Retrospective study of influencing factors on the outcomes of luteal phase stimulation in patients with dual stimulation

Study participants and data collection

From 8,460 cycles of fresh oocyte retrieval in Reproductive Medicine Center of the First Affiliated Hospital of Naval Medical University from January 2017 to March 2022, we screened out 78 groups of DS cases within one menstrual cycle. All patients had signed an Informed Consent Form for Dual Stimulation of Ovulation Induction Cycles.

These 78 cases were grouped according to the ovarian stimulation protocols during the follicular phase. This study analyzed the outcomes of two retrievals from each of these 78 cases, comparing the number of oocytes retrieved, and the number and ratio of available and high-quality embryos among subgroups.

Ovarian stimulation protocol

During the follicular phase, we performed a total of six ovarian stimulation options: GnRH-agonist short protocol (short protocol), GnRH-agonist long protocol (long protocol), GnRH-antagonist protocol (antagonist protocol), Progestin-primed ovarian stimulation protocol (PPOS protocol), mild-stimulation protocol and natural-cycle protocol. These protocols were performed as previously described (Schimberni et al., 2016; Xi et al., 2020; Xiao et al., 2019).

Briefly, for patients with short protocol, they received short-acting Triptorelin (Decapeptyl; Ipsen, Rome, Italy) 0.05 mg/day starting on day 1 of menstruation and subcutaneously injected with recombinant follicle stimulating hormone (rFSH) (Gonal-f; Merck Serono, Weiterstadt, Germany) or intramuscular injected with human menopausal gonadotropin (hMG) (Lizhu Pharmaceutical Factory, Zhuhai, China) 150–300 IU/day from day 2–3 of the menstruation until the hCG injection.

For patients with long protocol, they received pituitary down-regulation with a standard full dose (3.75 mg) of leuprorelin acetate microspheres (Lizhu Pharmaceutical Trading Co, Zhuhai, China) in the early follicular phase, usually day 3 of the menstruation. Patients came back 1 month later to check whether they had met the criteria for adequate suppression. The criteria for defining adequate pituitary–ovarian suppression were: E2 < 40 pg/mL, Prog < 1 ng/mL, LH < 3 IU/L, endometrial thickness <4 mm, and the absence of any growing follicle >6 mm. If the patient met the above criteria, they started controlled ovarian hyperstimulation (COH), which was performed by using rFSH or hMG at a dose of 150–300 IU/day.

Patients in the antagonist group got COH from day 2–3 of the menstruation using rFSH or hMG. When follicles with a mean diameter ≥14 mm were detected, a GnRH-antagonist (Cetrorelix; Pierre Fabre Medicament Production, Cahors, France) was injected subcutaneously at dose of 0.125–0.25 mg/day.

Patients in PPOS group were given rFSH or hMG at a dose of 150–225 IU/day and medroxyprogesterone acetate (MPA) (Zhejiang Xianju Pharmaceutical Co., Zhejiang, China) at a dose of 10 mg/day from day 3 of menstruation based on the results of the ultrasound and the blood tests.

For patients in mild-stimulation group, Clomiphene citrate (CC, Fadilan, Gaote Pharmaceuticals, Larnaca, Cyprus) was administered orally at 25–50 mg/day for 3–5 days, and some patients also injected with hMG every other day or every day until trigger day.

Follicular monitoring started 5 days later and was conducted every 2 to 3 days using a transvaginal ultrasound examination to record the number of developing follicles. Serum FSH, LH, E2, and Prog concentrations were measured using patient blood tests on the same days as the ultrasound exams. The start dose of rFSH or hMG was based on the patients age, hormone profile, AFC, body mass index (BMI) and previous treatment, and then adjusted according to E2 concentrations and ovarian responses.

For patients with natural cycle protocol, no Gn were used throughout the process, only trigger drugs were used. From day 8–10 of menstruation, transvaginal ultrasound was performed to monitor the growth of follicles.

When three dominant follicles reached 18 mm in diameter, the final stage of oocyte maturation was triggered by 0.25 mg rHCG (Ovidrel; Merck Serono, Weiterstadt, Germany) or 2000IU HCG (Lizhu Pharmaceutical Trading Co, Zhuhai, China). Transvaginal ultrasound-guided oocyte retrieval was carried out 34–36 h after the trigger.

Follicles ≤12 mm in diameter were left for the next retrieval. LPS was initiated 1–3 days after the follicular oocyte retrieval, except that GnRH-antagonist was not used to prevent LH surge.

