This study evaluated the effects of combining 150 IU of menotropin with follitropin delta on ovarian response during IVF stimulation in patients with serum AMH < 2.1 ng/mL. The administration of follitropin delta combined with menotropin has been reported in controlled analyses [21] and uncontrolled real-life studies [22]. However, to the best of our knowledge, this is the first controlled study to test follitropin delta combined with menotropin at these doses for this specific group of patients. Serum estradiol levels were significantly higher in the intervention group (2150 pg/mL vs. 1372 pg/mL, p < 0.001), indicating a higher ovarian response with menotropin combined with follitropin delta compared to the administration of follitropin delta alone. Furthermore, although the higher number of eggs in the study group did not reach statistical significance, this group had significantly more blastocysts and more good quality blastocysts to transfer, which, in the end, is more important to IVF success.

We used phase 1 and 2 studies with follitropin delta to develop this study, estimating that 12 μg of the drug would be equivalent to 150–200 IU [23,24,25]. This inference was confirmed a posteriori in a dose equivalence study published in 2020, which showed that 10 μg of follitropin delta is equivalent to approximately 150 IU of FSH [8]. Therefore, the maximum dose of follitropin delta for a first IVF cycle determined by the algorithm, 12 μg, is equivalent to approximately 180 IU of FSH daily, which we consider to be an underdose for patients at risk of low response. A systematic review from Cochrane Library, last updated in 2018, reported that using doses of 300–450 IU of recombinant FSH instead of 150 IU resulted in a higher number of eggs for poor responders [9]. Thus, it seems very plausible from a biological point of view that the improvement in ovarian response we observed may have occurred due to the addition of 150 IU of FSH activity for this group of patients, amounting to approximately 300–350 IU of FSH daily in the intervention group compared to 180 IU in the control group.

We believe that a group of patients with a very low ovarian reserve could benefit from using lower gonadotropins doses, so-called minimal and mild stimulation, or even modified natural and natural cycles extensively discussed in the literature under a well-known “less is more” concept [26,27,28,29]. In our study, 60% of women with AMH < 0.5 ng/mL had their cycle canceled due to lack of follicular growth, absence of mature eggs or embryos available for transfer. Despite being a small subgroup of five patients, we did not rule out the possibility they could benefit from low doses of gonadotropins stimulation or natural cycle if they choose to use their own eggs for IVF. This strategy can reduce treatment burden, side effects, and costs. The hypothesis that this very low ovarian reserve patients do not benefit by increasing gonadotropins dose was obviously not tested in this study, but future research could examine it.

A study by Bosch et al. published in 2019 reported that using higher doses of follitropin delta in the second and third IVF cycles (average daily doses of 12.0 ± 3.6 μg and 14.6 ± 5.1 μg, respectively) improved ovarian response in patients with low response in the first cycle, clearly indicating that there is room to administer doses > 12 μg of follitropin delta (or 180 IU of FSH) in patients at risk of low response [30]. We could simply have given higher doses of follitropin delta to patients with AMH < 2.1 ng/mL. However, we believe that the benefit of combining menotropin for ovarian response is not only in function of FSH activity but also because this drug contains LH acitivity, which would be an advantage over merely increasing the dose of follitropin delta. The addition of LH to FSH has been shown to improve ovarian response in women with low ovarian reserve [13, 31] by a mechanism that seems to involve an LH-induced increase in FSH responsiveness [32].

Our study also found improved fertilization rates, blastocysts number, and top-quality blastocysts number in the menotropin group which may have been a result of the addition of LH activity to the stimulation protocol. A Canadian study by Bissonnete et al., published in 2021, found that follitropin delta combined with menotropin increased blastocyst number and quality [21]. That study, unlike ours, included patients from the general population with no restrictions on AMH levels. In addition, the dose of menotropin was variable according to the dose of follitropin delta and the patient’s weight. Previous studies have shown a qualitative benefit of adding LH to the ovarian response for patients with low ovarian reserve (AMH < 1.2 ng/mL), and in those > 35 years of age [11, 12, 31]. In both, the study by Bissonnete et al. and ours, the improvement in the quantity and quality of blastocysts obtained may have been due to the addition of the LH contained in menotropin. In the Canadian study, 50% of the patients were > 35 years old. In our study, 47 and 54% of the women had AMH < 1.2 ng/mL and were over 35 years old, respectively. Although this beneficial effect of LH is plausible, a difference in performance between the IVF laboratories in the multicenter phase 3 study and the one in this study, as well as the use of ICSI in 100% of cases in the intervention group cannot be ruled out as the cause of these better fertilization rate, blastulation rate and better blastocyst quality observed.

