 2.36 [95% CI\n2.13-2.59]).\nIn this meta-analysis, six trials including 163 female patients aged 36+ years\nalso indicated increased numbers of oocytes after DHEA treatment (95% CI\n0.73-1.90,  p <0.0001, I 2 =72%). The study by  Tsui  et al . (2015) , for\nexample, reported that the mean of number of retrieved oocytes from controls was\n2.4,  versus  4.2 in the DHEA group (RR 1.80 [95% CI\n1.22-2.38]).\nAlthough statistically heterogeneous, three case-control trials, two RCTs, and\none prospective cohort study featured in this meta-analysis enrolling a combined\n1398 patients were performed to determine implantation rates. The implantation\nrate of patients treated with DHEA was significantly higher than the rates seen\nin untreated controls (RR 1.56, 95% CI 1.20-2.01,  p =0.0007).\nThe study by  Xu  et al. \n(2014)  reported that the implantation rate increased in DHEA group,\nfrom 27 to 47 patients (RR 1.82 [95% CI 1.19 -2.87]) ( Zhang  et al.,  2016 ).\nQin  et al.  (2017) \npublished a meta-analysis to evaluate the effect of DHEA therapy on the ovarian\nresponse and pregnancy outcomes of patients with diminished ovarian reserve.\nNine studies were included - four  RCTs, four retrospective studies, and\none prospective study. The patients were given 25 mg of DHEA three times a\nweek for a minimum of six or 12 weeks.  In this meta-analysis, two\nRCTs and three retrospective studies concluded that the number of retrieved\noocytes was not different between the DHEA and control groups. There was\nsignificant heterogeneity between the studies (RR -0.69, 95% CI:2.18-0.81)\nThe patients included in a randomized controlled trial by  Kotb  et al.  (2016)  were given 25 mg of DHEA\nthree times daily for three months before IVF. Women undergoing IVF with POR\nbased on the Bologna criteria with ages ranging from 20 to 45 years were\nincluded. Women with a body mass index > 35 kg/m 2 , individuals\nwith a single ovary, subjects allergic to DHEA, and females with diabetes were\nexcluded. The individuals given DHEA had statistically greater counts of\nretrieved oocytes (6.9  vs.  3.1,  p =0.039),\nsignificantly higher fertilization rates (62.3%  vs.  52.2%,\n p =0.03), fewer days on controlled ovarian hyperstimulation\n(11.6  vs.  12.6,  p =0.001), and lower\ngonadotropin doses (3383IU  vs.  3653IU,\n p =0.045). Interestingly, the number of metaphase II (MII)\noocytes, total number of embryos, and number of transferred embryos were not\ndifferent in the case groups ( Kotb  et\nal.,  2016 ).\nThe literature indicates the existence of a tendency in DHEA supplementation\nincreasing the number of oocytes retrieved and embryo implantation rates. More\nrandomized controlled trials are needed to confirm the actual benefits of DHEA\nfor patients with poor ovarian response undergoing ART treatment.\nYeung  et al.  (2014) \nconcluded that the number of follicles and oocytes retrieved were similar\nbetween groups, although the DHEA group had a non-statistically greater mean\nnumber of fertilized embryos (1  vs.  3,\n p =0.155), cleaved embryos (1  vs.  3,\n p =0.169), transferred embryos (1  vs.  2,\n p =0.430), and top-quality embryos (0  vs. \n1,  p =0.141).\nThe retrospective analysis by  Ferrario  et\nal . (2015)  found that AMH and DHEA sulfate levels were\npositively correlated with the number of mature oocytes (RR=0.784), fertilized\noocytes (RR=0.607), and developed embryos (RR=0,513). The authors\nretrospectively analyzed the data from 148 poor responders diagnosed based on\nthe Bologna criteria. Women with polycystic ovary syndrome and endometriosis\nwere excluded, along with normal responders and patients with severe male factor\n(cryptozoospermia or azoospermia).\nTherefore, few studies have investigated the correlation between DHEA and oocyte\nquality, and the findings are still controversial.\nIn the meta-analysis by  Li  et\nal.  (2015) , six studies evaluated the effects of DHEA on\nclinical pregnancy rates. DHEA significantly increased the clinical pregnancy\nrates of poor responders submitted to IVF compared with controls (RR 2.13, 95%\nCI 1.12-4.08;  p =0.02). However, no significant effect was\nobserved in an analysis including only RCTs. In addition, careful evaluation of\neach individual study included in the meta-analysis found that most of the\nstudies were not statistically sound.\nThe meta-analysis of  Zhang  et\nal.  (2016)  included eight randomized control trials, ten\ncohort studies, and three case-control studies designed to calculate clinical\npregnancy rates. The meta-analysis found statistically significant increases in\nthe pregnancy rates of patients treated with DHEA (RR 1.53, 95% CI 1.25-1.86,\n p <0.0001). However, most of the studies were not\nstatistically sound.\nThe meta-analysis cited above included four RCTs ( Moawad & Shaeer ,  2012 ;  Tartagni  et al.,  2015 ;  Wiser  et al.,  2010 ;  Yeung  et al.,  2014 ) and\ntwo prospective cohort trials ( Jirge  et\nal.,  2014 ;  Vlahos\n et al.,  2015 ) designed to calculate live birth\nrates. In the group analysis of these articles, there were significantly higher\nlive birth rates among the patients given DHEA (RR 1.87, 95% CI 1.22-2.88,\n p =0.004) ( Zhang  et\nal.,  2016 ). However, careful evaluation of each\nindividual study found that only one ( Tartagni\n et al.,  2015 ) reported statistically reliable\nhigher live birth rates in the DHEA group (RR 1.79, 95% CI 1.01-3.17). In this\nrandomized study, 109 infertile patients with ages ranging from 36-40 years were\nselected to undergo the long protocol and received 75 mg of DHEA once a day for\neight weeks before starting IVF.