IVF laboratory work

According to our standard protocol, all retrieved oocytes were fertilized with either conventional IVF or ICSI. Embryos were cultured individually and scored on the third day (D3). Embryos were graded according to ASEBIR embryo assessment criteria (Alpha Scientists in Reproductive Medicine & ESHRE Special Interest Group of Embryology, 2011). All high-quality cleavage embryos were cryopreserved on D3 and the other available embryos were cultured for blastocyst until the 5^th/6^th day (D5/D6). All available embryos including D3/D5/D6 from both FPS and LPS were cryopreserved with a commercial vitrification kit (Kitazato Biopharma Co., Shizuoka, Japan) on Cryotops. High-quality cleavage embryo was defined as normal fertilized embryos if they had 7–9 cleavages with less than 20% fragmentation. Available cleavage embryos on D3 were defined as normal fertilized embryos if they had five or more blastomeres and fragmentation <50%. The available blastocyst was defined as containing at most one “C” of scores for inner cell mass or trophoblast at stage 3 and above. High-quality blastocyst was defined as containing no “C” of scores for inner cell mass or trophoblast at stage 3 and above.

Statistical analyses

The primary outcome was the number of oocytes retrieved. The secondary outcomes included the number of available and high-quality embryos, and the rate of available and high-quality embryos.

All statistical analyses were performed using IBM^® SPSS^® Statistics version 26.0 software (SPSS Inc. Headquarters, Chicago, IL, USA). Statistical comparisons were carried out using one-way ANOVA for the normally distributed data, and the Mann-Whitney U test was used for non-normally distributed data. If the variances are homogeneous, LSD method was used, and if the variances are unequal, the Tamhane method was used. For correlation studies, correlation analysis and multiple linear regression were used and the linear regression equations were given out. P < 0.05 was considered as statistically significant.

Ethics

The study was approved by Reproductive Ethics Committee of the First Affiliated Hospital of Naval Medical University. The Ethical approval number is CHREC-2016-3. Informed Consent Form for Dual Stimulation of Ovulation Induction Cycles is obtained from the participants.

Characteristics of the study participants and comparisons among subgroups

Among 8,460 fresh IVF cycles in our center, 78 cases performed DS within the same menstrual cycle. These 156 cycles counted for 1.84 percent of all our fresh cycles in the past 5 years. We did not find triple or more stimulations and oocyte retrievals within the same menstrual cycle (Gemmell, Wright & Brady, 2020). For these DS cases, the first oocyte retrieval was in follicular phase and the second was in luteal phase.

During the follicular phase, patients received different ovarian stimulation protocols, including short protocol, long protocol, antagonist protocol, PPOS protocol, mild-stimulation protocol, and natural-cycle protocol. During the luteal phase, patients received luteal phase stimulation protocol.

Table 1 showed the number of cases and distribution of various FPS protocols in each year. For different groups in FPS, nine cases performed short protocol, five cases performed long protocol, 26 cases performed antagonist protocol, 26 cases performed PPOS protocol, seven cases performed mild-stimulation protocol, and five cases performed natural-cycle protocol (Table 1).

Oocyte and embryo quality comparison between FPS and LPS in dual stimulation

All the patients in this study were for assisted reproduction. After retrieval, all the oocytes were fertilized and embryo cultured. Since the fertilization types in this study included IVF and ICSI, in terms of oocyte outcomes, we only focused on the number of oocytes retrieved, instead of the maturity of the oocytes. In addition to the number of oocytes retrieved, we paid more attention to embryo quality, including the number and the ratio of available embryos and high-quality embryos.

Previous studies reported that DS doubled the number of oocytes when compared with FPS-only and the number of oocytes retrieved in LPS was significantly higher compared to FPS, although LPS yielded a lower rate of metaphase II oocytes (Polat et al., 2021; Zhang et al., 2018). Here we further evaluated the number and the rate of available and high-quality embryos.

When comparing the outcomes of LPS with FPS, we found that the number of oocytes retrieved, the number of available embryos, and the number of high-quality embryos were slightly increased. But the rate of available embryos significantly decreased, while the rate of high-quality embryos slightly decreased in LPS (Table 2).

Due to the lower rate of available embryos, we still recommend that the number of oocytes retrieving should be maximized during follicular phase to maximize the number of available embryos rather than simply rely on LPS when the patients perform DS regimen.

Comparison of different FPS protocols on LPS outcomes of DS

Firstly, whether there was a difference in the age of the women between FPS groups was studied. Among the six FPS groups, only the women in the antagonist group were significantly younger (−4.3 years) than those in the PPOS group, which was consistent with the choice of the routine COH programs in our center: the elderly patients are more inclined to choose the PPOS protocol due to the decreased ovarian reserve (Table 3).

Regarding the interval between two retrievals, the interval in the short and the long protocol group was significantly shorter than that in the antagonist group and the natural-cycle group, and there was no significant difference between other groups (Table 3).

When comparing Gn priming dose in LPS, it was found that Gn priming dose in LPS after the short protocol and the long protocol was significantly lower than that in the PPOS protocol, respectively (Table 3).

When comparing the duration of COH in LPS, it was found that the COH duration in LPS after the short protocol and the long protocol was significantly lower than that of the antagonist protocol and the natural-cycle protocol (Table 3).

When comparing the total amount of Gn used in LPS, it was found that the total amount of Gn in LPS after the short protocol and the long protocol was significantly lower than that of the antagonist protocol, the PPOS protocol and the natural-cycle protocol, respectively. This was consistent with the shorter stimulation duration in these two groups (Table 3).

