Oral Dydrogesterone Versus Vaginal Progesterone for Luteal Phase Support in Frozen-Thawed Embryo Transfer Cycles: A Systematic Review and Meta-Analysis of Randomized Controlled Trials.

The protocol is published in the International Prospective Register of Systematic Reviews (PROSPERO) with reference number (CRD42023489318), while no deviations from the original protocol will be presented in the current manuscript. The systematic review was conducted according to the guidelines of the Preferred Reporting Items of Systematic Reviews and Meta-analyses (PRISMA) [ 14 ] (PRISMA Checklist, Supplementary Material ). We performed a literature search for relevant studies from inception to 28 November 2024 across five databases: Embase (OVID), MEDLINE (OVID), APA PsycInfo (OVID), Global Health (OVID), and Health Management Information Consortium (HMIC) (OVID). Additional records were searched through registers, including Research Square and MedRxiv. The MedRxiv search was simplified according to database search functionality. The following search terms were combined in Ovid: “dydrogesterone”, “DYD”, “micronized progesterone”, “MVP”, “luteal phase support”, “ART”, “Assisted reproduction”, “IVF”, “In Vitro fertilization”, “Frozen thawed”, “FET”. Cross references were hand-searched to ensure that no study had been missed. We (K.S and M.I.K) independently reviewed the titles and abstracts and further assessed the full-text of the articles for compliance with the inclusion and exclusion criteria. Any disagreement between the authors was resolved by discussion. We applied the following eligibility criteria structured by PICOS items: Population: females undergoing frozen embryo transfer cycles; Intervention: oral dydrogesterone; Comparator: vaginal progesterone; Outcome: The reported primary outcome of interest was the ongoing pregnancy rate (OPR), defined as a viable pregnancy beyond 12 weeks of gestational age. The secondary outcomes considered were clinical pregnancy rate (CPR), miscarriage rate (MR), and live birth rate (LBR). Clinical pregnancy was evaluated by ultrasonographic visualization of ≥1 gestational sac or clinical signs of pregnancy (visualization of fetal heartbeat). Miscarriage was defined as a spontaneous loss of an intra-uterine pregnancy before week 12 of pregnancy. RCTs with women undergoing frozen embryo transfer cycles. Reviews, meta-analyses, cohort studies, case-control studies, case series, case reports, commentaries, editorials, and conference abstracts, and non-English language manuscripts were excluded. Data from selected publications were extracted on first author’s name, publication year, country, sample size, numbers lost to follow-up (post withdrawals), type of FET (e.g., programmed/artificial cycle or natural/mNC), numbers of patients in the two comparator groups (oral and vaginal progesterone), maternal age (in years as mean ± standard deviation (SD)) per group, Body Mass Index (BMI) (in kg/m 2 as mean ± SD) per group, progesterone dosage per group, and outcome data as number of cases per treatment group ( Table 1 ). As the outcome variables were dichotomous, namely positive pregnancy (either clinical, ongoing, live birth, or a miscarriage) or not, the odds ratio (OR), calculated from tables after raw data were extracted from each paper, was used as an effect measure with associated standard error. We applied a common effect model with inverse variance and assessed the heterogeneity across studies using the I 2 metric [ 15 ] In cases where an I 2 over 50% was estimated, indicating substantial heterogeneity between the study effects estimates, a random effects model was applied; otherwise, a common effects model was used. We investigated heterogeneity by sub-group analysis based on dydrogesterone clinical dosage. To evaluate the potential influence of the study by Ozer et al. [ 16 ], which utilized an mNC-FET, we performed a leave-one-out meta-analysis for each outcome reported by this study. All statistical analyses were performed using STATA software v13. The risk of bias was assessed by two independent reviewers (K.S and M.I.K), using the Risk of Bias 2 (RoB 2) tool for randomized studies [ 15 ]. A study could be of “high risk of bias”, “some concerns”, or “low risk of bias” [ 15 ]. Risk-of-bias assessment figures were created via the web-app RobVis (Risk-Of-Bias VISualization) [ 17 ]. We assessed the certainty of evidence using the GRADE framework, categorized as high, moderate, low, or very low [ 18 ]. Characteristics of the studies included in the meta-analysis. BMI: Body Mass Index; SD: standard deviation; P4: progesterone; NA: not applied

