br Methods br Results br Discussion LPD in terms

Methods
Results
Discussion LPD, in terms of insufficient secretion of progesterone by the CL, reflected by low serum progesterone levels can occur in IVF attempts using any kind of ovarian stimulation protocol (Fatemi et al., 2007, Garcia et al., 1981, Van der Linden et al., 2011). However, it is particularly frequent in ovarian stimulation cycles controlled by GnRH antagonist in which GnRH agonist is used as ovulation trigger (Bar-Hava et al., 2016, Leth-Moller et al., 2014). Bar-Hava et al. (2016) have shown recently that daily administration of GnRH agonist during the early luteal phase of IVF cycles using GnRH agonist, instead of hCG, for ovulation induction not only rescues CL function but is also sufficient to support 4-IPP implantation and further development without the need of any other kind of luteal phase support therapy. In the present study we used a similar approach in a particular group of patients, characterized by LPD occurring even after IVF using recombinant hCG for ovulation induction. These data show that LPD can occur in some patients independently of the type of ovulation trigger used, and that this problem can also be resolved by prolonged luteal phase administration of GnRH agonist. Unlike the study by Bar-Hava et al. (2016), luteal phase of patients involved in the present study was not supported uniquely by GnRH agonist, but vaginal progesterone administration was also used. This choice was motivated by ethical reasons, since it was not clear whether GnRH agonist would function, with this particular group of patients and ovarian stimulation protocol, in a similar way as in the study by Bar-Hava et al. (2016). However, this could not compromise the data interpretation because the same doses of progesterone were administered in all treatment cycles, both those using luteal phase GnRH agonist administration and those without the agonist. Because of the self-control design of this study, there is no reason to suppose that the basal level of serum progesterone, resulting from external progesterone administration were significantly different in the two successive attempts performed in the same patients. The values of serum progesterone were similar in both attempts on the day of embryo transfer, but they were significantly different on day 14 after oocyte recovery. This suggests that the very early CL (3 first days after oocyte retrieval) are not yet responsive to GnRH agonist action, and the serum progesterone measured was basically due to the external progesterone administration. Alternatively, a longer exposure to GnRH agonist may be needed to enhance CL progesterone secretion. It is important to note that the midluteal progesterone levels were significantly higher in the GnRH agonist-treated cycles even in those patients who did not become pregnant after this treatment, although the values of the pregnant patients were higher than those of the non-pregnant ones. This can be explained by an effect of the pregnancy itself on the CL progesterone secretion (Takaya et al., 2018), in addition to that of GnRH agonist. Anyway, the consistently higher progesterone levels both in the successful and unsuccessful cycles treated with luteal phase GnRH agonist as compared with the untreated cycles, in spite of the same dose of external progesterone administered, show clearly that GnRH agonist increased the patients’ own progesterone production. This increase was accompanied by a significant improvement of pregnancy outcomes in the GnRH agonist-treated IVF cycles, whereas other cycle characteristics, which could theoretically affect success rates, were comparable in both the GnRH agonist-treated and the untreated attempts. Even though the present data show clearly that luteal administration of GnRH agonist stimulates the production of progesterone by the corpus luteum in women with LPD, the mechanism of this action is not clear. In addition to its effect on the corpus luteum, GnRH agonists appear to target some other processes related to embryo implantation. It has to be reminded that the first report on a beneficial effect of GnRH agonist administration after embryo transfer was based on data obtained in egg donation cycles, after previous suppression of the recipients’ ovarian activity, a condition leading to a complete absence of the corpus luteum (Tesarik et al., 2004), thus excluding any action of GnRHa at this level. It was concluded that GnRH agonist acted through an enhancement of embryo developmental potential (Tesarik et al., 2004). However, later studies have shown that GnRH agonist also improved IVF/ICSI outcomes after ovarian stimulation (Pirard et al., 2015, Tesarik et al., 2006), thus adding the corpus luteum as another possible target, and this idea has been further corroborated by the finding that continuous GnRH agonist after embryo transfer can completely substitute for exogenous progesterone in GnRH antagonist-controlled and GnRH agonist-triggered ovarian stimulation regimens (Bar-Hava et al., 2016).