The Effects of Three Methods of Luteal Phase Support on Pregnancy Outcomes in Poor Ovarian Responders: A Randomized Clinical Trial.

The requirement of luteal phase support (LPS) in assisted reproduction cycles is well established, but the issues that remain controversial are the ideal drug, the appropriate route of administration, and the timing and duration of support ( 1 ). A Cochrane systematic review and meta-analysis confirmed that LPS enhances the in vitro fertilization (IVF) pregnancy outcome ( 2 ). The administration of progesterone is standard for LPS, whether it is a controlled ovarian stimulation (COS) cycle or a frozen embryo transfer (FET) cycle. However, LPS does not have as many options as the individualized COS protocols and endometrium preparation protocols ( 3 ). The progesterone production in the luteal phase is a key component that is necessary for the successful implantation of the developing embryo. Different route of progesterone administration was used including oral, intramuscular, subcutaneous and vaginal. Recently, different types of vaginal progesterone (suppositories, gelsand tablets) are common choices for LPS; however, those may be associated with different side effects and discomfort in patients ( 4 ). The need for simplified treatment approaches to reduce the treatment burden of IVF is evident, and clinical physicians in assisted reproductive technology (ART) medicine often recommend drug therapy with lower doses and repeated use; therefore, clinical physicians in ART medicine usually recommend drug therapy with lower doses and repeated use. For example, the use of only two doses of human chorionic gonadotropin hormone (hCG) injection for luteal phase support is more comfortable and tolerable than the administration of progesterone daily, either vaginally or by injection for patients. A number of clinical trials have compared the effect of hCG and progesterone for luteal phase support; the results of a recent meta-analysis showed that there were no statistically significant differences among different luteal phase support methods in terms of clinical pregnancy, miscarriage and ongoing pregnancy rates; therefore, it seems that progesterone alone is the best strategy due to lower risk of ovarian hyperstimulation syndrome (OHSS) ( 2 ). Recently, the use of micro-dose of hCG without exogenous progesterone for luteal phase support has been considered and it is hypothesized that corpus luteum (CL) produces other hormones than estrogen and progesterone, which are essential for endometrial preparation and optimization of the environment for embryo implantation and development ( 5 ). Due to the fact that in patients with poor ovarian response (POR) diagnosis, there is no risk of OHSS and on the basis of the novel concept that suggested two boluses of 1500 IU hCG with no additional luteal support revert the luteolysis after a gonadotropin-releasing hormone agonist (GnRHa) trigger in the normo-responder patient ( 6 ), designing a clinical trial study to evaluate this simple and patients friendly method for luteal phase support in POR patients is gaining interest. Therefore, this clinical trial was designed to investigate and compare the efficiency of different methods of luteal phase support (progesterone alone or hCG alone and the combination of progesterone with hCG) in these patients. It is hoped that the results of this study will be useful in improving the clinical pregnancy rate and in making decisions for the best method of luteal phase support in these women.

A total of 524 women were evaluated to participate in the study. Thirty-eight patients did not consent to participate, and 111 patients did not reach to ET stage. Finally, 375 patients were allocated into three groups randomly. In intervention group A, 125 patients received two boluses of 1500 IU hCG, while in intervention group B, 125 patients received three boluses of 1500 IU hCG, in addition to daily vaginal progesterone. The control group C consisted of 125 patients who received only daily vaginal progesterone ( Fig .1 ). The analysis of data showed that the three groups were comparable and there was no significant difference in terms of age, body mass index (BMI), duration of infertility, and other basic characteristics of patients among the three groups ( Table 1 ). Flowchart of the study sampling. hCG; Human chorionic gonadotropin. The ovarian stimulation outcomes are presented in Table 2. No difference existed among the three groups with respect to the type of controlled ovarian hyperstimulation (COH) protocol, the total dose of gonadotropins and the duration of the ovarian stimulation, the number of retrieved oocytes, fertilization rate, number of good quality embryos transferred, and endometrial thickness on the day of ovum pickup (P>0.05 for all). Comparison of demographic and basic characteristics of studied women in three groups Data are presented as mean ± SD or n (%). P≤0.05 was considered statistically significant. hCG; Human chorionic