Hum. Reprod. Advance Access published online on August 20, 2008
Human Reproduction, doi:10.1093/humrep/den319
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letter to the Editor |
Androgen priming before ovarian stimulation for IVF
The Center for Human Reproduction, New York, USA; Foundation for Reproductive Medicine, New York, NY, USA
1 Correspondence address. E-mail: ngleicher{at}thechr.com
We would like to congratulate Lossl and associates on an in principle very well-designed prospectively randomized study (Lossl et al., 2008
), but would at the same time like to point out some potentially important facts, which may somewhat modify the conclusions reached by the authors.
Considering the quality of their study design, there is little argument with the authors' principal conclusion that the short-term administration of aromatase inhibitors (AI) before controlled ovarian hyperstimulation (COH) failed to improve ‘in their study’ the number of top-quality embryos. Their data, however, raise a number of significant questions, among them whether their finding may not actually reflect their stimulation protocol, and why AI would even be expected to show an effect on ovarian function in women with normal ovarian reserve.
Without wishing to enter the controversies surrounding use of antagonist versus agonist protocols and their respective impacts on COH (Al-Inany et al., 2006; Huirne et al., 2007), it would seem reasonable to assume that in average patient populations, pre-stimulation antral follicle counts, stimulated follicle counts, peak estradiol (E2) levels and E2 levels per follicle should correlate with number of ultimately retrieved oocytes. Expectations also are that the number of retrieved oocytes statistically correlates with oocyte quality and, therefore baring differences in semen parameters and timing of ovulation induction, should reflect fertilization potential of oocytes and, ultimately, pregnancy outcomes (Roest et al., 1996). This study, however, paradoxically deviated from these expectations in all of these parameters. Although the AI-treated group demonstrated significantly more pre-ovulatory follicles (P = 0.014), peak E2 (P < 0.001) and E2 per follicle (P = 0.005), oocyte number were not higher and the fertilization rate was, indeed, significantly lower (P = 0.007). Moreover, ongoing pregnancies, though statistically similar, demonstrated a trend against the utilization of AI.
It, of course, is possible that different pharmaceutical agents, leading to increases in intraovarian androgen levels, may affect outcomes differently. Such a conclusion seems, however, unlikely under the proposed assumption that intraovarian androgen levels, indeed, can affect ovarian function. Why, as quoted by the authors, Casson et al. (2000) and we (Barad and Gleicher, 2005, 2006; Barad et al., 2007), after androgenization with dehydroepiandrosterone (DHEA), would see the expected correlation between pre-oocyte retrieval COH parameters and subsequent oocyte numbers, but Lossl et al. did not, is, therefore, difficult to understand and possibly suggests that their observation may be the consequence of ovarian stimulation, rather than evidence of absence of effects of AI on ovarian function. Adverse effects on oocyte numbers in antagonist cycles in comparison to short agonist cycle have, of course, been reported (Franco et al., 2006).
Our biggest concern about this study arises, however, from patient selection. As the authors note in Materials and Methods, prior poor ovarian response to COH was an exclusion criteria. In practical terms, this means that women with diminished ovarian reserve were excluded from their study. The quoted claims of prior reported efficacy of ovarian androgenization by Casson et al. (2000) and our group (Barad and Gleicher, 2005, 2006; Barad et al., 2007) relate, however, exclusively to women with severely diminished ovarian reserve. In humans, there really, therefore, is no prior reported support for beneficial effects of androgenization in women with normally functioning ovaries.
Indeed, our concern would be that, assuming a normal intraovarian milieu with normal intraovarian androgen levels, added additional androgenization could result in excessive levels and well recognized adverse effects of excessive androgenization, widely reported in the literature (Taniguchi et al., 2007). We, therefore, caution from interpreting the data of Lossl et al. as relevant to women with diminished ovarian reserve and object to any comparison of their study outcome to outcomes reported by Casson et al. (2000) and our group (Barad and Gleicher, 2005, 2006; Barad et al., 2007).
Paradoxically, in this case, the authors' ability to conduct such a well-designed prospectively randomized study offers further evidence of normal ovarian function in their patient cohort. In attempts to prospectively randomize women with severely diminished ovarian reserve between DHEA and placebo, we have twice failed in recruiting adequate patient numbers, willing to undergo potential randomization to placebo. A first attempt had to be abandoned in New York City in 2006 and a second trial, started in January of 2007 in collaboration with European investigators in a number of European countries, had later that same year to be stopped for the same reasons. In view of their severely limited reproductive life span, women with significantly diminished ovarian reserve, on both sides of the Atlantic, apparently are not willing to risk lengthy randomization to placebo.
