Pretreatment with oral contraceptive pills in women with PCOS scheduled for IVF: a randomized clinical trial.

Polycystic ovary syndrome (PCOS) is the most common cause of anovulatory infertility. A high level of circulating LH is an important biochemical feature of PCOS ( Baldani et al. , 2012 ; Liu et al. , 2012 ). In women with PCOS, the prevalence of elevated LH varies between 35% and 75%, especially in Asian populations ( Banaszewska et al. , 2003 ; Wang and Alvero, 2013 ; Endocrinology Subgroup and Expert Panel, Chinese Society of Obstetrics and Gyneocology, Chinese Medical Association, 2018 ). In a study by Chen et al. (2016) involving 1508 women with PCOS, the average ratio of LH to FSH was 1.5–1.6. Previous studies indicated that inappropriately high LH levels might affect oocyte maturation and fertilization rates, and impaired embryo quality, consequently resulting in higher rates of impaired pregnancy and miscarriage in women with PCOS ( Balen et al. , 1993a , b ; Homburg et al. , 1993 ; Shoham et al. , 1993 ; Jabara and Coutifaris, 2003 ; Urman et al. , 2004 ; Santos et al., 2010 ). However, recent evidence from retrospective studies has shown that high LH levels had no obvious negative effect on embryo quality in women with PCOS ( Chen et al. , 2016 ; Sun et al. , 2018 ; Singh et al. , 2021 ; Liu and Wang, 2023 ). Clinically, there are several methods used to reduce LH levels in women with PCOS. Pituitary downregulation using GnRH agonists significantly lowers the LH level in women with PCOS ( Homburg et al. , 1993 ; Balen et al. , 1993a ). However, the GnRH agonist long protocol increases the risk of ovarian hyperstimulation syndrome (OHSS) in women with PCOS compared with the GnRH antagonist protocol. OCPs, widely used in women with PCOS with irregular menstrual cycles, can also suppress LH levels. OCP pretreatment before the GnRH antagonist protocol is therefore a reasonable choice to lower LH levels before ovarian stimulation. The present study investigated whether not providing OCP to lower LH levels would affect oocyte/embryo quality in women with PCOS. We designed a single-center, open-label non-inferiority randomized controlled trial (RCT) to explore whether OCP pretreatment before the GnRH antagonist protocol in women with PCOS would improve embryo quality.

From February 2018 to August 2021, 281 infertile women with PCOS undergoing their first or second IVF cycle were enrolled. There were 39 participants excluded from the study for various reasons ( Fig. 1 ). A total of 242 participating participants were randomized to the OCP (n = 121) or non-OCP (n = 121) group. Flowchart of patient enrollment, allocation, follow-up, and analysis in a randomized clinical trial of pretreatment with oral contraceptive pills in women with PCOS undergoing IVF . OCP, oral contraceptive pills; FET, frozen/warmed embryo transfer. The baseline characteristics of the participants were comparable between the two groups ( Table 1 ). Hormonal and metabolic indices were similar between the two groups ( Supplementary Table S1 ). Baseline characteristics of patients in the OCP and non-OCP groups. Age is presented as median (minimum, maximum). Other values are presented as mean ± SD or n (%). OCP, oral contraceptive pill; half-ICSI, in cases with the risk of fertilization failure, half oocytes were inseminated by conventional IVF and half were inseminated by ICSI. The baseline characteristics of the participants were comparable between the two groups. The mean cycle day of gonadotrophin initiation was significantly lower in the OCP group than in the non-OCP group, mainly because of the random start protocol in the non-OCP group. As expected, the OCP group had lower LH levels than the non-OCP group on the day of gonadotrophin initiation ( Table 2 ). The LH/FSH ratio on the day of gonadotrophin initiation was also lower in the OCP group. The OCP group required significantly less gonadotrophin and a shorter duration of stimulation ( Table 2 ). Two of 101 women in the OCP group and 2 of 109 women in the non-OCP group had hCG 2000 IU administered 14–16 h after the dual trigger because of inadequate endogenous LH levels. The number of retrieved oocytes was significantly lower in the OCP group than in the non-OCP group. The metaphase II oocyte rate, the fertilization rate, and two pronuclei (2PN) rate in the ICSI cycle were similar between the two groups. However, the fertilization and 2PN rates in the IVF cycles were significantly higher in the OCP group than in the non-OCP group ( Table 2 ). Outcomes of controlled ovarian hyperstimulation in the OCP versus non-OCP group. Values are presented as mean ± SD, median (interquartile range), or n (%). In the OCP group, one patient did not have a blood sample for the detection