gonadotropin, BMI; Body mass index, LH; Luteinizing hormone, FSH; Follicle stimulating hormone, AMH; Anti-müllerian hormone, POR; Poor ovarian response, Subgroup I; age> 40+a history POR or risk factor, Subgroup II; One history of POR+abnormal ovarian reserve tests, Subgroup III; age> 40+abnormal ovarian reserve tests or risk factor, Subgroup IV; A history of POR age >40+abnormal ovarian reserve tests, Subgroup V; Two previous history of POR, and *; It was obtained by independent sample t test and Chi square test as appropriate. Comparison of ovarian stimulation cycle results among three groups Data are presented as mean ± SD or n (%). P≤0.05 was considered statistically significant. hCG; Human chorionic gonadotropin, COH; Controlled ovarian hyperstimulation, rFSH; Recombinant follicle stimulating hormone, hMG; Human menopausal gonadotropins, *; It was obtained by independent t test, and € ; It was measured by the Chi-square test. The twin pregnancy rate (3.2%) in group A (hCG-only) was higher than those of the other two groups (1.6 and 0%), although this difference was not statistically significant (P=0.060). Table 3 shows pregnancy outcomes in three groups. The analysis demonstrated that the implantation, clinical pregnancy, miscarriage and live birth rates were similar among the groups (P>0.05 for all). Comparison of IVF- ICSI/ ET cycle outcomes among three groups Data are presented as mean ± SD or n (%). P≤0.05 was considered statistically significant. *; It was obtained by independent sample t test and Chi square test as appropriate and hCG; Human chorionic gonadotropins.

The findings of the present study responded to the question of whether changing the type of LPS in patients with POR diagnosis improves the rates of clinical pregnancy and live births. The results of the study showed that the type of luteal phase support did not affect the clinical pregnancy and live birth rates in POR patients; however, the rate of twin pregnancy in the hCG-alone group was slightly higher than in the progesterone-alone group. Fundamentally, LPS plays a vital role in IVF cycles because ovarian stimulation disrupts the specific functions of the CL through various mechanisms. In the formation of several mature ovarian follicles, estradiol (E 2 ) levels reach supraphysiological ranges specifically during the follicular phase; moreover, progesterone levels increase in the early days of the luteal phase due to numerous CL structures affected by triggering doses of hCG. Thus luteinizing hormone (LH) secretion decompressed during the luteal phase afterwards, CL support by pituitary LH will not precede ( 10 , 11 ). Notwithstanding the foregoing, it is important to realize that the CL does not require supraphysiologic levels of LH/hCG to secrete high values of progesterone. During the natural menstrual cycle, the LH level in the luteal phase rarely rises over 5-10 IU/L and most commonly is capable of eliciting progesterone levels up to 25-35 nmol/L. When ten CL structures exist, each will secrete progesterone in amounts similar to the natural menstrual cycle when exposed to physiologic concentrations of LH/hCG ( 12 ). Collectively, this results in a high concentration of progesterone. However, this condition occurs when the ovarian response is normal or excessive, and therefore in POR patients due to limited antral follicle count (AFC) and growing follicles and the lack of existence of suitable CL, the possibility of hormonal defects in the LPS is higher. Several protocols have been developed to correct this hormonal inadequacy in IVF cycles, based on the substances used or the route of administration. The development of new regimens for LPS based on ‘‘low-dose’’ or ‘‘microdose’’ hCG supplementation was considered recently. In preliminary studies, four injections of 1500 to 2500 IU of hCG were administrated after final oocyte triggering with 10000 IU of hCG ( 13 , 14 ); since then, several studies have examined different methods of luteal phase support using hCG alone or with progesterone in normal “responders” or “oocyte donors” ( 9 , 15 - 17 ). With these regimes, the LH/ hCG concentration in the luteal phase often reaches more than five to ten times that observed in the normal menstrual cycle which makes powerful luteal phase support; however, significantly increases the risk of OHSS ( 5 ). A recent Cochrane review study reported that the administration of hCG alone or in combination with progesterone to support the luteal phase should be avoided due to an increase in the risk of OHSS ( 2 ). Moreover, a number of studies have suggested that in cases where GnRH agonist is used for the final oocyte triggering, very low doses of hCG of 100 to 150 IU per day could be an appropriate option for the LPS ( 18 , 19 ). Progesterone levels in the middle of the luteal phase in this procedure were similar to the use of progesterone with 6500 IU of hCG to induce ovulation ( 12 ). Haas et al. ( 20 ) reported that having 1500 IU hCG 3 days after oocyte pick-up caused