This raises our final criticism of the paper by Lossle et al.: the efficacy of androgenization, at least as it relates to DHEA supplementation, appears directly associated with length of use. Our data strongly suggest that beneficial effects of DHEA on ovarian function only peak after 4–5 months of supplementation, though significance is already reached at 
2 months (Barad et al., 2007). This observation also correspond with much milder effects observed by Casson et al. (2000) after only short-term DHEA supplementation in comparison with our much longer supplementation (Barad and Gleicher, 2005, 2006; Barad et al., 2007).
Lossl et al. used only short-term androgen priming and still, as earlier noted, observed a number of significant ovarian function improvements, such as larger follicle numbers, higher peak E2 levels and higher E2 levels per follicle. In addition, they confirmed the expected increases in intrafollicular E2 (P = 0.007) and androgen levels (P = 0.014). It, therefore, is entirely possible that not only their stimulation protocol may have prevented them from achieving significant results, but also an insufficiently short time of treatment with AI.
The one conclusion their study, therefore, in our opinion allows is that it may, after all, also be worthwhile to investigate longer-term androgenization, even in women with apparently normal ovarian function. This is a conclusion we recently also reached after recognizing in a collaborative study with Canadian colleagues that DHEA supplementation apparently decreases miscarriage rates in women age of >35 years with diminished ovarian reserve by up to 50–80% (Gleicher et al., 2008).
References
Al-Inany HG, Abou-Setta AM, Aboulghar M. Gonadotropin-releasing hormone antagonists for assisted conception. Cochrane Database Syst Rev (2006) 19:CD0011750.
Barad DH, Gleicher N. Increased oocyte production after treatment with dehydroepiandrosterone. Fertil Steril (2005) 84:756.[Medline]
Barad D, Gleicher N. Effect of dehydroepiandrosterone on oocyte and embryos yields, embryo grade and cell number in IVF. Hum Reprod (2006) 21:2845–2849.
Barad D, Brill H, Gleicher N. Update on the use of dehydroepiandrosterone supplementation among women with diminished ovarian reserve. J Assit Reprod Genet (2007) 24:629–634.[CrossRef]
Casson PR, Lindsay MS, Pisarska MD, Carson SA, Buster JE. Dehydroepiandrosterone supplementation augments ovarian stimulation in poor responders: a case series. Hum Reprod (2000) 15:2129–2132.
Franco JG Jr, Baruffi RL, Mauri AL, Petersen CG, Felipe V, Cornicelli J, Cavagna M, Oliveira JB. GnRH agonist versus GnRH antagonist in poor ovarian responders: a meta-analysis. Reprod Biomed Online (2006) 13:618–627.[Web of Science][Medline]
Gleicher N, Ryan E, Weghofer A, Oktay K, Blanco-Mejia S, Barad DH. Dehydroepiandrosterone (DHEA) supplementation in women with diminished ovarian reserve reduces miscarriages. Fertil Steril. (in press).
Huirne JA, Homburg R, Lambalk CB. Are GnRH antagonists comparable to agonists for use in IVF? Hum Reprod (2007) 22:2805–2813.
Lossl K, Yding Andersen C, Loft A, Freiesleben NLC, Bangsbøll S, Nyboe Andersen A. Short-term androgen priming by use of aromatase inhibitor and hCG before controlled ovarian stimulation for IVF. A randomized controlled trial. Hum Reprod (2008) doi:10.1093/humrep/den131.
Roest J, van Heusden AM, Mous H, Zeilmaker GH, Verhoeff A. The ovarian response as a predictor of successful in vitro fertilization treatment after the age of 40 years. Fertil Steril (1996) 66:969–973.[Web of Science][Medline]
Taniguchi F, Couse JF, Rodriguez KF, Emmen JM, Poirier D, Korach KS. Estrogen receptor-alpha mediates an intraovarian negative feedback loop on thecal cell steroidogenesis via modulation of Cyp17a1 (cytochrome P450, steroid 17alpha hydroxylase/17,20 lyase) expression. FASEB J (2007) 21:586–595.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||