of FSH, LH, and estradiol. In the non-OCP group, six patients did not have blood samples for the detection of FSH, five patients did not have blood samples for the detection of LH and estradiol. In the OCP group, three patients with half-ICSI, and in the non-OCP group six patients with half-ICSI were analyzed separately in the IVF cycle and ICSI cycle. rFSH, recombinant FSH; Gn, gonadotrophin; GnRH-a, GnRH agonist; MII, metaphase II; 2PN, double pronuclear; OCP, oral contraceptive pill. Thirteen participants in the OCP group and nine participants in the non-OCP group did not undergo FET ( Fig. 1 ). The FET outcomes of 106 participants in the OCP group and 109 participants in the non-OCP group were analyzed. One patient in each group was lost to follow-up after 12 weeks of gestation. There were no significant differences between the groups in terms of the FET endometrial preparation protocol, endometrial thickness, type of transferred embryos, or number of embryos transferred ( Supplementary Table S2 ). The number of good-quality embryos on Day 3 in the non-OCP group was not inferior to that in the OCP group (6.96 ± 4.51 versus 6.25 ± 4.88, respectively, absolute difference (AD) 0.71, 95% CI: −0.50 to 1.9), P  =   0.25 for the intention to treat (ITT) population, while in the per-protocol (PP) analysis, the number of good-quality embryos on Day 3 was 7.18 ± 4.39, versus 6.58 ± 4.93 (AD 0.61, 95% CI: −0.65 to 1.86, P  =   0.34, respectively). There were no significant differences between the two groups in the rate of blastocyst formation. Incidence of moderate or severe OHSS was 6.61% in the OCP group and 10.74% in the non-OCP group (relative risk 0.59, 95% CI: 0.24 to 1.48, P  = 0.36). The results were similar for the ITT and PP analyses ( Table 3 ). Primary, secondary, and safety endpoints for intention-to-treat analysis and per-protocol analysis. Unless otherwise stated, values are presented as mean ± SD or n (%). One patient in each group was lost to follow-up after 12 weeks of gestation. OCP, oral contraceptive pill. The clinical outcomes after the first FET cycle, including the implantation rate, clinical pregnancy rate, live birth pregnancy rate, and first trimester miscarriage rate, were similar between the two groups ( Table 3 ). There were no ectopic pregnancies in either group. Cumulative live births were also similar between the two groups ( Table 3 ). Neonatal gestational age, birthweight and fetal gender were also similar between the two groups ( Table 4 ). Newborns’ gestational age, birthweight, and male gender in the OCP and non-OCP groups. Unless otherwise stated, values are presented as mean ± SD or n (%).

We performed an RCT from 7 February 2018 to 31 August 2021, at the Reproductive Center of the First Affiliated Hospital of Sun Yat-sen University in Guangzhou, China. The study protocol was approved by the ethics committees for Clinical Research and Animal Trials of the First Affiliated Hospital of Sun Yat-sen University (No.: [2018]014). The study was conducted in accordance with the Declaration of Helsinki and good clinical practice guidelines. Our study was registered in the Chinese Clinical Trial Registry (registration number chiCTR1800014822) on 7 February 2018. All women provided written informed consent to participate in the study. Patients were eligible if they fulfilled the following inclusion criteria: diagnosis of PCOS according to the 2003 Rotterdam PCOS diagnostic criteria ( Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group, 2004 ); patient age 20–40 years; and scheduled for a first or second IVF/ICSI cycle. Indications for IVF included tubal pathology, male factor, combined tubal and male factors, or PCOS only. Women with endometriosis, endometrioma or submucous myoma, uterine malformations such as bicornuate uterus and uterine cavity adhesion, ovarian tumors, recurrent spontaneous abortion, and chromosomal abnormalities were not eligible. Women fulfilling the inclusion criteria were approached by their clinician and asked if they were willing to participate. Ultrasound was performed to confirm the absence of dominant follicle (follicles > 8 mm in size) development, and a urine pregnancy test was used to exclude pregnancy. After obtaining written informed consent, eligible women were randomized using random numbers generated by SPSS software (version 24.0, IBM Corp., Armonk, NY, USA). Randomization was conducted by a research assistant of the hospital who was not involved in the recruitment and clinical management of the participants. The research assistant provided sealed envelopes with treatment allocation (the OCP group and the non-OCP group) to the study staff. The obvious treatment differences between the two groups meant that the study was designed as an