This study categorized the patients into three separate groups differentiated by the method used for final oocyte maturation. Based on their own expertise, the clinicians would choose the most optimal regime for each patient by evaluating her baseline characteristics and responses to treatment. According to their experiences, hCG injections would be employed for patients with low E2 levels (usually <1500 pg/mL) on the day of triggering. On the other hand, if patients have high E2 (usually >3500 pg/mL) on the day of triggering and risks for developing OHSS[22], GnRH-agonist would be used. GnRH agonist trigger for final oocyte maturation significantly reduces the risk of ovarian hyperstimulation syndrome (OHSS) in in vitro fertilization (IVF) cycles [11].
For patients falling within the middle of the spectrum, dual trigger was the main modality used, with a few exceptions that would be explained later. Therefore, upon initial assessment, it could be expected that the baseline parameters of the patients among the three groups were distinctive. However, after redefining the inclusion criteria and comparing the patients by evaluating the hCG versus dual trigger group and GnRH agonist versus dual trigger group, the baseline characteristics of the study population showed no statistical differences.
After eliminating the numerous confounding factors associated with different baseline characteristics, we could further analyze the data by the two new established categories: “hCG versus dual trigger” and “GnRH agonist versus dual trigger”. In the “hCG versus dual trigger group”, if dual trigger was prescribed for POR patients with E2 <1000 pg/mL on the day of triggering, statistically higher rate of oocyte maturation was found. This observation was also previously described in GnRH-antagonist cycles, where the authors believed hCG and GnRH-agonist together could incite more oocyte maturation [6, 14]. GnRH-agonist has the ability to stimulate excretion of endogenous FSH and LH. LH has long been established as an important hormone for inciting final maturation of the oocytes; nonetheless, recent studies have also demonstrated the role of FSH in in vitro maturation of oocytes while animal studies have revealed the ability of FSH to induce ovulation, independently of the LH surge. It is theorized that FSH surge prompts the formation of more LH receptor on the luteinized granulosa cells, which then promotes the maturation of oocytes and expansion of cumulus cells [23–26].
For the “GnRH-agonist versus dual trigger group”, if patients had E2>6500 pg/mL on the day of triggering, use of either method for triggering did not show any statistically differences in the outcomes. Presumably, for these hyper-responders or patients with PPOS, utilizing hCG with higher efficacy or affinity would not make a significant difference in the final outcome. Therefore, for these patients, physicians should focus on lowering the risks of developing OHSS or other complications instead. In our study, none of the patients were found to have OHSS.
The PPOS protocol used in this study included the standard method (on menstrual day 3 till day of triggering )[6] and the flexible method (on menstrual cycle day 5 to day 7 or E2>200 or follicle >10mm till day of triggering ) [27]. Current studies have not concluded on the prognostic effects of either method. A comparison of flexible PPOS with GnRH antagonist protocol in women who donated oocytes showed no significant differences in the final outcomes [27].
The choice of progestin supplementation used for our PPOS protocols included Medroxyprogesterone 10mg once per day and Dydrogesterone 10 mg twice per day, with neither showing a significant advantage or disadvantage in the final outcomes in current studies [7, 28, 29].
With molecular structure more similar to the natural progesterone hormone, Dydrogesterone is widely used for hormone replacement, therapy, menstrual disorder treatment, endometriosis treatment, luteal support in pregnancy and threaten abortion [28]. However, use of Medroxyprogesterone is contraindicated in pregnancy [30] and breast cancer [31]. Dydrogesterone, meanwhile, appears to have fewer side effects and can be used for pregnant patients and those with history of breast cancer. However, its pricing is generally higher, and some studies have mentioned a higher rate of premature LH surge associated with its usage [29].
Researches and analysis focused on comparing live births resulted from PPOS protocols with those of GnRH-agonist have not revealed higher rates of congenital malformation, preterm labor, low body weight, and others [32].
This is the first retrospective study discussing whether different methods for triggering follicular maturation could produce different outcomes. We hypothesized that with the higher concentration of progesterone required for PPOS cycles, it could have various effects on the agents used for triggering. Hence, different agents used for follicular maturation could possibly produce distinctive outcomes from those obtained from GnRH-antagonist cycle. Surprisingly, for POR patients, similar results were seen while not much comparisons could be made for hyper-responders due to lack of published research so far.
A major limitation to our study is the non-randomized grouping of the patients. Since all of the patients were categorized into the three study groups based on the physicians’ clinical experiences, this created many differences in baseline parameters and other characteristics. For instance, most patients in the hCG group had limited number of embryos, so the retrieved follicles were all cryopreserved on Day 2 or day 3. We would not observe any blastocyst stage for that group of patients. In addition, during PPOS cycles, there were no standardization on the timing and type of medications prescribed, which could potentially affect the final results. Thirdly, due to the chemical structure of dydrogesterone, we could not accurately measure and monitor the levels of progesterone in the blood with the current diagnostic tests when using PPOS cycles [33]. Such data would have been helpful to include in the results. Lastly, longer study duration could have provided even more accurate analysis regarding the clinical pregnancy and live birth rates.