Clomiphene citrate priming increases sensitivity during ovarian stimulation in poor ovarian responders undergoing in vitro fertilization treatment: a retrospective cohort study

Abstract This single-centre retrospective cohort study aimed to investigate whether a clomiphene citrate (CC) priming protocol could increase ovarian sensitivity in poor ovarian responders. It included 294 patients (374 ovarian stimulation cycles). Of these, 193 cycles were treated by a CC priming antagonist protocol (study group) and 181 by the classical flexible gonadotropin-releasing hormone antagonist protocol (control group). Stimulation data and laboratory and clinical outcomes were compared between the groups. The results showed that in the study group, total gonadotropin dosage and dosage per follicle were considerably lower, the follicle-to-oocyte index was significantly higher, and the gonadotropin duration was shorter. After adjusting for potential confounders, multivariate regression analysis showed that cumulative ongoing pregnancy remained comparable between the groups (adjusted odds ratio: 0.761, 95% confidence interval: 0.300–1.933, p = 0.566). Age, body mass index, gonadotropin dosage per follicle, and the follicle-to-oocyte index were negatively associated with the reproductive outcomes. The result of the sensitivity analysis showed that patients in the study group were administered less gonadotropin at a lower gonadotropin dosage per follicle and for a shorter duration. In conclusion, the CC priming antagonist protocol offered a convenient and patient-friendly way to increase ovarian sensitivity during ovarian stimulation in poor ovarian responders.


Introduction
Since ovarian stimulation was introduced as an assisted reproductive technology (ART), poor ovarian response (POR) management has proved challenging for practitioners.Such a response to ovarian stimulation usually indicates reduced follicular response and/ or oocyte number, resulting in a low live birth rate.The most universally accepted definition of POR is the Bologna Criteria (Ferraretti et al., 2011).Multiple strategies have been suggested to enhance the outcomes of POR patients.A simple approach is to vary the gonadotropin (Gn) dosage or the stimulation initiation time, but the pregnancy rate remains very low (Klinkert et al., 2005).Lekamge et al. (2008) found that higher gonadotrophin stimulation did not improve the in vitro fertilization (IVF) outcomes in POR patients.Another commonly used stimulation regimen is the administration of gonadotropin-releasing hormone (GnRH) agonist and gonadotropins during the follicular phase (the flare-up protocol).Non-stimulated (natural cycle) IVF treatment has also been attempted.All these methods have had limited success, and the results achieved remain controversial (Ashrafi et al., 2018;Robert et al., 2017;Sfakianoudis et al., 2020).Guidance on the optimal management of patients with poor response and/or low sensitivity to ovarian stimulation is still lacking.
Clomiphene citrate (CC) is a nonsteroidal triphenylethylene derivative that exhibits oestrogen agonist and antagonist properties (Chaube et al., 2005).It increases the pulse frequency of hypothalamic gonadotropin-releasing hormone (GnRH) and pituitary luteinizing hormone (LH) and follicle-stimulating hormone (FSH) by inhibiting the negative feedback effect of circulating oestradiol (Kamath et al., 2017).Previous studies suggested that CC administration could reduce FSH use, and thus the costs incurred to the patient during ovarian stimulation (Bechtejew et al., 2017;Pilehvari et al., 2016).However, few studies have compared ovarian sensitivity to CC-related protocols with gonadotropin plus GnRH protocols.On the other hand, many methods have been proposed to evaluate the ovarian response and/or ovarian sensitivity to gonadotropin stimulation alone, such as follicle output rate (FORT) and follicle-to-oocyte index (FOI).The optimal index for assessing ovarian sensitivity to gonadotropin stimulation remains controversial.We proposed a new biomarker, namely, Gn dosage per follicle.This index is calculated as the ratio between the total FSH dosage administered and the number of preovulatory follicles.Our unpublished data demonstrate that the index could reflect ovarian sensitivity and is positively related to IVF outcomes.
In the present study, we evaluated the use of a CC priming antagonist protocol in POR patients to discover whether it could increase ovarian sensitivity among POR patients over what could be achieved with the classical flexible antagonist protocol.