significantly higher progesterone levels compared with not having hCG after oocyte retrieval. The pregnancy rates were similar between groups, and no severe OHSS was reported. Castillo et al. ( 21 ) explained the effect of low-dose hCG administration periodically after GnRH-a trigger in women at high risk of OHSS. They were given 1000 IU, 500 IU, or 250 IU hCG every 3 days from oocyte pick-up day. The clinical pregnancy rate was 43.4% and the incidence of moderate and severe OHSS was 4.1 and 3.6%, respectively. A limitation of this strategy that hinders its application in routine procedures is the difficulty of diluting hCG over time with normal saline to achieve a dose of 100 to 150 units and the requirement for daily injections for the patient ( 5 ). Given the absence of the risk of OHSS in POR patients, the investigation of this new LPS regimen with low-dose hCG to improve pregnancy rates appears valuable. Furthermore, women appear to prefer low-dose hCG administration to exogenous P administration, especially by the vaginal route. In the present study, when comparing the twin pregnancy rates among groups, it was higher in the hCG-only group than in the P-only and hCG+P groups, although it approached statistical significance. In this regard, Var et al. reported that multiple pregnancy rates were significantly higher in the hCG+P group than in the P-only and E 2 +P groups ( 9 ). Similar to their findings, Ludwig et al. ( 15 ) also found that multiple pregnancy rates were higher in the hCG+P group than in the P-only group. However, Ghanem and colleagues reported no statistical difference among the groups in terms of multiple pregnancies; they determined a higher tendency in the E 2 +P and hCG+P groups compared with the P-only group ( 22 ). The use of hCG for LPS is likely involved in increasing the rate of multiple pregnancies, and more studies are necessary in this regard. The main limitation of the study is that we did not consider the good quality euploid blastocysts and unique COH protocol as inclusion criteria. Although in data analysis the number and quality of transferred embryos and the proportions of two COH protocols were similar among groups, it is recommended that these points be considered in future studies to eliminate these confounding factors. On the other hand, it would have been better if the same COH protocol was performed for all patients, but due to the conditions of the patients, a single protocol was not chosen, of course, the percentage of different protocols in the three groups was not statistically different and the three groups were comparable, it is suggested that this issue be considered in future studies. In addition, there is another limitation point that sperm quality was not taken into account in this study, although the percentage of patients with the etiology of male factor infertility was similar among groups, it is recommended to be investigated in this regard in subsequent studies.

The present study is the first clinical trial to compare the effect of low-dose hCG with the routine protocol for LPS in patients with POR diagnosis on the basis of the Bologna criteria. We found the same pregnancy and live birth rates in different regimes for LPS. Interestingly, administration of two bolus of low-dose hCG 1500 was associated with a slight increase in the rate of multiple pregnancies. In the cases of contraindication to the use of progesterone or sensitivity to the usage of vaginal drugs, we recommend this effective method which is easier and patient-friendly as an alternative treatment.

This randomized clinical trial was carried out to investigate the efficacy of different luteal phase support methods in patients with POR undergoing intracytoplasmic sperm injection (ICSI) cycles at the Royan Institute between November 2015 and June 2019. The study protocol is approved by the Institutional Review Board and Ethics Committee of Royan Institute (IR.ACECR.ROYAN.REC.1394.121). The study is conducted according to the Declaration of Helsinki for medical research. All participants provided informed consent after receiving an explanation of the purpose of the study. The trial protocol was registered prospectively in the Clinicaltrials.gov site ( NCT02798653 ). All the patients who were diagnosed as POR based on the Bologna criteria ( 7 ) were eligible for participation in this study. In order to define the poor response in IVF, at least two of the following three features must be present: i. Advanced maternal age (over 40 years), ii. A previous POR (total retrieved oocytes less than three oocytes using conventional protocols), and iii. An abnormal ovarian reserve test (ORT) (antral follicle count of less than 5 on menstrual cycle day 2-3, and/or serum anti-Müllerian hormone level less than 1 ng/ml). Two episodes of POR after maximal stimulation are sufficient to define a patient as a poor responder in the absence of advanced maternal age or abnormal ORT. The advanced maternal age over 45 years old, cigarette, alcohol