Does hormonal contraception prior to in vitro fertilization (IVF) negatively affect oocyte yields? - A pilot study

Because the response to ovarian stimulation, to a large degree, depends on choice of stimulation protocols, any definition of poor response in association with in vitro fertilization (IVF) has to be considered relative. Poor response is, however, universally seen as the production of fewer than expected oocytes and embryos [1].

While by many considered a standard protocol in IVF, long agonist stimulations, because of suppressive effects on ovaries, are often considered inappropriate for poor responders [2]. Instead, hormonal contraceptives (HCs) are regularly utilized in such patients in preparation for ovulation induction [3, 4]. HCs are also used to reduce ovarian cyst formation [5] and to allow convenient scheduling of cycle starts [6, 7]. Results, utilizing HCs, have, however, not always been favorable [8‐10]. A recent Cochrane review found that pretreatment with combined oral contraceptives (OCs) led to poorer pregnancy outcomes [11].

The primary mode of action for HCs is thought to be suppression of pituitary gonadotropin secretion, secondarily leading to suppression of ovarian function [12, 13]. So-called combined HCs are combinations of estrogen (generally ethinyl estradiol, EE) and a variety of progestins. Differences among HCs, therefore, primarily reside with their progestational agents. Classical HCs have progestins, derived from estranes and gonanes, all to various degrees androgenic [14].

Synthetic progestins interact not only with progesterone receptors but also with other intranuclear steroid receptors [15]. Affinities for the various steroid receptors depend on the molecule from which a given synthetic progestin is derived [16]. In addition to progestational activity, steroid derivatives may also have androgenic, anti-androgenic, estrogenic, glucocorticoid or mineralocorticoid activities [16].

Consequently, one also has to conclude that different OCs, based on progestational content, should affect follicle maturation differently. Assuming this to be the case, different OCs could also be expected to affect functional ovarian reserve differently and, therefore, potentially oocyte yields.

Progestins may be classified into old progestins (norethisterone, levonorgestrel, gestodene) and new progestins (drospirenone, dienogest, trimegestone) [17]. Newer progestins are designed to be less androgenic. Like spironolactone, from which it derives, drospirenone functions as a competitive androgen antagonist [18, 19]. Dienogest is a 19- nortestosterone derivative but it differs significantly in structure from other nortestosterone derivatives [20]. An OC containing dienogest was recently approved for use in the United States [21].

AMH is considered a good marker of ovarian aging [22]. AMH levels generally correlate well with day 2/3 FSH levels [23], and are good predictors of ovarian response to ovulation induction [24‐26].

Suppressive effects of HCs on antral follicles and AMH levels have been reported. Women with polycystic ovarian syndrome (PCOS) who for 6 months took OCs, containing 35 ug of EE and 2 mg of cyproterone acetate, experienced significant decreases in antral follicles but no change in anti-Müllerian hormone (AMH) [27]. Cyproterone acetate is anti-androgenic but also suppresses gonadotropin secretion [28, 29].

Whether OC utilization suppresses AMH has remained controversial, with some studies supporting such an association [30, 31], and others disagreeing [32, 33]. This divergence suggests that different types of OCs, containing different progestins, and maybe different utilization length, could affect results.

Oocyte donors are selected to be healthy young women, without evidence of infertility and/or abnormal functional ovarian reserve. They often use OCs or other HCs. Especially in association with utilization of the so-called “contraceptive ring” (NuvaRing™, Organon USA), which contains, and continuously releases, etonogestrel/ethinyl estradiol, we noted in some of our egg donor applicants abnormally low age-specific AMH levels. Others noted diminished numbers of small follicles amongst women using contraceptive rings in comparison to those utilizing OCs, containing EE and levonorgestrel [34].

Considering all of these observations and reports, we in here presented study decided to formally investigate the effects of progestins on follicle maturation by determining how different HCs affect oocyte yields in young oocyte donors, stimulated with standard ovarian gonadotropin protocols.

For the whole study group mean age was 24.2 ± 4.0 years; mean AMH 4.4 ± 2.9 ng/mL; and mean oocyte yield was 15.6 ± 7.7. Time elapsed between last AMH and IVF cycle start was 8.9 ± 3.6 weeks.

There was no difference in age, age of menarche, BMI and AMH among the three groups (Table 2). Mean AMH was non-significantly higher without HC use but did not differ significantly between the three groups. This analysis was further limited by the fact that not all study subjects had AMH values available for analysis.

Oocyte yields among donors, who utilized more androgenic HCs, like estrane and gonane derived HCs, were lower 11.3 (95% CI 8.3 – 14.3) than those of either donors using no HCs at all 16.6 (95% CI 14.7 -18.4) (p < 0.05) or than those using anti-androgenic contraceptives 19.0 (95% CI 12.2-25.8) (p < 0.01) (Figure 1). Comparing androgenic HCs to the non-androgenic and no HC combined the omnibus test of significance was (p=0.018) after adjusting for donor age age and total gonadotropin dosage the significance was (p = 0.03).

In the introduction section of this manuscript we laid out in detail why, depending on included progestational agent, HCs should be expected to affect functional ovarian reserve in different ways. Investigating young oocyte donors, this study presents support for this concept, suggesting that recent use of HCs affects response of healthy young women to ovulation induction, resulting in significantly different oocyte yields, depending on whether HCs were utilized or not, and depending on what HCs were used. HC-use before IVF, therefore, quite apparently reduces the response to ovarian stimulation, producing lower than expected oocyte numbers. HCs, therefore, under a recently summarized consensus definition of poor ovarian response, qualify as inducers of relatively poor ovarian response to stimulation [1].

Since our center, at most, transfers two embryos in donor/recipient cycles, a diminution in oocyte yields, as here reported, likely, will not affect immediate, fresh-cycle pregnancy rates. Since total available oocytes, and especially embryos, do, however, reflect cumulative pregnancy chances, it is reasonable to conclude that any diminution of oocyte yields will also negatively affect cumulative pregnancy chances. Especially for recipients, desirous of more than one child, this may be of relevance.

The utilization of HCs in women with less favorable functional ovarian reserve than young oocyte donors, as represented by many women under standard infertility treatments with use of autologous oocytes, who may produce only relatively small oocyte yields, can, however, also be expected to affect fresh-cycle pregnancy rates since in such cycles the transfer of larger embryo numbers may be indicated.

Here presented data, while in view of a relatively small data set, still to be considered as preliminary, nevertheless, warrant a careful reevaluation of current practice patterns, involving the use of HCs in association with IVF.

This study, therefore, in summary, suggests that HCs, containing progestins derived from androgenic estranes and gonanes, suppress functional ovarian reserve, most likely via gonadotropin suppression, which interrupts the normal synergism between androgens and FSH at small follicle growth stages, in turn impacting oocyte yields. Since many fertility centers routinely use OCs in preparation for IVF cycles, such a practice, even in young women with normal functional reserve, appears to have negative consequences on oocyte numbers, as here demonstrated in oocyte donors.

An even more profound negative impact from androgenic HCs can, however, be expected in women with diminished functional ovarian reserve. Further investigations are needed to determine whether they should be utilized at all. At minimum, the conclusion from this preliminary study should, however, be that androgenic HCs should, likely, be avoided in women with evidence of low functional ovarian reserve.