When comparing the number of oocytes retrieved among each group, there was no difference in either FPS or LPS (Table 4).

We further compared the number of available embryos, and the results showed that there was no difference between the groups in FPS. However, in LPS, the number of available embryos in short protocol was slightly lower (−1) than that in antagonist protocol. In addition, there was also no difference in the rate of available embryos between groups in FPS or LPS (Table 4).

For the number and the rate of high-quality embryos, compared with the natural-cycle group, the antagonist group and the PPOS group showed better results in follicular phase. However, in the luteal phase there was no significant difference between the groups (Table 4).

According to the above comparisons of LPS outcomes among various FPS groups, when considering DS, antagonist or PPOS protocol was more recommended according to the patients’ condition in the first cycle. For younger patients with better ovarian reserve, antagonist protocol was the first choice. For patients with reduced ovarian reserve, PPOS protocol was recommended as FPS program. However, we did not recommend natural-cycle protocol in FPS when patient consider the second retrieval in the same menstrual cycle.

Study on the influencing factors of LPS outcome in DS

After comparing the LPS outcomes among different FPS protocols, we also evaluated the effects of female age, LPS-Gn duration, LPS-total Gn dose, and two retrievals interval on LPS outcomes.

A total of nine DS cases in short protocol group. Only a 44-year-old patient who stimulated with Gn for 6 days (with an interval of 8 days between two retrievals) obtained one available and high-quality embryo in the second cycle. This patient was also the one who received the longest LPS in this group.

A total of five DS cases in long protocol group, and two obtained available and high-quality embryos in luteal phase. One was a 28-year-old woman, who got 16 oocytes by LPS, resulting in four available and high-quality embryos. Another 44-year-old woman got one available and high-quality embryo, and this patient also received the longest stimulation (5 days) in this group.

Similar conclusions were found in antagonist protocol, PPOS protocol, mild-stimulation protocol, and natural-cycle protocol groups. The number of oocytes retrieved after LPS was negatively correlated with the woman’s age, and positively correlated with the duration and the total amount of Gn in LPS. The number and the rate of available and high-quality embryos was positively correlated with the interval between two retrievals, the duration and the total amount of Gn in LPS, and the strongest correlation was the duration of LPS (Table 5).

Next, linear regression on LPS outcomes and influencing factors was performed. Firstly, regardless of the FPS protocols, the outcomes were regressed on the age of the woman, the interval between two retrievals, the duration of stimulation and the total amount of Gn (Table 6).

The regression equation for the number of oocytes retrieved and the female age (years old) and total Gn dose (IU) in LPS is:

$$\mathrm{O}\mathrm{o}\mathrm{c}\mathrm{y}\mathrm{t}\mathrm{e}\mathrm{s}=6.525-0.14\ast \mathrm{a}\mathrm{g}\mathrm{e}+0.001\ast \mathrm{G}\mathrm{n}$$

The regression equations for the number of available and high-quality embryos (LPS-AE/LPS-HQE) and the duration of Gn administration (GnD) in LPS were:

$$\mathrm{L}\mathrm{P}\mathrm{S}‐\mathrm{A}\mathrm{E}=0.072+0.145\ast \mathrm{G}\mathrm{n}\mathrm{D}$$

$$\mathrm{L}\mathrm{P}\mathrm{S}‐\mathrm{H}\mathrm{Q}\mathrm{E}=0.048+0.114\ast \mathrm{G}\mathrm{n}\mathrm{D}$$

When including FPS regimens, linear regression on LPS oocytes retrieved and influencing factors showed that the long protocol had significant protective effects compared with the antagonist protocol, the total amount of LPS-Gn was a protective factor while the age of the woman was a risk factor. As to the number of available and high-quality embryos, Gn duration of LPS was the only protective factor, and the effect of FPS protocol selection was not obvious (Table 7).

The above regression analysis showed that for LPS in DS, sufficient duration of stimulation should be ensured to maximize available and high-quality embryo acquisition.

Analysis of DS pregnancy outcomes

Due to the limited sample size, we preliminarily analyzed pregnancy outcomes in the samples of this study. There were no successful pregnancy cases in the groups of short, long, and natural-cycle protocols. In the mild-stimulation group, FPS and LPS each obtained one clinical pregnancy, but early spontaneous abortion happened in the case of LPS. In the antagonist group, there were nine transplants in FPS and LPS respectively, including four and two live births, two and four cases of early spontaneous abortion, and three cases of non-pregnancy respectively. In the PPOS group, there were 12 and seven transplants in FPS and LPS respectively, including four and one live birth, one case of early spontaneous abortion respectively, and seven and five cases of non-pregnancy (Table 8). Overall, similar to embryonic outcomes, pregnancy outcomes of FPS were better than LPS in DS. However, LPS can still lead to successful live births, and the success cases were all from antagonist and PPOS groups. Therefore, based on the preliminary comparisons of clinical live birth outcomes, we still recommend the use of antagonist and PPOS protocols for DS.