Materials and methods
This was a retrospective, observational single-centre cohort study.The patients were recruited from the Medical Centre for Human Reproduction, Beijing Chao-Yang Hospital, Capital Medical University from 1 January 2017 to 31 October 2020.The present study was approved by the Ethics Committee of Beijing Chao-Yang Hospital, Capital Medical University with the number 2021-SCI-61.Written informed consent was waived due to the retrospective nature of the study.
An electronic database was interrogated and only women who fulfilled the Bologna Criteria were selected.The exclusion criteria included: (1) body mass index (BMI) higher than 36 kg/m 2 ; (2) age over 45 years; (3) severe endometriosis, autoimmune or metabolic disorders; and (4) severe azoospermia for the partner.A total of 374 cycles performed in 294 patients were included.They were divided into two groups based on the stimulation protocols: CC priming GnRH antagonist protocol (study group, n ¼ 193) and classical flexible GnRH antagonist protocol (control group, n ¼ 181).The choice of treatment protocol was based on the doctors' judgement.
In the study group, ovarian stimulation was initiated with 100 mg/day CC for 5 days, from day 3 to 7 of the menstrual cycle.A recombinant FSH (rFSH; Gonal F, Merck Serono, Germany) dose of 225-300 IU was administered daily, starting from day 6 of the cycle.Gn dosage was adjusted according to hormone levels and follicular development.A daily dose of 0.25 mg of GnRH antagonist (Cetrotide, Merck Serono, Germany) was administered when the dominant follicle reached 14 mm in diameter.The co-treatment continued until the trigger day.Women in the control group were stimulated with the flexible GnRH antagonist protocol.A daily dose of 225-300 IU rFSH was administered from day 3 of the menstrual cycle.The rFSH dosage adjustments and GnRH antagonist administration were similar to the study group.
In both groups, 250 lg recombinant human chorionic gonadotropin (rhCG; Ovidrel, Merck Serono, Germany) was administered when the leading follicle reached a diameter of 18-20 mm or when the diameter of at least two follicles reached 17-18 mm.Ovum pick-up was performed 36 h after rhCG administration.Retrieved oocytes were fertilized by IVF or intracytoplasmic sperm injection (ICSI).Up to two embryos were transferred on day 3 after ovum pick-up.Goodquality embryos not transferred were vitrified.Fresh embryo transfer was cancelled if the patient had an unfavourable endometrium, progesterone level 1.5 ng/mL on the trigger day, or when no good quality embryos developed.At the treating clinician's decision, the endometrium was prepared through a natural or artificial cycle regimen for frozen embryo transfer.The indices used to evaluate ovarian sensitivity were FORT, FOI, and Gn dosage per follicle.

Statistical analysis
An independent samples t-test was used to compare continuous variables that were normally distributed, while the Kruskal-Wallis test was applied for variables with skewed distribution.Categorical data were assessed by the chi-squared test or Fisher's exact test.Cumulative ongoing pregnancy per cycle was assessed both crudely and using multivariate logistic regression analysis.The decision to add the potential confounders evaluated to the model was based on previous scientific evidence.
Fifty-five patients were treated in different cycles by both protocols.A sensitivity analysis was performed to assess if the results obtained with the full sample set were biased by the inclusion of multiple ovarian stimulation cycles in the same patient.
Statistical analysis was performed using the IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp, Armonk, NY, USA).All statistical tests were two-sided.Differences with a p-value <0.05 were considered statistically significant.