Over the last decades, frozen embryo transfer (FET) has been increasingly adopted in modern fertility units. The Human Fertilisation and Embryology Authority highlighted that between 2013 and 2018, fresh embryo transfers decreased by 11%, while FET cycles nearly doubled and accounted for almost 40% of all in vitro fertilization (IVF) cycles in 2018 [ 1 ]. In the USA, Australia, and New Zealand, the number of thawed transfers started to prevail over fresh cycles [ 2 ]. The European fertility units, although clearly delayed, followed this trend by increasing the rate of FET cycles from 28% to 34% between 2010 and 2016 [ 3 ]. Several factors contribute to the extreme rise of FET cycles over the years. First, a more efficient and safe warming and vitrification technique has been established. Furthermore, it has been proposed that FET cycles are linked to a lower likelihood of ovarian hyperstimulation syndrome (OHSS) associated with controlled ovarian stimulation (COS) [ 4 , 5 , 6 ]. Lastly, the routine use of pre-implantation genetic testing for aneuploidies (PGT-A) and the subsequent better genetic counseling for fertility couples contributed further to the extreme rise of FET cycles. In these cases, the embryos could remain frozen until chromosomal abnormalities have been excluded [ 2 ]. In FET cycles, progesterone can be administered via five different routes, including vaginal (PV), oral, intramuscular (IM), rectal, and subcutaneous. To date, no specific guidelines have been established to indicate the superiority of any of the aforementioned routes of progesterone administration in terms of reproductive rates. Nevertheless, various studies have compared other treatment routes with the most popular one, which is the vaginal route. In a three-arm randomized control trial (RCT) published in 2021, Devine and colleagues assessed the use of PV alone (200 mg twice daily), the combination treatment of PV (200 mg twice daily) with IM injections every 3rd day (50 mg), and the use of IM injections alone (50 mg daily), in terms of live birth rates among 1125 women planning vitrified-warmed transfer of high-quality non-biopsied blastocysts. The authors concluded that the PV-only treatment was associated with poor pregnancy rates, which was reflected by the high miscarriage rates, leading to the exclusion of the PV-only treatment group from the rest of the study. However, higher doses of 400 mg of progesterone daily, as performed in many European studies, might have yielded improved reproductive outcomes [ 7 ]. When comparing the most widely used methods of administration, it is essential to quantify and consider their adverse effects. Vaginal irritation, discharge, and bleeding are the most common side effects described for the vaginal administration, whilst intramuscular injections are associated with local pain, discomfort, and skin inflammation [ 8 , 9 ]. In this context, the oral route of progesterone administration seems an option for luteal phase support, as it is considered more patient-friendly and easy to use [ 10 ]. In fresh cycles, prospective studies have indicated that oral dydrogesterone is as effective as vaginal treatment with comparable patient satisfaction and tolerability [ 11 , 12 , 13 ]. In FET cycles, however, the role of oral dydrogesterone compared to the gold-standard, the vaginal route, remains inconclusive. Based on the above, the aim of this current systematic review and meta-analysis is to compare oral dydrogesterone with vaginal progesterone in FET cycles regarding pregnancy outcome rates.

The initial search retrieved 2372 studies. After duplicates were removed, 1748 studies remained for screening. Following title and abstract screening, 26 studies remained eligible for full-text screening. Finally, after full-text examination, 21 studies were excluded. In this stage, the three main reasons for exclusion were (i) non-randomized design (n = 10), (ii) reporting data on fresh cycles only (n = 7), and (iii) no direct comparison of oral versus vaginal progesterone (n = 4). References for the excluded studies are listed in Table A1 , Appendix A . Therefore, five studies remained suitable for inclusion in the present systematic review and meta-analysis ( Figure 1 ). A total of 636 women who underwent a frozen–thawed embryo transfer cycle were included in the present meta-analysis ( Table 1 ). Of the included studies, two were conducted in Turkey [ 16 , 20 ], two in Iran [ 21 , 22 ], and one in Brazil [ 19 ]. Regarding route of administration and daily dosage, three studies administered 40 mg of oral dydrogesterone [ 19 , 20 , 22 ], one study administered 30 mg of dydrogesterone [ 16 ], and another 20 mg orally [ 21 ]. As for the vaginal route, three studies administered 800 mg PV [ 19 , 20 , 21 ], one study 180 mg vaginal gel [ 18 ], and another one 8% vaginal gel [ 16 ] ( Table 1 ). For total ongoing pregnancy events, studies did not estimate statistically significant differences between patients receiving oral dydrogesterone and patients receiving progesterone vaginally for luteal phase supplementation. Τhe pooled odds ratio (OR) was 0.90 (95% confidence interval [CI] 0.59, 1.35), indicating than women receiving oral dydrogesterone had 10% lower odds of ongoing pregnancy compared to women receiving vaginal progesterone, but the result did not reach the nominal level of statistical significance. There was low heterogeneity as indicated by I 2 = 8.7% attributed to Zarei et al. (2016) [ 21 ] ( Figure 2 ). Overall, studies did not identify a statistically significant difference regarding miscarriage events between the two groups of progesterone luteal phase support. More specifically, regarding MR, oral dydrogesterone was associated with higher odds of miscarriage than vaginal progesterone, but the result was not statistically significant (OR = 1.41, 95% CI: 