We have read with great interest the article by Witz et al. describing the results of the Menopur in GnRH Antagonist Cycles with Single Embryo Transfer – High Responder (MEGASET-HR) trial which compares highly purified human menotropin (HP-hMG) and recombinant human follicle-stimulating hormone (r-hFSH) in high responders undergoing intracytoplasmic sperm injection (Witz et al. 2020). With a standard starting dose of 150 IU for either r-hFSH or HP-hMG, Witz et al. report that HP-hMG was non inferior to r-hFSH in regards to ongoing pregnancy (primary endpoint) after fresh transfer (35.5% [HP-hMG] vs 30.7% [r-hFSH]; difference 4.7%, 95% CI, –2.7%, 12.1%).

We believe there are major methodological flaws in study design, statistical analysis, data collection and reporting, and interpretation of the results which raise serious questions regarding the validity of the authors’ conclusions.

Please read here the reasons and research for these findings.

Salvatore Longobardi, MD, PhD, Senior Medical Director Global Medical Affairs Fertility, Merck, Darmstadt, Germany Thomas D’Hooghe, MD, PhD, Vice-President and Head, Global Medical Affairs Fertility, Merck, Darmstadt and Professor Reproductive Medicine, Department of Development and Regeneration, Group Biomedical Sciences, KU Leuven (University of Leuven), Belgium The relevance of randomized controlled trials (RCTs) has long been established. Such trials significantly contribute to improve our knowledge, answer questions, and definitely change the standards of care. Thus, great attention and scrutiny should be used when evaluating such studies in terms of design, statistics and results according to strict criteria and guidelines for designing, implementing, and reporting RCTs. In addition, the paucity of public funds in our field increases the importance of properly designed high quality RCTs sponsored by pharmaceutical companies. Recently, Andersen NA et al (1) published the results of the ESTHER-1 trial (Evidence-based Stimulation Trial with human rFSH in Europe and Rest of world 1). The Authors concluded that “the present trial demonstrates that an individualized follitropin delta dosing is non inferior to conventional follitropin alfa with respect to ongoing pregnancy rate, ongoing implantation rate, and also live births, with a concomitant reduction in iatrogenic complications, including preventive interventions of OHSS”. This publication also reported that individualized follitropin delta resulted in more women with target response (8–14 oocytes) (43.3% vs. 38.4%), fewer poor responses (fewer than four oocytes in patients with AMH <15 pmol/L) (11.8% vs.17.9%), fewer excessive responses (>15 or >20 oocytes in patients with AMH >15pmol/L) (27.9% vs. 35.1% and 10.1%vs. 15.6%, respectively), and fewer measures taken to prevent ovarian hyperstimulation syndrome (2.3% vs. 4.5%), despite similar oocyte yield and similar blastocyst numbers, and less gonadotropin use (90.0 + 25.3 vs. 103.7 + 33.6 mg). In our opinion, major limitations in study design do not support the claims made regarding non-inferiority for primary outcomes ongoing pregnancy and implantation rates, or superiority for other outcomes as listed above. Firstly, a fair comparative assessment is impossible since patients in the follitropin delta arm and patients in the comparator arm receiving follitropin alpha were not subject to the same dosing regimens in the ESTHER-1 trial. Follitropin delta was used in an individualized dosing regimen based on body weight & anti-Müllerian hormone (AMH), whereas follitropin alfa with a fixed and inappropriate starting dose of 150IU and with dose adjustments as of day 6.It is well known that the starting dose is a very important determinant of the total number of oocytes retrieved (2), and that significant proportions of patients require different starting doses than 150 IU to reach optimal stimulation (3-6). This is the reason why the Summary of Product Characteristics (SmPC) for GONAL-f recommends a starting dose for follitropin alfa between 150 IU and 225 IU (7). However, in the ESTHER-1 trial, investigators were only allowed a follitropin alfa starting dose of 150 IU, preventing them from using the full range of follitropin alfa starting doses (150–225 IU) provided in its label. As a result of this design, patients with AMH <15 pmol/L received a higher starting dose of follitropin delta (12 μg) than those treated with follitropin alpha (11 μg contained in 150 IU of follitropin alpha) (7). On the contrary, the total gonadotropin dose received was significantly lower for follitropin delta (90.0 + 25.3 μg) than for follitropin alpha (103.7 + 33.6 μg, p<.001) in all patients. Secondly, only patients in the follitropin delta arm (not those in the follitropin alpha arm) were allowed the benefit of using clinical markers (body weight) and biomarkers (AMH) for determination of the starting dose, whereas these markers are used worldwide in real clinical practice for determination of the starting dose, regardless of the type of gonadotrophin used (4). Claims of non-inferiority can only be made if both products were used in exactly the same way in both arms, requiring a design where either both products were individually dosed based on AMH levels and body weight, or both products were used with a fixed starting dose that could be modified based on ovarian response. Thirdly, as described in the Rekovelle/follitropin delta Assessment Report of the Committee for Medicinal Products for Human Use of the European Medicines Agency in October 2016 document (www.ema.europa.eu/docs/en_GB/document_library/.../WC500220237.pdf) (8), the overall non-inferiority