open-label study for women and doctors. However, embryologists and IVF technicians who evaluated the oocytes and embryos were unaware of the treatment allocation. Participants allocated to the OCP group were provided with OCPs containing ethinyl estradiol (0.03 mg) and drospirenone (3 mg). OCPs were administered daily for 21 days to induce menstruation, followed by 7 days of washout. Recombinant FSH (rFSH) was started on Day 7 of the washout period. In participants allocated to the non-OCP group, rFSH was started immediately regardless of the day of the menstrual cycle. Otherwise, the groups were treated comparably. All participants underwent a fixed protocol of GnRH antagonist ovarian stimulation. First, an ultrasound was performed to confirm the absence of dominant follicles (follicle size > 8 mm). After confirmation, controlled ovarian stimulation (COS) was initiated using human rFSH (Puregon, Merck Sharp & Dohme, Ravensburg, Germany) daily for 3–4 days. The starting dose of rFSH was fixed according to the BMI of the patient (150 IU for BMI 19–25 kg/m 2 ; 225 IU for BMI > 25.1 kg/m 2 ). Transvaginal ultrasound was performed, and serum LH, estrogen, and progesterone levels were assessed to monitor follicle growth. The rFSH dose was adjusted according to the ovarian response. A GnRH antagonist (Ganirelix; Orgalutran, N. V. Organon, New Jersey City, NJ, USA) was administered s.c. at 0.25 mg daily on Day 5 or 6 of rFSH stimulation in the two groups. Triggering of the final oocyte maturation was performed when two follicles ≥18 mm or three follicles ≥17 mm were noted under ultrasound; a dual trigger [2000 IU hCG plus 0.2 mg of triptorelin (Decapeptyl, Ferring, Kiel, Germany)] or 250 μg of recombinant hCG [Ovidrel; Merck Serono, Modugno (Bari), Italy] was used. Blood samples were obtained 8–12 h after the dual trigger to determine LH, estradiol, and progesterone levels. If the LH level was <15 IU/l or the progesterone level was <3 ng/ml ( Chang et al. , 2016 ), hCG 2000 IU was re-administered to the patient 14–16 h after the trigger to supplement the insufficient endogenous LH that could lead to reduced oocyte retrieval. Oocyte retrieval was performed 34–36 h after the trigger using the transvaginal ultrasound-guided puncture of follicles. Standard laboratory protocols for conventional IVF and ICSI ( Magli et al. , 2008 ) were performed. ICSI was only performed for male factor infertility, according to ESHRE guidelines ( ESHRE Capri Workshop Group, 2007 ). Evaluation of embryo quality on Day 3 was performed according to the modified Istanbul consensus ( Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011 ). Good-quality embryos were defined as 7–9 cells with 0–20% cytoplasmic fragments and an even size on the third day after fertilization. The blastocyst score on Day 5 or 6 was assessed according to the Gardner morphological criteria ( Gardner et al. , 2000 ), based on the degree of expansion and the development of the inner cell mass and trophectoderm. A freeze-all strategy was applied for all participants. A hormone replacement therapy-based protocol was performed for endometrial preparation in the frozen/warmed embryo transfer (FET) cycle. On Days 2–3 of the second menstrual cycle (either spontaneous or induced by progesterone) after oocyte retrieval, women received oral estradiol valerate (Progynova, Delpharm Lille, Lys-Lez-Lannoy, France) for endometrial preparation. Intramuscular progesterone (40–60 mg) or vaginal progesterone gel (Crinone, Merck Serono, SP, Brazil; 90 mg/day) with oral dydrogesterone (10 mg twice daily) was administered when endometrial thickness reached 8 mm, and estradiol valerate was used for ≥10 days. Embryos were transferred on the fourth (cleavage-stage embryo) or sixth (blastocyst) day after progesterone initiation. Twelve days after the embryo transfer, serum β-hCG was measured. If biochemical pregnancy was achieved, clinical pregnancy was confirmed on the basis of the detection of gestational sacs by ultrasonography at 4–5 weeks after the embryo transfer. The luteal-phase support continued until 10 weeks after conception. All pregnancy and neonatal outcomes were obtained through review of medical records. Basal FSH, LH, and estradiol levels were determined in the morning of Days 2–5 of the menstrual cycle (either spontaneous or induced by progesterone). Serum FSH, LH, estradiol, and progesterone levels were determined using an automated Elecsys immunoanalyzer (Abbott Ireland Diagnostics Division, Longford, Ireland). Intra- and inter-assay coefficients of variation were <4% and <5%, respectively, for LH; <5% and <6%, respectively, for FSH; <7% and <8%, respectively, for estradiol; and <6% and <7%, respectively, for progesterone. The assays were carried out in batches. The primary