Comparison of Oral Dydrogesterone with a Micronized Vaginal Progesterone in Fresh Embryo Transfer in IVF ± ICSI

Introduction: Luteal insu�ciency corresponds to a progesterone de�ciency affecting women who receive treatment for in vitro fertilization (IVF). Different routes of progesterone administration exist and have varying degrees of acceptability to patients. The aim of this study was to compare two luteal phase support (LPS) treatments: oral dydrogesterone versus micronized vaginal progesterone on the clinical pregnancy rates after fresh embryo transfert. Material and Methods: This study was a retrospective, monocentric and observational study carried out in the reproductive medicine department at the University Hospital, Femme Mère Enfant in Lyon. 580 consecutive women between 18 and 43 years old, who completed an IVF cycle with or without ICSI, followed by fresh embryo transfer on the second or third day after oocyte retrieval (D2 or D3) or at the blastocyst stage (D5 or D6) between July 2019 and July 2020 were included. Results: In the univariate analysis, the clinical pregnancy rate per transfer was comparable between the MVP and OD groups (29.7% and 27.6% respectively with p = 0.6460). In the multivariate analysis, OD also appeared to be associated with a similar pregnancy rate compared to MVP, with a non-signicant difference (p > 0.05) (OR [95% CI]): 0.922 [0.626; 1,358] with p = 0.6817. The use of OD compared to MVP did not signi�cantly in�uence the clinical pregnancy rate in any age group (p > 0.05) (OR [95% CI]): 0.954 [0.657; 1.386] with p = 0.8057. There was no signi�cant difference between the two groups in the clinical pregnancy rate, whether the patients belonged to the reference population of the center or not (p > 0.05) (OR [95% CI]): 2.367 [0.568; 3.568] with p < 0.0001. Conclusion: This is the largest retrospective study comparing these two routes of progestogens in LPS during IVF and it reinforces the use of the oral


Introduction
Luteal phase de ciency affects women undergoing in vitro fertilization (IVF) procedures.It is a progesterone insu ciency leading to an inadequate endometrium with a negative impact on embryo implantation and development.Several theories may explain this de ciency during ovarian stimulation.
Negative hypothalamic-pituitary feedback on LH secretion with supra-physiological doses of gonadotropins delivered (1) and high serum concentrations of steroids secreted by the large number of maturing follicles, could be one of these hypotheses (2).The mechanical withdrawal of granulosa cells, useful for the progesterone production, during oocyte retrieval could be another one (3).In addition, most IVF cycles currently using GnRH modulators (agonists or antagonists) disrupt the physiological pulsatility of GnRH.This may result in a dysfunctional corpus luteum shortening the progesterone production (3).
Progesterone, GnRH agonists or human chorionic gonadotropin (hCG) are used in luteal phase support (LPS).However, the increased risk of ovarian hyperstimulation syndrome (OHSS) generated by hCG and the heterogeneity of results concerning GnRH agonists (1,4) led to a consensus of progesterone-only use.
Moreover, administration of progestogens during luteal phase in IVF is associated with an improvement in the live birth rate (1,2,5,6).
Progesterone can be administered orally, vaginally, subcutaneously or intramuscularly.The optimal duration of treatment remains controversial and may be continued until the rst positive hCG plasma test, to the day of the rst ultrasound, or as late as the 12th week of pregnancy (7).Oral administration of progesterone is associated with a low bioavailability due to the hepatic rst-pass effect (8) but seems well tolerated (5).High absorption in vaginal administration is permitted by the vaginal rst-pass effect (9).However, it can be lowered after sexual intercourse (10,11) and can cause irritations, vaginal discharges and bleedings (10).Daily intramuscular or subcutaneous injection of progesterone has good bioavailability, nevertheless may be responsible for local pain or even abscesses (10,12).Finding the optimal progesterone treatment combining e cacy and high tolerance with the best compliance seems crucial to obtain and remains a challenge in IVF.
Dydrogesterone is an oral retroprogesterone used in the treatment of threatened or recurrent miscarriage (with proven progesterone de ciency) and repeated implantation failure (13,14).Recently, a positive association has been found between early luteal serum progesterone level and live birth rate in IVF using oral dydrogestérone (OD) for luteal support (15).Dydrogesterone has a good and almost exclusive a nity for progesterone receptors.Its high bioavailability and the progestogenic nature of the metabolites allows it to be used at lower oral doses than progesterone (16).Dydrogesterone safety in pregnancy, as well as tolerability during treatment, have already been described (17,18).Dydrogesterone half-life varies between 5 to 7 hours and involves 3 intakes per day to avoid serum concentration uctuations (19).
In a study by Tournaye et al. in 2017, a non-inferiority of OD compared to micronized vaginal progesterone (MVP) was found on the clinical pregnancy rate in IVF, with good tolerance and higher patient satisfaction.Similar results have been obtained in intrauterine insemination after controlled ovarian stimulation and support the use of dydrogesterone (20,21).Thus, we changed our habits in order to provide patients with the best possible comfort.
The aim of this study was to compare the pregnancy rate between OD and MVP in LPS after IVF or intracytoplasmic sperm injection (ICSI) and fresh embryo transfer.