Regarding ovarian response, we found a statistically significant increase in measures to prevent hyperstimulation, in the intervention group, such as GnRHa trigger and freeze-all. Administering a higher dose of gonadotropins, a known risk factor for high ovarian responses and OHSS, seems to have caused this response [33,34,35]. All the cases with a high (> 14 eggs) or very high (≥20 eggs) ovarian response had AMH levels > 1.0 ng/mL and ≥ 1.5 ng/mL, respectively. So, consider adding lower daily doses of menotropin, such as 112.5 IU or 75 IU, to patients with AMH levels between 1.0 and 2.1 ng/mL, may be a good alternative in a scenario where fresh embryo transfer is a priority. In our country, however, the use of GnRHa trigger and freeze-all are frequent. The freezing of all embryos is routine, not only for the prevention of OHSS, but also to allow preimplantation genetic test as well as a strategy to increase endometrial receptivity and reduce obstetric complications, as previously demonstrated in the literature [36,37,38]. Therefore, in the trade-off between having a fresh transfer or the highest possible number of eggs, the latter is more desirable in our daily practice. The intervention group had more blastocysts (3.10 vs. 2.42, p = 0.030) and better-quality blastocysts (1.80 vs. 1.37, p = 0.017) available for transfer, but this fact did not increase pregnancy rates after the first embryo transfer. However, we cannot rule out a long-term change in cumulative pregnancy rates after two or more embryo transfers.

Our study was not designed to demonstrate possible differences in pregnancy, implantation, miscarriage and live birth outcomes. The only gestational outcome with a difference was twin pregnancies (16.7% vs. 0, p = 0.014), due to the greater number of couples who opted for double embryo transfer in the intervention group (25% vs. 5.05%, p < 0.001). We understand those who advocate the use of a single live birth at term as the ideal primary outcome from the patient’s point of view in assisted reproduction studies [39, 40], but this does not seem to us to be the appropriate or even feasible endpoint for studies of ovarian stimulation interventions. There are plausible arguments for this: (i) between ovarian stimulation and the birth of the baby there are many variables that in no way depend on stimulation, such as seminal quality, IVF laboratory conditions, the couple’s intention to transfer two embryos or do PGT and even prenatal care [41]; (ii) the number of individuals to be allocated per group to demonstrate differences in the live-birth outcome in a low-reserve population (e.g., AFC < 10) can be as high as 2000 [41]. The multicenter phase 3 trial that served as the control group for this study, for example, had 297 patients with AMH < 2.1 ng/mL. Since robust evidence indicates that retrieving more eggs after ovarian stimulation increases pregnancy rates, we believe that using ovarian response endpoints, such as the number of eggs is a valid and appropriate strategy to avoid these difficulties. Sunkara et al. studied more than 400,000 IVF cycles with fresh embryo transfers and found that the more eggs in the range between 1 and 15 were obtained, regardless of age, the greater the chances of a live birth [42]. According to the findings of Law et al., with > 220,000 IVF cycles, and Polyzos et al., with approximately 15,000 IVF cycles, considering all the embryo transfers from the same stimulation and not just the fresh transfer, the more eggs were obtained, regardless of the number, the greater the chances of having a live birth, in all the age groups [43, 44].

Assessing the cost-effectiveness of any medical intervention is a crucial point to discuss. Considering Brazilian scenario, we estimated that the additional cost per patient (medication plus endometrial preparation due to the freeze-all strategy) was approximately 1167 euros in the intervention group. If we consider that cases with surplus frozen embryos were more frequent in the intervention group (61.9% versus 51.9% in control arm), there could be a cost reduction of around 282 euros per patient due to IVF cycles that would not be needed in case of not reaching the pregnancy after the first transfer. Consequently, the inclusion of only direct costs would have resulted in an additional expenditure of approximately 885 euros per patient. However, one cannot ignore the great physical and emotional burden of carrying out a new IVF cycle for patients who did not become pregnant after the first cycle or who planned to have more than one child and did not have surplus embryos, a situation that was more prevalent in the control group (48.1% versus 38.1% in the intervention arm). The last scenario, involving family planning for an additional child, could lead to future savings in the intervention group. However, calculating these savings is challenging, given the ovarian aging that will occur in 2 years or more. When these patients, who are currently 35 years old, return for a new treatment, the absence of frozen embryos for use adds to the complexity of the calculation.

A limitation of our study is that it is not randomized, with a retrospective historical control group. Although we used the same inclusion and exclusion criteria as the phase 3 study from which the control group originated, the intervention group had a lower antral follicle count than the control group (9.5 vs. 11.5, p = 0.007), which is important for a study on ovarian stimulation and response. Thereby one way we found of improving the comparison between the responses was to use indices that relativized the number of eggs obtained as a function of the initial antral follicles. One index, for example, suggested by Alviggi et al., is the Follicle-to-oocyte index (FOI = number of mature eggs × 100/number of antral follicles), with a FOI > 50% being considered normal [45]. Our control group, which used 12 μg of follitropin delta, had an average FOI of 66%, while the intervention group, using the combination of follitropin delta and menotropin, had an average FOI of 94%, a statistically significant difference (p < 0.001), which reinforces the positive effect of the intervention on the ovarian response. It should be remembered that this and all the other comparisons in this study were carried out with grouped results and not paired on a case-by-case basis due to the limited access to individual data from the phase 3 study for ethical reasons.