outcome of the study was the number of good-quality embryos on Day 3 after insemination. Secondary outcomes included the rates of blastocyst formation, implantation, clinical pregnancy, live birth pregnancy during the first FET cycle, and cumulative live births rates and neonatal gestational age, birthweight, and fetal gender. The blastocyst formation rate was defined as the ratio of the number of blastocysts to the number of Day 3 embryos cultured for blastocyst formation. Clinical pregnancy was defined as the presence of an intrauterine gestation sac at 6–7 weeks of gestation. Live birth was defined as delivery of any viable infant at 28 weeks or more of gestation. Cumulative live births were defined as live birth resulting from pregnancies that occur within 24 months from the day of ovarian stimulation. The safety outcomes included moderate to severe OHSS ( Practice Committee of the American Society for Reproductive Medicine, 2016 ), miscarriage, and ectopic pregnancy. The sample size was calculated using PASS15 software (NCSS, LLC, East Kaysville, UT, USA). We hypothesized that immediate start with ovarian stimulation was non-inferior as compared to pretreatment with OCP. Non-inferiority was assumed to be proven if the lower boundary of the two-sided 95% CI did not exceed −2.0 good-quality embryos on Day 3. Assuming that the actual distribution was normal and the SD was 5.0, a total of 210 participants (in a 1:1 ratio) were required to achieve 80% power to detect non-inferiority with a one-sided significance level of 0.025. The margin of non-inferiority was −2.0, and the true difference between the means was assumed to be 0. Considering a 15% dropout rate in the study, we planned to include two groups of 121 women (242 women). All randomized participants were included in the intent-to-treat analysis. Values are reported as the median (interquartile range) for non-normally distributed variables, the mean ± SD for normally distributed variables, or as numbers (percentages) for categorical variables. We compared baseline data between the two groups. After confirming that the baseline data between the two groups were balanced and comparable, the outcomes of COS and pregnancy outcome were compared. The differences in continuous variables between the intervention and control groups were analyzed using the independent sample Student’s t- test if the data followed normal distributions; otherwise, Mann–Whitney U tests were applied. Categorical variables were analyzed using the chi-squared test. SPSS version 24.0 (IBM, Armonk, NY, USA) was used for the data analysis. A P -value of <0.05 indicated statistical significance.

Our study showed that not suppressing LH was non-inferior to pretreatment with OCP before IVF in the GnRH antagonist protocol.

In this RCT study, we confirmed that in women with PCOS scheduled for IVF, no pretreatment was non-inferior to pretreatment with OCP. Apparently, high LH levels on the day of gonadotrophin initiation did not have an obvious negative impact on embryo quality. Lowering the LH level with OCP before ovarian stimulation might be unnecessary. PCOS frequently presents with abnormally high basal LH levels, especially in Asian populations. The impact of high basal LH levels on reproductive outcomes remains unclear. Previous studies suggested that high basal LH levels have a deleterious effect on embryo quality and implantation rates, and may be a causal factor in early pregnancy loss ( Homburg et al. , 1988 ; Tesarik and Mendoza, 2002 ). However, others came to different conclusions. Studies showed that suppressing LH levels using GnRH analogs did not decrease the miscarriage rate in women with PCOS ( Clifford et al. , 1996 ; Ludwig et al. , 1999 ). Recent studies showed that a high basal LH level and/or a high LH/FSH ratio in women with PCOS had no significant impact on the number of top-quality embryos, the clinical pregnancy rate ( Sun et al. , 2018 ; Singh et al. , 2021 ; Fu and Kuang, 2023 ; Liu and Wang, 2023 ), or even the miscarriage rate ( Sun et al. , 2018 ; Fu and Kuang, 2023 ). However, all the above studies were retrospective analyses. Prospective RCT studies are lacking on this issue, begging the question: is it necessary to lower LH levels in women with PCOS before initiating the GnRH antagonist protocol? Oral contraceptives are commonly used to schedule the day of initiation in an IVF cycle and can lower the LH level. Clear consensus is lacking on whether pretreatment with combined oral contraceptives in COS protocols has a negative effect on pregnancy outcome. To control the possible adverse impact of oral contraceptives on oocyte quality, we initiated COS after 7 days of washout in our study. As expected, the LH level on the day of gonadotrophin initiation