and drug addiction, diagnosis of ovarian failure including basal follicle stimulating hormone (FSH) above 20 IU/l or no antral follicle by ultrasound examination, severe male factor (azoospermia) diagnosis, severe endometriosis and the presence of hydrosalpinges, uterine factor (polyps, myoma and previous myomectomy,…), the patients with cardiovascular disease and/or uncontrolled systemic or endocrine diseases and repeated implantation failures and repeated miscarriages cases were excluded from the study. The ovarian stimulation is performed with “Stop GnRHagonist” or “conventional GnRH-antagonist” protocols for the eligible patients as explained in detail previously ( 8 ). In both protocols, COS was started on day 2 of the menstrual cycle with 225 IU recombinant FSH (Gonal-F®; Serono Laboratories Ltd., Geneva, Switzerland) and 75 IU human menopausal gonadotropin (hMG, Menopur®; Ferring). The doses of gonadotropins were adjusted as ovarian response in the ultrasound monitoring and final oocyte triggering was done with 10000 IU of hCG (Choriomon®; IBSA). If there were one or more dominant follicles, oocyte retrieval was performed under transvaginal ultrasound guidance 32-34 hours after hCG administration. On the day of oocyte pickup, patients were allocated randomly (by the blocked randomization method) into three groups to receive three different luteal support protocols. Permuted block (a block size of 6) randomization was prepared by the methodological advisor according to a computer-generated list. The patients’ enrolment and assignment to different groups were carried out by a researcher midwife in the clinic. Each patient participated in the study only once and if she had written consent. The researcher who followed the results of patients’ treatment and the researcher who analyzed the data were uninformed regarding the type of LPS regimen. In the first group (A), 1500 IU of hCG IM on the embryo transfer (ET) day, as well as 4 days after that were administrated for luteal phase support. In the second group (B), the patients received 1500 IU of hCG IM on the ET day, as well as 3 and 6 days after the ET along with vaginal progestrone suppositories (Cyclogest ®, Actavis, UK) 400 mg twice daily. For the third group (C), only vaginal progesterone suppositories (Cyclogest ®, Actavis, UK) 400 mg twice daily were given for luteal support from the day of oocyte pick up until the pregnancy test day. ICSI was performed for all metaphase II oocytes. Embryos were cultured in a commercially available culture medium until the day of transfer. The quality of embryos was graded from 1 to 3 under inverted microscope 3 days after the IVF/ICSI procedure. Embryos with even-sized blastomeres and/or ≤10% fragments were classified as grade 1 (excellent or good quality [A and AB]). Grade 2 embryos (moderate or fair quality [B and BC]) had blastomeres with slightlymoderate size differences and/or 10-20% fragments. Grade 3 embryos (poor quality [C and CD]) had markedly different- sized blastomeres and/or >20% fragments. According to the standard, endometrial thickness on ovum pick day equal to or more than 7 mm with a three-line view was a necessary condition for ET. The obtained embryos at the cleavage stage were transferred by an ET catheter (Guardia ™, Access ET Catheter, Cook Medical), three days after oocyte retrieval. The serum ß-hCG level was checked 2 weeks after ET to confirm a positive pregnancy test. The vaginal progesterone was administrated in all patients who become pregnant in all three groups until the 10th week of pregnancy. The pregnancy outcomes are compared among the three study groups. The primary outcome was implantation (the number of embryos transferred for each patient), chemical pregnancy (only positive ß-hCG test) and clinical pregnancy (the presence of a gestational sac with fetal heart beat on vaginal ultrasound) rates. The secondary outcomes included early miscarriage (the spontaneous loss of a clinical pregnancy under 12 weeks of gestation) late miscarriage (the spontaneous loss of a clinical pregnancy between 12 and 20 weeks of gestation) and live birth (the delivery of a living child irrespective of the duration of pregnancy) rates. The sample size was calculated on the basis of a previous related study ( 9 ) by using the Power Analysis and Sample Size (PASS) software version 11 (NCSS, LLC. Kaysville, Utah, USA). A sample size of 125 patients was required in each group, considering an effect size of 0.163 for implantation rate and α error of 0.05, with a power of 80%. Statistical analysis was done using the Statistical Package for Social Sciences (SPSS) version 23 (IBM Corp., Armonk, NY, USA). The continuous variables were compared among groups by the one-way analysis of variance (ANOVA) and were presented as mean ± standard deviation (SD). The chi-square test was applied for comparing the categorical variables among groups and the results were reported as numbers/percentages. The statistical significance level was set at P<0.05.