Results
Baseline patient characteristics and demographic data are summarized in Table 1.The antral follicle count (AFC) was lower, and the basal FSH higher in the study group.This suggests that the ovarian reserve may have been lower in these patients.As shown in Table 2, some of the ovarian stimulation characteristics were comparable.However, the total Gn dosage was considerably lower, and the Gn administration duration shorter in the study group.The endometrium was thinner on the trigger day, as expected from the anti-oestrogen effect of CC on the endometrium.Significantly less Gn was administered to obtain a preovulatory follicle, and FOI was much higher in the study group.Although the number of oocytes retrieved was similar between the groups, more available embryos were obtained in the study group.
Seventeen (8.8%) cycles in the study group and 47 (26.0%) in the control group were cancelled as no oocyte was retrieved nor embryos were available for transfer (p < 0.001, Table 3).At the time of statistical analysis, embryos remained cryopreserved without transfer for 25 cycles in the study group and 22 in the control group.The numbers of transferred cycles were 151 and 112 in the study and control groups, respectively.The crude reproductive outcomes were similar between the groups (Table 3).
Table 4 summarizes the results of a multivariate regression analysis.After adjusting for potential confounders, the results showed that cumulative ongoing pregnancy was still comparable between the   treatment groups (adjusted odds ratio: 0.761, 95% confidence interval: 0.300-1.933,p ¼ 0.566, Table 4).Age, BMI, Gn dosage per follicle, and FOI were negatively correlated with cumulative ongoing pregnancy.Details of the sensitivity analyses are presented in Supplemental Tables 1 and 2. With comparable numbers of retrieved oocytes and available embryos, patients were administered less Gn and the duration of Gn administration was shorter when stimulated by the CC priming antagonist protocol, suggesting a lower cost per treatment cycle.Gn dosage per follicle was considerably lower after the CC priming antagonist protocol, possibly suggesting improved ovarian sensitivity.The reproductive outcomes following the two protocols were similar.