\nThe meta-analysis by  Qin  et al. \n(2017)  included four RCTs ( Wiser\n et al.,  2010 ;  Kara  et al.,  2014 ;  Zhang  et al.,  2014 ;  Yeung  et al.,  2014 ), three retrospective studies\n( Barad  et al.,  2007 ;\n Xu  et al.,  2014 ;\n Fusi  et al ., 2013 ),\nand one prospective study ( Vlahos  et\nal.,  2015 ) designed to analyze clinical pregnancy rates.\nClinical pregnancy rates were significantly higher in the DHEA group (OR 1.47,\n95% CI: 1.09-1.99). However, subgroup analysis based on RCTs revealed that there\nwas no significant difference between the groups (OR 1.08, 95% CI:\n0.67-1.73).\nThe randomized controlled trial by  Kotb  et\nal . (2016)  described significant benefits from DHEA to\npoor responders in the form of increased clinical pregnancy (11\n vs.  23,  p =0.09) and ongoing pregnancy (9\n vs.  20,  p =0.036) rates.\nThe randomized controlled trial by  Yeung\n et al . (2014)  failed to identify statistically\nsignificant differences in clinical pregnancy (18.8%  vs.  25.0%,\n p =0.380), ongoing pregnancy (18.8%  vs. \n12.5%,  p =0.326), live birth (12.5%  vs.  12.5%,\n p =1.0), or miscarriage (0  vs.  12.5%,\n p =0.326) rates.\nThe randomized prospective study by  Wiser\n et al . (2010)  described higher clinical\npregnancy (2  vs.  4,  p =0.25) and live birth (1\n vs.  3,  p =0.20) rates among individuals\ngiven DHEA. However, the differences were not statistically significant.\nThe literature describes a tendency toward increased clinical pregnancy and live\nbirth rates in females given DHEA, albeit not statistically significant.\nIn addition to DHEA, the agents more commonly used in daily practice to improve\novarian response in IVF cycles are transdermal testosterone, clomiphene citrate,\naromatase inhibitors, recombinant LH, and recombinant human chorionic\ngonadotropin.\nThe meta-analysis by  González-Comadran\n et al.  (2012)  looked into the effects of\ntransdermal testosterone in women with poor ovarian response undergoing IVF. The\nauthors focused on transdermal testosterone, an agent known to produce powerful\nsystemic androgenization and subsequent greater action of FSH compared with\nother androgen-modulating agents. Three randomized controlled trials were\nincluded. The dose of transdermal testosterone used in the intervention group\nvaried between the articles as follows: gel, 10 mg for 15-20 days; gel, 12.5 mg\nfor 21 days during pituitary desensitization; or 2.5 mg patches per day for five\ndays.\nThe authors also found that women on transdermal testosterone achieved\nsignificantly higher live birth rates (RR 1.91, 95%CI 1.01-3.63,\nI 2 =0%) ( González-Comadran\n et al.,  2012 ).\nThe systematic review by  Kamath  et\nal.  (2017)  analyzed the effectiveness of oral medication\nfor ovulation induction (clomiphene citrate, 100mg daily for five days;\nletrozole, 5 mg daily for five days) versus gonadotropin-only regimens for\ncontrolled ovarian stimulation in IVF. The authors were unable to find\nconclusive evidence indicating that clomiphene citrate or letrozole with\ngonadotropins differed from gonadotropin-only in terms of live birth or\npregnancy rates, either in the general population of women undergoing IVF cycles\nor in poor responders.\nLarge, high quality controlled randomized trials are needed to provide input on\nothers therapies designed to improve the IVF outcomes of poor responders.\n\nThis review aimed to evaluate whether DHEA supplementation might increase the ovarian\nresponse of females undergoing IVF cycles. The first obstacle to attaining the goal\nof this review was to find consensus over the definition of POR. Authors define POR\nbased on different principles and use of the Bologna criteria is not universal.\nSecondly, the populations selected in each study were not similar in matters such as\nage, cause of infertility, stimulation protocols, or total dose of ovulation\ninduction medication. And lastly, there was no standardization over the time of\nadministration or dose of DHEA in use in the studies.\nDespite substantial heterogeneity between studies, we have decided to review the\ncurrent literature on DHEA and ovarian response. And based on this review, DHEA\nsupplementation appears to improve the IVF outcomes and level of ovarian response of\npoor responders. DHEA is a simple-to-use, side-effect-free medication probably\nlinked to fewer days of stimulation and lower gonadotropin doses. Furthermore,\naccording to our review, DHEA showed a tendency toward increased numbers of\nretrieved oocytes and higher embryo implantation, clinical pregnancy, and live birth\nrates. Further randomized controlled trials are needed to confirm the actual\nbenefits of DHEA to patients with poor ovarian response.\n\nFew RCTs have looked into DHEA supplementation as adjuvant therapy in IVF cycles.\nFurther multicenter RCTs are needed to confirm whether supplementation with DHEA\nimproves the outcomes of poor responders.","source_license":"CC-BY-4.0","license_restricted":false}