Study design and participants
This retrospective, monocentric and observational study was carried out in the reproductive medicine department at the University Hospital, Femme Mère Enfant in Lyon.This study compared two types of LPS treatments.All consecutive women between 18 and 43 years old, who completed an IVF cycle with or without ICSI, followed by fresh embryo transfer on the second or third day after oocyte retrieval (D2 or D3) or at the blastocyst stage (D5 or D6) were included.Exclusion criteria were: oocyte retrieval complications (hemoperitoneum), the absence of collected oocytes, oocyte dysmaturity, unsuitable endometrium for transfer (thickness < 7 mm), the absence of embryo due to a fertilization or embryo culture failure, a risk of OHSS, patient who did not take LPS treatment or who did not show the day of transfer.Subjects with adenomyosis, untreated intracavitary broid or polyp, untreated symptomatic hydrosalpinx, stage IV endometriosis, or with a history of more than 3 recurrent early miscarriages were also excluded.If the embryo transfer was di cult, meaning a change of catheter or a technical di culty during the transfer, it was speci ed in the clinical le.Either single embryo transfer (SET) or dual embryo transfer (DET) were done.The transferred blastocysts were discriminated in 3 categories according to Gardner's classi cation (22).Embryos classi ed as "good" included blastocysts AA, AB and BA; embryos classi ed as "average" included BB.Embryos classi ed as "bad", all the other types of blastocysts were not transferred.The quality of the best embryo was retained in case of DET (23).845 patients were pre-selected between July 2019 and July 2020.

Ethical approval
Written consent for the use of personal medical and research data was collected for each patient prior to inclusion.The institutional review board of the Hospices Civils de Lyon gave its approval.

Study procedures
GnRH agonist or antagonist protocols were used for ovarian stimulation according to the center's habits.
LPS with progesterone began on the day of the oocyte retrieval (D0) and was continued until 12 weeks of gestation.In the MVP group, patients were treated as follows: 1 vaginal tablet of 200 mg of micronized progesterone in the morning and 2 vaginal tablets in the evening (i.e.600 mg per day).In the OD group, the regimen was: 1 oral tablet of 10 mg of dydrogesterone, taken three times daily (TID) (i.e. 30 mg per day).On Day 15 ± 3 (2 weeks after embryo transfer), a pregnancy test (serum β-hCG) was performed to determine whether treatment should be continued in cases of ongoing pregnancy until 12 weeks of gestation.The main investigator collected patient's data retrospectively using the services software (Médi rst® and Easily®).Data for the MVP group came from the year 2019 and data for the OD group from the year 2020, according to the chronology of the change in practice previously decided.

Types of infertility
Male infertility was de ned by abnormal sperm characteristics according to the World Health Organization's criteria (24).Infertility by poor ovarian reserve was de ned by the Bologna criteria (25) with antral follicle count of less than 5-7 follicles or an anti-Müllerian hormone (AMH) dosage < 0.5-1.1 ng/ml.
Infertility by anovulation or dysovulation was de ned as oligomenorrhoea or amenorrhoea and polycystic ovary syndrome (PCOS) was de ned according to the Rotterdam criteria of 2018 (26).