Discussion
POR patients undergoing IVF/ICSI usually suffer from a limited number of oocytes, poor embryo quality, and a low pregnancy rate per cycle.Multiple stimulation protocols and various Gn and/or starting doses have been reported in attempts to address these problems (Err azuriz et al., 2019;Huang et al., 2019;Lainas et al., 2015;Lin et al., 2018) and although there is insufficient evidence to support their administration, supplements, such as growth hormone, dehydroepiandrosterone, coenzyme Q10, and multi-nutrients have all been used in an attempt to improve oocyte quality (Narkwichean et al., 2017;Nouri et al., 2017;Safdarian et al., 2019;Xu et al., 2018).To date, the most effective protocol for POR patients remains controversial, and the management of these patients is still a challenge.
It is increasingly believed that an efficient, patientfriendly regimen that can improve ovarian response and decrease the costs involved is needed for POR patients, such as oral ovulation induction medications, for example, CC.The present study aimed to compare ovarian response and clinical outcomes in POR patients treated by a CC priming protocol vs. a flexible GnRH antagonist protocol.
Many methods have been proposed to evaluate the ovarian response to gonadotropin stimulation.For example, FORT was first introduced in 2011 (Genro et al., 2011).However, this index does not assess the actual number of oocytes retrieved, which is strongly associated with live birth rates (Sunkara et al., 2011).Alternatively, the FOI was proposed (Alviggi et al., 2018) which, although seemed to be useful in evaluating ovarian sensitivity, had drawbacks.For example, technical aspects, such as triggering the final oocyte maturation and ovum pick-up were related to oocyte retrieval and could influence FOI results such that this index alone was not robust enough to assess the dynamical aspect of the ovarian response.In the present study, we also use a new biomarker; Gn dosage per follicle, as a form of outcome measure, which showed promising results in our unpublished data.The results indicated that the Gn dosage required to obtain a pre-ovulatory follicle was considerably lower in the study group.Growing evidence suggests that increased Gn stimulation cannot improve clinical outcomes, but increases the treatment costs for poor ovarian responders (van Tilborg et al., 2017).Thus, the lower Gn dosage per pre-ovulatory follicle in the study group suggested a better ovarian sensitivity.Taken together, we conclude that patients stimulated by the CC priming antagonist protocol achieved a better ovarian response.With poorer ovarian responsiveness, the number of available embryos per cycle was much less in the GnRH antagonist group.Moreover, more cycles were cancelled as no oocyte was retrieved or no embryos were available for transfer (Table 3).
Such an improved ovarian response could have several causes.First, through a negative feedback mechanism.CC may have occupied the hypothalamic oestrogen receptors for a longer period of time than oestrogen (Oktem et al., 2015), increasing GnRH release, and thus the endogenous gonadotropin levels.FSH could stimulate follicular development and synthesis of oestrogen.Luteinizing hormone (LH) was also essential for normal folliculogenesis and oocyte maturation.POR patients, in whom LH activity was usually insufficient, may have benefitted from the elevation of LH (Klinkert et al., 2005;Lehert et al., 2014).Secondly, CC has a relatively long half-life of 5-7 days.Residual CC continues to work for some time after its administration is terminated.In this case, CC increases the endogenous FSH and LH content over a long period, improving the ovarian response to the stimulation process.
Previous studies have reported a negative effect of CC on the endometrium (Mehdinejadiani et al., 2019).In the present study, endometrial thickness in the study group on the trigger day was significantly lower.Recent studies have suggested that under CC treatment, an endometrial thickness cut-off value of 8 mm at midcycle is associated with improved outcomes (Gaba et al., 2019).It is well known that timeto-pregnancy should be considered when making decisions related to infertility treatment, especially for POR patients.We, therefore, propose that when sufficient endometrial thickness is observed on the trigger day, the negative effect of CC during ovarian stimulation will be avoided, and a good pregnancy outcome may be achieved by fresh embryo transfer, resulting in a shorter time-to-pregnancy.
The study group had a significantly lower total Gn dosage and shorter Gn administration time.Such results imply lower costs and more convenient and patient-friendly treatment.These aspects are important, considering that POR patients often suffer from economic and time stresses due to repeated ovarian stimulation.
There were no differences in the reproductive outcomes between the two groups.The cumulative ongoing pregnancy rate was comparable even after adjusting for possible confounders.Considering the limited size of the present study and the limited number of embryos available for transfer per cycle in POR patients, the comparison of reproductive outcomes between the two protocols needs further research.
We performed a sensitivity analysis of 55 patients stimulated by both protocols in separate cycles as a comparative self-control study.Most stimulation markers and laboratory outcomes were in line with the data on the entire POR patient group.FORT, FOI, and the number of available embryos were higher in the study group, although the differences were statistically insignificant which could, at least partially, be due to the limited sample size available for the comparison.
Guidance on how to manage POR patients optimally is still lacking.Many clinicians may be keener on developing new ovarian stimulation protocols, but old protocols including drugs may still have new uses, such as the CC priming protocol in the present study.In all such research, it is essential to retain the confidence and co-operation of the patients, in shortening the time-to-pregnancy.
The study has some limitations arising from its retrospective nature and small sample size.Thus, any conclusions from the current study should be viewed with caution.Secondly, the patients were from a single centre with similar ethnicities, and large scale multicentre randomized clinical trials are warranted to confirm the utility of this CC priming antagonist protocol.

Table 1 .
Baseline patient characteristics and demographic data.

Table 2 .
Ovarian stimulation and laboratory outcomes.
FORT: follicular output rate; FOI: follicle-to-oocyte index.Note: Values are presented as mean ± standard deviation/median (min-max).p-Values were calculated using the chi-squared or Fisher's exact test for categorical data.An independent samples t-test was used to compare continuous variables that were normally distributed, while the Kruskal-Wallis test was applied for variables with skewed distribution.
p-Values were calculated using the chi-squared or Fisher's exact test.a For 25 cycles in CC priming antagonist group and 22 cycles in GnRH antagonist group, embryos remained cryopreserved without transfer until statistical analysis.

Table 4 .
OR for cumulative ongoing pregnancy by multivariate regression analysis.