was significantly lower in the OCP group than in the non-OCP group, although the levels were comparable before OCP intervention. Our study aimed to determine whether lowering LH levels would benefit oocyte quality; therefore, we set the number of good-quality cleavage-stage embryos as our primary endpoint. The results showed that the number of Day 3 good-quality embryos was similar between the two groups. In the first FET cycle after COS, there were no significant differences between the two groups in terms of the implantation, clinical pregnancy, miscarriage, and live birth rates. Our study is in line with previous retrospective studies, which showed that elevated LH levels in women with PCOS undergoing IVF did not have a negative effect on oocyte and embryo quality ( Sun et al. , 2018 ; Singh et al. , 2021 ; Fu and Kuang, 2023 ; Liu and Wang, 2023 ) or pregnancy outcomes after FET ( Sun et al. , 2018 ). In addition, our results showed that cumulative live birth rates did not differ between the two groups. Our results implied that lowering the LH level before COS in women with PCOS might be unnecessary. We did not expect that the duration of ovarian stimulation would be significantly shorter in the OCP group than in the non-OCP group. In women with a normal response, oral contraceptives prolong the duration of stimulation and increase the total consumption of gonadotrophin ( Rombauts et al. , 2006 ). The discrepancy between women with a normal response and our women with PCOS might reflect the degree of LH level reduction after OCP intervention. Rombauts et al. (2006) reported that oral contraceptive-scheduled group stimulation with rFSH commenced 2 days after OCP was discontinued; however, in our study, gonadotrophin stimulation was initiated 7 days after OCP discontinuation. More importantly, the serum LH level was 0.9 IU/l in the Rombauts et al. (2006) study, indicating a more profound pituitary suppression than that noted in our study (mean LH level = 5.32 IU/l). In other words, the relatively normal LH levels in the OCP group in our study might not have prolonged the duration of COS. We found that the fertilization rate in the IVF cycles was significantly lower in the non-OCP group. Singh et al. (2021) also observed that the fertilization rate was lower in the group with higher basal LH levels, although the reason for this was unclear. We speculated that women with high LH levels might have smaller follicles after ovulation induction, resulting in a lower fertilization rate. However, nearly 70% of our women underwent IVF; therefore, it was impossible to compare the actual oocyte maturation rate between the two groups. This interesting phenomenon will be the subject of our future research. Although we used the freeze-all strategy, a moderate/severe OHSS rate was still notable in both groups, especially in the non-OCP group (10.74%). This might have occurred for several reasons. First, we used a fixed initial dose according to patient BMI. The FSH doses were adjusted according to the ovarian response after 3–4 days. Although women with PCOS usually have a higher BMI, the initial dose might still be too high because of ample ovarian antral follicles. Second, the numbers of retrieved oocytes in both groups were high in our study. Finally, most women in our study were triggered with a short-acting GnRH agonist accompanied with small-dose hCG (2000 IU) in case of inadequate endogenous LH surge after GnRH agonist administration, and the latter might have increased the risk of OHSS. To the best of our knowledge, this is the first RCT to focus on the possible benefits of lowering high LH levels in women with PCOS. Our study provides solid evidence that not suppressing LH was non-inferior to pretreatment with OCP. In addition, pretreatment with OCP may increase the risk of thrombo-embolic events in PCOS patients who will end up with grossly elevated serum levels of estradiol after COS. Furthermore, our control group was designed to have a random start (median 6 days of menstruation) unlike the traditional initiation time. Our study showed that withdrawal bleeding before COS was unnecessary for women with PCOS if the freeze-all strategy was used. There are several limitations to our study. Only women with PCOS in Southern China were recruited; therefore, caution should be exercised when generalizing our results to all women with PCOS. Also, since a freeze-only strategy was used, the results of this study only are mainly applicable when infertile women with PCOS undergo the freeze-only method. Furthermore, because of the obvious treatment difference between the two groups, the study was designed as an open-label study for women and doctors. However, the study had a randomized controlled design that minimized bias.