Study objectives
The primary objective of this study was to compare the clinical pregnancy rate between the MVP and OD groups, assessed by the presence of fetal heartbeats on transvaginal ultrasound between 6 and 12 weeks of gestation (weeks 4 to 10 of treatment).The secondary objectives were the comparison of clinical pregnancy rate between the MVP and the OD group among two subgroups: the reference population of the center (de ned as follows: age < 35 years; number of previous IVF attempts < 3; number of oocytes retrieved > 6; embryo freezing done) and by different age groups (< 30 years, 30 ≤ -< 35 years, 35 ≤ -< 40 years, ≥ 40 years).

Statistical analysis
Statistical analysis was carried out with the software R (v4.02).Quantitative data are represented as mean ± standard deviation (M±SD) and qualitative data as: number (percentage).In the bivariate analysis, quantitative variables were compared with a Student's t-test and qualitative variables with the χ² test.The general linear model (multivariate logistic regression) was used for the multivariate analysis, the mixed general linear model (mixed multivariate logistic regression) was used to account for the repetition of the same patient within the same sample.A test was considered signi cant when p was less than 0.05.We de ned therefore a required number of subjects of 580 (290 in each group) to ensure a study power of 80%.

Results
Study population 845 subjects were eligible and 580 were included in the statistical analysis, with 290 in each group (Figure 1).
Clinical and biological characteristics of patients and IVF cycles are summarized in Tables 1 and 2. The mean age was 34.1 ± 4.6 years, with a majority of nulliparous patients (70.5%).In the univariate analysis, there was a signi cant difference between the 2 groups on body mass index (BMI), smoking habits, PCOS's patients and the number of transferred embryos (Table 1).The groups were comparable on all other criteria, including age, gravidity and parity, type of IVF protocol received, number of oocytes retrieved, sperm characteristics, number of embryos, stage and embryo quality at the time of transfer (Table 2).Only 22.1% of the population analyzed corresponded to our reference population.Table 2 shows that the majority of embryo transfers were transfers of a single good quality blastocyst.In 94.7% of the time, the transfer was easy.Regarding di cult transfers, the difference between groups was not signi cant.
Table 3 shows that the clinical pregnancy rate per transfer was comparable between the MVP and OD groups (29.7% and 27.6% respectively with p = 0.6460).

Multivariate analysis
Data for the multivariate analysis are summarized in Table 4. Variables signi cantly associated with a decrease in the clinical pregnancy rate between 6 and 12 weeks of gestation regardless of treatment were (p < 0.05) (OR [95% CI]): age ≥ 35 years (0. 414 [0.260; 0.749] with p = 0.0001), history of ≥ 3 previous IVF attempts (0.407 [0.218; 0.739] with p = 0.0042), the embryonic stage D2-D3 at the time of transfer (0.597 [0.341; 1.017] with p= 0.0631).All variables with a signi cant difference in univariate analysis turned out to be insigni cant when analyzed in the multivariate model.Finally, OD appeared to be associated with a similar pregnancy rate compared to MVP, with a non-signi cant difference (p > 0.05) (OR [95% CI]): 0.922 [0.626; 1,358] with p = 0.6817.5b illustrates that belonging to the reference population was positively associated with the clinical pregnancy rate.There was no signi cant difference between the two groups in the clinical pregnancy rate, whether the patients belonged to the reference population or not (p > 0.05) (OR [95% CI]): 2.367 [0.568; 3.568] with p < 0.0001.
An additional mixed analysis considering repeated IVF attempts within the inclusion period for some patients did not provide any additional information and did not change the results.