0.63, 3.13, I 2 = 0, Figure 3 ). Moreover, regarding clinical pregnancy events, the pooled analysis resulted in an OR of 0.94 (95% CI: 0.62, 1.42), p = 0.11, I 2 = 49.2%, Figure 4 ) between oral hydrogesterone vs. vaginal progesterone administration. The moderate heterogeneity (I 2 = 49.2%, Figure 4 ) found in the analysis of clinical pregnancy events was further explored by a clinically relevant sub-grouping based on the dosage of progesterone in the experimental group. As such, two clinically relevant sub-groups were created, one with the studies where oral progesterone daily dosage was ≤ 20 mg and one with studies that have reported oral progesterone > 20 mg for luteal phase support. This sub-group was chosen since large RCTs that have established the use of oral dydrogesterone in fresh cycles have used higher dosages than 20 mg daily for the supplementation of the luteal phase [ 13 , 23 ]. Sub-grouping by dosage of oral progesterone administration explained the heterogeneity, as only one study has given the low dosage, and the heterogeneity in the remaining studies (group of studies with oral progesterone dosage > 20 mg daily) was 0% ( Appendix A , Figure A1 ). Omitting this study resulted in an increase of the overall OR, suggesting favorable results for the experimental group ( Appendix A , Figure A2 ). Sub-group analysis was conducted exclusively for the outcome of CPR, as this was the only outcome exhibiting high heterogeneity among the studies. Finally, regarding live birth events, patient samples regarding oral dydrogesterone in comparison to those receiving PV for luteal phase support did not statistically differ in the common effects model (OR: 1.08; (95% CI: 0.67, 1.75, p = 0.99, I 2 = 0) ( Figure 5 ). To explore the influence of different FET protocols on the overall pooled analysis, a leave-one-out meta-analysis was performed. Excluding the study by Ozer et al. [ 16 ], the pooled OR for ongoing pregnancies decreased from 0.90 to 0.80, and for clinical pregnancies from 0.94 to 0.84, while the OR for miscarriages remained unchanged ( Figure A4 ). Two studies were considered to be of high risk of bias [ 21 , 22 ], two as low [ 16 , 19 ], and one of some concern [ 20 ] ( Figure 6 A,B). A comprehensive analysis of the RoB2 assessment is presented in Table A2 , Appendix A . Overall, the certainty of evidence using the GRADE approach was deemed moderate to low. The downgrading was mainly due to the risk of bias (two included studies were of high risk of bias), small sample sizes, the number of studies, and the wide confidence intervals reported ( Figure A3 , Appendix A ).

In the present systematic review and meta-analysis, we aimed to explore the efficacy of oral dydrogesterone in FET cycles compared to vaginal progesterone for luteal phase support in terms of reproductive outcomes. Overall, this pooled analysis did not yield significant differences in terms of pregnancy rates between the two groups and heterogeneity was minimal, indicating that oral dydrogesterone could be used similarly to vaginal progesterone in FET cycles. Low heterogeneity was expected, as only RCTs were included in the present analysis. Dydrogesterone is an orally given synthetic progesterone that has been used since 1960 for various menstrual disorders, for post-menopausal hormone replacement, cycle irregularity, as well as for the treatment of endometriosis [ 24 ]. Compared to natural progesterone, dydrogesterone has better oral bioavailability, with an elimination half-time of 5–7 h [ 25 ]. Several RCTs and meta-analyses have examined the use of oral dydrogesterone for luteal phase support in fresh IVF cycles. At first, oral dydrogesterone was used as an empirical supplementation for luteal phase support (LPS). The first studies that have compared dydrogesterone with vaginal progesterone have taken place in India [ 23 , 26 ]. In 2005, Chakravarty et al. compared the use of 20 mg dydrogesterone versus 600 mg micronized progesterone daily and concluded that similar reproductive outcomes were achieved [ 26 ]. In 2015, an updated Cochrane review, which compared the two routes of progesterone administration, reported that for CPR, the synthetic form produced better reproductive outcomes. For LBR, OPR, and MR, similar pregnancy rates were reported. Nevertheless, a substantial risk of bias and low-quality evidence was highlighted. Later, in 2020, the LOTUS II, a multi-center, open-label RCT assigned 1034 patients to receive either 30 mg dydrogesterone or 90 mg vaginal gel daily. Similar efficacy and safety were reported by the research group, suggesting that dydrogesterone may replace vaginal progesterone for LPS in the future [ 27 ]. These findings were confirmed by a later meta-analysis of nine RCTs, which included both fresh and FET IVF cycles. The researchers reported that “good quality evidence from RCTs suggests that oral dydrogesterone provided at least similar reproductive outcomes than vaginal progesterone” [ 28 ]. Nonetheless, this meta-analysis could be generalized primarily to fresh cycles, as only two RCTs regarding FET cycles were included due to the limited RCTs on FET cycles back then. As a result, data and evidence regarding FET cycles are still scarce. Regarding adverse