for ongoing pregnancy claimed in the ESTHER-1 trial (1) can be explained by some heterogeneity of responses related to female age. Indeed, non-inferiority for ongoing pregnancy rate was driven by only 15% of the study population (aged 38 or older), and in fact non-inferiority was not demonstrated for women aged 37 years or less, who represented the large majority (85%!) of the ESTHER-1 trial participants (8). These women aged 37 years or less also represented a good prognosis ART population, especially taking into account the in- and exclusion criteria used (1) and likely included most of the 629 participants (mITT population) (8) with serum AMH levels of 15 pmol/l or more. They were expected to benefit most from AMH determined dose setting for follitropin delta but ended up with a lower ongoing pregnancy rate after ovarian stimulation with follitropin delta than after ovarian stimulation with follitropin alpha. We hypothesize that this observation can be explained by higher egg quality obtained after ovarian stimulation with follitropin alpha then after ovarian stimulation with follitropin delta. This hypothesis is supported by a trend towards higher fertilization rates (62% versus 53-56%) and higher blastocyst/oocyte ratios (35% versus 29-31%) after ovarian stimulation with follitropin alpha 11 microgram (equal to 150 IU) when compared to follitropin delta doses 8.6 microgram, 10.3 microgram and 12.1 microgram, regardless of serum AMH levels below and above 15 pmol/l, as published in the Phase II Rekovelle trial (9). Fourthly, the study was only powered to demonstrate non-inferiority for the primary outcome of ongoing pregnancy and implantation rates, and any conclusion regarding differences in pre-specified (1) secondary outcomes (live birth rates, targeted ovarian response, extreme ovarian response, embryology data, proportion of women with early and late OHSS and/or preventive intervention for early OHSS) can only be considered to be hypothesis generating. Furthermore, the authors failed to differentiate which analyses were planned in the Statistical Analysis Plan and which analyses were performed posthoc. We hypothesize that all the following variables, that were not mentioned as primary/secondary outcome variables in the published paper (1) or on clinicaltrials.gov (10), were analyzed posthoc: number of oocytes retrieved, women with dose adjustments implemented, ovarian response stratified by AMH subdividing women with AMH <15 pmol/L (at risk of hypo-response) versus women with AMH >15 pmol/L (at risk of hyper response), proportion of women with blastocysts cryopreserved. Furthermore, authors need to report specifically, and separately for each category, the proportion of patients with “extreme responses” predefined as <4, >= 15 and >=20 oocytes recovered (now represented as an aggregate result in Table 3). For the reasons mentioned above, the design of the ESTHER-1 trial does not allow to conclude non-inferiority for follitropin delta versus follitropin alpha regarding primary outcomes ongoing pregnancy and implantation rates, or to claim any superiority based on secondary outcome or posthoc analysis.   References 1. Andersen N.A., Nelson S.M., Fauser B., García-Velasco J.A., Klein B.M. and Arce JC, for the ESTHER-1 study group. Individualized versus conventional ovarian stimulation for in vitro fertilization: a multicenter, randomized, controlled, assessor-blinded, phase 3 non inferiority trial Fertil Steril 2017;107:387-396.e4. doi: 10.1016/j.fertnstert.2016.10.033. Epub 2016 Nov 29. 2. La Marca A, Grisendi V, Spada E, Argento C, Milani S, Plebani M, Seracchioli R, Volpe A. Reference values in ovarian response to controlled ovarian stimulation throughout the reproductive period. Gynecol Endocrinol 2014; 30: 66-9. 3. Popovic-Todorovic, B, Loft A, Bredkjaeer H.E, Bangsbøll S, Nielsen I.K, Andersen A.N. A prospective randomized clinical trial comparing an individual dose of recombinant FSH based on predictive factors versus a ‘standard’ dose of 150 IU/day in standard’ patients undergoing IVF/ICSI treatment. Hum. Reprod 2003; 18, 2275–2282. 4. La Marca, A, Sunkara S.K. Individualization of controlled ovarian stimulation in IVF using ovarian reserve markers: from theory to practice. Hum. Reprod. Update 2014;20, 124–140. 5. Papaleo E, Zaffagnini S, Munaretto M, Vanni VS, Rebonato G, Grisendi V, Di Paola R, La Marca A. Clinical application of a nomogram based on age, serum FSH and AMH to select the FSH starting dose in IVF/ICSI cycles: a retrospective two-centres study European Journal of Obstetrics & Gynecology and Reproductive Biology 2016; 207; 94–99. 6. Allegra A, Marino A, Volpes A, Coffaro F, Scaglione P, Gullo S, La Marca A. randomized controlled trial investigating the use of a predictive nomogram for the selection of the FSH starting dose in IVF/ICSI cycles. RBMOn line 2017;34: 429-438. doi: 10.1016/j.rbmo.2017.01.012. Epub 2017 Jan 23. 7. www.ema.europa.eu/docs/en_GB/document_library/EPAR.../WC500023748.pdf 8. European Medicines Agency. Committee for Medicinal Products for Human Use. Rekovelle/follitropin alpha Assessment Report, 13th October 2016, EMA/11072/2017 (www.ema.europa.eu/docs/en_GB/document_library/.../WC500220237.pdf). 9. Arce JC, Andersen AN, Fernández-Sánchez M, Visnova H, Bosch E, García-Velasco JA, Barri P, de Sutter P, Klein BM, Fauser BC. Ovarian response to recombinant human follicle-stimulating hormone: a randomized, antimüllerian hormone-stratified, dose-response trial in women undergoing in vitro fertilization/intracytoplasmic sperm injection.Fertil Steril. 2014; 102:1633-40. 10. https://clinicaltrials.gov/ct2/show/record/NCT01956110