Discussion
Our study found no signi cant difference between DO and MVP for the LPS after fresh embryo transfer in IVF ± ICSI, for the main objective which was the presence of fetal cardiac activity between 6 and 12 weeks of gestation.The use of OD compared to MVP also did not signi cantly in uence the clinical pregnancy rate by transfer in the center's reference population, nor by age group.Our results reinforce the interest in the use of dydrogesterone for LPS in IVF with or without ICSI in the local population of our center and support the results of similar published studies.Four prospective studies in the meta-analysis of Barbosa and colleagues (27) are comparable to our study and nd similar e cacy on the clinical pregnancy rate in IVF +/-ICSI between the use of OD and MVP (28-31).The RCT, published by Tournaye and colleagues (28) showed the non-inferiority of OD compared to MVP on the pregnancy rate with 974 randomized double-blind patients with su cient statistical power.In the four studies mentioned above, the daily dose of dydrogesterone ranged from 20 to 40 mg per day, but 10 mg twice daily was shown to reduce endometrial development compared to MVP given at a dose of 200 mg three times daily (32).We, therefore, chose the 10 mg dose TID on validated data from the Tournaye's study.Chakravarty's study has a selection bias because of a higher proportion of women aged 40 and over in the dydrogesterone group (29).As our study was retrospective, we included all eligible patients consecutively in order to minimise selection bias (33).Our study power is also guaranteed by the large sample size calculated beforehand.An attrition bias was found in the Saharkhiz et al. study, as 10.3% of participants were excluded after randomisation (31).This led to an imbalance regarding loss of sight (LOS) between groups: 17.9% LOS in the OD group and 2.6% in the vaginal progesterone group.We had no LOS and only one missing data concerning embryo transfer characteristics in our cohort.
The results of our study support previous studies, con rming that maternal age and embryonic stage are predominant elements in the prognosis for successful IVF.Pregnancy and live birth rates in IVF are increased before the age of 35 and when embryos are preferentially transferred at the blastocyst stage (34)(35)(36).It has to be mentioned that there were only 22% of good prognosis patients in our center explaining our low pregnancy rate.
Like most published studies, we chose the clinical pregnancy rate as the primary endpoint because we did not have the necessary hindsight to use the live birth rate during the analysis.However, the latter seems to be a criterion of higher clinical interest because it is more objective in evaluating the effectiveness of a technique.Only one study comparing oral dydrogesterone and vaginal progesterone after fresh embryo transfer was interested in the live birth rate, but it contains the biases mentioned above (29).An additional study with live birth records would therefore be necessary to support our ndings.
The retrospective nature of our study did not allow us to collect relevant data such as patient compliance or satisfaction toward treatments.Therefore, we were not aware of possible therapeutic discontinuities.Nevertheless, the Chakravarty ( 29) study seems to report a patient preference for oral dydrogesterone compared to vaginal micronized progesterone and Carp's meta-analysis (37) did not describe any signi cant adverse effect of dydrogesterone over the 22 studies studied.A speci c study looking at patient compliance with these treatments would be of interest.
According to several studies, the supra-physiological concentrations of progesterone used vaginally and the possible manual contamination during administration would alter the local microbiota, which is currently of growing interest in the eld of IVF (38,39).Indeed, the success of an embryo transfer depends on many factors, and in particular on utero-cervical microbial colonization (40).This may in uence the IVF pregnancy rate by the potential upward colonization of the endometrium or embryo during the embryo transfer through the cervix.
Finally, few studies compare oral and vaginal dydrogesterone in frozen embryo transfers.Unlike fresh transfer, the absence of corpus luteum in frozen embryo transfer, makes the endometrium entirely dependent on exogenous progesterone supplementation (41).Although two randomized studies have examined the use of OD in LPS in frozen cycles (42,43), they were small and larger additional studies are needed to establish its effectiveness in this context.
To conclude, this is the largest retrospective study comparing these two routes of progestogens in LPS during IVF and it reinforces the use of the oral form to improve patients' comfort.

Declarations
Tables

Figure 1 Patient
Figure 1

Table 5a
shows that the use of OD compared to MVP did not signi cantly in uence the clinical pregnancy rate in any age group (p > 0.05) (OR [95% CI]): 0.954 [0.657; 1.386] with p = 0.8057.Table

Table 1 :
Demographics and baseline (full analysis sample)

Table 3 :
Pregnancy rate per embryo transfer according to treatment MVP: micronized vaginal progesterone; OD: dydrogesterone.

Table 4 :
Factors in uencing clinical pregnancy rates in multivariate analysis in a population of women who underwent fresh embryo transfer FiguresPage 18/18