effects, the LOTUS trial reported a comprehensive overview of possible effects on maternal health both for dydrogesterone and vaginal progesterone. From the whole trial, 12.4% of patients in the dydrogesterone group and 16.0% of females in the micronized vaginal progesterone group suffered treatment-emerging adverse events that resulted in research termination. The liver enzyme analysis of almost every patient in both groups was normal [ 13 ]. Other research groups have outlined that vaginal progesterone could be associated with vaginal discharge and irritation [ 26 ], as well as vaginal bleeding and perineal irritation [ 29 ]. Furthermore, regarding fetal risks, malformation rates were reported to be very low when dydrogesterone was used [ 30 ]. The LOTUS I trial reported <2% congenital, familial, or genetic disorders in both the dydrogesterone and the vaginal group [ 13 ], suggesting no significant safety concerns. Due to the increased rise of FET cycles and the consequent “freeze-all strategy” during the last decade [ 31 ], the research community has focused on the improvement of both the FET process and the freeze-all approach. Indeed, it has been reported that FET cycles are associated with a lower likelihood of OHSS occurrence and higher pregnancy rates [ 32 ]. Among the benefits of FET include the ability to preserve fertility by egg freezing in situations like cancer, and the ability to examine chromosomal abnormalities and select the best embryo for transfer through pre-implantation technology [ 2 ]. Luteal phase support is a key factor in the success of a FET cycle, and researchers have focused on finding the ideal progesterone duration, route of administration, and dosage to achieve high reproductive outcomes with minimal costs and patient satisfaction. The emphasis is currently on FET cycles in order to validate the high pregnancy rates that have been reported in fresh cycles, as dydrogesterone has been used more and more as LPS in fresh cycles in recent years. The best way to achieve this is to compare oral dydrogesterone with the gold-standard, the vaginal approach. A meta-analysis conducted by Barbosa et al. in 2018 [ 28 ] concerned mostly fresh cycles, as only two RCTs regarding FET cycles were included. Thus, the present systematic review and meta-analysis are of great significance in shedding light on the role of oral dydrogesterone in cryopreserved artificial cycles. Our study has several limitations. First, we included a relatively small sample size, as only a small number of studies were included for each outcome. More specifically, for ongoing pregnancies, miscarriages, and clinical pregnancies, four studies were included, whereas for live births, only three studies were deemed suitable to be included in the pooled analysis. Furthermore, four of the included studies were conducted in the Middle East, and one in Brazil, which limits the generalizability of our findings to the broader infertile population worldwide. Furthermore, a sensitivity analysis or sub-group analysis according to the different forms of vaginal supplementation was not feasible, due to the small number of RCTs available. Additionally, we performed a leave-one-out meta-analysis to assess the influence of the study by Ozer et al. [ 16 ], which utilized an mNC-FET. The analysis showed slight differences in the pooled analysis. Specifically, for miscarriage rates, the OR remained unchanged. Future research based on the FET protocol regarding the use of oral dydrogesterone is needed with a larger number of included patients to better evaluate this aspect. The main strength of our study was the inclusion of randomized trials, which, due to their inherent study design characteristics, could provide robust evidence regarding the research question. As explained above, RCTs, if successfully designed and executed, are not susceptible to confounding, which could falsely demonstrate an association between an exposure and an outcome. Additionally, the scope of our research is restricted to FET embryo transfer cycles, which have unique protocols and different endocrinological profiles than fresh cycles, such as the presence or absence of a corpus luteum. Therefore, we believe that the inclusion of only FET cycles could provide a deeper knowledge of LPS in these contexts. Despite the inherent limitations, this study is the first meta-analysis to demonstrate the noninferiority of reproductive outcomes between oral dydrogesterone and vaginal progesterone for LPS in FET cycles. Future research should focus on direct comparisons of oral dydrogesterone with other LPS administration routes and establish standardized dosing and duration to draw more definitive conclusions.

Oral dydrogesterone and vaginal progesterone may be associated with similar pregnancy rates in patients undergoing FET cycles. Nevertheless, based on the few RCTs conducted so far, further research on oral dydrogesterone and FET cycles is required to assess the effectiveness, tolerability, and/or possible superiority of oral dydrogesterone over other progesterone forms for LPS. Clarifying this issue will help to establish the best possible protocol for LPS in FET cycles, in terms of effectiveness, patient safety, and satisfaction.